South Asia Economic Focus | Fall 2019 Making (De)centralization Work © 2019 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW, Washington, DC 20433 Telephone: 202-473-1000; Internet: www.worldbank.org Some rights reserved 1 2 3 4 23 22 21 20 This work is a product of the staff of The World Bank with external contribu- tions. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guar- antee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Nothing herein shall constitute or be considered to be a limitation upon or waiver of the privileges and immunities of The World Bank, all of which are specifically reserved. Rights and Permissions This work is available under the Creative Commons Attribution 3.0 IGO li- cense (CC BY 3.0 IGO) http://creativecommons.org/licenses/by/3.0/igo. Under the Creative Commons Attribution license, you are free to copy, distribute, transmit, and adapt this work, including for commercial purposes, under the following conditions: Attribution—Please cite the work as follows: World Bank. 2019. “Rethinking Decentralization” South Asia Economic Focus (October), World Bank, Washington, DC. Doi: 10.1596/978-1-4648-1515-7. License: Creative Commons Attribution CC BY 3.0 IGO Translations—If you create a translation of this work, please add the fol- lowing disclaimer along with the attribution: This translation was not cre- ated by The World Bank and should not be considered an official World Bank translation. The World Bank shall not be liable for any content or error in this translation. Adaptations—If you create an adaptation of this work, please add the fol- lowing disclaimer along with the attribution: This is an adaptation of an orig- inal work by The World Bank. Views and opinions expressed in the adaptation are the sole responsibility of the author or authors of the adaptation and are not endorsed by The World Bank. Third-party content—The World Bank does not necessarily own each com- ponent of the content contained within the work. The World Bank therefore does not warrant that the use of any third-party-owned individual compo- nent or part contained in the work will not infringe on the rights of those third parties. The risk of claims resulting from such infringement rests solely with you. If you wish to re-use a component of the work, it is your respon- sibility to determine whether permission is needed for that re-use and to obtain permission from the copyright owner. Examples of components can include, but are not limited to, tables, figures, or images. All queries on rights and licenses should be addressed to World Bank PHOTO BY: JAN S/SHUTTERSTOCK Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; e-mail: pubrights@worldbank.org. ISBN (electronic): 978-1-4648-1515-7 DOI: 10.1596/978-1-4648-1515-7 Cover design: Alejandro Espinosa/sonideas Fall 2019 Economic Focus South Asia 4 PHOTO BY: POLTU SHYAMAL/SHUTTERSTOCK T his report is a joint product of the Office of the Chief Economist for the South Asia Region (SARCE) and the Macroeconomics, Trade and Investment (MTI) Global Practice. Its preparation was led by Robert Beyer (Economist, SARCE) under the oversight of Hans Timmer (Chief Economist, South Asia Region), in close collaboration with Manuela Francisco (Practice Manager, MTI). The core team consisted of Sebastián Franco-Bedoya, Ishita Dugar, Milagros Chocce, and Rucheta Singh (all SARCE). Substantive contributions were made by Florian Blum and Nyda Mukhtar (both Economists, MTI), as well as by Sabiha Mohana (SARCE). The report greatly benefitted from inputs from Temel Taskin and other colleagues in the Prospects Group (EPGDR) under the supervision of Ayhan Kose (Director, EPGDR). Useful comments and suggestions were provided by numerous colleagues from the Office of the Chief Economist for the South Asia Region, the Macroeconomics, Trade and Investment Global Practice, the Poverty Global Practice, participants of an internal World Bank workshop on decentralization, with pre- sentations by Urmila Chatterjee (Senior Economist, ESAPV), Hideki Higashi (Senior Economist, HSAHN), David Ryan Mason (Urban Development Specialist, SSAU), Akmal Minallah (Senior Financial Management Specialist, ESAG1) and Roland White (Lead Urban Specialist, SURDR), as well as by participants of the 4th South Asia Economic Policy Network Conference on ‘Subnational Public Finance and Local Service Delivery’ in Dhaka, Bangladesh. Data and technical inputs were contributed by Thanh Bui (Research Analyst, EMFMD), Erik Feyen (Lead Financial Sector Economist, EFNDR), Hideki Higashi (Senior Economist, HSAHN), Massimo Mastruzzi (Senior Governance Specialist, EMFTX), Kirk David Schmidt (Governance Specialist, ESAG1), Temel Taskin (Economist, EPGDR), Marius Vismantas (Lead Financial Sector, ESAF1), and Roland White (Lead Urban Specialist, SURDR). Colleagues providing information for the country briefs coordinated by Mona Prasad (Lead Economist, MTI) include Kishan Abeygunawardana, Afroza Alam, Zehra Aslam, Florian Blum, Kene Ezemenari, Rangeet Ghosh, Adnan Ashraf Ghumman, Poonam Gupta, Bernard Haven, Tobias Haque, Zahid Hussain, Fernando Gabriel Im, Yoichiro Ishihara, Nayan Krishna Joshi, Nazmus Sadat Khan, Aurelien Kruse, Tae Hyun Lee, Nyda Mukhtar, Sayed Murtaza Muzaffari, Habiburahman Sahibzada, Saurabh Shome, and Muhammad Waheed (all MIT) under supervision of Manuela Francisco (Practice Manager, MTI). Alejandro Espinosa at Sonideas was responsible for the layout, design, and typesetting and William Shaw edited the chapters. Elena Karaban (Senior Communications Officer) and Yann Doignon (Online Communications Officer) at South Asia External Communications coordinated the dissemination, Gonzalo Alberto Villamizar De La Rosa created an accompanying video, and Neelam Chowdhry provided valuable administrative support. South Asia as used in this report includes Afghanistan, Bangladesh, Bhutan, India, Maldives, Nepal, Pakistan and Sri Lanka. The cutoff date for this report was October 7, 2019. 5 South Asia Chief Economist Office South Asia Economic Focus Macroeconomics, Trade and Investment Global Practice Fall 2019 Table of contents Recent economic developments 8 Overview 9 Global growth moderated 10 South Asia is losing its shine 11 Just another industrial production cycle? 13 Don’t blame the trade 14 Moderate inflation but rising food prices 16 What is going on in India? 17 References 21 South Asia economic outlook 22 Overview 23 Uncertain global environment 24 Uncertain domestic environment 25 Growth revised downward, but rebound expected 28 References 34 Appendix 35 Making (de)centralization work 36 Overview 37 South Asia is further decentralizing 38 One size doesn’t fit all 39 Partial decentralization 40 Decentralization and centralization belong together 43 Optimal allocation of central resources is key to success 45 Better geospatial data can improve evidence-based policy designs 48 Conclusions 50 References 52 Appendix 54 South Asia country briefs 58 PHOTO BY: DON MAMMOSER/SHUTTERSTOCK Recent economic developments 70 Outlook 70 Risks and challenges 70 7 South Asia at a glance 84 South Asia Economic Focus Fall 2019 RECENT ECONOMIC DEVELOPMENTS 8 South Asia Economic Focus Fall 2019 Overview vv Global GDP growth is decelerating, while trade and industrial production are stagnating. The slowdown started early 2018 and is strongest in in- vestments and other cyclical components of GDP. Uncertainty caused by trade tensions, Brexit, and oil supply is weighing on global confidence. vv In recent quarters, South Asia was no longer the fastest growing region in the world. In most South Asian countries, growth is expected to be below long- run averages this year. The growth slowdown is re- flected in weak stock markets and tensions in finan- cial markets. vv There is significant diversity across countries, which is evident in the high-frequency data of industrial production (IP). India’s IP cycle is more pronounced than in other parts of the world, which might explain the recent remarkably weak GDP data (relative to recent averages). In Pakistan, IP contracted much earlier than in the rest of the world, as the coun- try suffered a macroeconomic crisis. In Bangladesh, despite slowing somewhat IP remains surprisingly strong, as the country’s garment industry benefitted from the trade tensions between the United States and China. vv Current account deficits have declined in the re- gion, as is often the case during economic down- turns. The sharp declines in import volumes reflect the weakness in investments. The only moderate de- celeration of export volumes signifies that the slow- down in South Asia does not primarily reflect weak external demand, but rather weak domestic demand, affected by global and domestic uncertainties. vv Inflation remains near target in most countries, but food price inflation is picking up. Inflation be- low target offers room for monetary policy easing in some countries. After declining last year, food prices picked up recently. Empirical evidence suggests that food prices in South Asia are driven by rainfall and oil prices, and not by international food prices. vv The remarkable weakness of Indian economic ac- PHOTO BY: LUMENITE/SHUTTERSTOCK tivity during the first half of 2019 is largely driven by external and cyclical factors. However, during this downturn several structural problems have come 9 to the surface. One of these problems is related to vul- nerabilities in the financial markets that have con- South Asia strained credit supply. Financial sector reforms are Economic Focus needed to bring India back to a rapid growth path. Fall 2019 Recent Figure 1: Global GDP growth moderated and trade and industrial production are stagnating. economic Global GDP growth slowed from an average of 3.0 percent in 2018 to 2.4 percent in the second quarter of 2019 and world trade developments and industrial production growth, both more cyclical variables, decelerated strongly. Real GDP growth World trade and industrial production Percent change, y-o-y Percent change, y-o-y 5 6 5 4 4 3 3 2 2 1 1 0 0 -1 2018 Q2 2018 Q3 2018 Q4 2019 Q1 2019 Q2 2017 2018 Q1 2018 Q2 2018 Q3 2018 Q4 2019 Q1 2019 Q2 Developing Countries United States World OECD (excluding United States) World trade (volume) World IP Source: World Bank. Source: Netherlands Bureau for Economic Policy Analysis (CPB). Global growth moderated industrial production even contracted. Both trade and industrial production are related to investment demand and their hefty slowdowns point to weak investor confi- In the second quarter of this year, global GDP growth dence, as the trade tensions between the US and China, declined for the sixth quarter in a row, trade growth was below 1.0 percent, and global industrial produc- the possibility of a no-deal Brexit, and the Japan-South tion even contracted (Figure 1). Global GDP growth Korea dispute all weigh on investment sentiments. In slowed to 2.4 percent in the second quarter of 2019, addition, the rising tensions between Iran and Saudi compared to 3.3 percent a year ago. The slowdown was Arabia/USA loom over the oil price and consequently broad based and driven both by developing and devel- the prospects for the global economy. oped countries. For the former, growth declined to 3.7 percent, well below the average growth rate of 4.3 per- This year, US and emerging market stock prices co- cent over the last five years. Growth in OECD countries, move strongly, signaling a global cycle (Figure 2). excluding the United States, was only 1.0 percent and hit Strong GDP growth in the US last year led to expecta- a 6-year low. In the United States, a crucial export mar- tions of rising US interest rates, which tightened exter- ket for South Asian countries (World Bank 2019), growth nal financing conditions for emerging markets (World slowed to 2.3 percent in the second quarter of this year Bank 2018a) and resulted in a negative correlation of US compared to 3.2 percent a year ago. World trade and in- and emerging market stock price movements between dustrial production are usually more volatile, and in fact March and October 2018. Since then, however, the cor- their growth decelerated even faster. World trade grew relation between the two stock price indices has been below one percent in the second quarter of this year and relatively high at 0.7. This is the result of global factors Figure 2: Co-movement between US and emerging market stocks point to a global cycle. After opposing trends last year, US and emerging market stock prices are positively correlated again, signaling a global cycle. After US and emerging market stocks recovered in early 2019, both lost again more recently. United States and emerging market stocks Index, 100=January 2, 2018 100 80 60 40 10 20 0 Jan-18 Feb-18 Mar-18 Apr-18 May-18 Jun-18 Jul-18 Aug-18 Sep-18 Oct-18 Nov-18 Dec-18 Jan-19 Feb-19 Mar-19 Apr-19 May-19 Jun-19 Jul-19 Aug-19 Sep-19 Oct-19 South Asia Economic Focus S&P 500 FTSE Emerging All-Cap Source: Haver Analytics. Fall 2019 Note: The last observation is October 4, 2019. Figure 3: The oil price remains very volatile but is expected to be below 60 USD per barrel in the Recent medium-run. economic At the end of last year, the oil price was 64 USD per barrel. It increased by nearly 40 percent until April of this year and subsequently developments fell by 25 percent until September. Recently, the price jumped to nearly 70 USD per barrel following attacks on Saudi oil production. However, it started falling again and currently stands at around 60 USD per barrel. Oil price, Oil intra-year uctuation price, intra-year uctuation (Brent) (Brent) Brent futures Brent curves futures curves USD/barrel USD/barrel USD per USD barrel per barrel 80 80 +39% since +39% since 80 80 January January 75 75 -19% since -19% since the peak the in April peak in April 70 70 70 70 65 65 60 60 60 60 55 55 50 50 45 45 50 50 2019 2019 2020 2020 2021 2021 2022 2022 Jan-19 Feb-19 Mar-19 Apr-19 May-19 Jun-19 Jul-19 Aug-19 Sep-19 Jul 31, Jul 2019 31, 2019 Dec Dec 31, 31, 2018 2018 Sep 23, Sep 2019 23, 2019 Source: Haver Analytics. Sources: Bloomberg and World Bank. Note: The last observation is September 30, 2019. Notes: Futures curves as reported on the closing prices of the respective day. The last observation is December 2022. driving both advanced and emerging markets jointly – nearly 70 USD per barrel again. The subsequent tensions the world is experiencing a global cycle. Both US and between Iran and Saudi Arabia have increased uncer- emerging stock markets gained early in the year, but US tainty about future oil price movements. Higher oil pric- stocks outperformed those in emerging markets. Stock es could further worsen the global economic downturn. prices from January to May increased by 13.1 percent In a survey conducted for this report (see Box 2), South in the US and by 7.9 percent in emerging markets. Since Asian economists expressed more concern about oil pric- then, however, both indices have fluctuated strongly es than about global GDP and trade growth decelerations. without a clear trend, illustrating the heightened uncer- However, oil price futures suggest that prices will remain tainty in the global economy. close to 60 USD per barrel. Oil price volatility remains high and oil price futures suggest prices may remain around 60 USD per barrel (Figure 3). Over the course of this year, the price of oil has South Asia is losing its shine fluctuated between a low of 53 USD per barrel and a high of 75 USD per barrel. After rising until mid-April, when it In line with global developments, economic activ- reached values above 70, it fell again in line with concerns ity is moderating in many countries in South Asia about weakening global demand. However, after the dis- (Figure 4). In India, quarterly growth has declined for ruption of Saudi oil production in September, caused by five quarters in a row. The 5 percent (y-o-y) growth in the attacks on two major oil facilities, the oil price jumped to second quarter of this year was the lowest since the first Figure 4: Economic activity in South Asia is slowing down. Quarterly growth declined strongly in India and has been weak in Sri Lanka in the second quarter of this year. In Bangladesh and Pakistan, high-frequency indicators signal a slowing economic activity. Quarterly real GDP growth Pakistan Bangladesh Percent change, y-o-y 9 60 60 8 50 50 Monthly y-o-y change in percent: Monthly y-o-y change in percent: 7 40 40 most recent observation most recent observation CC 6 30 30 5 20 20 EP 4 EXP MP 10 PSC 10 PSC 3 EXP 2 0 0 IMP 1 0 2018 Q1 2018 Q2 2018 Q3 2018 Q4 2019 Q1 2019 Q2 -10 -20 -20 0 MP EP 20 IMP 40 60 -10 -20 -20 0 20 40 60 11 India Sri Lanka Monthly y-o-y change in percent: a year ago Monthly y-o-y change in percent: a year ago South Asia Sources: World Bank, Trading Economics, National Authorities, and Haver Analytics. Economic Focus Notes: EXP= Exports, IMP= Imports, MP= Manufacturing Production, PSC= Private Sector Credit, EP= Electricity Production, CC= Consumer Confidence. Fall 2019 Recent Figure 5: Many South Asian countries are growing below medium-term averages and potential. economic In six out of the eight countries in South Asia, growth projections for fiscal year 2019 are below medium-run averages and developments potential. Projected real GDP growth in calendar year 2019 Percent change, y-o-y 8 7 6 5 4 3 2 1 0 Bangladesh Nepal India Maldives Bhutan Pakistan Sri Lanka Afghanistan Average 2013-2018 Potential Source: World Bank. Notes: Bars represent forecasts for 2019, horizontal line markers represent the average growth between 2013 and 2018, and triangles represent the potential growth based on the World Bank’s macroeconomic and fiscal model (Burns et al. 2019). Real GDP growth is in calendar years for all countries. Figure 6: South Asia is not the fastest growing region in the world anymore. With 5.1 percent, South Asia’s growth has been below growth in East Asia and Pacific in the second quarter of this year. Quarterly growth peaked in 2016 with 7.5 percent and since then has decelerated strongly. Regional quarterly real GDP growth Percent change, y-o-y 8 7 6 5 4 3 2 1 0 2018 Q1 2018 Q2 2018 Q3 2018 Q4 2019 Q1 2019 Q2 East Asia & Paci c Europe & Central Asia Middle East & North Africa Sub-Saharan Africa Latin America & Caribbean South Asia (India & Sri Lanka) South Asia (approximated) Source: World Bank. Notes: Data is for countries that report quarterly data. The grey line is for all South Asian countries, the quarterly data for which was approximated using linear interpolation. quarter of 2013. In Sri Lanka, growth fell to just 1.6 per- in Sri Lanka has been below long-run averages for some cent in the second quarter of this year due to the Easter time and consistent with that potential growth is below attacks. While this was the slowest growth in over 4 years, its long-run average as well. In Pakistan, measures to re- it was still higher than many expected. For Pakistan and store macroeconomic stability weigh heavily on growth, Bangladesh, the second and third largest economies in which is expected to have dropped to 3.3 percent in the region, unfortunately no up-to-date quarterly GDP fiscal year 2018/19. Growth is also expected to decel- data is available. But high-frequency indicators in these erate below long-run trends and potential in Bhutan, countries, suggest slowing economic activity as well. Maldives, and India this year. None of these countries Imports were stable in July 2019 in Bangladesh and grew below long-run trends or potential last year. In even decreased in Pakistan in August 2019, while they in- Afghanistan, growth this year benefits from improved creased over 18 percent (y-o-y) a year ago in Bangladesh farming conditions and is expected to be the long-run and over 30 percent (y-o-y) in Pakistan. Manufacturing average, which is equal to potential. The two exceptions production in Pakistan contracted by over 3 percent are Nepal and Bangladesh, which are leading growth in 12 (y-o-y) in March 2019, while it was still growing a year earlier. Other indicators like private sector credit also South Asia and are expected to have grown above po- tential and long-run averages in fiscal year 2018/19. point to a slowdown in these countries. South Asia In the last two quarters, South Asia has not been the Economic Focus Most South Asian countries are expected to grow fastest growing region in the world anymore (Figure Fall 2019 below long-run averages this year (Figure 5). Growth 6). India’s GDP accounts for 84 percent of regional GDP Figure 7: Stock markets in South Asia mostly turned downward. Recent While stock prices remained stable in India and gained momentum in Sri Lanka following an initial slowdown due to the terror economic attacks, they declined significantly in Bangladesh and Pakistan. developments 6-month change in South Asian stock market indexes Percent 5 0 -5 -10 -15 -20 Karachi Dhaka Bombay Colombo Sources: Haver Analytics and staff calculations. Notes: Growth rates are calculated using daily price data. The last observation is October 4, 2019. and hence drives regional growth. But strong changes in growth in other countries also can have a significant Just another industrial impact on regional growth. The only two countries re- production cycle? porting up-to-date quarterly GDP growth in South Asia are India and Sri Lanka. But whether approximations Global industrial production cycles are strongly of the other countries’ quarterly growth rates are in- correlated with GDP, but they are more pronounced cluded in the regional growth rate or not, the region (Figure 8). The correlation between global industrial has not been the fastest growing in the world for the production and GDP growth from 2009 to 2018 was 0.97. last two quarters. It had been growing faster than all The strong and positive correlation shows that when other regions since 2015, but the region’s 5.1 percent industrial production grew fast, GDP tended to grow (y-o-y) growth in the second quarter of this year was fast as well (and the other way around). The correlation below growth in East Asia and Pacific. The continued between industrial production and GDP is especially deceleration in the second quarter resulted in a larger strong in China, where industrial production accounts gap with East Asia and Pacific. for more than half of the economy. But in the US the relationship is also strong, despite industrial production In line with macroeconomic and financial de- accounting for less than a fifth of the economy. One rea- velopments, stock prices in South Asia have fall- son is the spill-overs from industrial production to ser- en (Figure 7). Over the course of the last six months, vices. In South Asia, the share of industrial production stock prices have declined strongly in Pakistan and in overall economic activity is close to the world average Bangladesh, remained nearly stable in India and gained and much smaller than in China. Yet, the correlation be- somewhat in Sri Lanka. In Pakistan, the decline was tween industrial production and GDP is large as well, triggered by the economic slowdown and macroeco- especially in India and Bangladesh. In Sri Lanka and nomic adjustment polices and was particularly sharp Pakistan, this correlation is somewhat weaker. between May and August. Pakistan’s stock market in- dex dropped to its lowest level in August 2019 amid The relationship between industrial production cy- the rising tensions between Pakistan and India over cles in different countries is stronger than between the Kashmir dispute. Over the last month, however, the GDP cycles. Despite South Asian countries not export- KSE has been bullish and went up by over 6 percent. ing a lot of merchandise goods (World Bank 2019), in- In Bangladesh, the stock market continues to tumble; dustrial production in South Asia is strongly correlated stock prices have fallen by 10 percent since March. The with industrial production elsewhere. The relationship Dhaka Stock Exchange (DSEX) index hit a 33-month low between industrial production in South Asia and in the in September. Reasons for the downturn despite con- US, China, and the rest of the world is strong (Figure 9). tinued high growth include financial sector stability Industrial production refers to the output of industrial concerns and liquidity constraints of banks and non- establishments but also covers sectors such as mining bank financial companies (NBFCs). In India, the SENSEX and gas, and hence commodity price booms may con- has been stable since March. In Sri Lanka, stock prices tribute to the strong relationship. Therefore, given the have gone up by 3 percent compared to six months ago. weakness in industrial production in both advanced While the Colombo Stock Exchange was on a strong economies and emerging markets around the world, downward trend until May due to the economic leth- weakness in South Asia is not a surprise. argy after the terror attacks, it has reversed course in recent months. The recent decline in industrial production across 13 South Asian countries follows global developments, but it is more pronounced and idiosyncratic factors South Asia are at play as well. In the rest of the world, industrial Economic Focus production started to slow earlier than in India, which Fall 2019 Recent Figure 8: Industrial production cycles and GDP are correlated. economic Industrial production is correlated with GDP around the world. In South Asia, the correlation is stronger in India and Bangladesh developments compared to Pakistan and Sri Lanka. Global IP and GDP momentum Percent, quarter-on-quarter Share of IP Correlation 4 Country in GDP IP/GDP 2 Bangladesh 33 0.75 0 China 54 0.98 -2 India 26 0.93 -4 Sri Lanka 25 0.30 -6 Pakistan 18 0.52 -8 United States 19 0.88 2019 Q1 2017 Q2 2015 Q3 2013 Q4 2012 Q1 2010 Q2 2008 Q3 2006 Q4 2005 Q1 2003 Q2 2001 Q3 1999 Q4 1998 Q1 1996 Q2 1994 Q3 1992 Q4 1991 Q1 World 26 0.97 Industrial production GDP Source: World Bank and staff calculations. Notes: The growth rates, shares, and correlations are computed using annual data of IP and GDP in 2010 USD constant prices. The share refers to 2018 and the correlation has been computed from 2009 to 2018. The correlation is the ratio of total common variation (covariance) between industrial production and GDP growth to a measure of the total variation (the product of the standard deviations). Figure 9: Industrial production cycles are correlated across the world. The decline in growth is in line with global developments, as industrial production slows. Industrialproduction Industrial production Relationshipbetween Relationship betweenGDP GDPand andIP IPin SouthAsia inSouth andthe Asiaand world theworld Percent, Percent, y-o-y, y-o-y, 12-month 12-month moving moving average average Regression Regression coe cient coe cient 14 14 0.45 0.45 12 12 0.40 0.40 10 10 0.35 0.35 88 0.30 0.30 66 0.25 0.25 44 0.20 0.20 22 0.15 0.15 00 0.10 0.10 -2 -2 0.05 0.05 -4 -4 0.00 0.00 Dec-11 Dec-11 Apr-12 Apr-12 Aug-12 Aug-12 Dec-12 Dec-12 Apr-13 Apr-13 Aug-13 Aug-13 Dec-13 Dec-13 Apr-14 Apr-14 Aug-14 Aug-14 Dec-14 Dec-14 Apr-15 Apr-15 Aug-15 Aug-15 Dec-15 Dec-15 Apr-16 Apr-16 Aug-16 Aug-16 Dec-16 Dec-16 Apr-17 Apr-17 Aug-17 Aug-17 Dec-17 Dec-17 Apr-18 Apr-18 Aug-18 Aug-18 Dec-18 Dec-18 Apr-19 Apr-19 Aug-19 Aug-19 with with China China with with US US with with rest rest of of the the world world India India China China World World United United States States Industrial Industrial production production GDP GDP Source: World Bank. Source: World Bank and staff calculations. Note: The last observation is August 2019. Notes: Solid bars represent coefficients that are significant and patterned bars represent coefficients that are not significant. All GDP and IP series are in growth rates. The frequen- cy of data is annual and covers years 1999 to 2018. The countries included in South Asia are India, Pakistan, and Bangladesh. Rest of the world was calculated using the world in- dustrial production less South Asia, China, and United States. experienced an industrial production boom in 2016 and early 2017. After some moderation industrial production Don’t blame the trade accelerated again in 2018 but this year, in line with the Exports grew faster than imports in the first two rest of the world, Indian industrial production is slowing quarters of 2019, suggesting weak domestic demand (Figure 10). Industrial production in Pakistan has con- (Figure 11). Thus, weakening global conditions do not tinued its collapse that began at the beginning of last seem to be affecting South Asia through the trade chan- year, and now industrial production is even contracting. nel. Apart from Pakistan, South Asian exports continued Bangladesh, on the other hand, benefits from its strong growing fast in the first quarter of this year, but export garment sector and from trade diversion arising from the growth moderated strongly in India in the second quar- trade tensions between the US and China. Despite some ter. Import growth, on the other hand, has declined se- slowdowns recently, its industrial production is growing verely across countries in South Asia, and imports even 14 faster than the overall economy, making it the only coun- try in South Asia that continues to industrialize. However, contracted between 15 and 20 percent (y-o-y) in Pakistan and Sri Lanka. The different developments for exports the growth is coming primarily from the ready-made- South Asia and imports have led to a reduction of the trade im- garments sector, while other major exporting sectors are Economic Focus balances for all countries. In addition, the stronger de- not performing as well and hence the industrialization is clines in imports suggest that the trade channel is not Fall 2019 neither broad based nor sufficiently diversified. Figure 10: Industrial production in South Asia shows idiosyncratic developments as well. Recent The recent decline in industrial production growth is in line with global developments, but at the same time there are important economic idiosyncratic developments in Pakistan and Bangladesh. developments Industrial production Percent, y-o-y, 12-month moving average 18 15 12 9 6 3 0 -3 -6 Dec-11 Apr-12 Aug-12 Dec-12 Apr-13 Aug-13 Dec-13 Apr-14 Aug-14 Dec-14 Apr-15 Aug-15 Dec-15 Apr-16 Aug-16 Dec-16 Apr-17 Aug-17 Dec-17 Apr-18 Aug-18 Dec-18 Apr-19 Aug-19 India Pakistan Bangladesh World Source: World Bank. Note: The last observation is May 2019 for Bangladesh and August 2019 for the rest. Figure 11: Imports are falling while exports keep growing, suggesting weak domestic demand. Nominal exports are growing faster than imports, suggesting weak domestic demand. In Pakistan and Sri Lanka, imports are contracting sharply. Growth of exports Growth of imports Percent change, y-o-y Percent change, y-o-y 30 30 20 20 10 10 0 0 -10 -10 -20 -20 2017 2018 Q1 2018 Q2 2018 Q3 2018 Q4 2019 Q1 2019 Q2 2017 2018 Q1 2018 Q2 2018 Q3 2018 Q4 2019 Q1 2019 Q2 Bangladesh India Pakistan Sri Lanka Bangladesh India Pakistan Sri Lanka Sources: International Monetary Fund (IMF) and World Bank. Sources: International Monetary Fund (IMF) and World Bank. Notes: Exports are in current USD. The year 2017 represents the average of 2017. Notes: Imports are in current USD. The year 2017 represents the average of 2017. responsible for the current economic slowdown in South capital flows reversed, while credit default spreads in- Asia. Oil prices increased between mid-2017 and October creased, and currencies depreciated. This challenging situ- 2018, contributing to rising imports back then. Since ation bottomed out six months ago (World Bank 2019) and then, the oil price has decreased, which contributed to has further improved since then. In addition to a closing the decline in imports. However, non-oil imports show trade gap, the current account was supported by strong the same dynamics. This indicates that after strong do- remittances inflows, which grew in all countries except mestic demand growth in recent years (see World Bank Sri Lanka. FDI flows, on the other hand, were mostly flat, 2019), domestic demand has started to moderate. with Pakistan being the only exception with a continuous fall since 2018. In line with these developments, current Closing current account balances are easing concerns account imbalances are closing. Sri Lanka registered a sur- about external imbalances (Figure 12). Last year, dou- plus of 2 percent of GDP in the first quarter of the year, ble-digit import growth and tighter financial conditions but the current account was again in deficit in the second quarter due to the Easter attacks. Increased external stabili- 15 made it more difficult for South Asian countries to finance the increasing current account deficits. In some countries, ty is reflected in mostly stable exchange rates in South Asia. South Asia Economic Focus Fall 2019 Recent Figure 12: Current account balances are closing, easing concerns about external imbalances. economic Helped by the closing of the trade gap, current account deficits in South Asia are closing. developments Current account balance Percent of GDP 3 2 1 0 -1 -2 -3 -4 -5 2017 2018 Q1 2018 Q2 2018 Q3 2018 Q4 2019 Q1 2019 Q2 Bangladesh India Pakistan Sri Lanka Sources: Haver Analytics and National Authorities. Notes: Quarterly GDP for Bangladesh and Pakistan were derived from annual GDP and assumed to be constant for all four quarters. The year 2017 represents the average of 2017. Moderate inflation but eased its monetary policy stance and has lowered rates twice this year already. In August, the Central Bank of Sri rising food prices Lanka announced a 50-basis points rate cut to boost cred- it flows to support growth. Pakistan, on the other hand, Inflation is below target in India and Sri Lanka, but has now increased its main policy rate nine times since above in Pakistan, and monetary policies are adjust- the beginning of last year. The last increase took place in ing (Figure 13). In line with lower economic activity, in- July, when the State Bank of Pakistan increased its rate by flation rates in India declined from 5.0 percent a year ago 100 basis points to 13.25 percent, due to high inflationary to 3.2 percent at the end of the second quarter of this year. And in line with growth below potential, inflation was and external pressures. The central banks in Nepal and below the mid-point of the Reserve Bank of India’s (RBI) Bangladesh have left interest rates unchanged since mid- inflation target band. The RBI used this room to become 2018. The Bangladesh’s central bank kept its 12.5 percent the first central bank in the Asia-Pacific region to begin target for broad money growth unchanged in July. an easing cycle. It shifted the policy stance from “neutral” to “accommodative” and reduced the repo rate by 135 Across South Asia, food prices have been increasing basis points (year to date) to 5.1 percent. Sri Lanka also over the last few months (Figure 14). Food prices in Figure 13: Inflation is below target in Sri Lanka and India but above in Pakistan, and monetary policy is adjusting accordingly. In line with growth below potential, inflation is below target in Sri Lanka and India. In Bangladesh inflation is at the target, and in Pakistan strong exchange rate adjustments resulted in strongly increasing prices. Consequently, Sri Lanka and India started easing cycles and Pakistan sharply increased its policy rate. In ation and target in ation O cial interest rate (policy instrument/base rate) Percent Repo rate (EOP, percent) Repo rate (EOP, percent) 12 8.5 15 11 8.0 10 13 9 7.5 8 11 7 7.0 6 6.5 9 5 4 6.0 3 7 5.5 2 1 5.0 5 2017 Jan-18 Feb-18 Mar-18 Apr-18 May-18 Jun-18 Jul-18 Aug-18 Sep-18 Oct-18 Nov-18 Dec-18 Jan-19 Feb-19 Mar-19 Apr-19 May-19 Jun-19 Jul-19 Aug-19 Sep-19 Oct-19 0 16 Sri Lanka India In ation (August 2019) Bangladesh In ation target Pakistan Bangladesh India Nepal Sri Lanka Pakistan (RHS) South Asia Sources: Inflation target data is from Haver Analytics and National Authorities. Current in- Sources: Haver Analytics and National Central Banks. Economic Focus flation data is from World Bank. Notes: The year 2017 represents the average of 2017. The last observation is October 2019 for India, July 2019 for Nepal, and September 2019 for the rest. Notes: Sri Lanka is moving toward an inflation-targeting regime; this figure uses the im- Fall 2019 plicit target defined as the midpoint of the target band between 4 and 6 percent. Figure 14: In contrast to last year, food prices recently increased. Recent In Pakistan, Afghanistan, Nepal, and Bangladesh, food prices are around a 6 percent higher than a year ago. economic developments Food consumer price in ation Percent, y-o-y 14 12 10 8 6 4 2 0 -2 -4 -6 Pakistan Afghanistan Nepal Bangladesh India Maldives Sri Lanka May-Aug 2019 average 2018 Sources: Afghanistan, Bangladesh, India, Pakistan and Sri Lanka National Authorities. Maldives data series from MMA / Haver Analytics. Notes: The year 2018 represents the percent change between Dec-2017 and Dec-2018. The last observation is July 2019 for Nepal, June 2019 for Afghanistan, and August 2019 for the rest. 2018 were stable in Pakistan, Afghanistan, and Nepal and The slowdown is mostly due to a deceleration in fell in India, Maldives, and Sri Lanka. However, food prices domestic demand. After years of contributing to high recently have been increasing in all countries except Sri growth rates (World Bank 2018a), domestic demand Lanka. In Pakistan, Afghanistan, Nepal, and Bangladesh, slipped and contributed the most to the disappointing food prices are more than 6 percent higher than a year performance in the last quarter. Private consumption ago. Food prices are especially important for poverty and investment both grew slower than overall GDP in reduction. On the one hand, food expenditures repre- the second quarter of this year. Investment grew 4.0 per- sent a substantial share of consumption for poor peo- cent (y-o-y) in the second quarter, compared to 13.3 per- ple (Attanasio et al. 2013). On the other hand, sustained cent a year ago, while private consumption grew 3.1 per- increases in food prices have often benefited the poor cent, compared to 7.3 percent a year ago. One reason for and likely contributed to faster global poverty reduction, slowing private consumption is the strong contraction of since their incomes depend positively on food prices car sales that started in mid-2018, driven in part by high- (Headey and Martin 2016). In South Asia the negative ef- er insurance premia, new emission norms, uncertainty fect prevails, and the region is vulnerable to increasing about GST cuts, and the squeeze in the non-bank finan- food inflation especially because of the large segment of cial companies (NBFC) sector (see Chapter 2). However, the population living near the poverty line. If food prices global sentiments may also play a role since car sales increase by 10 percent, the poor population increases by are down in Europe and China as well. Together, con- over 2.1 percentage points on average; however, differenc- sumption and investment grew 6.0 percentage points es within the region are large. Sri Lanka is least affected slower than a year ago. In line with weakening domes- by rising food prices, and rural India and Bangladesh are tic demand, import growth fell from 11.0 percent a year most affected (Carrasco and Mukhopadhyay 2012). Food ago to only 4.2 percent in the second quarter of this year. prices partly depend on the monsoon patterns (see Box With a growth rate of 8.6 percent, government consump- 1). This year, the monsoon started late and many areas tion has become the fastest growing expenditure com- in South Asia received less rain than typical at the begin- ponent. The recent slowdown in India is not surprising ning of the season. In mid-July, however, heavy rainfalls given global economic developments, and even slow- helped push rainfall very close to the 50-year average. downs driven by domestic demand can be caused by ex- ternal shocks through capital flows or global sentiment, but the downturn is more pronounced than elsewhere. What is going on in India? In such a weak economic environment, structural is- India’s cyclical slowdown is severe. Quarterly GDP sues surface and the weak financial sector is becom- growth slowed for 5 quarters in a row, declining from a ing a drag on growth. Despite high economic growth peak of 8.1 percent in the first quarter of 2018 to only in the last decade, India’s banking system still has a sig- 5.0 percent in the second quarter of this year (Figure 15). nificant level of non-performing assets of close to 10 per- Growth decelerated by 3 percentage points in the last cent of total assets (Figure 16, left panel), some of which year and growth in the second quarter of this year was were generated or revealed during the global financial the lowest in over six years. Manufacturing growth fell crisis. The introduction of the 2016 bankruptcy code and from over 10 percent a year ago to below 1 percent in re-capitalizations of state-owned banks were necessary the second quarter of 2019. This drop follows the global steps, but not enough to resolve the weakness. The high trend but is more pronounced. Services and construc- share of non-performing assets in the banking sector tion also started decelerating over the last quarters, suggesting that the slowdown is not related to idiosyn- did not immediately result in a significant fall in overall credit, because non-banking financial companies (NBFCs) 17 cratic factors related to a specific sector. Export growth increased their credit supply to the real sector. Recently recently declined – in line with slowing world growth NBFCs came under liquidity stress and NBFC funding has South Asia and weak external demand – but cannot alone explain contracted severely. Before that, NBFCs financed 40 per- Economic Focus India’s sharp downturn. cent of car sales, and hence the squeeze contributed to Fall 2019 Recent economic Box 1: The drivers of food price inflation in South Asia developments Food is an essential and large component of the consumption basket in South Asia and strongly impacts aggregate infla- tion. In addition to the direct effect, rising food prices translate into higher non-food prices because workers demand higher wages, raising the cost of production and by default the prices of non-food items as well. Around three quarters of the variation in consumer price inflation in Afghanistan, India, Sri Lanka, and Pakistan can be attributed to food price fluctuations (World Bank 2019). In recent months, food prices in South Asia have increased sharply. Among the many factors that may impact food prices, the following three seem the most obvious ones: global food prices, oil prices, and unusual rainfall patterns. Both global factors (e.g. food and oil prices) and regional factors (e.g. rainfall patterns, which are correlated across South Asian countries) have been shown to matter for food prices in South Asia (World Bank 2019). To understand how important the different factors are, we compile monthly data from January 2013 to December 2016 for food prices in Bangladesh, India, Pakistan, and Sri Lanka, global food prices (FAO food price index), oil prices (average of Brent, Dubai, WTI) and rainfall for Bangladesh, India, Pakistan, and Sri Lanka (data from World Bank). We compute both extreme deficient and surplus rainfall. As a measure of unusual rainfall patterns, we define deficient or surplus rainfall as the percent deviation from country-specific and month-specific averages by the following formula: From this series we create dummies to capture months of extreme rainfall. The first dummy captures months of deficient rain, which are defined as months with rain at least one standard deviation below the typical rain in this country in this month. The second one captures months of surplus rain, defined as months with rain at least one standard deviation above the typical rain in this country during this month, but not above twice the standard deviation. In contrast to deficiency, there are many months of extreme surplus rain, so a dummy variable is included for months when rainfall is above twice the standard deviation. Table 1: Food prices are impacted by oil prices and deficient rainfall, but not by global food prices. Deficient Surplus Extreme Global food Fixed # of Within   Oil price Constant rain rain surplus price effects countries country R2 Food prices 1.2** 0.3 1.4 5.4*** 0.01 YES YES 4 0.46 Sources: Food and Agriculture Organization of the United Nations, World Bank, Haver Analytics, and staff calculations. Notes: Coefficients are from a pooled panel regression of Bangladesh, India, Pakistan, and Sri Lanka from January 2013 to December 2016. Standard errors in parentheses; ** p<0.05, *** p<0.01. Oil prices and deficient rainfall affect South Asian food prices, but global food prices do not (Table 1). We run a pooled panel regression of food prices in South Asia (Bangladesh, India, Pakistan, and Sri Lanka) on global food prices, oil prices, and extreme rainfall. The within-country fit of the regression, with a coefficient of determination of 0.46, is remarkable. Higher oil prices in- crease agricultural production costs, for example, through fertilizers and the cost of gasoline to run agricultural equipment. And indeed, oil prices significantly affect food prices in South Asia. Food prices also tend to be higher in months of extreme surplus rain, but the relationship is not statistically significant (the same result is found when surplus rain and extreme surplus rain are combined). Deficient rainfall, however, has a statistically significant effect on food prices. In months in which the rainfall is at least one standard deviation below normal, food prices are 1.2 percentage points higher. Across South Asia, most of the year’s total rain falls during the monsoon and large parts of the total cropped area is not irrigated, making the farmers heavily depen- dent on rainfall. It seems likely that the weak start of the monsoon in 2019 has contributed to the higher food prices observed in recent months. Afghanistan, following a severe drought in 2018, experienced more precipitation and snowfalls in early 2019, which is expected to improve farming conditions and agricultural production. The effect of rainfall deviations can also explain the co-movement in inflation rates amongst the South Asian economies (Blagrave 2019). Global food prices are often alluded to as a significant factor explaining movements in food prices in South Asia. While they are correlated with food prices in South Asia, they are not a significant factor in our regression, suggesting that both are commonly driven by oil prices. the recent consumption slowdown. There are some signs could deteriorate further if the recent slowdown is not of improvement in the financial sector. The non-perform- properly addressed and contained. ing assets ratio decreased from March 2018 to March 2019 for both public and private banks and both in industry This critical situation demands decisive policy ac- 18 and services (Figure 16, left and middle panels). And de- spite new regulatory and supervisory efforts from the RBI that resulted in greater market discipline, overall credit tions, and initial government steps point in the right direction. The deceleration in growth in the context of slowing global GDP and trade growth, as well as an un- South Asia growth in the economy picked up again in July (Figure certain external environment, bears resemblance to Economic Focus 16, right panel). But these positive developments leave 2008 and 2012/13 when GDP growth in India slumped. Fall 2019 no room for complacency, as financial sector conditions Both monetary and fiscal policy measures are needed Recent Figure 15: India’s strong slowdown is driven by domestic demand. economic India’s slowdown is broad-based but mostly driven by domestic demand. Apart from government consumption, all demand developments components have been weakening. Manufacturing growth fell from over 10 percent a year ago to below 1 percent in the second quarter of this year. Growth in agriculture and services has been muted as well. India GDP by expenditure components India GDP by industry Percent, y-o-y Percent, y-o-y 25 12 20 9 15 6 10 3 5 0 0 -3 2017 Q1 2017 Q2 2017 Q3 2017 Q4 2018 Q1 2018 Q2 2018 Q3 2018 Q4 2019 Q1 2019 Q2 2017 Q1 2017 Q2 2017 Q3 2017 Q4 2018 Q1 2018 Q2 2018 Q3 2018 Q4 2019 Q1 2019 Q2 GDP Agriculture Mining GDP C G I X M Manufacturing Construction Services Sources: Central Statistics Office and Haver Analytics. Sources: Central Statistics Office and Haver Analytics. Notes: C=Private consumption, I=Investment and G=Government consumption, Note: GDP growth by industry is in basic prices. X=Exports, M=Imports. GDP growth by expenditure is in market prices. Figure 16: Credit has been sluggish and the uptick in July leaves no room for complacency. Credit growth slowed down from February to June but increased again in July. Non-performing loans continue to be high, especially related to industry and more in public than in private banks. Gross NPAs to total advances ratio Non-performing assets India domestic bank credit growth Percent Percent of total assets Percent, y-o-y 16 24 14 14 20 12 13 16 10 8 12 12 6 8 4 11 4 2 0 0 10 All SCBs Public Private Agriculture Industry Services Jan-19 Mar-19 May-19 Jul-19 Mar-17 Mar-18 Mar-19 Mar-17 Mar-18 Mar-19 Sources: Reserve Bank of India and CEIC. Notes: NPAs stand for non-performing assets and SCBs stand for scheduled banks, which means banks included in the Second Schedule of the Reserve Bank of India Act, 1934. to contain the downturn, and both the RBI and the gov- for bank recapitalization. In addition to giving the RBI reg- ernment of India have already reacted. The RBI began an ulatory authority over the housing financial sector and easing cycle at the beginning of the year, and below-tar- NBFCs, the government is planning to give partial credit get inflation will allow for further easing. Despite surplus guarantees to public sector banks and included measures inter-bank liquidity, however, monetary transmission in the budget to address NBFCs funding needs. All these has been weak and bank lending rates have not reacted measures will help to contain the downturn, but also raise much. India’s government has recently launched a stimu- concerns about fiscal space (World Bank 2018b). Right be- lus package that will support the ongoing expansionary fore the announcement of the stimulus, the RBI commu- monetary policy. While monetary policy will continue to nicated that the margin left for fiscal stimulus was not act as the main countercyclical tool, the fiscal measure is too wide. The current situation illustrates how important expected to have a positive economic impact. The stim- ulus package came as a surprise and will amount to 0.7 it is to preserve fiscal space in good times to be able to manage economic downturns. In South Asia, however, 19 percent of GDP. It includes the rollback of some previous procyclical public spending and a positive expenditure corporate tax increases and some additional tax cuts. As multiplier imply that fiscal policy often amplifies boom- South Asia part of the package, more mergers of the publicly-owned and-bust cycles rather than smoothing them (Beyer and Economic Focus banks are planned and around USD 32bn will be available Milivojevic 2019). Fall 2019 Recent economic Box 2: Views from the South Asia Economic Policy Network developments The South Asia Economic Policy Network, launched by the office of the regional Chief Economist at the World Bank in 2017, represents an attempt to engage more strongly with thinkers and doers across South Asia. The objective is to be more pro- active in nurturing the exchange of ideas and to learn more systematically from colleagues and counterparts in the region. The Network currently focuses broadly on macroeconomics and includes nearly 400 researchers and practitioners from the region. Network members include researchers from seven South Asia countries, selected based on peer recognition, recent conference presentations, and research outputs. Many of them are academics at renowned universities, others are researchers in central banks and think tanks, and some are affiliated with policy-making institutions. Figure 17: Over 100 researchers and practitioners shared their views. Survey among South Asia Policy Network Number of experts 40 35 30 25 20 15 10 5 0 India Pakistan Bangladesh Sri Lanka Nepal Afghanistan Undisclosed country Source: South Asia Economic Policy Network. As for the last three editions of this report, a short opinion survey was conducted of Network members. The objective was to take the pulse of informed and influential experts about economic developments in their countries. We also used the survey to gather their views on fiscal decentralization. With 105 filled-in questionnaires from 6 countries, over 27 percent of the invited experts responded to the survey. Almost all respondents identified themselves as academics and around 80 percent as macroeconomists. Almost one third of the respondents are involved in policy making and almost two thirds in policy advising. Responses regarding the economic situation are summa- rized here. The views on decentralization are reported throughout the third chapter. Figure 18: A diverse range of experts participated in the survey. Self-assessment of respondents Distribution Academic Macroeconomist Policy advisor Policy maker 0 10 20 30 40 50 60 70 80 90 100 Yes No Source: South Asia Economic Policy Network. The expectations of Network members regarding economic developments over the next six months are summarized in a sin- gle number, using so-called diffusion indices. For any indicator, a value above 50 indicates that an increase is expected, whereas 20 a value below 50 corresponds to an expected decrease. The farther away the number is from 50, the greater the consensus among Network members that an important change is under way. South Asia Respondents anticipate different GDP growth developments in different countries. In Pakistan and India, there is a strong consen- Economic Focus sus that the growth rate will come down. In the other countries, the diffusion index signals stable growth, with a slight expectation Fall 2019 Recent of increased growth in Bangladesh. Network members expect that inflation will pick-up across all countries, but there still is an economic expectation of declining interest rates across all countries. Especially Indian economists agree that the easing cycle will continue. In developments Pakistan, on the other hand, most economists (though not by a large margin) expect an end of the tightening cycle. In India imports and exports are strongly expected to decrease, which seems pessimistic but is in line with the most recent developments and the experts’ expectation of lower growth. In Pakistan, while imports are expected to decrease, exports are expected to increase, showing some optimism regarding continued external re-balancing. In all other countries the volume of imports and exports are both expect- ed to increase. And in all countries except Pakistan, Network members strongly agree that fiscal deficits will increase over the next six months. Most Network members also agree that their currencies will be under pressure over the next six months. Finally, across the region there is a consensus that financial sector stress will rise, and the situation seems most worrying in Bangladesh. Figure 19: Views on the outlook are not very optimistic. What do you expect to happen in your country whitin the next six months? Di usion index 100 90 80 Increase 70 60 50 Decrease 40 30 20 10 0 Real GDP growth Headline in ation Interest rates Volume of imports Volume of exports Fiscal de cit Exchange rate Financial sector stress India Pakistan Bangladesh Others Sources: South Asia Economic Policy Network and staff calculations. Notes: The index is calculated as follows: Index=(P1*100) + (P2*50) + (P3*0), where P1 is the proportion of responses that report that the variable will increase, P2 is the proportion of responses that report that the variable will remain unchanged, and P3 is the proportion of responses that report that the variable will decline. There are some notable differences in expectations compared to six months ago. As forecasted by the Pakistani experts six months ago, interest rates in the country indeed increased, and the fiscal deficit worsened. Different from six months ago, they now expect lower interest rates and a lower fiscal deficit. In India, amid surprising trade data, experts have adjusted their views on future import growth and now expect a continuation of the decline. The survey offered room to express general views on the economy. In Bangladesh, economists seem very concerned about the balance of payments and the financial sector, and in both Bangladesh and India experts raised concerns about the quality of growth and its measurement. Experts in India also express apprehension regarding the global environment which may lead to further stagnation of the economy. Pakistani respondents expressed optimism that after an initial deceleration of growth, the macroeconomic adjustment policies will result in higher potential growth, partly through an improved export performance. They do however express concern regarding the risks associated with oil price variations which may in turn affect inflation and growth. In Sri Lanka, experts argued that in light of the upcoming elections, the economic situation may worsen before improving again. References Attanasio, O., Di Maro, V., Lechene, V., and Phillips, D. Carrasco, B., and Mukhopadhyay, H. (2012). Food (2013). Welfare consequences of food prices in- Price Escalation in South Asia—A Serious and creases: Evidence from rural Mexico. Journal of Growing Concern. Asian Development Bank No. 10 Development Economics, Volume 104, 136-151. February. Burns, A., Campagne, B., Jooste, C., Stephan, D., and Headey, D. D., and Martin, W. J. (2016) The Impact of Food Bui, T. (2019). The World Bank Macro-Fiscal Model Prices on Poverty and Food Security. Annual Review Technical Description. World Bank Policy Research of Resource Economics 8:1, 329-351. Working Paper (No. 8965). World Bank. (2018a). Global Economic Prospects: The Beyer, R. C. M., & Milivojevic, L. (2019). Fiscal Policy and Turning of the Tide?. The World Bank. Washington, DC. Economic Activity in South Asia. World Bank Policy 21 World Bank. (2018b). South Asia Economic Focus Fall 2018: Research Working Paper (No.8792). Budget Crunch. The World Bank. Washington, DC. Blagrave, P. (2019). Inflation Co-Movement in Emerging World Bank. (2019). South Asia Economic Focus, and Developing Asia: The Monsoon Effect. IMF Spring 2019: Exports Wanted. The World Bank. South Asia Working Papers 19/147. International Monetary Fund. Washington, DC. Economic Focus Fall 2019 SOUTH ASIA ECONOMIC OUTLOOK 22 South Asia Economic Focus Fall 2019 Overview vv Uncertainty in global markets and a worsening global outlook have become more important driv- ers of the forecast of economic growth in South Asia. vv Macro-financial risks have increased in South Asia and high-frequency data suggest a sharp downturn in the investment cycle. There is a dan- ger that slowing growth and deteriorating balance sheets of banks and corporates are reinforcing each other. vv Growth forecasts for South Asia are revised down- ward. Compared to six months ago, GDP growth is revised downward by 1.1 percentage points for this year and by 0.8 and 0.4 percentage points for the next two years. With 5.9 percent, growth in 2019 is PHOTO BY: AMIT KG/SHUTTERSTOCK now forecast to be 0.7 percentage points lower than growth in 2018. Compared to earlier forecasts, less pri- vate consumption and more government consump- 23 tion is expected. The projected modest recovery to 6.3 percent in 2020 and 6.7 percent in 2021 is tentative South Asia as forecasts under current circumstances, particularly Economic Focus for investment, are highly uncertain. Fall 2019 South Asia Figure 20: The US yield curve suggests a US recession may be ahead, reflecting a worsening outlook. economic The US yield curve inverted recently and hence models predicting the probability of a recession in the US based on the yield curve outlook predict an increasing probability, while others do not. US recession probability Percent Percent 50 120 40 100 80 30 60 20 40 10 20 0 0 Jan-90 Feb-91 Mar-92 Apr-93 May-94 Jun-95 Jul-96 Aug-97 Sep-98 Oct-99 Nov-00 Dec-01 Jan-03 Feb-04 Mar-05 Apr-06 May-07 Jun-08 Jul-09 Aug-10 Sep-11 Oct-12 Nov-13 Dec-14 Jan-16 Feb-17 Mar-18 Apr-19 US yield curve model Dynamic-factor markov-switching model (RHS) Sources: Sources: Federal Reserve Bank of New York, Federal Reserve Bank of St. Louis (FRED), and National Bureau of Economic Research (NBER). Notes: The orange line is the smoothed US recession probabilities derived from the FRED. It is obtained from a dynamic-factor Markov-switching model applied to four monthly coincident vari- ables: non-farm payroll employment, the index of industrial production, real personal income excluding transfer payments, and real manufacturing and trade sales. The blue line is the probability of a recession in 12 months derived from the US yield curve model of the Federal Reserve Bank of New York. The shaded areas represent past recessions. Last observation is August 2019. Uncertain global environment recession in major economies such as the US and the Eurozone, could have severe spillover effects to South Asia through trade, financial, and confidence channels. The global outlook is rapidly worsening, and in the current global economic environment South Asia’s growth outlook is especially unclear. Some major Global growth spillovers matter for South Asia. economies are decelerating substantially in 2019 and Despite South Asia not being integrated into world mar- are projected to slow down further in 2020 (World Bank kets as much as most other regions (World Bank 2019), 2019). Nine out of ten countries are expected to grow it is not isolated from global developments. Unexpected slower this year compared to last (IMF 2019). The trade changes in GDP growth in the US, the Eurozone and tensions between China and the US have not been re- China, for example, all have strong implications for solved and the US yield curve recently inverted, which growth in South Asia. These can be quantified with a means that the interest rates on short-term bonds be- Structural Bayesian Vector Autoregression Model as de- came higher than the interest rates on long-term bonds. scribed in Appendix A1 (Almansour et al. 2015). Real Some investors interpret such a switch as a signal that GDP shocks in the US and the Eurozone affect India im- an economic downturn is ahead. Models that predict mediately and the effect is rather persistent (Table 2). the possibility of a recession in the US based on the yield After two years, the cumulative impact from a one curve show indeed a rising probability. The probability percentage point GDP shock in the US on India is 1.4 of a recession in the next 12 months derived from the US percentage points, and from a shock of the same size in yield curve model of the Federal Reserve Bank of New the Eurozone it is 1.1 percentage points. This is 30 per- York, for example, has tripled since the beginning of last cent and 25 percent of the effect in the destination, the year and was 15 percent as of August 2019 (Figure 20). US and the Eurozone respectively. The spillovers from However, the probability is much lower in other mod- Chinese GDP shocks follow a different pattern: while els that consider other variables (Figure 20). A sharp- there is no effect on impact, it builds slowly over time er-than-expected slowdown, including a possibility of and after two years the impact is even larger than for Table 2: Real GDP shocks in the US, the Eurozone, and China transmit to India. Unexpected changes in real GDP in the US and the Eurozone have an immediate and persistent effect on economic activity in India. While Chinese shocks take some time to affect India, their effect builds up strongly.   Transmission of real GDP shock (percentage points)   US-US US-India EU-EU EU-India China-China China-India On impact 1.00 0.66 1.00 0.98 1.00 -0.09 24 End of first year 3.41 1.61 3.59 2.04 3.11 0.20 South Asia End of second year 4.47 1.36 4.38 1.12 4.90 1.74 Economic Focus Sources: Haver, IMF database, FRED, World Bank, and staff calculations. Fall 2019 Note: Spillovers based on Bayesian VAR of Almansour et al. 2015 (see Appendix A1). Figure 21: External factors have contributed substantially to the recent GDP growth slowdown in India. South Asia A decomposition of Indian growth deviations with a Bayesian VAR model unveils a large contribution of external factors. economic outlook Historical decomposition of real GDP deviations Percentage points 1.0 0.5 0.0 -0.5 -1.0 -1.5 -2.0 -2.5 -0.3 2018 Q1 2018 Q2 2018 Q3 2018 Q4 2019 Q1 2019 Q2 External Internal Sum of external and internal Sources: Haver Analytics, IMF, FRED, World Bank, and staff calculations. Note: Spillovers based on Bayesian VAR of Almansour et al. 2015 (see Appendix A1). shocks in the US or the Eurozone. This pattern is found could further deteriorate if an unexpected tightening in in other emerging markets as well and attributed to global financial conditions elevates borrowing costs to lagged terms of trade effects (Almansour et al. 2015). For unsustainable levels and leads to capital outflows from Sri Lanka, the model suggests much weaker spillovers. To South Asia. In addition, geopolitical risks remain high quantify the spillovers to Bangladesh and Pakistan, one and the recent attacks on Saudi Arabia’s oil production needs to approximate their quarterly GDP, which can have intensified concerns about the increasing tensions be done using industrial production as a high-frequen- between Saudi Arabia and Iran that could hike oil prices cy indicator (Appendix A1). A shock in the US seems to and worsen the global outlook. have a similar effect in Pakistan to the one in India, but no effect can be identified for Bangladesh. A shock in the Eurozone has a very strong but short-lived effect on Uncertain domestic environment Pakistan, and a more persistent effect in Bangladesh. A shock in China has a large initial and persistent effect High-frequency data point at a sharp decelera- on growth in Pakistan, and a small initial impact in tion in consumption and industrial production. Bangladesh that then grows over time. Taken together, Compared to a year ago, private consumption and in- these results confirm that growth developments in oth- vestment growth have been much lower so far this year er countries matter for South Asia. (Figure 22, left panel). In India, private consumption de- clined by 4.2 percentage points in the second quarter of External factors have worsened over the last year. this year, compared to a year before. And in Sri Lanka, The methodology used to asses growth spillovers can private consumption declined by 4.0 percentage points also be used to decompose growth deviations into those in the second quarter of this year, compared to a year caused by external factors and those caused by domes- before. In addition, industrial production growth has tic factors. The external factors considered are real GDP decelerated strongly in Pakistan and Bangladesh (Figure growth in the US, the Eurozone and China (to capture 22, right panel). The correlation between industrial pro- global demand shocks), US inflation (to approximate duction and investment has been high in both coun- advanced economies supply shocks), changes in the tries, with 0.56 in Pakistan and 0.75 in Bangladesh since 10-year US Treasury bond rate (to capture the stance 1992. In Pakistan, industrial production grew 6 percent- of monetary policy in advanced countries), the EMBI age points slower in the last three months compared to Global yield (to capture changes in emerging market fi- the same months a year earlier and even in Bangladesh, nancing conditions), and terms-of-trade growth (to cap- where industrial production growth remained strong, it ture factors other than those from external demand or came down by 4 percentage points recently. This indi- financing conditions). From 1998 to 2018, the external cates that steep decelerations in GDP growth and invest- variables explain 30 percent of the variation in Indian ment during this and next year are possible. GDP, while the remaining variation is explained by do- mestic factors (in the model captured by Indian GDP, in- Investment cycles are more pronounced than those flation, short-term interest rates and the exchange rate in GDP and private consumption. In the past, invest- of the Indian rupee versus the USD). While the external ment growth has been highly correlated with private factors were still beneficial for India (and presumably the rest of South Asia) at the beginning of last year, they consumption growth and investment has been much more volatile (Figure 23). High-frequency data relat- 25 turned negative at the end of it (Figure 21). For the dis- ed to consumption suggests that private consumption appointing second quarter in 2019 (calendar year), 80 may slow down in a similar magnitude as in 2008 and South Asia percent of the deviation from more normal growth 2012. The last two times private consumption dropped Economic Focus was caused by external factors. The external conditions by a similar magnitude, investment growth decelerated Fall 2019 South Asia Figure 22: High-frequency data suggest the recent investment cycle may have ended. economic High-frequency indicators show that private consumption and investment growth are strongly decelerating across South Asia. outlook Deceleration of private consumption and investment growth Deceleration of industrial production growth Percentage points Percentage points 0 1 0 -10 -1 -15 -2 -3 -20 -4 -5 -25 India Sri Lanka -6 Private consumption Investment Pakistan Bangladesh India Sri Lanka Source: World Bank. Sources: World Bank. Notes: The most recent observation is 2019Q2 and growth in these quarters is compared Notes: The growth in the last three months with available data is compared to a year ago. to growth a year before. The last observation for Bangladesh is May 2019 and for the rest it is July 2019. Figure 23: Private consumption and investment growth co-move, and the latter is more volatile. Growth in private consumption and investment are highly correlated but investment cycles are much more pronounced. Private consumption and investment growth in South Asia Percent 16 14 12 10 8 6 4 2 0 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 Private consumption Investment Source: World Bank. strongly. In 2008, private consumption slowed by two Deviations from the equilibrium are typically ex- percentage points, and investment growth by over 10 plained by movements in the real rental rate of capital percentage points. In 2012, private consumption growth and an accelerator effect. The latter is related to over- declined again by 2 percentage points and investment proportionate changes in investment growth following growth dropped by 7 percentage points in that year and changes in GDP. Depending on the model specification, by another 3 percentage points the next year, despite a we find an accelerator in India between 1.3 and 2.0, recovery of consumption growth. Thus, the most recent close to the ones found in other countries (Burns et al. data on consumption indicate the potential for a severe 2019). In addition to the standard variables, one can slowdown in investment. add domestic economic uncertainty (Baker, Bloom and Davis 2016). As expected, higher uncertainty weakens The growth of investment in South Asia is deter- investment growth in India, which may be particular- mined both by internal and external conditions. ly important at the current juncture. In addition, one One way to analyze the determinants of private invest- can add external variables to test whether the issues 26 ment growth is by means of an error-correction mod- el, which explains deviations of private investment outlined above effect private investment. And indeed, both global industrial production and world trade growth are positively related to investment in India. growth around an equilibrium growth rate determined South Asia by the growth of GDP. For India, private investment The current slowdown in these variables is hence likely Economic Focus growth does follow this long-run relationship (as con- to bring private investment growth down. Interestingly, Fall 2019 firmed by the negative sign in the first row of Table 3). once these external conditions are considered, the Table 3: Private investment growth depends on internal and external conditions. South Asia economic Change of private investment in India outlook Model 1 Model 2 Model 3 Deviations from long-run trend (error correction) - - - GDP growth in short-run (accelerator) + (.) (.) Change in cost of capital - - - Economic policy uncertainty (lagged) - - - World industrial production growth + World trade growth + Sources: World Bank; CPB; and Baker, Bloom, and Davis (2016). Notes: + and – signs mean that the terms have a positive/negative impact and are statistically significant at least at the 10 percent level. (.) means that the variable is not statistically signif- icant at the 10 percent level. See Appendix A2 for details on the model and estimation. Figure 24: Rising financial tensions may harm investment growth. Macro-financial risks in South Asia continued to increase over the last twelve months. Global EMDEs and South Asia EFI macro- nancial risk indices Percent, regional country averages 40 35 30 25 20 15 10 5 0 2006 Q1 2007 Q1 2008 Q1 2009 Q1 2010 Q1 2011 Q1 2012 Q1 2013 Q1 2014 Q1 2015 Q1 2016 Q1 2017 Q1 2018 Q1 2019 Q1 South Asia Global EMDEs Source: World Bank. Notes: EMDE=Emerging Market and Developing Economies. EFI=Equitable Growth, Finance, and Institutions Global Practice at the World Bank. The risk indices are the number of risks signaling as a percent of nine risks and are calculated as four-quarter moving averages. A risk is considered “signaling” if it exceeds a rating of 7 which corresponds to the 80th percentile. The nine risk dimensions captured are- 1. Spillover risk from the external environment, 2. Macroeconomic risks, 3. Banking risks, 4. Public sector risks, 5. Corporate sector risks, 6. Household risks, 7. Market and liquidity risks, 8. Monetary and financial conditions, and 9. Risk appetite. The index covers 44 EMDEs and 3 South Asian economies. accelerator becomes statistically insignificant. In mod- banks (SOBs) and the credibility of its Letter of Credit els for Bangladesh and Pakistan, we find the acceler- guarantees in international markets remain a worry. ator to be smaller or not statistically significant even While an adequate reserve coverage provides a buffer without adding these variables, which may signal data against contagion for now, reserves have declined re- issues in these countries. cently and are already below the levels of comparable emerging market economies. Financial market developments could interfere vv In India, non-bank financial companies in India re- with investment growth across South Asia. In con- main vulnerable to financial stress, despite liquid- trast to other EMDEs, macro-financial risks in South ity enhancing measures. New defaults in this sector Asia are rising (Figure 24). As regional growth is decel- could trigger a broader liquidity crunch. And the erating, non-performing assets are likely to rise further, sector’s significant share in total loans and its link- from already high levels. And any further deteriora- ages with the banking sector through liabilities, pose tion in balance sheets of banks and corporates would broad-based contagion risks. These are further exac- further constrain domestic and foreign investment. erbated by the over-leveraged balance sheets of the Weak financial sectors and rising macro-financial risks corporate sector. hence threaten to cut off any rebound in South Asia’s vv In Pakistan,  increased pressures on the asset quali- investment: ty and capital adequacy buffers due to the economic slowdown and inflationary environment could hold 27 vv In Bangladesh,pressures on loanable funds, con- back the forecast rebound in growth, especially when strained by low deposit growth rates and increasing strong short-term deposit mobilization (due to recent South Asia NPLs, could impact credit growth more than antici- increases in policy rates) continues to be intermediat- Economic Focus pated. The undercapitalization of its state-owned ed mostly towards government securities. Fall 2019 South Asia Figure 25: Growth forecasts for South Asia are revised downward but growth is still expected to economic remain strong. outlook South Asia’s GDP growth is revised downward compared to six months ago by 1.1 percentage points this year and by 0.8 and 0.4 percentage points for the next two years. For 2019, growth is now forecast to be lower than last year. Nevertheless, growth in 2019 could remain the highest in the world and growth dynamics could turn around again from the next year onward. Revisions to real GDP growth in South Asia Real GDP growth Percentage points Percent 0.0 7 -0.2 6 5 -0.4 4 -0.6 3 2 -0.8 1 -1.0 0 South Asia East Asia Europe and Sub Saharan Middle East Latin and Paci c Central Asia Africa and North America and -1.2 Africa Caribbean 2019 (f) 2020 (f) 2021 (f) 2018 2019 (f) 2020 (f) 2021 (f) Source: World Bank. Source: World Bank. Notes: (f)= forecast. Revisions are the percentage point change in projections from the Note: (f)= forecast. Spring 2019 edition of the report. Other domestic risks remain severe. Within the region, 6.3 percent in 2020 and to 6.7 percent in 2021 (Figure 25, recent tensions between major economies remain a con- right panel). Despite growing slower than East Asia and cern and a re-escalation could further impair confidence Pacific in the first half of this year (see Chapter 1), the and weigh on investment in the region. In addition, there expected annual growth would push the region slightly are country-specific risks. In Sri Lanka, reduced tourism above East Asia and Pacific for the whole year. receipts, following the recent attacks, will exert pressure on external accounts, despite reduced import demand. Growth of private consumption and investment has Since the fiscal balance may deteriorate amid contracting been downgraded and more government consump- revenues, large refinancing needs, weak fiscal buffers, and tion is expected (Table 4). Compared to six months high debt make the country vulnerable to rollover risks. ago, the expected contribution from private consump- Maldives needs to strike an appropriate balance between tion has been revised downward by 1.0 percentage making large investments needed to close infrastructure point this year and by 0.8 percentage points the follow- gaps, potentially boosting tourism, increasing resilience to ing year, and the expected contribution of investment climate change and easing constraints in service delivery, has been revised downward slightly by 0.1 percentage versus managing the rapid accumulation of public debt. points this year and 0.2 percentage points next year. And in Nepal risks to the outlook primarily arise from ca- Consequently, only a weak recovery is predicted for 2020. pacity constraints, especially at the subnational level (see The contribution from investment is expected to be 2.6 Chapter 3), and delays in reform implementation. percentage points this year and 2.7 percentage points over the forecast horizon. The contribution from private consumption is expected to rise slightly from 3.1 per- centage points this year to 3.2 percentage points in 2020 Growth revised downward, (Figure 26). The contribution of government consump- but rebound expected tion, on the other hand, has been revised upwards by 0.1 percentage points for this year and by 0.2 percentage Growth forecasts for South Asia are considerably re- points in the next year in expectation of countercyclical vised downward and yet the outlook remains posi- fiscal policies. While these would support growth, such tive. The recent economic developments (see Chapter 1) policies could exacerbate concerns about missing fiscal have clouded the outlook for South Asia, and growth space and the sustainability of budget deficits in South forecasts have been revised downward compared to Asia (World Bank 2018). In line with a weakening glob- six months ago (World Bank 2019). For 2019, expected al trade outlook, expected export growth in South Asia growth is now 1.1 percentage points below the expec- has been revised downward as well. However, anticipat- tation in March due to downward revisions in Bhutan, ed import growth has been revised downward more, so 28 India, Maldives, Pakistan, and Sri Lanka. The downward revision for this year carries over into the next two years, that the forecast contribution of net exports to regional growth is now expected to be less negative. with a downward revision of 0.8 percentage points in South Asia 2020 and 0.4 percentage points in 2021 (Figure 25, left Investment growth and turning points are difficult Economic Focus panel). Nevertheless, regional growth is projected to to forecast. Even over the last two years, in which re- Fall 2019 pick-up again, from an estimated 5.9 percent this year to gional investment and private consumption growth Table 4: The projected contribution of private consumption has been weakened strongly for this South Asia year and the next. economic Compared to earlier forecasts, expected contributions from private consumption, investment and exports have been weakened, outlook while higher government consumption and lower export will support growth. Private Government Gross fixed Exports Imports GDP consumption consumption investment 2018 0.2 0.0 0.0 0.0 (0.2) (0.3) 2019 (f) (1.0) 0.1 (0.1) (0.6) (1.3) (1.1) 2020 (f) (0.8) 0.2 (0.2) (1.0) (1.4) (0.8) 2021 (f) (0.1) 0.0 (0.1) (1.1) (1.4) (0.4) Source: World Bank. Notes: (f)= forecast. Revisions are the percentage point change in projections from the Spring 2019 edition of the report. Numbers in parenthesis refer to downward revisions. Figure 26: A slight rebound in investment and private consumption is predicted for next year. The contribution of private consumption is expected to soften to 3.1 percentage points and to average 3.6 percentage points over the forecast horizon. The contribution of investment is expected to decline slightly to 2.6 percentage points this year and to average 2.7 percentage points subsequently. Future growth is expected to continue to be driven by private consumption and investment. Contributions to growth in South Asia Percentage points 12 10 8 6 4 2 0 -2 -4 2018 2019 (f) 2020 (f) 2021 (f) -6 Private consumption Government consumption Gross xed investment Imports Exports Real GDP growth (percent) Source: World Bank. Note: (f)= forecast. Figure 27: Since investment is more volatile than other demand components, it is difficult to forecast. Uncertainty in global and domestic markets makes investments very volatile, and hence difficult to forecast. Average of absolute forecast errors in 2017 and 2018 Volatility of demand components from 2000 to 2018 Percentage points Standard deviation 25 25 20 20 15 15 10 10 5 5 0 0 Imports Investment Exports Government Private GDP Imports Investment Exports Government Private GDP consumption consumption consumption consumption Bangladesh Bhutan India Nepal Bangladesh Bhutan India Nepal Pakistan Sri Lanka Average Pakistan Sri Lanka Maldives Average Source: World Bank. 29 Source: World Bank. Note: The forecast errors are from the third release prior to the release of first estimates. Note: The volatility of the demand components is calculated from 2000 to 2018. were rather stable, the investment forecast errors have imports and exports, investment also tends to be the only been exceeded by the forecast errors of imports most volatile component of GDP (Figure 27, right panel). South Asia (Figure 27, left panel; for a detailed assessment of the The average standard deviation of investment growth in Economic Focus forecasts presented in this report see Box 1). Apart from South Asia from 2000 to 2018 was twice as high as that Fall 2019 South Asia Table 5: The outlook for South Asian countries is mixed. economic GDP forecasts for South Asian countries are mixed, with accelerated growth in some countries, and some deceleration in others. outlook Real GDP growth at market prices in percent Revision to forecasts from April 2019 2018 2019 (e, f) 2020 (f) 2021 (f) 2019 (e, f) 2020 (f) 2021 (f) Afghanistan (CY) 1.8 2.5 3.0 3.5 0.0 (0.2) 0.0 Bangladesh (FY) 7.9 8.1 7.2 7.3 0.8 (0.3) 0.1 Bhutan (CY) 4.6 5.0 7.4 5.9 (0.4) 2.0 0.7 India (FY) 6.8 6.0 6.9 7.2 (1.5) (0.6) (0.3) Maldives (CY) 6.7 5.2 5.5 5.6 (0.6) 0.3 0.3 Nepal (FY) 6.7 7.1 6.4 6.5 1.1 0.3 0.3 Pakistan(FY, factor prices) 5.5 3.3 2.4 3.0 0.0 (0.3) (0.9) Sri Lanka (CY) 3.2 2.7 3.3 3.7 (0.8) (0.3) 0.0 Source: World Bank. Notes: (e)= estimate; (f)= forecast. CY= calendar year, FY= fiscal year. In Bangladesh, Bhutan, Nepal, and Pakistan, 2019 refers to FY2018/19 and ended in June 2019 (mid-July 2019 for Nepal). For India, 2019 refers to FY2019/20 and will end in March 2020. Numbers in parenthesis refer to downward revisions. of private consumption growth. Forecasting investment Figure 28: Turning points are difficult to project. growth is hence always a challenge, but in the current The momentum of regional annual growth changed six times environment future investment growth is especially un- since 2006. While in three instances a turning point was certain. Since 2006, the momentum of regional annual projected, it only materialized for the strong recovery in 2010. growth changed six times (Figure 28). However, over the same period, the forecasts predicted a change of the mo- mentum in only three instances. Since forecasts tend to Turning points in regional real GDP growth since 2006 Number of experts uphold recent trends, few trend changes are projected. 7 In addition, of the three projected turning points, only one materialized and the other five were missed. The 6 only one that was correctly expected was the strong re- 5 covery in 2010, which did indeed seem very likely at the end of 2009. For this year, no downturn was predicted in 4 January, but instead projections foresaw an acceleration 3 of 0.3 percentage points. It hence remains particularly 2 uncertain whether the growth momentum will indeed swing back next year already. 1 0 For some countries in South Asia growth is project- Number of turning Forecasted turning Correctly forecasted points points turning points ed to accelerate, but for others the growth rate is expected to decline (Table 5): Source: World Bank and staff calculations. Notes: A turning point is defined as a year in which the growth momentum changes direction. vv In Afghanistan,with improved farming conditions Stable growth is considered a turning point if the momentum changes the following year. and assuming political stability after the elections, growth is expected to recover and reach 3 percent in percent on the base of strong tourism growth and in- 2020 and 3.5 percent in 2021. However, the outlook creased revenue from the existing power plants. is highly vulnerable and may be affected by deteri- vv In India, a  fter the broad-based deceleration in the orating confidence due to uncertainty around inter- first quarters of this fiscal year, growth is projected national security assistance, election related violence, to fall to 6.0 percent this fiscal year. Growth is then and peace negotiations with the Taliban. expected to gradually recover to 6.9 percent in fiscal vv In Bangladesh,  GDP is projected to moderate to 7.2 year 2020 and to 7.2 percent in the following year. percent this fiscal year and 7.3 percent the following vv In Maldives,  growth is expected to reach 5.2 percent one. The outlook is clouded by rising financial sector in 2019, due to a slowdown in construction following vulnerability, but the economy is likely to maintain the completion of the international airport and a growth above 7 percent, supported by a robust mac- connecting bridge. However, with support from new roeconomic framework, political stability, and strong infrastructure investment and the expansion of tour- 30 public investments. vv In Bhutan,GDP growth is expected to jump to 7.4 ism, growth is expected to pick up again to an average of 5.6 percent over the forecast horizon. percent this fiscal year with the commissioning vv In Nepal, G DP growth is projected to average 6.5 per- South Asia of Mangdechhu, a new hydro power plant, and the cent over this and next fiscal year, backed by higher Economic Focus completion of the maintenance of Tala, another one. investment and public consumption and strong ser- Fall 2019 Growth in fiscal year 2021 is forecast just below 6 vices due to rising tourist arrivals. South Asia Box 3: South Asia Economic Focus forecasting performance economic outlook The forecasts reported in this publication are generated by country economists within the World Bank’s Macroeconomics, Trade and Investment Global Practice and are based on the World Bank’s macroeconomic and fiscal model (MFMod). Although the functional forms of the 181 individual country models are similar, the parameters are estimated at the country level. The forecasts across countries are linked and the export market growth of each country is calculated as a trade-weighted average of imports of each of its trading partners. Other cross-country linkages come through balancing remittances flows, the real effective exchange rate, and export and import prices, which are a function of world commodity prices and local cost considerations (Burns et al. 2019). To assess our forecasting performance, we computed the forecast error for real GDP growth and its components for 2017 and 2018. Given the different fiscal years in South Asia, we compare countries relative to the release of first estimates after the end of the (fiscal) year and not based on the chronological order of forecast releases. The growth reported in the latest available release is considered to be the actual growth and may refer to preliminary estimates (first release), revised estimates (second release), or actual data (third release). The forecast error is defined as actual growth minus forecasted growth, which makes uses of the most up-to-date data but of course assumes that the latest reported data is true. Due to measurement issues reflected in large data revisions in South Asia, forecast errors may hence change with new data releases. Figure 29: Forecast errors in South Asia depend on country characteristics. Forecasts for larger countries with lower volatility of growth rates have been more precise. Real GDP growth volatility and forecast error Standard deviation 10 8 MDV 6 AFG BTN 4 IND 2 PAK LKA BGD NPL 0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 Average real GDP absolute forecast error in 2017 and 2018 Sources: World Bank and staff calculations. Notes: AFG=Afghanistan, BGD=Bangladesh, BTN=Bhutan, IND=India, MDV=Maldives, NPL=Nepal, PAK=Pakistan, SLK=Sri Lanka. The size of the bubbles refers to the size of the economy measured as 2018 real GDP in 2010 USD. The forecast errors of the five releases prior to the release of first estimates are averaged. The data in the latest available report is considered the actual. The forecasts have been more precise for larger countries, whose growth is less volatile (Figure 29). The forecasts were most precise for India, Pakistan, and Bangladesh. In these countries, real GDP growth tends to be less volatile than in smaller countries, and for all these economies the state of the economy can be assessed based on numerous meaningful high-frequency indicators like industrial production and import growth. For some of the smaller countries – Sri Lanka, Bhutan, and Nepal – the average forecast error has been larger. Growth is especially volatile for Afghanistan and Maldives and hence particularly difficult to forecast. While in 2017 and 2018 the forecasts for Afghanistan were close to the actual, the forecast error for Maldives was by far the largest. Figure 30: GDP forecasts in South Asia were not biased… Forecast errors for South Asian countries have been normally distributed around 0. Distribution of GDP growth forecast errors Kernel density 0.25 0.15 0.10 0.05 0.00 -8 -6 -4 -2 0 2 4 6 8 31 Forecast errors Normal distribution (sigma=1.75) South Asia Asia absolute SouthWorld Source: forecast Bank and staff error calculations. Economic Focus Percentage Notes: points shows five forecast errors (related to different releases) from all South Asian countries for 2017 and 2018. The density 2.5 Fall 2019 2.0 0.05 South Asia 0.00 economic -8 31: … and Figure -6 became better -4 as the -2 end of the 0 fiscal year approached. 2 4 6 8 outlook On average, forecasts improved as the fiscal years progressed. Forecast errors Normal distribution (sigma=1.75) South Asia absolute forecast error Percentage points 2.5 2.0 1.5 1.0 0.5 0.0 -5 -4 -3 -2 -1 Periods prior to release of rst estimates 2017 2018 Source: World Bank and staff calculations. Notes: The forecast errors are simple averages of all South Asian countries. The numbers refer to the release before the release of first estimates (-1), the second release before (-2) and so on. The forecast errors have been normally distributed around 0 (Figure 30) and improved as the fiscal years progressed (Figure 31). That fact that forecast errors have a mean close to zero implies that there was no regional bias in the forecasts: real GDP growth has neither been systematically overpredicted nor systematically underpredicted. In addition, too low and too high forecasts have been equally distributed around 0. After the fiscal years had started and the first high-frequency data became available, the aver- age forecast error declined. Declines have been especially strong in Pakistan, where the average error over the two years declined from 0.7 percentage points a year before the last forecast to only 0.1 percentage points before the release of first estimates, and in Sri Lanka, where it declined from an average of 2.3 percent to 1 percentage point. Figure 32: Import forecast errors in South Asia For the last two years, errors in import forecasts have been large, but they are mostly explained by getting other components wrong. South Asia absolute forecast error Percentage points 12 10 8 6 4 2 0 -2 -4 2017 2018 2017 2018 2017 2018 2017 2018 2017 2018 2017 2018 Bangladesh Bhutan India Sri Lanka Nepal Pakistan Explained by forecast errors in components Unexplained Forecast error Source: World Bank and staff calculations. Import growth forecast errors are large but often related to forecast errors in other GDP components. For the last two years, average import forecasts have been too low in India, Nepal, and Pakistan, but far too high for Bhutan (Figure 32). Import forecasts can be too low if the other GDP components have forecast errors, or if the relationship with other demand components like invest- ment and consumption has played out differently than expected, for example due to changing terms of trade. In many countries in South Asia, the relationship between import growth and the growth of other components is rather stable due to stable import intensities of the different components. In India and Pakistan, for example, a model predicting import growth based on the growth 32 of the other GDP demand components from 2000 to 2018 can explain 88 percent of the variation in import growth. With such a model one can decompose the forecast error into contributions explained by forecast errors in the other GDP components and South Asia unexplained parts. In India, nearly the full forecast error is explained by forecast errors in components and in Pakistan and Bhutan, Economic Focus they are also the major contributor. Fall 2019 South Asia The forecast errors of imports are often intuitive. In Pakistan Figure 33: Forecast error of oil prices in 2017 economic and Nepal, an underprediction of imports coincided with an and 2018 outlook underprediction of growth in 2017. In both countries, invest- For 2018, oil prices were underestimated, which ment growth rates were stronger than expected, in the former contributed to large import forecast errors. related to the China-Pakistan Economic Corridor and in the Overestimation and underestimation of oil prices latter to the rebound after the end of the trade disruption. Percent It was the other way around in Bhutan due to delays in the 5 construction of hydropower projects. In India, on the other hand, imports were underpredicted in the same year, while 0 growth was overpredicted, and the unexpectedly high import growth was partly related to the demonetization in November -5 the year before. In 2018, Nepal imported more than expected partly due to the construction of subnational government of- -10 fices reflected in higher than expected investment. Overall, the link between forecast errors in other components and import growth was somewhat less strong in 2018 compared to 2017, -15 as the share of errors explained by other components went down from 55 percent to 45 percent. One reason may have -20 2017 2018 been higher than expected oil price, which resulted in higher Source: World Bank and staff calculations. costs of the oil imports and potentially affected the import in- Note: The forecast errors are those of the third release prior to the release of first tensity of the different demand components (Figure 33). estimates. Box 4: Growth expectations from within the region Figure 34: Growth forecasts are broadly in line with expectations of regional experts. South Asian experts expect higher growth than forecasted by the World Bank in Pakistan, Sri Lanka, Bangladesh, and Nepal, -0.6 lower growth than forecasted for India. and Di erence between expected growth by experts and World Bank forecasts Percentage points 1.2 1.0 0.8 0.6 0.4 0.2 0.0 -0.2 -0.4 Pakistan Sri Lanka Bangladesh Nepal India Sources: South Asia Economic Policy Network, World Bank, and staff calculations. Notes: The difference between the mean expected growth and World Bank Forecasts is calculated. The mean is calculated aggregating the responses from the South Asia Economic Network Survey for each country. There were 5 responses from Nepal, 7 from Sri Lanka, 10 from Bangladesh, 15 from Pakistan and 32 from India. For Bangladesh, India, Nepal, and Pakistan expected and forecasted growth is for the current fiscal year, whereas for Sri Lanka it is for the next fiscal year. Compared to expectations of regional exports, growth forecasts are somewhat lower in most countries but higher in India. For the first time, respondents to the survey conducted for this report (see Box 2 in Chapter 1) were asked what they expect the growth rate in their countries for the current fiscal year to be (next fiscal year for Afghanistan, Maldives and Sri Lanka). Results are only reported for countries with at least 5 responses (Figure 34). In Pakistan, Sri Lanka, Bangladesh, and Nepal, respondents predict that growth will exceed the World Bank forecast. In Pakistan, regional experts project 1.0 percentage point higher growth and in Sri Lanka and Bangladesh the mean forecast is 0.7 percentage points above the World Bank forecasts. In Nepal, the mean forecast is 0.4 percentage points higher. India is the only country for which experts expect lower growth. The average expectation is 0.3 percentage points lower. Indian economists were further asked whether they consider the recent slowdown a structural or a cyclical phenomenon. Over 65 percent said that both structural and cyclical factors are causing the slowdown, 25 percent see only structural factors at play and only around 10 percent consider the slowdown purely cyclical. This contrasts slightly with our explanation of the downturn that emphasizes cyclical patterns (see Chapter 1). Looking at the distribution of the experts’ growth expectations confirms that these are broadly aligned with the forecasts pre- sented in this report. To do so, one can compare the difference of the World Bank’s forecast not to the mean of the responses, but to 33 the 25th percentile, if it is lower, end the 75th percentile, if it is higher. In Sri Lanka, Nepal, and India, the World Bank forecasts fall with- South Asia in this range. Only in Pakistan and Bangladesh, less than a quarter of the experts expect a lower growth than the World Bank forecasts. Economic Focus Fall 2019 South Asia vv In Pakistan,  growth is projected to deteriorate fur- economic ther to 2.4 percent this fiscal year, as monetary policy outlook remains tight and the planned fiscal consolidation will compress domestic demand. The program signed with the IMF is expected to help growth recover from fiscal year 2021-22 onwards. vv In Sri Lanka, growth softened to 2.7 percent in 2019. However, supported by recovering investment and ex- ports, as the security challenges and political uncer- tainty dissipate, it is projected to reach 3.3 percent in 2020 and 3.7 percent in 2021. More details on each of the country forecasts are provid- ed in Chapter 4 featuring country briefs. Box 4 compares the World Bank forecasts to the growth expectations of regional experts. References Almansour, A., Aslam, A., Bluedorn, J., and Duttagupta, R. (2015). How vulnerable are emerging markets to external shocks? Journal of Policy Modeling,  37(3), 460-483. Baker, S. R., Bloom, N., & Davis, S. J. (2016). Measuring eco- nomic policy uncertainty. The quarterly journal of economics, 131(4), 1593-1636. Burns, A., Campagne, B., Jooste, C., Stephan, D., and Bui, T. (2019). The World Bank Macro-Fiscal Model Technical Description (No. 8965). World Bank Policy Research Working Paper. IMF (2014). World Economic Outlook: Recovery strength- ens, remains uneven. International Monetary Fund. IMF (2019). World Economic Outlook: Still Sluggish Global Growth. International Monetary Fund. World Bank. (2018). South Asia Economic Focus, Fall 2018: Budget Crunch. World Bank, Washington, DC. World Bank. (2019). South Asia Economic Focus, Spring 2019: Exports Wanted. World Bank, Washington, DC. 34 South Asia Economic Focus Fall 2019 Appendix South Asia economic outlook Appendix A1: A Bayesian VAR analysis quantifying the effect of external factors on Indian GDP The analysis is an update of the analysis by Almansour et al. (2015) and presented first in IMF (2014). It uses a stan- dard structural vector autoregression (SVAR) model of the following form: in which ys is a k × 1 vector, where k is the total number of endogenous variables, A(L) is a k × k matrix polynomial of lag operator L with lag length p, and εt is a k × 1 vector of contemporaneously correlated, mean-zero reduced-form errors. The contemporaneous relationships across variables are disentangled by mapping εt to a k × 1 vector of mu- tually orthogonal, mean-zero, structural shocks, ut, through the k × k matrix A0. The vector yt includes the following variables: US, Eurozone, and China real GDP growth, US inflation, the nominal 10-year U.S. government bond rate, the J.P. Morgan Emerging Markets Bond Index EMBI global yield, India’s terms-of-trade growth, Indian real GDP growth, Indian inflation, the rate of real appreciation of the Indian rupee vis-à-vis the USD, and the Indian short-term interest rate. The first seven variables constitute the external block, and the remaining variables the domestic block. Structural shocks are identified based on (i) the restriction that shocks to the external block are exogenous to shocks to the internal block contemporaneously, (ii) a recursive (Cholesky) scheme within the external block (in the order listed above). The lag structure and Bayesian estimation strategy are exactly as in Almansour et al. (2015). We updated the original dataset used by Almansour et al. (2015) and updated it with data from Haver, IMF, FRED, and the World Bank. We estimated different specifications in which we replaced GDP with industrial production and India with Pakistan and Bangladesh. For Pakistan and Bangladesh, we approximated quarterly GDP growth using industrial production with the Denton method. We are grateful to Rupa Duttagupta and Aqib Aslam (both with the IMF) for sharing their estimation code. Appendix A2: Estimating an investment equation for India We follow Burns et al. (2019) and use an error-correction form to model the non-stationary but cointegrated relation- ship between investment and GDP in India. Our model captures the long-run cointegrated relationship and short-run fluctuations in the following form: This equation relates the growth rate of private investment, , to GDP, , and rental cost of capital, . We augment the model to account for the lagged economic policy uncertainty, , and growth of world industri- al production and trade, World IP and World trade. This way we extend the number of factors that can affect investment in the short-run. They are particularly relevant for the economic context of India today and turn out to be statistically significant at least at the 10 percent level. The term captures the long-run cointegrated rela- tionship; the model ensures that the long-term relationship between investment and economic growth holds, i.e., the cointegrating relationship is always verified in the long-term. This cointegrated relationship is derived from the long-run equation , where the elasticity is equal to one and the expected value of is zero. Finally, the parameter measures the speed of adjustment in the absence of additional shocks. We estimate similar models also for Bangladesh and Pakistan. The data for investment, GDP, the rental rate of capital, and the labor share are from the World Bank. The Economic Policy Uncertainty Index comes from Baker et al. (2016), and the world industrial production and world trade growth rates come from the Netherlands’ CBP economic outlook. The model is estimated with annual data from 2003 to 2018, restricted by the availability of the Economic Policy Uncertainty Index. 35 South Asia Economic Focus Fall 2019 MAKING (DE)CENTRALIZATION WORK 36 South Asia Economic Focus Fall 2019 Overview vv In many countries in South Asia, further decen- tralization is a high policy priority. In Bhutan, the new five-year plan includes a prominent de- centralization pillar. In India, the share of spending by states in public spending has increased and the 15th Financing Commission is working on new rec- ommendations to recalibrate competitive federalism and improve local self-governance. In Maldives, the new government promised to empower local govern- ment councils again. In Nepal, the transition into a federal republic and the operationalization of provin- cial and local governments is in full swing. And in Pakistan, the National Financing Commission is cur- rently discussing the revenue sharing agreement be- tween provinces and states, and some provinces are enacting strong local government reforms. vv These policies are part of a global decentraliza- tion trend, which aims to improve local service delivery. As economies are becoming more complex, central governments find it increasingly difficult to manage service delivery. Locally elected officials, di- rectly accountable to citizens, are better placed to supply services according to local preferences. The assumption is also that local service delivery can re- duce transaction costs. vv Empirical evidence of the effectiveness of decen- tralization is mixed, a result which is often at- tributed to partial decentralization. Public service delivery in South Asia remains disappointing. Nearly half of the people still lack access to basic sanitation services and the region accounts for most of the world’s uneducated people. The explanation of the lack of results is often that decentralization has not gone far enough. Key decisions are still taken by high- er levels of government. The allocation of resources to local governments is unpredictable. Few opportuni- ties exist for local governments to raise own revenues. In this environment of partial decentralization local governments lack institutional capacity. Although these are all critical constraints, they are not a suffi- cient explanation for South Asia’s disappointing per- formance in service delivery. vv Successful development requires both decen- tralization and centralization at the same time. A core task of an effective central government is to create integrated markets in which local communi- ties compete and to facilitate mobility across local boundaries. Central governments can also address equity concerns and support disadvantaged regions PHOTO BY: SUYASH.DWIVEDI/ WIKIMEDIA (CC BY-SA 4.0) to give all citizens equal opportunities, irrespective of where they are born. Finally, central governments can set standards for education, health care, the en- vironment, and other services. Without an effective central government decentralization can degenerate into fragmentation. With fragmentation, local econo- mies are not competitive and opportunities of elite capture by local officials emerge. The middle level of 37 government (states or provinces) often has a strong own identity. Preserving that historical identity is im- South Asia portant, but it should not prevent empowerment of Economic Focus both the local and central level. Fall 2019 Making Figure 35: Many South Asian countries are further decentralizing. (de)centralization Subnational government expenditure work Percent of total expenditure 60 50 40 30 20 10 0 11th plan 12th plan 2013-14 2018-19 2016-17 2018-19 2009-10 2017-18 Bhutan India Nepal Pakistan Sources: Reserve Bank of India (Handbook of Statistics on the Indian Economy Tables 98 and 104); Ministry of Finance for Pakistan, Nepal, and Bhutan. Notes: For Bhutan, the 12th plan refers to planned expenditure for 2018-2023 and the total expenditure excludes the flagship programs. For Nepal, this is budgeted expenditure for FY2018/19 is defined as the sum of provincial plus local level expenditure (found in the budget speech annex 3 and 4) and for FY2016/17 is defined as the sum of grants to local bodies and local allowances. For India the expenditure is state level and for Pakistan it is provincial. vv In the interplay between central and local gov- policy uniformity, and due to an emphasis on national ernments, the allocation of resources plays a interests and security (Singh 2018). Similarly, in Pakistan crucial role. This allocation should balance equity there was initially little emphasis on local governments, concerns with an efficient allocation of resources and partly because these were tied strongly to the British be- the right incentives to improve the efficiency of ser- fore independence (Cheema, Khwaja, and Khan 2005). vice delivery at all local levels. Preliminary empirical results presented in this chapter suggest that South Most countries in South Asia actively pursue fiscal Asia is more successful in addressing equity concerns decentralization. In recent years, there is once again than in achieving efficient solutions. a push for further decentralization. This is reflected in an increased share of subnational spending in overall vv A lack of geospatial data on expenditure and de- government expenditure (Figure 35). In some cases, this velopment outcomes remains a major constraint. is driven by political reasons. In Nepal, for example, the More and better data is needed for the design of policies new Constitution ended a civil war and the federal re- that can create an environment in which autonomous structuring was fundamental to sustained peace. In oth- local communities can compete in integrated markets er cases, the main objective is to bring service delivery with equal opportunities. Evidence-based policy designs closer to the citizens. require also understanding of economic mobility: what determines where people want to live, where people want to work, and where firms want to invest? vv In Bhutan, the Local Government Act enacted in 2009 implemented a program of decentralization and devolution of power and authority. The Act tasks all local governments with specific objectives, South Asia is further including promoting Gross National Happiness, pro- decentralizing viding democratic and accountable government, pre- serving culture and tradition, promoting development, Political and fiscal decentralization has a long and protecting public health. Local governments are history in South Asia. The Panchayat system of local administrative divisions prohibited from making self-governance dates to at least 250 CE, but plenty of laws, but they are empowered to make rules and reg- evidence suggests that self-governing villages have exist- ulations. The new five-year plan introduced this year ed much longer. And the debate about decentralization includes a prominent decentralization pillar, and lo- goes back as long. In the Mauryan period, for example, cal government expenditure is expected to reach one local officials became more powerful under the reign of half of total expenditure in the coming years (Figure King Ashoka. But it was during the subsequent Gupta pe- 35). The government published the final draft of the riod that the government’s centralized power gradually National Decentralization Policy in July of this year. eroded while provincial governments gained power and vv In India, the 14th Financing Commission recom- autonomous governments in several cities of northern mendations resulted in a larger fund allocation India emerged (Chakrabarti 1996). During the British to the 29 states and gave them more spending rule, there was little interference with the existing autonomy. The share of state spending increased 38 Panchayati systems, but British Colonial rule replaced other institutional arrangements with more centralized from 51 percent in 2013-14 to 58 percent in 2018-19 (Figure 35). The main responsibilities of the state gov- ones. The British also created local governments, but ernments include public order, police, and adminis- South Asia these were fully controlled by the imperial bureaucracy tration of justice, public health, agriculture, local gov- Economic Focus (Cheema, Khwaja, and Qadir 2003). After independence, ernment, as well as jointly deciding with the Centre Fall 2019 India’s federation was considerably centralized to foster on areas such as education and the environment. The 73rd and 74th Constitutional Amendment, passed in estimated to have spent over 20 percent of total bud- Making 1992, aimed to replicate the center and state mod- geted expenditure for fiscal year 2018-19 (Figure 35). (de)centralization el for state-local relations and gave recognition and vv The 18th Amendment to Pakistan’s Constitution, work protection to local governments as institutions of passed by Parliament in 2010, provided the four self-governance. However, devolution remains partial, provinces with strong legislative and financial au- even in the front runners among the states. In many tonomy. The provincial share in the divisible revenue cases, political devolution of powers to local govern- pool has increased significantly under the 7th National ment has not been accompanied by the requisite Finance Commission award in FY10/11 (Figure 35). Local devolution of fiscal powers. In addition, some respon- Governments saw their powers diminish once the Local sibilities like town planning that could be devolved to Government Ordinance 2001 was rolled back in 2010. the over 250,000 local bodies remain at the state level. This coincided with the 18th Constitutional Amendment, vv In the Maldives, as envisioned in the 2008 consti- resulting in a re-concentration of administrative and fi- tution, the Decentralization Act of 2010 expanded nancial powers at the provincial level. Previously, local local governance structures and introduced atolls, governments were responsible for education and health islands, and cities as administrative divisions each service delivery, but now their powers are mainly concen- run by a local council. The democratically elected trated on municipal services. Their financing has been councils were supposed to provide basic services such reduced as the Provincial Governments control most as road maintenance, preschool and vocational educa- key service delivery aspects. However, different provinc- tion, social services and pest control. After regress in es have different approaches to local governments with the recent past, decentralization is now back on the some devolving services more than others. agenda and the new government promised to empow- er local government councils again. vv In Nepal, the new Constitution, which came into effect in 2015, restructured Nepal into a feder- One size doesn’t fit all al republic and introduced provincial and local governments. Some of the responsibilities of pro- As countries develop and become more complex, a vincial and local governments include town police, larger share of government spending becomes sub- basic health and sanitation, local taxes (e.g. wealth national. The further decentralization that is unfolding tax, house rent tax), basic secondary education, local in South Asia is not unique to that region. Many coun- roads, rural roads, irrigation, collection of statistics tries have decentralized during the last three decades and records, and water supply. Local governments also (Bardhan 2002). Data on the share of subnational spend- formulate their own budgets, specify tax rates and ing in overall public spending suggest that the more com- collect revenue. However, the constitution states that plex a country becomes, the more difficult it is to manage the imposition of taxes and collection of revenue on it from a central government. The share of subnational matters that fall within the jurisdiction of the federal government spending is positively correlated both with government, and those that have not explicitly been the size of the population and with GDP per capita, two given to lower tiers, shall be as determined by the proxies for the complexity of countries and economies. federal government. Revenues received by the federal In cross-sectional analyses, both are statistically signifi- government from excise taxes and value added taxes cant factors explaining a larger subnational spending (VAT) on domestic goods are deposited in a divisible share and the strength of these effects is very similar. pool and then shared among the federal, provincial The relationship can hence be shown relative to real GDP, and local governments. While spending by subnation- which combines information on the population size and al entities was almost nonexistent before the new GDP per capita (Figure 36). In some countries, complex- Constitution, provinces and local governments are ity is added by strong cultural and ethnic diversity. As Figure 36: The more complex a country, the more decentralized its spending. SNG expenditure as a share of total expenditure Percent 100 CHN 80 60 IND 40 BTN PAK NPL 20 LKA BGD 39 0 9 9.5 10 10.5 11 11.5 12 12.5 13 13.5 Log of real GDP (USD 2010) Sources: Reserve Bank of India; Ministry of Finance for Bangladesh, Pakistan, and Bhutan; Nepal Ministry of Finance (Budget speech 2018/19); and SNG-WOFI (World Observatory of Subnational Finance and Investment 2019) for all other countries. Population and per capita GDP are from the WDI. South Asia Notes: AFG=Afghanistan, BGD=Bangladesh, BTN=Bhutan, IND=India, MDV=Maldives, NPL=Nepal, PAK=Pakistan, SLK=Sri Lanka. Data are for latest available year for each country. The Economic Focus vertical axis measures subnational government (SNG) expenditure as a percentage of total expenditure. In Nepal, the SNG expenditure as a share of total expenditure was calculated as the sum of provincial and local spending over total expenditure as budgeted for FY2018/19. Fall 2019 Making Figure 37: The region experiences urbanization and spatial transformation. (de)centralization Growth in share of population in urban areas from 2001 to 2018 Share of urban area in Indian districts work Percent Percent 16 100 14 80 12 10 60 2011 8 40 6 4 20 2 0 0 0 20 40 60 80 100 Bhutan Bangladesh Nepal Maldives India Pakistan Sri Lanka Afghanistan 2001 Sources: UN population projections and staff calculations. Source: Spatial Database for South Asia. countries and economies become more diverse, the need and there is a vast theoretical literature discussing which to adjust policies and spending to local conditions grows. services should be decentralized under which circum- Conditional on their size, spending in most South Asian stances (Hart, Shleifer, Vishny 1997; Aghion and Tirole countries is roughly as decentralized as in other coun- 1997). In general, decentralization brings service delivery tries. The two exceptions are Bangladesh, where the share closer to citizens, which allows for adjusting policies and of subnational spending is lower, and Bhutan, where it is public spending to different needs and preferences. In ad- higher. However, Bhutan’s share of subnational spending dition, the expected strengthening of accountability and overstates the level of decentralization of expenditure transparency, especially of local governments that are since it only considers budgeted expenditure and thus democratically elected, is likely to result in greater spend- excludes the significant central spending on hydro in- ing efficiency. Greater local spending could also reduce vestments. A notable outlier in the world is China, where transaction costs. Ahmad and Brosio (2006 and 2009) pro- 90 percent of the spending is subnational. It is often un- vide rich overviews of the impact of decentralization on derappreciated that along with the opening of the econ- service delivery. The World Bank has also embraced de- omy and market reforms, China conducted a large-scale centralization as a major governance reform to improve political decentralization and empowered sub-provincial service delivery (World Bank 2000). In a survey of econo- governments, which contributed to its strong economic mists in South Asia conducted for this report (see Box 2 in performance (Lin and Liu 2000; Singh 2009). chapter 1), three quarters expect more decentralization to improve service delivery in their countries. The degree of decentralization is often the outcome of deep historical, political and cultural reasons and In addition, urbanization in South Asia has increased complex policy choices that involve many objectives. the need for spatially differentiated policies and In India, for example, many states feature a strong cultural spending. With manufacturing production and services and linguistic homogeneity, as well as a sense of belong- becoming more important relative to agriculture, cities ing to one distinct polity. Policy objectives determining the are becoming drivers of growth. Since 2000, the share of degree of centralization can include improved service de- population living in urban areas increased by more than livery, maintaining peace, local inclusion and engagement ten percent in Bhutan, Bangladesh, Nepal, and Maldives in decision making, addressing preference heterogeneity, and by over 5 percent in India and Pakistan (Figure 37, and ensuring national unity. In addition, decentralization left panel). This is an effect both of people migrating to takes very different forms (Rondinelli 1990; Schneider cities and of rural areas becoming more urban (Figure 37, 2003). Administrative decentralization in its weakest form, right panel). The 2011 Census in India, for example, re- referred to as deconcentration, only shifts responsibilities corded the emergence of 2500 additional cities since 2001. from officials at the center to those based in subnation- al entities, without empowering them. Delegation goes a step further and transfers decision-making power, in ad- dition to administrative functions. Devolution transfers Partial decentralization authority over specific geographical territories, including fiscal responsibilities. Political decentralization usually Empirical evidence of the effectiveness of decen- goes hand in hand with devolution of responsibilities and tralization is mixed, and the limited benefits are often features elected local government. In federal coun- often attributed to incomplete decentralization. tries, decentralization can either favor the second-tier of This holds both for decentralization efforts in the world 40 government (states or provinces) or local governments. (Prud’homme 1995; Bardhan 2002; Martinez-Vazquez, Lago-Peñas, and Sacchi 2017) as well as in South Asia One of the main objectives of decentralization is (see Box 5). Despite decentralization efforts in South South Asia improved service delivery. For the delivery of some Asia, public service delivery remains disappointing. Economic Focus services (for example education and health) a more de- Nearly half of the people still lack access to basic sani- Fall 2019 centralized provision is more appropriate than for others, tation services, low secondary school completion rates Making Box 5: The impact of decentralization on growth and service delivery in South Asia – a review (de)centralization of the literature work The empirical evidence regarding the benefits of decentralization is mixed. Martinez-Vazquez, Lagos-Peñas, and Sacchi (2017) provide a comprehensive and balanced overview of the academic literature on the impacts of decentralization in the world. Evidence on the impact of decentralization on service delivery in South Asia is limited, partly because major decentralization ef- forts are more recent than in other regions and because service delivery has often only been devolved to a limited extent (Robinson 2007). In addition, a lack of good data hampers such analyses. Cross-country analyses including some South Asian countries find benefits from decentralization only if certain conditions are met. Sow and Razafimahefa (2018) show in a study covering 64 advanced, emerging and developing countries (including Bhutan, India, Maldives, and Pakistan) that fiscal decentralization can improve the efficiency of public service delivery but does so only under specific conditions. These conditions include autonomy of local governments, strong accountability, good governance, and strong capacity at the local level. In addition, a positive effect requires a sufficient degree of expenditure and revenue decen- tralization. If these conditions are not fulfilled, the efficiency of public service delivery can go down. In line with this argument, Iqbal and Ahmed (2015) using a sample of 46 developing and transitional economies (including Bangladesh, India, Nepal, Pakistan and Sri Lanka) for the period 1974-2004 find that political decentralization significantly reduces the number of total deaths due to natural disasters only if it is accompanied by fiscal decentralization. Studies for South Asia provide conflicting reports on the success of decentralization efforts regarding improved service de- livery and growth: •  Mahal et al. (2000) use survey data from 33,000 households in villages across India to show that the decentralization of public service delivery in primary healthcare and education services is positively correlated with improved child mortality and school enrolment. Similarly, Asfaw et al. (2007) find a significant role of fiscal decentralization in reducing rural infant mortality rates in India between 1990-1997. Crook and Sverrisson (2003) argue that decentralization of expenditures for basic services has taken place on a significant scale only in the states of West Bengal and Kerala, where substantial untied funds at the discretion of local village councils for developmental purposes have been introduced. In West Bengal, they find evidence for improved access to administrative and justice systems and water provision in some areas. On the other hand, Kalirajan and Otsuka (2012) make the case that decentralization to rural local bodies has been dismal and not achieved any significant results across Indian states. Similarly, Raghunandan et al. (2016) argue that local government functions are hampered in Karnataka, due to unfunded mandates which account for 25 of the 29 functions devolved to these local government units. Along the same lines, Kumar and Managi (2009) study the mechanisms to compensate local governments for the public provision of environmental services in India and confirm that simply assigning functions at appropriate levels does not ensure optimal provision of environmental services, but it needs to be backed with appropriate compensation as well. Aslam and Yilmaz (2011), using data from over 180 villages in Pakistan, show that the decentralization reforms implemented in 2001 increased the provision of street paving, water canals, sanitation sewer lines, and school facilities. Panta (2016) stresses obstacles to decentralized service delivery in Nepal, which are mainly related to low revenue raising power, unclear assignments, as well as poor accountability and transparency. •  Malik et al. (2006) find that fiscal decentralization has had a positive impact on economic growth in Pakistan but argue that in earlier stages of development decentralization may have a negative impact if the central government is in a better position to ensure fiscal sustainability. Khattak, Ahmad and Khan (2010) find a positive impact of fiscal decentralization on growth in Pakistan only in the short run, while the accumulated effect overtime is found to be negative. They attribute the negative long- run effect to poor provincial capacity to efficiently generate own resources and provincial dependence on federal transfers. Lamichhane (2016) finds a positive contribution of local government expenditures on economic growth in Nepal. Herath (2009) shows that in Sri Lanka the actual degree of decentralization after the establishment of provincial councils has been low and has not had any impact on per capita incomes. For a summary of new research on subnational public finance and local service delivery presented at the 4th South Asia Economic Network Conference see Appendix A7. have resulted in a large and growing stock of uneducated governments is unpredictable and heavily earmarked, adults, water is an acute problem for many in rural areas, limiting flexibility in resource utilization at the local and stunting is higher than in most parts of the world level. Few opportunities exist for local governments to (World Bank 2018). These challenges underline the im- raise own revenues. For decentralization to improve ser- portance of policy reforms to improve service delivery. vice delivery, all three constituent components – fiscal decentralization, administrative decentralization, and 41 The explanation of the lack of results from decen- political decentralization – need to go hand in hand tralization is often that decentralization has not and downward accountability has to be strong. In an South Asia gone far enough. Key decisions are still taken by higher environment of partial decentralization, however, local Economic Focus levels of government. The allocation of resources to local governments lack institutional capacity. Fall 2019 Making Figure 38: In Nepal, a difficult transition to federalism and weak local government capacity (de)centralization exacerbates already low budget execution work Estimated budget execution rates in Nepal Unused capital expenditure of selected local governments Percent Percent 140 Pokhara-Lekhnath 120 Mechinagar Urlabari 100 Triyuga 80 Shuklagandaki Dhankuta 60 Itahari 40 Birtamod Baglung 20 Byas 0 Sundar Haraicha Province 1 Province 2 Province 3 Province 4 Province5 Province 6 Province 7 Province 8 Rajbiraj Recurrent expenditure Capital expenditure Total expenditure General government (2016-18) 0 20 40 60 Source: Budget Speeches of Provincial Ministry of Economic Affairs and Planning. Source: Strategic Assessment of Municipalities in Nepal (SAMN). Note: The provincial data is for fiscal year 2018/19. Decentralization in South Asia indeed remains environmental protection, water supply, street lighting work in progress. In Bhutan, the implementation of and land use planning. However, many of these responsi- the decentralization pillar included in the 12th five-year bilities have not been decentralized and despite consid- (2018-2023) plan will take some time. In India, the 15th erable efforts in the past, no effective local government Financing Commission has the mandate to re-think system emerged (Sarker 2003, 2006). Decentralization competitive federalism and to re-balance responsibili- seemed to progress when the Upzilla (sub-district) lev- ties and revenue sharing. States and local governments el election took place in 2009. However, that transition remain highly dependent on devolved resources and was unsuccessful, partly because many Members of underperform in direct tax collection (Jaitley 2018). Parliament were unwilling to accept a weakening of their More importantly, while states were given more resourc- position. For example, the Upzilla level chairmen are re- es and autonomy, neither the 73rd and 74th constitu- quired to take advice from their respective Members of tional amendments providing for the empowerment Parliament in decision making, weakening this level of of urban local bodies nor the recommendations by government. At Union Parishad level, elections are held the 14th Financing Commission for devolution to local regularly but very few effective steps have been taken to government bodies have been properly implemented. strengthen them. Bangladesh now has an administra- Devolution of funds remains very limited and state fi- tively decentralized system with various levels of elected nancing commissions are typically ineffectual and far be- local governments institutions, but these local govern- hind schedule. In Nepal, the federal transition is difficult ments have very little decision-making and financial au- and capacity constraints at the subnational level remain thority, in addition to a low level of resources and very severe. Three of the seven provinces spent less than half weak governing capacities (Mansur and Ahsan 2019). of the budgeted expenditure in fiscal year 2018-19 and some local governments seem still dysfunctional (Figure In Afghanistan, the central state is too weak to pre- 38). One of the main challenges is to reassign capable civ- side over an efficient decentralization process and il servants from central agencies in Kathmandu to local in Sri Lanka the decentralization process is still governments across the country. In Pakistan, tax autono- halted. In Afghanistan, rebuilding a functioning cen- my by provinces has resulted in a fragmentation of the tral government has been the primary objective, but market, the different economic size of provinces has re- the Constitution entails the aspiration to move toward sulted in power imbalances, and the empowerment of a state structure where some resources and authorities provinces has disempowered local governments. To ad- are shared with sub-national governments in the future. dress the latter, Punjab in Pakistan just enacted a local The centralized state structure coexists with a decentral- government reform that still has to be implemented. ized traditional society, and many areas outside of Kabul are dominated by regional and local commanders. Past In Bangladesh, the current institutional arrange- attempts for decentralization or de-concentration may ments lag the more decentralized system envi- have partly been driven by political motives of regional sioned in the Constitution. The first article of the elites rather than a desire for better service delivery. Sri Constitution defines Bangladesh as a unitary state. But Lanka, on the other hand, has a well-established legal sys- the Constitution does support local government and tem of local governments at the provincial and sub-pro- allows local government to impose taxes, prepare their vincial levels. The 13th amendment to the Constitution 42 own budget and maintain funds. The envisioned devo- lution of responsibilities to districts, sub-districts, Union provided the constitutional provision to establish the Provincial Councils, which were given powers and func- of Villages, Municipalities and city corporations include tions. Yet, years of conflict halted the decentralization South Asia civil administration, fire protection, development and process in Sri Lanka. But despite the armed conflict being Economic Focus operation of markets, roads and infrastructure, traffic over for ten years, there has not been meaningful prog- Fall 2019 and urban transportation, local economic development, ress in decentralization in the last decade. While poor service delivery outcomes in South Asia devolved to municipalities. Such a simultaneous central- Making are only partly related to imperfect decentraliza- ization and decentralization creates an environment in (de)centralization tion, completing the implementation of decentral- which autonomous local governments engage in benefi- work ization in South Asia could support better service cial competition in an integrated world. This encourages delivery. Past decentralization efforts have often been innovation, experimentation, and learning from rivals, half-hearted, and across the region the implementation while everybody has to comply with the same general of the decentralization agenda remains very incomplete. rules and same general standards. Such a competitive en- In the survey conducted for this report, nine out of ten vironment cannot be created by decentralization alone. economists in South Asia see a misalignment between the de facto decentralization in their country and the The academic literature provides some evidence that one envisioned de jure. To make decentralization work, a strong central government is needed for successful local governments need a clear mandate, need to be decentralization. Enikolopov and Zhuravskya (2007), both empowered and accountable, and need sufficient using data from 95 countries (including Bangladesh, and reliable resources. Other challenges include insuf- India, Pakistan, and Sri Lanka), show that the strength ficient auditing and performance measurement and a of national political parties improves the results of fis- lack of coordination between different levels of govern- cal decentralization. Blanchard and Shleifer (2000) ar- ments. The latter can result in deadlocks between differ- gue that decentralization was more successful in China ent levels of government, especially if governments be- than in Russia, because in China the central government long to different parties and institutional arrangements was able to reward and to punish local administrations, allowing higher tiers to block lower tiers. But while which reduced both the risk of local capture and the these are all critical constraints, they are not a sufficient scope of competition for rents. And Malik et al. (2006) explanation of the disappointing contribution of decen- find that too much decentralization can slow growth. tralization to better service delivery in South Asia. Shah (2010) argues that the information revolution has led to a diminished economic relevance of intermediate levels of governments and enhanced the need for em- Decentralization and powered local governments. He describes how conform- centralization belong together ing to federalism of provinces and states in countries like Australia, India, Mexico, and Pakistan has blocked local Partial decentralization is not good enough, but governments. Evidence from countries like China, Japan, full decentralization is not the solution. The previ- Korea, and the Nordic countries, on the other hand, sug- ous section has argued that partial decentralization or gests that where local governments are given a proper insufficient empowerment of local communities can role, they contribute to competitiveness and growth. prevent optimal service delivery. However, there is am- ple evidence that too much decentralization can impair If the central government has delegated service de- efficiency. In Pakistan, for example, the existence of five livery to local governments, it still has the important independent tax jurisdictions has led to double taxation task to provide the right incentives and to exercise and hampers commerce between provinces (Box 6). Too quality control. Decentralization is not a zero-sum game much decentralization can lead to inefficient fragmen- in which power is just re-distributed across different tiers tation. Like multiple tax systems, also multiple product of government. Instead, replacing central provision of ser- or environmental standards would limit economic in- vices with decentralized provision by local governments tegration within a country. Moreover, decentralization introduces a new relationship of accountability between does not necessarily reduce corruption (Bardhan and national and local policymakers (Ahmad, Devarajan, and Mookherjee 2006). Decentralization has been found to Shah 2005). Roles change and designing a proper system decrease corruption in some cases (Fisman and Gatti of responsibilities and interactions between different 2002) and to increase it in others (Fan, Lin, Treisman tiers of government is crucial. Often the critical question 2009). More corruption is likely if the functional devo- is not whether a first, second, or third tier of government lution is unclear and if accountability to communities can best provide a service, but how to organize the joint is weak (Véron et al. 2006). In the survey conducted for production of services (Prud’homme 1995). In Europe, the this report, nearly nine out of ten South Asian econo- subsidiarity principle, which became the general princi- mists agreed that unclear responsibilities of local gov- ple of European Union law, states that a central authority ernments foster corruption at the local level in their has a subsidiary function, performing only those tasks country. In unchecked, autonomous local communities that cannot be performed at a more local level. But one the probability of corruption and elite capture can be key function of the highest level is ensuring a common higher than in larger, integrated communities. market. That common market creates conditions under which local communities can provide services. Similarly, Efficient centralization can be the missing element. in India more decentralization to states in recent years Over the past 40 years, rapid decentralization in many went hand in hand with the introduction of the Goods countries has been accompanied by strong market inte- and Services Tax (GST), which subsumed local and state gration (through opening up to the world market) and a consequent loss of some of their national sovereign- taxes to foster the flow of goods and services within the country. Moreover, the highest level also plays a crucial 43 ty. This process requires strong central rules to facili- role in redistributing resources, in setting standards to tate market integration. The clearest example is Europe, prevent a race to the bottom (for example in regulatory South Asia where some national government responsibilities were standards and taxes), and in providing quality control. In Economic Focus entrusted to the European Union and some power was South Asia, creating competition between local entities Fall 2019 Making (de)centralization Box 6: The case for a harmonized sales tax in Pakistan work The current sales tax regime fragments Pakistan into five competing tax jurisdictions resulting in potential double taxation and high compliance costs for businesses. Federal and Provincial Governments acknowledge that the current sales tax regime needs improvement. The base is broken up into goods and services with the federal government taxing the former and provincial governments the latter. The base is further broken up spatially, as each province has the power to tax services supplied within its jurisdiction and levy its own tax rates on these services. This fragmented nature of the base has caused inter-provincial and Federal- Provincial jurisdictional conflicts resulting in potential double taxation, exporting of taxes to other provinces, and consequently high costs of compliance for businesses. Unfortunately, no common legislative or administrative forum exists to address these is- sues. It is therefore important that a harmonized sales tax is introduced, along with a national forum that has the power to legislate and address these issues related to taxation. The sales tax has always been a shared tax, but after the 18th Constitutional Amendment provinces have established their own tax collecting agencies to collect sales tax on services without systems or procedures to effectively coordinate with one another. Constitutionally, the Sales Tax on Goods (STG) is a Federal tax while the Sales Tax on Services (STS) is a Provincial tax. Until 2010, both taxes were collected by the Federal Government, through the Federal Board of Revenue (FBR) for a nominal fee, and the sales tax on services was remitted back to the provinces. After the 7th National Finance Commission award (2009) and the 18th Constitutional Amendment (2010), all provinces established their own revenue authorities to collect sales tax on services and claimed a substantial increase in collection. Yet, total sales tax collection by all provinces is still only 0.5 percent of GDP. On the other hand, the fragmentation of the base and resulting complexity and jurisdictional issues have significantly raised the costs of compliance for businesses. With five different revenue collecting agencies on Sales Tax on Goods and Services, the following issues have emerged: With different tax collection agencies, and five different sales tax legislations, the tax bases are different in • Fragmented bases:  each province. Issues also arise between the Federal and Provincial governments on tax bases, as the definition of goods and services is not clear, and complaints have been raised by both the Federal and Provincial Governments that other agencies are impinging on their tax bases. Issues between provinces have also arisen concerning principles of taxation unrelated to the tax • Different taxation principles:  base, especially on sales tax on services. For example, Sindh province charges sales tax on many services on the origin principle, while the rest of the provinces charge the tax on the destination principle. Given the nature of Sindh as the province with a developed port, taxation at origin of services has led to substantial double taxation when other provinces levy the sales tax on the same service according to the destination principle. Retaliation by other provinces, either by also using the origin principle selectively or not providing an input tax credit on cross-border purchases, have also been reported. • High administrative and compliance costs:With five different tax collecting agencies charging sales tax, businesses working across different geographical areas within Pakistan must file up to 60 different tax returns (one per month for each agency). This increases compliance costs for firms, hampering overall economic activity. All these issues significantly reduce the incentive for businesses to purchase from other provinces or sell to other provinces, essentially fragmenting Pakistan into five markets. To address these issues, Pakistan could consider converting the current sales tax into a harmonized sales tax on goods and services that allows for effective revenue raising while not negatively impacting economic activity and investments in the country. is not easy, as internal migration is low and firms and be devolved to the middle tier of government. However, capital are not very mobile. Therefore, higher levels of past decentralization efforts in Pakistan and India have government play a crucial role in stimulating innovation strengthened the second tier, not the third. In India, and in rewarding good performance. Creating competi- states were in turn expected to devolve functions to lo- tion between local governments is fostered by granting cal governments, but even the most decentralized states autonomy to generate own resources. In addition, higher are lagging in this respect. Interestingly, in Pakistan lo- levels of government have the role of providing incen- cal government institutions have usually been strength- tives that promote accountability of local officials to citi- ened by non-representative regimes that consolidated zens to ensure services provided match local preferences. central power (Cheema, Khwaja, and Khan 2005). The 18th Constitutional Amendment agreed upon by dem- Decentralization to states and provinces may com- ocratically elected representatives, on the other hand, plicate a further devolution of responsibilities to not only devolved powers from the central government 44 local governments. The potential for improving basic service delivery is strongest when service delivery is lo- to the provinces, but also centralized local responsibil- ities at the provincial level. Efforts to reverse some of the negative consequences of the weakening of local calized. At the same time, only economic rules and stan- South Asia dards set at the highest level ensure market integration. governments have led to new reforms in KP and Punjab Economic Focus From a perspective of better service delivery in integrat- in Pakistan to strengthen local governments again. In Fall 2019 ed markets, only limited decision-making power should a survey among South Asian economists conducted for Figure 39: The relationship between economic complexity and local spending is relatively weak. Making Local government expenditure share (de)centralization Percent work 100 80 60 40 20 0 9 10 11 12 13 14 Log of real GDP (USD 2010) Sources: Reserve Bank of India; Ministry of Finance for Bangladesh, Pakistan, Nepal, and Bhutan; and SNG-WOFI and SNG-WOFI (World Observatory of Subnational Finance and Investment 2019) for all other countries. Population and per capita GDP are from the WDI. Notes: Data are for latest available year for each country. The vertical axis measures local government expenditure as a percentage of total expenditure. In Nepal, the local expenditure as a share of total expenditure is as budgeted for 2017/18. Figure 40: The share of local spending is correlated with political decentralization. Local government expenditure share Percent 70 60 50 40 30 20 10 0 0.0 0.2 0.4 0.6 0.8 1.0 Political decentralization index Sources: Reserve Bank of India; Ministry of Finance for Bangladesh, Pakistan, Nepal, and Bhutan; and SNG WOFI and SNG-WOFI (World Observatory of Subnational Finance and Investment 2019) for all other countries. The political decentralization index is from Ivanyna and Shah (2014). Notes: Data for local spending shares are for latest available year for each country. The vertical axis measures local government expenditure as a percentage of total expenditure. In Nepal, the local expenditure as a share of total expenditure is as budgeted for 2017/18. this report, a lack of political will to devolve power to lower levels was ranked the biggest challenge for effi- Optimal allocation of central cient local government institutions. resources is key to success Insufficient devolution of power to local govern- Even with empowered local governments and a ments and consolidation of power at the second strong and effective central government, challenges tier of government might well be a global phe- remain. The central challenge is how to allocate central nomenon. Figure 36 earlier in this chapter showed a resources over very diverse districts. Some areas expe- strong correlation between economic complexity and rience faster development than others and living stan- total subnational spending. The level of local spend- dards between rural areas and cities diverge. In South ing, however, is much lower than total subnational Asia, subsidiary farming nowadays takes place next to spending, and the relation with economic complexity world-class IT programming and space engineering. In is weaker (Figure 39). Especially the large federal states all countries the living standards in the most advanced of South Asia tend to have limited empowerment of cities are decoupled from the living standards in the local governments, as the states and provinces control most backward regions. Spatial variations in living stan- a relatively large part of public spending. The measure dards within countries are often larger than the varia- of local spending used even overestimates the deci- tion across countries. Each country in South Asia faces sion-making power of local governments, as spending decentralization is purely administrative in some cas- a geography of opportunity and of despair. Sustaining high growth rates and reducing poverty are both local 45 es. However, the share of local government spending challenges and require not only investing in people, but is highly correlated with political decentralization in people at the right place. Decentralization brings South Asia (Figure 40) and other indicators of closeness of govern- questions of the appropriate spatial distribution of pub- Economic Focus ment to its citizens (Ivanyna and Shah 2014). lic spending to the surface. Fall 2019 Making Figure 41: The appropriate distribution of resources depends on the perspective. (de)centralization work Political decentralization index Above-Above Above-Below Below-Above Below-Below No data Sources: Global Data Lab, Reserve Bank of India, World Bank, and staff calculations. Notes: The need/outcome/performance refer to expected years of schooling and are compared to the corresponding country average. Spending on education per capita is also compared to country averages. Thus, for example, above-above means that the need/outcome/performance indicator is above average and per capita spending on education is above average. See Appendix A6 for details. Governments try to achieve several objectives with prosperous cities in South Asia urgently need more pub- the spatial allocation of funds. As countries decide lic investment. The return on investments in cities can be where and on whom to spend their resources, they face a multiple of the return in rural areas. Third, rewarding potentially very difficult trade-offs between equity con- good performance can improve incentives, but also results cerns and the return on their spending (Bardhan 2002). in higher spending for those areas improving their out- First, the amount of public resources needed to ensure an comes. Balancing the need for equity across the country, equal (or at least similar) level of basic service delivery the need for investments in high-yielding projects, and for all citizens varies with local conditions. For example, the need for the right incentives to increase the efficiency providing basic education and health services is more of service delivery everywhere, is a daunting task (see Box expensive in sparsely populated areas. And a promise of 7 for a discussion of these trade-offs in India). universal basic service delivery can result in very high costs even for individuals. In Bangladesh, for example, the Evaluating the spatial distribution of spending de- attempt to connect everyone to the grid meant high ex- pends on the perspective. All these different issues penditure on those individuals living very far away from result in very different perspectives on the spatial dis- others, like isolated fishermen at the coast. Second, there tribution of public spending. In Balochistan, for exam- are good reasons to provide lagging areas with additional ple, the need for education spending (measured by low resources to allow them to catch-up. All countries in South expected years of education) is higher than in other Asia are concerned about spatial inequalities and pro- provinces in Pakistan, but so is per capita spending on vide special funding for lagging areas, often in the form education. Thus, spending is well targeted in terms of of equalization grants. On the other hand, even the most needs (Figure 41 , left panel), but poor outcomes mean Figure 42: In South Asia, equity concerns are important. Distribution of provincial public Human capital index and per Transfers per capita in Nepal expenditures in Afghanistan capital health spending in Sri Lanka Percent Index LKR NPR 100 0.7 6000 10000 Cumulative share of expenditures 0.6 5000 80 Kabul 0.5 4000 8000 60 0.4 3000 6000 0.3 40 2000 0.2 20 0.1 1000 4000 0.0 0 2000 Southern Western Sabaragamuwa North Western North Central Uva Central Northern Eastern 0 0 20 40 60 80 100 o Cumulative share of poor people Municipalities Sub-metropolitan Metropolitan Share of social spending Equalization transfers Share of NGO expenditure Human capital index Conditional transfers 46 Share of INGO expenditure Share of other government spending Per capita health spending (RHS) Linear (Per capita health spending (RHS) Sources: World Bank and staff calculations (left); Sri Lanka Human Capital Report (World Bank 2019a); Managing the federal transition to support sustainable urbanization in Nepal. World South Asia Bank (2019b). Data for Afghanistan is 2016, for Sri Lanka it is 2018, and for Nepal it is as budgeted for FY2017-18. Economic Focus Notes: For the left panel, provinces are ranked from low levels of economic activity to high levels measured by nightlight intensity and rural population (World Bank 2017). Social spending is defined as spending by the Ministry of Education, Ministry of Higher Education, Ministry of Public Health, Ministry of Refugees and Repatriates, Ministry of Women Affairs, Ministry of Fall 2019 Martyrs, Disabled and Social Affairs and the Ministry of Counter Narcotics. Making Box 7: Design aspects of a fiscal transfer system – evidence from India (de)centralization work The design of a fiscal transfers’ system is a challenging task. On the one hand, the fundamental idea of decentralization is to increase subnational governments’ discretion over expenditure to allow subnational governments to deliver the services demanded by citizens, while assuring that service providers are held accountable if services do not meet citizens’ expectations. On the other hand, increasing subnational governments’ discretion over the allocation of expenditure can mean that services are not provided to a level deemed satisfactory by the central government. One option to resolve this trade-off is for the central government to impose conditions linked to transfers on subnational government. Two types of conditions are especially common: incentives grants, which tie the disbursement of a transfer to the achievement of an outcome, and earmarked grants, which tie funds to specific uses. This box presents two examples from India on when such instruments work, and when they do not. Other work investigating Finance Commission transfers in India includes Rajaraman and Gupta (2016), who investigate properties of transfers intended for devolution to the local level, and Rajaraman and Vasishtha (2000), who highlight that the transfer of unconditional grants can reduce local government tax effort. Figure 43: Incentive grants linking fiscal prudence to debt relief had the intended effect. Treatment e ect of the debt relief incentive scheme on revenue de cit Estimates 400 200 0 -200 -400 -600 2002 2003 2004 2005 2006 2007 2008 2009 2010 Sources: RBI, Finance Commission, and staff calculations. Notes: Estimates represent coefficient estimates based on empirical model explained in Appendix A5. Error bars represent 90 percent confidence intervals. India’s 12th Finance Commission introduced a working incentive grant that linked fiscal prudence to debt relief. Starting in 2006, states that improved their revenue balance were eligible to write off repayments on central loans, effectively making federal assistance conditional on fiscal performance. Evidence suggests that the scheme has been effective at reducing revenue deficits at the state level (the detailed methodology to obtain these estimates is described in the appendix A5). Figure 43 plots the effect of the scheme on the revenue deficit by year, showing the reduction in the revenue deficit of states whose repayment obligations cover 1 percent of total expenditure compared to states with no repayment obligations. From 2006 onwards, revenue deficits dropped noticeably more in states with a 1 percent repayment obligation than in those without, peaking at a reduction of over INR 200 crore in 2008. Scaling these estimates through back of the envelope calculations implies that at the peak of its effectiveness, an increase in payment obligations to the center by 3 percentage points of total expenditure led to a reduction in the revenue deficit by 40 percent. The scheme was thus effective in inducing fiscal discipline. Once these incentives ceased to exist, deficits went back to previous levels. Figure 44: Additional grants earmarked for education increased non-education spending more than education spending. Treatment e ect of receiving earmarked grants on education Treatment e ect of receiving earmarked grants on other expenditure expenditure Estimates Estimates 0.2 .015 .01 1 .005 0 0 -0.1 2007 2008 2009 2010 2011 2012 2013 2014 2015 -.005 2007 2008 2009 2010 2011 2012 2013 2014 2015 47 Sources: RBI, Finance Commission, and staff calculations. South Asia Notes: Estimates represent coefficient estimates based on empirical model explained in Appendix A5. Error bars represent 90 percent confidence intervals. Economic Focus Fall 2019 Making (de)centralization Around ten percent of all grants in India are earmarked for a specific purpose, typically for spending in social sectors or on work agriculture, but earmarking does not always achieve the intended results. Evidence suggests that earmarking spending is ef- fective in raising social sector spending: estimates highlight that a 5-percentage point increase in the share of earmarked grants between Finance Commissions increases the pass-through of one rupee transferred to total expenditure by 35 percent – of which about one third is accounted for by increased spending on education, health, and agriculture. Conditional grants, however, do not exclusively raise spending in targeted sectors. The 13th Finance Commission earmarked grants for primary education equivalent to approximately 8 percent of baseline education expenditure. These grants were inframarginal, which means that they did not exceed the amount that states would have spent on education anyway. As a result, states responded to the introduction of the grant in 2012 by expanding non-education expenditure. By contrast, the increase in education spending after 2012 was less steep and statistically not distinguishable from zero (Figure 44). This occurred because it was possible for states to use the funds from the grant to cover the spending on education they would have incurred anyway and divert the resulting savings to other uses. Taken together, these results suggest that conditional grants can work in raising targeted sector spending, but if the amounts distributed are too small, conditional grants have effects that are comparable to untied grants. that spending is not well targeted in terms of efficien- larger population sizes, receive lower levels of grants cy (Figure 41, middle panel). And looking at chang- per capita than sub-metropolitan or other smaller mu- es over time results in yet another perspective. Since nicipalities (Figure 42, right panel).  The federal govern- Balochistan improved education outcomes more than ment allocates greater amounts of grants (in per capita the average province in Pakistan, increases in resourc- terms) to those municipalities that are smaller in size es that are higher than average may be reasonable and less urbanized (World Bank 2019b). (Figure 41, right panel). The appropriate distribution of resources is hence not straightforward. In South Asia, the Also globally, the allocation of resources to decen- distribution of revenue between the higher and lower tralized governments seems to reduce spatial dif- levels of government is usually determined by financing ferences. In principle, more decentralization can both commissions that design revenue sharing formulas. In increase or decrease spatial variation within a country. addition to some standard variables like the population, If decentralization is driven by demands for more au- they usually include a measure of backwardness. That tonomy by the richest areas of a country, it is likely that can introduce some counter-intuitive incentives. For ex- these areas will be able to keep more of the resourc- ample, when a higher share of people living in pover- es generated by them after decentralization. And with ty results in larger transfers, subnational governments more resources staying in the richest areas, spatial in- that fight poverty successfully are penalized. Instead, it equalities are likely to increase. On the other hand, if may be more useful to define needs and backwardness decentralization is driven by areas feeling left behind, independent from the outcome policy makers target. and such forces are driving the decentralization pro- cess in Nepal, a more decentralized system can have an In South Asia, equity concerns appear to be an equalizing effect. Globally, fiscal decentralization has important determinant of the spatial allocation resulted in less spatial heterogeneity. Data on spend- of funds. Available data suggests that equity con- ing of 82 countries between 1990 and 2016 show that a cerns result in pro-poor social spending in South Asia higher share of spending by subnational governments (Figure 42). In Afghanistan, ranking provinces accord- was associated with a lower within-country variation ing to a measure of economic development based on of income, education outcomes (measured by the ex- night light intensity and rural population (World Bank pected years of education) and health outcomes (mea- 2017) and then comparing the cumulative share of sured by the life expectancy) (Table 6). The effect of de- poor people to the cumulative share of expenditure centralization on the spatial variation was not linear shows that social spending is benefiting lagging prov- (Figure 45). For education and health outcomes, spatial inces more than others (Figure 42, left panel). Other variation only decreased if a substantial share was de- government spending is not. Surprisingly, NGO spend- centralized. In line with these results, a higher subna- ing and especially international NGO spending is going tional spending share is argued to have increased the over-proportionally to Kabul and does not favor lagging efficiency of spending as well (Sow and Razafimahefa provinces. A similar analysis is possible using income 2015). Many countries experiencing convergence have per capita for regions from the Global Data Lab. The succeeded by promoting an economic union and by en- results using the alternative measure and regions in- suring minimum standards in basic services across the stead of provinces are very similar. In Sri Lanka, health country (Shankar and Shah 2001). spending is often ad hoc and largely driven by human resources, which accounts for around half of health spending. Yet, aggregating provincial health spending from different sources and contrasting it with human Better geospatial data capital shows that per capita spending is higher in can improve evidence- 48 provinces with lower human capital (Figure 42, mid- dle panel) providing evidence that the formula used to based policy designs South Asia benefit lagging areas is working at least to some extent Better geospatial data of public expenditure is fun- Economic Focus (World Bank 2019a). In Nepal, there is a clear pattern damental. Often fiscal incidence is analyzed across that larger metropolitan cities, as characterized by households, but spatial variation may account for a big Fall 2019 Table 6: Globally, fiscal decentralization has Figure 45: … and the effect of decentralization Making reduced spatial variation… on inequality is non-linear. (de)centralization E ect of decentralization on spatial inequality work Income Health Education Percent 20 10 Subnational -0.87*** 0.25 0.64*** expenditure (lagged) (0.21) (0.42) (0.23) 0 0 Squared subnational 0.82*** -1.20** -1.68*** expenditure (lagged) (0.30) (0.81) (0.44) -20 -10 Controls YES YES YES -40 -20 Observations 1332 1244 1298 -60 -30 R2 overall 0.95 0.86 0.96 -80 -40 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Sources: IMF’s Fiscal Decentralization Dataset, Global Data Lab, and staff calculations. Notes: The data covers 82 countries from 1990 to 2016. Spatial variation is measured Health Education Income as coefficient of variation (CV) within the country. GDP per capita is used as a control, and the fiscal decentralization variable is lagged to address endogeneity concerns. Distribution of subnational spending shares The analysis includes country fixed-effects. The figure shows the impact of changes in Kernel density fiscal decentralization (subnational expenditure over total government expenditure) on the within-country inequality in the outcome variables of income (gross national 2.0 income per capita in PPP), education (expected years of schooling) and health (life expectancy). The impact of moving from one level of fiscal decentralization to another is obtained by subtracting the corresponding inequality levels in the y-axis. For in- stance, doubling the level of fiscal decentralization from 20 to 40 percent leads to a 1.0 7.7 percent reduction in within-country income inequality, a 14.8 percent reduction in health inequality and a 7.4 percent in education inequality. See the Appendix A3 for details on the data and estimation method. Standard errors in parenthesis; * p<0.1, ** p<0.05, *** p<0.01. 0.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 portion of it. In addition, urbanization is arguably the a World Bank initiative to facilitate access to budget data. most important transformation faced by developing BOOST data is available at different levels of detail and for countries, but urbanization is a spatial phenomenon, different years for Afghanistan, Pakistan, and Bangladesh. boosted or hindered by local public spending. And, Unfortunately, none of this data is publicly available yet. finally, successful resource decentralization is critical- For India, state level spending is publicly available, but ly important to attain development outcomes but re- the distribution of local spending is not, and it is very quires accountability and performance measurement cumbersome to collect. This is partly because the flows that only work with better data. Without subnational of public funds are not all available from treasuries but expenditure data, decentralization cannot be man- involve different schemes as well (see Box 8). The absence aged, policies cannot be evidence-based, spending can- of district-level expenditure data in India is problematic not be targeted, and it is impossible to understand the given the large size of some of the states and the impor- effect of spending on outcomes. Investments in regular- tance of local governments in providing services. Even ly published geospatial data would help track spatial recording and publishing all on-budget expenditures at a development patterns and could guide policy designs. granular spatial level does not provide a full picture, since For that, authoritative administrative areas need to be substantial public spending is undertaken by self-ac- defined and used across tiers of government. These al- counting entities like municipalities and state owned-en- low for accurate data collection, aggregation, and sub- terprises (see Box 8 for an example in Pakistan). sequent communication. Linking subnational spending data with subnation- South Asia lags far behind in recording and publish- al outcome data deepens understanding of the rela- ing subnational fiscal data. This is reflected in such data tionship between the two. Incentive structures and in- not being as easily available as that of other countries in stitutional arrangements are crucial for the efficiency of data efforts by different international organizations. For service delivery, and in many cases spending more will example, the only South Asian countries covered by the have a weaker effect than spending well (World Bank IMF dataset on fiscal decentralization are Afghanistan, 2018). But better data is fundamental to better under- Maldives, and Sri Lanka. The World Observatory of stand the relationship of spending and outcomes. For Subnational Finance and Investment (2019), led by the example, while the relationship between spending and OECD and the United Cities and Local Government, pro- vides some general characterization of some South Asian learning outcomes is often weak, there is some evidence that higher spending is beneficial in Pakistan. Spending 49 countries but very limited data. The maps presented in on pre-school and primary education per student in Figure 41 are based on an attempt to provide a regional public schools varies a lot between districts in Pakistan South Asia overview of subnational public spending. They combine (Figure 46). After controlling for different variables Economic Focus data from different sources and build strongly on BOOST, that may impact the costs and quality of education, Fall 2019 Making Figure 46: Spending on primary education Table 7: … and higher spending has a positive (de)centralization varies across districts in Pakistan… impact. work Spending per primary student (USD) (1) (2) (3) <100 Years in School Learning 101 - 125 school satisfaction outcomes 126 - 150 ln(spending) 0.52*** 0.06*** 0.08*** 151 - 200 (0.06) (0.005) (0.02) > 200 No data Controls YES YES YES Observations 124607 206267 41310 R2 overall 0.77 0.22 0.20 Sources: PSLM 2012-13, PIFRA database, and staff calculations. Note: Controls are district fixed-effects, population density, GDP per capita, share of public students, agriculture employment, average years of schooling of the household head, and share of population with access to electricity. Standard errors in parenthesis; * p<0.1, ** p<0.05, *** p<0.01. higher district spending raises the years students in pub- Bhutan, new decentralization efforts require building lic schools remain in school and the satisfaction of their local capacity. In India, the 15th Financing Commission parents with the schools. More importantly, with higher can properly balance responsibilities and revenue shar- spending the probability of good learning outcomes in- ing and ensure fiscal prudence at the state level. Most creases (Table 7; see Appendix A4 for more details). importantly, India could benefit by giving some of the responsibilities accumulated at the state level to local Better data and more research can lead to better pol- governments. In Nepal, easing tensions between differ- icy advice. There has been more progress in generating ent levels of government and improving capacity at the spatially granular outcome data, compared to spatially local level would support a successful federal transi- granular expenditure and revenue data. Subnational tion. In addition, the transition will need time to deliv- outcome data for South Asian countries is, for exam- er better services and, in the meantime, managing the ple, available from the South Asia Spatial Database (Li expectations can be helpful. In Pakistan, the power of et al. 2015) and from the Global Data Lab (Smits and provinces could be more balanced, and a new revenue Permanyer 2019). In addition, there are different initia- sharing formula can create incentives better aligned tives at the country level. But more subnational public with improved service delivery. In addition, reversing finance data is fundamental for a deeper understand- the market fragmentation between provinces would ing and well-functioning decentralized systems. Linking benefit economic growth. outcome and spending data results in useful insights, and better data will lead to more research on subnation- Transforming South Asia’s cities into dynamic en- al public finance and local service delivery in the region. gines of growth requires increasing revenues for And since decentralization is work in progress across those cities, but that comes with the risk of widen- South Asia, more research will be valuable to further ing spatial inequality. South Asian countries would strengthen evidence-based policy advice. benefit from investing more in their cities (World Bank 2016) and to raise lagging places at the same time. Across South Asia, subnational entities depend on cen- Conclusions tral government grants rather than on fees or own tax revenue. In India, municipal revenue is only one percent Decentralization in South Asia has yet to deliver on of GDP. Tariffs for many basic services, like water supply its promises. Despite a long history of decentralized in India or solid waste collection in Sri Lanka, do not systems, public service delivery in South Asia remains reflect cost recovery or are even free. Higher tariffs and disappointing. To benefit from more decentralization, fees could provide some additional resources to local South Asian countries need to avoid that decentraliza- governments. More importantly, however, South Asian tion turns into fragmentation. It is crucial that effec- cities are generating much less revenue from land than tive central governments create integrated markets in cities in other regions in the world. More autonomy in which local communities compete and facilitate mobil- property and land taxation, and in re-zoning, could be ity across local boundaries. Central governments’ role is crucial. In addition to funding city development, higher 50 also to address equity concerns and support disadvan- taged regions to give all citizens equal opportunities, revenue generation from land in cities could also free resources to invest in lagging areas. Land policies touch irrespective of where they are born. In addition, South upon vested interests and elite capture and are difficult South Asia Asian countries have specific priorities. In Bangladesh, to change. While decentralization can potentially rein- Economic Focus urban management can be improved, and empower- force local elites, democratically elected, empowered, Fall 2019 ing local governments can improve service delivery. In and accountable mayors could bring about positive Making Box 8: Tracking on- and off-budget expenditure at the district level in Pakistan (de)centralization work A complete picture of public spending requires including self-accounting public entities. Across most countries there is a strong focus on public financial management information systems, which provide information on spending at local levels. And there is no doubt that strengthening these is important. However, some of the most important government agencies have their own accounts, and their expenditures are often not included in analyses of government expenditures. Expenditure tracking sur- veys have been conducted to address this issue, but these are costly and are mainly focused on education and thus are not well suited to track municipal spending or infrastructure investments. The World Bank hence started a pilot for a comprehensive public expenditure database in Pakistan, where state-owned en- terprises (SOEs) play a crucial role in service delivery. On-budget data has been collected from PIFRA, which tracks federal, provincial, and district expenditure at the district level and connects expenses to the thematic functions of the government using very detailed descriptions. Off-budget expenditures by SOEs and TMAs have been collected individually. These efforts built heavily on pre-existing personal relationships, substantial trust building, and were very time intensive. The SOEs included in the pilot, cov- ering years 2012/13 and 2013/14, are the Electricity Distribution Companies (DISCOs), Oil and Gas Development Company Limited (OGDCL) and the National Highway Authority (NHA). Merging data from the SOEs with the on-budget government data requires three key steps. First, the expenditures need to be made compatible across different accounting procedures. This requires using the Cash Flow Statement of the SOEs included in the accrual-based accounts, published as per the requirement of International Financial Reporting Standards, to convert their spending to a cash basis. Second, SOE expenditures need to be distributed at the district level. Road expenditure of the NHA, for example, has been distributed using kilometers of national roads per district and DISCO expenditure has partly been distributed using the number of consumers or users across districts. Third, SOE expenditure categories need to be merged with PIFRA categories. PIFRA utilizes very detailed descriptions to connect expenses to thematic functions of the government, and some of the categories are opaque to economic and development questions. Hence, expenditures are re-categorized into eight categories: health, education, infrastructure, social protection, pensions, debt servicing, operational expenses, and others. However, the filter we use is easily adjustable and expenditures can be categorized differently. Development and current expenditures are considered separately, and population data from the 2017 Census is used to approximate per capita expenditures. Figure 47: Merged off-budget expenditure amounts to more than a quarter of total public spending. Public spending o - and on-budget Shares O -budget On-budget Table 8: Overall infrastructure spending is benefiting richer districts more than only on-budget expenditure. (1) (2) Infrastructure, total per capita Infrastructure, PIFRA per capita ln(GDP per capita) 1.12*** 0.40** (0.25) (0.20) Controls YES YES Observations 103 102 R2 overall 0.52 0.71 51 Sources: PIFRA, World Bank, and staff calculations. Notes: GDP per capita is based on nightlights and rural population (World Bank 2017) and controls include dummies for Islamabad, provincial and division capitals. Standard errors South Asia in parenthesis; * p<0.1, ** p<0.05, *** p<0.01. Economic Focus Fall 2019 Making (de)centralization Off-budget expenditure is substantial and matters. The tracked off-budget expenditures amount to over a quarter of total work spending. In addition, off-budget expenditure is important across all provinces and nearly all districts, in some of which more than half of the expenditure is not on-budget. Using a district level measure of GDP per capita based on nightlights and population (Beyer et al. 2018), it is possible to analyze the relationship between the level of development of a district and its spending pat- terns. Districts with higher per capita GDP spend significantly more per person on education but not on health. Richer districts also spend more on infrastructure. But considering only on-budget spending (column 2 in Table 8) suggests a much weaker relationship between per capita GDP and infrastructure spending than when off-budget expenditure is included. This is just one example of how providing the full picture of public spending can change our understanding. change. Currently, even large and economically im- empowerment of cities is crucial, without generating portant cities across the region have no elected may- more revenue that can free resources to finance de- ors, and where they exist, they are often weak. In India, velopment in lagging regions, it may increase spatial even town planning is still at the state level. While the heterogeneity. References Aghion, P., and Tirole, J. (1997). Formal and real au- Causes Harvard University, John F. Kennedy School of thority in organizations. Journal of political econo- Government. Research Working Paper Series RWP05-034. my, 105(1), 1-29. Crook, R.C. and Sverrisson, A.S. (2003) Does Ahmad, E., and Brosio, G. (2006). Handbook of Fiscal Decentralisation Contribute to Poverty Reduction? Federalism. Edward Elgar Publishing. Surveying the Evidence. Houtzager and M. Moore Ahmad, E., and Brosio, G. (2009). Decentralization and (eds.) in Changing Paths: International Development local service provision: what do we know? Asia and the New Politics of Inclusion. University of Research Centre Working Paper 27 Michigan Press, Ann Arbor. 233–59. Ahmad, J., Devarajan, S., Khemani, S., and Shah, S. Enikolopov, R., and Zhuravskaya, E. (2007). (2005). Decentralization and service delivery. The Decentralization and political institutions. Journal World Bank. Washington, DC. of Public Economics, 91(11-12), 2261-2290. Asfaw, A., Frohberg, K., James, K. S., and Jütting, J. (2007). Fan, C. S., Lin, C., and Treisman, D. (2009). Political decen- Fiscal decentralization and infant mortality: em- tralization and corruption: Evidence from around pirical evidence from rural India. The Journal of the world. Journal of Public Economics, 93(1-2), Developing Areas, 17-35. 14-34. Aslam, G., and Yilmaz, S. (2011). Impact of decentral- Fisman, R., and Gatti, R. (2002). Decentralization and ization reforms in Pakistan on service delivery— corruption: evidence across countries. Journal of An empirical study. Public Administration and Public Economics, 83(3), 325-345. Development, 31(3), 159-171. Hart, O., Shleifer, A., and Vishny, R. W. (1997). The proper Bardhan, P. (2002). Decentralization of governance and scope of government: theory and an application to development. Journal of Economic Perspectives, prisons.  The Quarterly Journal of Economics, 112(4), 16(4), 185-205. 1127-1161. Bardhan,P.,and Mookherjee,D.(2006).Decentralization,cor- Herath, T. N. (2009). Decentralization of governance and ruption and government accountability. International economic development: the Sri Lankan experience Handbook on the Economics of Corruption, 6, 161-188. after establishment of provincial councils. South Asia Economic Journal, 10(1), 157-185. Beyer, R., Chhabra, E., Galdo, V., and Rama, M. (2018). Measuring Districts’ Monthly Economic Activity IMF (2019). Fiscal Decentralization Dataset. from Outer Space. World Bank Policy Research Iqbal, K., and Ahmed, M. (2015). Are Decentralized Working Paper No. 8523. Governments More Effective in Mitigating Disaster Blanchard, O., and Shleifer, A. (2001). Federalism with Risks? Bangladesh Development Studies, 38(3), 1-24. and without political centralization: China versus Ivanyna, M., and Shah, A. (2014). How Close Is Your Russia. IMF Staff Papers, 48(1), 171-179. Government to Its People? Worldwide Indicators on Chakrabarti, K. (1996). The Gupta Kingdom. History of Localization and Decentralization. Economics: The Civilizations of Central Asia (The crossroads of civili- Open-Access, Open-Assessment E-Journal, 8(3), 1-61. zations: AD 250 to 750). 3, 188-210. Jaitley, A. (2018). Economic Survey 2017: Volume I, Chapter 4: Reconciling Fiscal Federalism and Accountability: Is 52 Cheema, A., Khwaja, A.l., and Qadir, A. (2003). Local gov- ernment reforms in Pakistan: context, context and there a Low Equilibrium Trap?. Government of India causes’. Harvard University, John F. Kennedy School of Kalirajan, K., and Otsuka, K. (2012). Fiscal decentral- South Asia Government. ization and development outcomes in India: An Economic Focus Cheema, A., Khwaja, A. I., and Khan, A. (2005). exploratory analysis. World Development, 40(8), Fall 2019 Decentralization in Pakistan: Context, Content and 1511-1521. Khattak, N. U. R., Ahmad, I., and Khan, J. (2010). Fiscal Sarker, A.E. (2006). The political economy of decen- Making decentralisation in Pakistan. Pakistan Development tralized governance: an assessment of rural local (de)centralization Review, 49(4), 419. government reforms in Bangladesh. International work Kumar, S., and Managi, S. (2009). Compensation for Journal of Public Administration, Vol. 29 No. 13, pp. 1285-1309. environmental services and intergovernmen- tal fiscal transfers: The case of India. Ecological Schneider, A. (2003). Decentralization: Conceptualization Economics, 68(12), 3052-3059. and measurement. Studies in comparative interna- tional development, 38(3), 32-56. Lamichhane, H. R. (2016). Contribution of Expenditure of Local Bodies to Economic Growth in Nepal. Shankar, R., and Shah, A. (2001). Bridging the economic Journal of Fiscal Federalism, 99-113. divide within nations: a scorecard on the perfor- mance of regional development policies in reduc- Li, Y., Rama, M., Galdo, V., and Pinto, M.F. (2015). A ing regional income disparities. World Bank Policy Spatial Database for South Asia. The World Bank. Research Working Paper (No 2717). Washington, DC. Shah, A. (2010). Empowering States and Provinces or Lin, J., and Liu, Z. (2000). Fiscal Decentralization and Unshackling Local Governments: Does It Matter for Economic Growth in China. Economic Development Peace, Order, Good Government, and Growth?. The and Cultural Change, 49(1), 1-21. Pakistan Development Review, 333-362. Mahal, A., Srivastava, V., and Sanan, D. Smits, J., and Permanyer, I. (2019). The Subnational Human (2000). Decentralization and public sector delivery Development Database. Scientific Data, 6, 190038. of health and education services: The Indian ex- perience (No. 20). ZEF discussion papers on develop- Singh, N. (2009). Fiscal Decentralization in China ment policy. and India: Competitive, Cooperative or Market Preserving Federalism?. Public Finance & Malik, S., Mahmood-ul-Hassan, and Hussain, S. (2006). Management, 9(1), 97-136. Fiscal Decentralisation and Economic Growth in Pakistan. The Pakistan Development Review, 845-854. Singh, A. K. (2018). Dynamic De/Centralization in India, 1950–2010. Publius: The Journal of Federalism, 49(1), Mansur, Dr., and Ahsan, T. (2019). Bangladesh local gov- 112-137. ernment public financial management systems as- sessment. The World Bank. Washington, DC. Sow, M., and Razafimahefa, M. I. F. (2015). Fiscal decen- tralization and the efficiency of public service Martinez-Vazquez, J., Lago-Peñas, S., and Sacchi, A. delivery (No. 15-59). International Monetary Fund (2017). The impact of fiscal decentralization: A sur- Working Paper. vey. Journal of Economic Surveys, 31(4), 1095-1129. Sow, M., and Razafimahefa, M. I. F. (2018). Fiscal decen- Panta, K. R. (2016). Elite Capture and Corruption: tralization and the efficiency of public service Obstacles to Local Public Service Delivery. Journal delivery.  Kim, J. and S. Dougherty (eds.) in Fiscal of Fiscal Federalism, 114-137. Decentralization and Inclusive Growth. OECD Prud’homme, R. (1995). The dangers of decentral- Publishing, Paris. ization.  The World Bank research observe, 10(2), Véron, R., Williams, G., Corbridge, S., and Srivastava, M. 201-220. (2006). Decentralized corruption or corrupt de- Radboud University. (2019). Global Data Lab. Institute for centralization? Community monitoring of pov- Management Research. erty-alleviation schemes in Eastern India. World Development, 34(11), 1922-1941. Raghunandan, T., Iyengar, S., Bhatikar, T., Narayanan, U. and Janakiraman, P. (2016). Paisa for Panchayats: World Bank. (2000). Entering the Twenty-First Century. Tracking Fiscal Devolution to Local Governments: Oxford and New York: Oxford University Press. A case study from Kolar district, Karnataka. Centre Washington, DC. for Policy Research. World Bank. (2016). Leveraging Urbanization in South Rajaraman, I. and Gupta, M. (2016). Preserving the Asia. The World Bank. Washington, DC. Incentive Properties of Statutory Grants. Economic World Bank. (2017). South Asia Economic Focus Fall & Political Weekly, 35(33): 79-84. 2017 – Growth out of the Blue. The World Bank. Rajaraman, I. and Vasishtha, G. (2000). Impact of Grants Washington, DC. on Tax Effort of Local Government. Economic & World Bank. (2018). World Development Report 2018: Political Weekly, 35(33):2943-48. Learning to Realize Education’s Promise. The World Robinson, M. (2007). Does Decentralisation Improve Bank. Washington, DC. Equity and Efficiency in Public Service Delivery World Bank. (2019a). Spending more and spending bet- Provision? IDS Bulletin, 38(1), 7-17. ter: Sri Lanka Health Financing System Assessment. The World Bank. Washington, DC. 53 Rondinelli, D. (1990). Decentralization, Territorial Power, and the State: A Critical Response. Development World Bank. (2019b). Managing the federal transition and Change, 21(3), 491-500. to support sustainable urbanization in Nepal. The Sarker, A.E. (2003). The illusion of decentralization: evi- World Bank. Washington, DC. South Asia dence from Bangladesh. The International Journal World Observatory of Subnational Finance and Economic Focus of Public Sector Management, 16(7), 523-548. Investment (2019). SNG-WOFI Database. Fall 2019 Making Appendix (de)centralization work Appendix A3: Global analysis of the distributional effect of decentralizing spending This analysis covers 82 countries from 1990 to 2017 as a result of merging two different databases. Outcome data was collected from the Global Data Lab. Education outcomes were measured as the expected years of schooling, and health outcomes were approximated by life expectancy. Gross National Income per capita (PPP, 2011 USD) was used as the income variable. Fiscal decentralization is measured as subnational expenditure over total government expenditure. This variable comes from and the IMF’s Fiscal Decentralization Database. Estimating the impact of fiscal decentralization on the spatial variation of income, education, and health. The impact of fiscal decentralization on the three outcome variables is estimated using a global pooled regression (Sow and Razafimahefa 2015). To explain the spatial variation within a country we compute the year and coun- try-specific coefficient of variation across regions and regress it on the fiscal decentralization variable of interest. To allow for possible non-linear effects of fiscal decentralization on the spatial variations of the outcomes, the squared term of fiscal decentralization was included. Country-fixed effects were added to control for time-invariant unex- plained factors at the country level, and GDP per capita (WDI) was added to avoid omitted variable bias. Finally, fiscal decentralization has been lagged to address endogeneity concerns. Income, education, health Variable of interest: fiscal Country decentralization with FEs cuadratic term Appendix A4: The effect of higher spending on education in Pakistan This analysis covers 110 Pakistani districts during FY 2012/13. Data was collected from the Pakistan Social and Living Standards Measurement 2012/13 (PSLM); the Data for Pakistan initiative; the Annual Status of Education Report database 2012 (ASER); and spending data from the Project to Improve Financial Reporting and Auditing da- tabase 2012/13 (PIFRA). Estimating the effect of education spending on education outcomes. To estimate the effect, the education out- come variables were defined as years of schooling, school satisfaction and learning outcomes in language and math. Two different specifications were used, and both models included district fixed effects, to control for unexplained factors at a district level, and other control variables related to education spending, as well as to individual and household characteristics. Years of schooling, school satisfaction Appendix A5: Estimating the impact of fiscal transfer design aspects in India This analysis is based on a state-year panel that covers the fiscal years from 2001 to 2017 and a balanced sample of 25 states. Data was collated from different sources, including the Finance Commission reports the Reserve Bank of India’s state reports database. Estimating the Effect of the Debt Relief Incentive Scheme under the 12th Finance Commission. The effect of the debt relief scheme is estimated using a difference-in-difference approach. The temporal variation for this approach 54 comes from the comparison of the 12th and 11th Finance Commission. The spatial variation is generated as states with a higher debt burden have a larger incentive to shoulder the cost of revenue deficit reductions in response to the incentive scheme. The regression we estimate is thus: South Asia Economic Focus Fall 2019 Making (de)centralization work The outcome is the revenue deficit in a given state s in year t. measures the percentage of total expenditure allocated to repaying loans to the center in the baseline year 2005. FC_12t is a dummy variable for the 12th Finance Commission period and μs and gt capture state and year fixed effects respectively. The coefficient of interest is b1, which measures the differential change between the 12th and 11th Finance Commission between states with higher and lower repayments to center, relative to total expenditure. Estimating the Impact of Conditional Finance Commission Grants. Estimates are obtained through a 2-SLS pro- cedure which first predicts transfers received by whether a state has special category status and how this was treated under a given Finance Commission through the following equation: The variable SCSs denotes a dummy for a special category state, and the variable FCit denotes an indicator that takes the value 1 if a given year falls under the i-th Finance Commission period, and 0 otherwise. The predicted values from this regression are then used to estimate the parameters of interest in the following specification: The outcome variable, yst, captures state-level expenditure normalized to represent expenditure in per capita terms and the term captures the share of transfers received in a given state and given year that were earmarked for a spe- cific purpose. The coefficient of interest is b2 which captures the difference in the effect of transfers on fiscal outcomes from marginally increasing the earmarked share. Estimating the Effect of Education Grants under the 13th Finance Commission. The estimation of the effect of the tied education grants on spending under the 13th Finance Commission leverages a difference-in-difference design, where the temporal variation comes from the start of the 13th Finance Commission period and the pre-treatment period is the 12th Finance Commission period. To obtain special variation, we define a continuous treatment variable measured as the share of education grant allocation to baseline education expenditure in 2001. This series is scaled by a factor 0.2 to make the coefficient interpretable as the treatment effect of receiving a grant that covers 20 percent of baseline education expenditure (compared with receiving no grant): The primary outcome variables we consider are (log) education expenditure and (log) non-education expenditure. FC_13 denotes a time dummy that takes the value 1 for all observations within the 13th Finance Commission period. The re- gression also includes state and year fixed effects. The coefficient of interest is b1, which measures the differential change between the 13th and 12th Finance Commission between states with higher and lower education grant allocations. Appendix A6: Maps on subnational education expenditure and outcome variables The analysis covers five South Asian countries: Afghanistan, India, Pakistan, and Sri Lanka for 2016 and Bangladesh for 2010. Data on education outcomes (expected years of schooling) was collected from the Global Data Lab for all countries except for Sri Lanka, for which it comes from the Household and Income Survey. The education expenditure data comes from the World Bank Boost Initiative (except for India and Sri Lanka) and makes use of the COFOG functional and economic classifications to identify what corresponds to education expenditure. In the case of India, expenditure data comes from the Reserve Bank of India and has been linked to the COFOG classification. For Sri Lanka, it comes from a recent World Bank (2019) report on human capital. Subnational education expenditure is put in per capita terms using population data from the Global Data Lab. Expenditure data disaggregation varies by country; for India it is at the state level but for other countries it is more disaggregated. To carry out the analysis, three perspectives were adopted: need, outcome and performance. For each of them the regional expenditure per capita is identified to be either above or below the country average and the education 55 needs, the education outcome and the change in the education outcome are identified as above or below the country average. This way we present three different perspectives with four categories that mix the situation in the expendi- ture and the need/outcome side. Needs and outcomes are exactly the opposite of each other. South Asia Economic Focus Fall 2019 Making Appendix A7: New research on decentralization: A summary of the 4th South Asia Economic (de)centralization Network Conference work In preparation for this report, the World Bank co-organized a regional academic conference with the South Asia Network on Economic Modelling (SANEM). For this two-day event, young researchers, practitioners and ex- perts from five South Asian countries came together to discuss challenges in subnational finance and local service delivery in their countries. The role of local governance institutions, their capacity to function and their fiscal autonomy are at the center of the discussion of fiscal decentralization and efficient service delivery in South Asia. Monzur Hossain (Bangladesh Institute of Development Studies) does not find a strong impact of improved administrative governance on local economic development in Bangladesh. However, training programs for local government officials, the num- ber of parishad meetings, and the total number of schemes implemented over the previous year positively affected perceived outcome indicators. Dr. Ahsan Mansur (Executive Director, Policy Research Institute Bangladesh) highlight- ed the strong centralization of power in Bangladesh and emphasized that local governments lack not just resources, but also authority to perform their assigned functions. The Minister of Planning for Bangladesh, MA Mannan, shared similar sentiments in the inaugural session and agreed that one of the biggest challenges hindering progress in Bangladesh is the absence of well-functioning local government institutions. But these challenges are not restricted to Bangladesh only. Manish Gupta (National Institute of Public Finance and Policy, India) concluded that India’s State Finance Commissions have so far had very little impact on improving state-local fiscal relations, addressing vertical and horizontal fiscal imbalances, or increasing access to public amenities. Dr. Ishrat Husain (Advisor for Institutional Reforms and Austerity to the Prime Minister of Pakistan) presented Pakistan’s efforts to bring decision making closer to the people in KP and Punjab in Pakistan, for example by establishing separate rural and urban governance bodies as well as municipal and town bodies with locally elected representatives. Historically, the empowerment of local gov- ernments in Pakistan was mostly done by military regimes. He also discussed the need to bring competition among districts for efficiency of service delivery and accountability of local governments. Even for access to local governments, social networks and personal access play an important role. Atonu Rabbani (University of Dhaka, Bangladesh) finds that personal contacts and connections to local representatives have a significant impact on access to social pension benefits and other targeted programs. He confirms a strong mistargeting for social pensions, which is partly explained by representatives being unaware of eligibility criteria and allocating pensions to individuals with higher social capital, as well as higher income and assets. Similarly, Saheli Bose (Jadavpur University, India) shows that social connections, network strength and links with influential house- holds matter for the allocation of household public goods through local governments in West Bengal. In addition to the location, a household’s own network strength as well as those of its peers and neighbors play a significant role in obtaining a higher number of benefits from the local government. Panchali Banerjee (Jadavpur University, India) suggested that an increase in fiscal devolution to local governments led to a significant decrease in corruption in pub- lic service delivery, whereas an increase in tax devolution led to a significant increase in corruption in public service provisioning in India. Bushra Yasmin (Fatima Jinnah Women University, Pakistan) suggested that decentralization in Pakistan is not yet strong enough to reduce public corruption effectively and that the current system characterized by an absence of strict monitoring and evaluation of processes still allows for rent-seeking behavior. With continued urbanization in South Asia, the issues of urban planning and urban service delivery become more important. Bilesha Weeraratne (Institute of Policy Studies, Sri Lanka) highlighted that solid waste manage- ment at the local level in Sri Lanka is punctured with financial issues intertwined with low integrity of elected offi- cials, low transparency, and high corruption. Leena Bhattacharya (Indira Gandhi Institute of Development Research, India) argued that the inclusivity of slums in urban planning policies has done little to equalize living conditions and that the conditions in slums related to sanitation, water supply and housing remained dismal despite various initia- tives for equitable sustainable development, such as the Jawaharlal Nehru National Urban Rural Mission (JNNURM) and Smart City Mission. In a special lecture, Dr. Isher Judge Ahluwalia (Chairperson, Indian Council for Research on International Economics) discussed underrepresentation of urban voters and an inadequate power devolution to urban governments reflected, for example, by town planning remaining at the state level. In addition, few majors in India are elected and even those are not sufficiently empowered. Two of the most decentralized function in South Asia are education and health, two areas crucial to devel- opment outcomes. Mutawakkil Ahmad Abbasi (PIDE School of Public Policy, Pakistan) showed that devolution of education in Khyber Pakhtunkhwa resulted in better student performance, increased enrollment in government schools, and a significant move of students from private schools to public schools. Ashani Abayasekara (Institute of 56 Policy Studies, Sri Lanka) found evidence that additional funds for schools to be spent at their discretion had some benefits for teaching and learning outcomes, but that the effective use of these funds was negatively affected by delays in receiving them, difficulties in following the rules of the Ministry of Education, issues in procurement of South Asia goods, and a lack of clarity on what the funds could be used for. Gopi Khanal (National Natural Resource and Fiscal Economic Focus Commission, Nepal) found no evidence that more own revenue generation, in contrast to more equalization funds, Fall 2019 result in higher subnational spending on health and education. Megha Rao (Indian Institute of Management, India) identified a strong variation in the degree of decentralization in different health subsectors within Karnataka, with Making high levels of local discretion over funds for nutrition but abysmal financial jurisdiction over most other spheres of (de)centralization public health. In addition, she provided evidence for poor allocative efficiency across districts. work Researchers and policy experts seemed to agree that a further empowerment of local governments can improve local service delivery. One important tool mentioned frequently was the need to expand the taxing power of local governments. Dr. Ahsan Mansur emphasized that tax revenues account for only a small share of total resources of local governments, and that the top priority should be to strengthen the system of property taxation through proper and computerized land records, proper land and property valuation, and sensible tax rates. Participants also seemed to agree that elected local government representatives should oversee public service delivery and resource manage- ment instead of professional civil servants, as elected officials are accountable to the community. Dr. Selim Raihan pointed to the lack of political will in Bangladesh to empower local governments and an inclination towards deal- based fiscal transfers as opposed to rule-based ones. Dr. Ishrat Husain and Dr. Isher Ahluwalia pointed to the need for a complementary relationship between the federal and the local governments instead of an adversary one to ensure efficient service delivery. And Dr. Hans Timmer emphasized the importance of competition for efficient service deliv- ery, the need for structured land policies, and the issues of lacking data to assess progress effectively. Finally, everyone seemed concerned with growing elite capture in the region and many acknowledged that further decentralization may come into conflict with vested interests. Papers Presented: Ashani Abayasekara,  Nisha Arunatilake and Priyanka Jayewardena (all Institute for Policy Studies, Sri Lanka): Formula Based School Funding and Improving Education Service Delivery in Sri Lanka Atonu Rabbani(University of Dhaka, Bangladesh), Viola Asri, Kumar Biswas, Sebastian Fehrler, Urs Fischbacher (all University of Konstanz) and Katharina Michaelowa (University of Zurich): Why contacts matter: Local governance and the targeting of social pensions in Bangladesh Bilesha Weeraratne (  Institute for Policy Studies, Sri Lanka): Subnational Financing and Local Service Delivery: The Case of Solid Waste Disposal in Sri Lanka Bushra Yasmin  (Fatima Jinnah Women University, Pakistan): Role of Fiscal Decentralization in Corruption: An Empirical Evidence from Pakistan Gopi K Khanal ( National Natural Resources and Fiscal Commission, Nepal) and Bishal K. Chalise (Niti Foundation, Nepal): Role of local government finance in improving public service delivery in Nepal Leena Bhattacharya(Indira Gandhi Institute of Development Research, India): Has urban planning improved living conditions in slums? Evidence from million plus cities in India Manish Gupta and Pinaki Chakraborty (  both National Institute of Public Finance and Policy, India): Strengthening Local Government Finances and Service Delivery - Role of State Finance Commissions in India Megha Rao, A  rnab Mukherji and Hema Swaminathan (all Indian Institute of Management, India): Trends in Rural Fiscal Decentralization in Karnataka: A Focus on Public Health Monzur Hossain(Bangladesh Institute of Development Studies, Bangladesh) and Paritosh Kumar Roy (Institute of Statistical Research and Training, Bangladesh): Does democratic local governance facilitate local economic develop- ment? Evidence from Bangladesh Mutawakkil Ahmad Abbasi ( PIDE School of Public Policy, Pakistan), Iftikhar Ahmad (PIDE, Pakistan) and Tariq Rahim (University of Agriculture, Pakistan): Devolution of Education in Pakistan; A case study of Khyber Pakhtunkhwa Panchali Banerjee  (Jadavpur University, India) and Kausik K. Bhadra (University of California, San Diego): Fiscal Decentralization and Corruption in Public Service Delivery: Evidence from India Saheli Bose,  Vivekananda Mukherjee and Malabika Roy (all Jadavpur University, India): Social networks and political participation: Does ‘neighborhood effects’ influence the allocation of household public goods? 57 South Asia Economic Focus Fall 2019 SOUTH ASIA COUNTRY BRIEFS 58 South Asia Economic Focus Fall 2019 PHOTO BY: ALEKSANDAR TODOROVIC/SHUTTERSTOCK 59 Fall 2019 South Asia Economic Focus Afghanistan South Asia country briefs After a slowdown in 2018, growth is expected to slight- 2018 ly recover in 2019 with improved farming conditions. Population, million 36.4 However, the economic outlook remains highly suscepti- GDP, current USD billion 19.5 ble to elevated political and security risks from the pres- idential elections and the uncertain prospects of peace GDP per capita, current USD 536.9 talks. Any rapid decline in aid flows would pose risks to School enrollment, primary (percent gross)a 54.5 fiscal and external sustainability, as well as the capac- Life expectancy at birth, yearsa 64.0 ity to maintain basic services. Accelerated reform and Sources: WDI, World Bank, and official data. improved security conditions are critical to accelerate Note: (a) Most recent WDI value (2017) growth, mobilize economic potential, and help alleviate poverty from its current high levels. Figure 1: Real GDP growth and contributions. Percent, percentage points 4 3 2 1 0 -1 2016 2017 2018 2019(f) 2020 (f) 2021 (f) Agriculture Industry Services Real GDP growth Source: World Bank. Notes: (f) = forecast. Afghanistan’s fiscal year is the calendar year. Figure 2: Change in poverty rate. Percent, percentage points 70 60 50 40 30 20 10 0 2007-12 2013-17 2018-19 Change in GDP Change in poverty Percentage point change in poverty 60 Real GDP growth (average) Source: World Bank. Agricultural growth (average) Change in agricultural GDP Agriculture contribution to growth (average) Note: Afghanistan’s fiscal year is the calendar year. South Asia Economic Focus Fall 2019 Recent economic developments Outlook South Asia country briefs Afghanistan faced severe economic headwinds in 2018, Improved weather conditions are expected to help with the economy growing by an estimated 1.8 per- growth recover slightly in 2019, to around 2.5 percent. cent. Severe drought had a strong negative impact on Performance of the industry and services sectors will agriculture. Confidence deteriorated significantly in the likely remain subdued, given heightened political insta- context of elevated uncertainty around: i) the level of bility surrounding the upcoming presidential elections international security assistance; ii) the outcome of up- and elevated uncertainties over peace negotiations and coming presidential elections in September and the po- international security support. Over the medium term, tential for election-related violence; and iii) uncertain assuming a stable political transition following the prospects for peace negotiations with the Taliban. presidential election and a subsequent improvement in investor confidence, growth is expected to accelerate to Despite drought conditions, benign global food inflation 3 percent in 2020 and 3.5 percent in 2021. Inflation is and increased food imports kept inflation at a moder- expected to increase to 3.1 percent in 2019 and stabilize ate 0.6 percent (period average). Inflation accelerated in at around five percent in the medium term. 2019 to 4 percent y-o-y in July, reflecting higher prices of cereals and dry fruits. The trade deficit is expected to improve slightly over the medium term but to remain at around 30 percent of GDP. The trade deficit widened in 2018 to 35.3 percent of GDP The current account is however expected to deteriorate due to weaker export growth and a moderate increase gradually due to declining international grants, result- in imports. As a result, the current account surplus nar- ing in a deficit of around two percent of GDP by 2021-22. rowed to 0.6 percent of GDP. Grants, although lower International reserves will decline reflecting the deteri- than in 2017, financed almost all of the trade deficit. In orating external position but will still stay at a comfort- the first quarter of 2019, export growth accelerated to 14 able level at slightly less than eight months of imports. percent y-o-y while imports declined by 7 percent. A small fiscal deficit is expected in 2019. Revenue mo- The exchange rate depreciated by nine percent against bilization is expected to slow during the second half of the USD during 2018, mainly driven by the strengthen- the year, reflecting: i) limited room for further revenue ing of the USD against major currencies. However, the mobilization through measures such as the tax amnes- nominal appreciation of the afghani against other ma- ty scheme implemented in 2018; and ii) declining cus- jor trading partners’ currencies contributed to lower im- toms revenues in the context of political uncertainty ported inflation. Depreciation of the afghani against the and weakened governance. Both security and civilian USD further accelerated over the first half of 2019 amid grants are expected to decline substantially, leading to continued strengthening of the USD in the region and increased fiscal pressures. With limited access to debt declining confidence in the context of upcoming elec- financing, the overall debt-to-GDP ratio is expected to tions. Gross international reserves increased slightly to remain low at around 7 percent. USD 8.3 billion, enough to cover around 12 months of imports. Despite improved farming conditions and its potential for employment gains, poverty is unlikely to decline in Fiscal performance was strong in 2018, with a surplus of 2019, as growth rates and incomes in agriculture would 0.7 percent of GDP. Domestic revenue collection grew by need to increase substantially to realize welfare gains. 12 percent in 2018, reaching a record high of Afs 189.6 Stronger economic growth, accompanied by improved billion (13.4 percent of GDP). Higher revenues were sup- security conditions, would be needed to lift substantial ported by improved tax administration, with estimated numbers of Afghans out of poverty.  arrears collection of Afs 10.5 billion and a surge in non- tax revenues from state-owned enterprises. Revenue growth continued strong in the first half of 2019, with a 22 percent y-o-y increase over the first seven months. Risks and challenges Budget execution also improved, to 92 percent in 2018 from 83 percent in 2017. Progress continued in 2019, as Economic prospects are subject to substantial down- budget execution reached 47 percent (of the total year’s side risks. In the short-term, election-related political expenditure) by end-July, largely because of improved instability and disruptions to revenue collections could development budget execution. further undermine growth and fiscal stability. Over the medium-term, any rapid decline in international aid Poverty in 2018 is likely to have increased from an al- flows would result in difficult fiscal and external adjust- ready high poverty headcount of 55 percent in 2016-17. ments and undermine the capacity of the government The rate of economic growth substantially lagged pop- to maintain basic services. ulation growth, leading to declining per capita income. The 2018 drought adversely affected rural livelihoods and increased food insecurity, especially since 40 per- Political and security risks will remain elevated in the context of the recent cancellation of peace negotiations 61 cent of the population relies on agriculture as an income in September. While the future course of peace negoti- source. Reflecting widespread hardship, drought-in- ations is subject to substantial uncertainty, sustained South Asia duced displacement reached the record level of 298,582 and comprehensive improvement in security, achieved Economic Focus individuals, mainly to urban areas in adjacent provinces.  through political settlements, could boost confidence Fall 2019 South Asia and bring economic dividends. The impact of any polit- potential, especially in agriculture and extractives. country briefs ical settlement, however, will depend on whether peace Continued international assistance in security and de- can be sustained and how the broader post-conflict in- velopment is critical to preserve development gains stitutional environment evolves. achieved over the last seventeen years. A clear commit- ment to sustained support from international partners Reforms are required to both improve general invest- would help to reduce current levels of uncertainty and ment confidence and mobilize the country’s economic raise confidence and investment. Table 2: Macro poverty outlook indicators (annual percent change unless indicated otherwise) . 2016 2017 2018 2019 (f) 2020 (f) 2021 (f) Real GDP growth, at constant market prices 2.4 2.7 1.8 2.5 3.0 3.5 Private consumption -0.2 4.3 1.2 2.0 2.3 2.2 Government consumption 0.3 1.5 4.2 4.9 1.5 2.3 Gross fixed capital investment -6.0 6.4 0.5 3.1 2.8 2.9 Exports, goods and services -0.3 7.0 5.0 10.0 12.0 15.0 Imports, goods and services 25.8 8.0 1.0 3.6 2.5 2.4 Real GDP growth, at constant factor prices 2.0 2.2 1.8 2.5 3.0 3.5 Agriculture 6.0 3.8 0.8 4.5 3.0 5.5 Industry -0.8 0.4 2.5 2.5 3.0 3.0 Services 2.0 2.5 1.8 1.9 3.1 3.1 Inflation (consumer price index) 4.3 4.7 0.6 3.1 5.0 5.0 Current account balance (percent of GDP) 5.6 1.0 0.6 -1.6 -2.7 -2.4 Net foreign direct investment (percent of GDP) -0.1 -0.1 -0.1 -0.1 0.1 0.1 Fiscal balance (percent of GDP) 0.1 -0.5 0.6 -0.9 -0.6 -0.3 Debt (percent of GDP) 6.1 5.9 6.9 7.1 7.1 7.1 Primary balance (percent of GDP) 0.2 -0.4 1.7 0.0 0.3 0.6 Source: World Bank. Note: (f) = forecast. 62 South Asia Economic Focus Fall 2019 Bangladesh South Asia country briefs Rising exports and remittances underpinned manufac- 2018 turing-led growth in FY19, thus reducing the current Population, million 166.9 account deficit. Inflation, driven by growing domestic GDP, current USD billion 274.1 demand, was tempered by strong agricultural output. The fiscal deficit persisted as revenues continued to un- GDP per capita, current USD 1642.0 derperform. Over the medium-term, growth is expected International poverty rate (USD 1.9)a 14.8 to remain above 7 percent, supporting continued pover- Lower middle-income poverty rate (USD 3.2)a 52.9 ty reduction. Risks to the outlook include rising financial Gini indexa 32.4 sector vulnerability, fiscal pressures, and loss of external competitiveness as the real exchange rate appreciates. School enrollment, primary (percent gross)b 111.1 Life expectancy at birth, yearsb 72.8 Sources: WDI, World Bank, and official data. Notes: (a) Most recent value (2016), 2011 PPPs; (b) Most recent WDI value (2017). Figure 1: Real GDP growth and contributions. Percent, percentage points 10 8 6 4 2 0 -2 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 (e) 2020 (f) Final consumption Gross xed investment Net export Statistical discrepancy GDP growth rate Sources: Bangladesh Bureau of Statistics (BBS) and staff calculations. Notes: (e) = estimate; (f) = forecast. Figure 2: Actual and projected poverty rates and real GDP per capita. Poverty rate Real GDP per capita Percent Constant LCU 70 80000 60 70000 50 60000 50000 40 40000 30 30000 20 20000 10 10000 63 0 0 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 (e) 2020 (f) 2021 (f) International poverty rate Lower middle-income poverty rate Real GDP per capita Source: World Bank. South Asia Notes: (e) = estimate; (f) = forecast. Calculations based on SAR-POV harmonization, using 2010 -HIES and 2016-HIES. Actual data: 2016. Nowcast: 2017-2018. Forecast are from 2019 to 2021. Economic Focus Fall 2019 South Asia Recent economic developments Rohingya crisis, the July floods and recent fires in the country briefs slums of Dhaka underscore the continued vulnerability of large swathes of the population and demonstrate the Real GDP growth is estimated at 8.1 percent in FY2019, up case for a spatial approach to poverty reduction. from 7.9 percent in FY2018. On the demand side, growth was underpinned by higher net exports and increased consumption. Diversion of garment export orders from China boosted exports while record remittances of USD Outlook 16.4 billion in FY2019 boosted private consumption. On the supply side, growth was driven by industry and services. Bangladesh is likely to maintain GDP growth above 7 Manufacturing contributed 3.2 percentage points to growth percent, supported by strong macroeconomic funda- while services contributed another 3.3 percentage points. mentals, faster implementation of public investments in megaprojects, higher domestic demand aided by remit- Inflation decreased to 5.5 percent in FY2019 from tances and continued export growth as production shifts 5.8 percent the previous year, remaining within the further from China. Private investment may also rise in Bangladesh Bank (BB) target. The decline was primar- response to the ongoing Doing Business reforms and op- ily driven by falling rice prices. The BB Monetary Policy erationalization of the new economic zones. Inflation is announcement of July 2019 kept a 12.5 percent target projected to rise due to an increase in natural gas prices for broad money growth. The main policy interest rate and possible crop failures due to the recent floods. (repo rate) was unchanged at 6 percent, and the reverse repo rate remained at 4.75 percent. A modest increase in the current account deficit is likely, as strong investment-driven import growth is likely to Non-performing loans (NPLs), which rose to 11.9 percent outweigh export and remittance growth. FDI is expected of bank assets at end-March 2019 from 10.8 percent a to rise due to regulatory reforms, infrastructure develop- year ago, have constrained financing for private invest- ments and stable political conditions. Increased invest- ment. Recent policy changes have allowed defaulters to ment will support job creation and contribute to poverty re-schedule loans on easy terms, while the easing of loan reduction. Revenue underperformance together with in- classification standards has obscured accurate measure- creased spending on subsidies (on exports, remittances ment of NPLs. In addition, commercial lending rates have and LNG) and megaprojects is likely to increase the fiscal averaged 9.5 percent in FY2019 despite government ef- deficit in FY2020. Public debt was 32.8 percent of GDP at forts to reduce rates below 9 percent. As a result, private the end of FY2019, with a low risk of debt distress. sector credit grew only by 11.3 percent during FY2019. In contrast, credit to the public sector grew by 19.1 percent, driven by the resumption of central government borrow- ing from domestic banks to finance the deficit. Risks and challenges The current account deficit (CAD) declined from 3.5 per- Downside risks include financial sector vulnerability, cent of GDP in FY2018 to 1.7 percent in FY2019, supported reform reversals, fiscal pressures and loss of external by higher net exports and record remittances. A lower CAD competitiveness. Increased NPLs, enforcement of a blan- was coupled with increased foreign direct investment (FDI) ket ceiling on lending rates and increased government resulting in a small balance of payment surplus. With a borrowing from banks could crowd out credit to the pri- strengthening dollar, the BB sold USD 2.3 billion in the for- vate sector. Lack of progress in modernizing tax admin- eign exchange market in FY2019 to stabilize the taka-dollar istration may result in revenue shortfalls while higher exchange rate. As a result, the real effective exchange rate spending and donor fatigue in response to the Rohingya appreciated by 5.6 percent and gross foreign exchange re- crisis could add to fiscal pressures. serves declined from 6.2 months equivalent of imports at the end of FY2018 to 5.8 months at end-June 2019. External risks are balanced. Tariff escalation by the US against China may provide a further boost to exports The fiscal deficit remains high, estimated at 4.4 percent in the short run if Bangladesh can capture some of the of GDP in FY2019. Both revenues and spending underper- trade diversion. However, recession in European and US formed during the year. Revenues remained modest, at 10.3 export markets and appreciation of Bangladesh’s real percent of GDP due to a narrow tax base and limited im- exchange rate would adversely impact export demand plementation of administrative reforms. After a seven-year and remittances. delay, the implementation of a new VAT law began in July 2019 but with multiple rates for different types of goods and Moving forward, resolving fragile banks, accelerating on- services, the complexity of the VAT regime has increased. going business regulatory reforms, addressing exchange The Annual Development Program (ADP) execution rate rate overvaluation and deepening fiscal reforms are im- improved in FY2019 and domestic financing of the deficit mediate priorities. Key structural challenges that need to has shifted to bank borrowing, a departure from the prac- be addressed include reducing the infrastructure deficit, 64 tice of relying on expensive national savings certificates. enhancing human capital, improving urban manage- ment, and managing climate change risks. Emerging Strong growth in garment exports, buoyant remittanc- spatial inequalities warrant greater attention. Progress South Asia es, faster growth in nominal wages relative to inflation, in poverty reduction in the (north) west needs signifi- Economic Focus particularly in agriculture and services, and a decline cant acceleration through improvements in human de- Fall 2019 in food inflation have sustained poverty reduction. The velopment and structural transformation outcomes. Table 2: Macro poverty outlook indicators (annual percent change unless indicated otherwise). South Asia country briefs 2016 2017 2018 2019 (e) 2020 (f) 2021 (f) Real GDP growth, at constant market prices 7.1 7.3 7.9 8.1 7.2 7.3 Private consumption 3.0 7.4 11.0 5.4 6.2 6.4 Government consumption 8.4 7.8 15.4 8.0 8.1 10.6 Gross fixed capital investment 8.9 10.1 10.5 8.2 9.1 8.7 Exports, goods and services 2.2 -2.3 8.1 14.9 8.0 8.3 Imports, goods and services -7.1 2.9 27.0 4.8 8.1 8.5 Real GDP growth, at constant factor prices 7.2 7.2 7.9 8.3 7.2 7.3 Agriculture 2.8 3.0 4.2 3.5 3.0 3.1 Industry 11.1 10.2 12.1 13.0 9.0 10.0 Services 6.2 6.7 6.4 6.5 7.0 6.5 Inflation (consumer price index) 5.9 5.4 5.8 5.5 5.9 5.7 Current account balance (percent of GDP) 1.9 -0.5 -3.5 -1.7 -2.0 -2.2 Net foreign direct investment (percent of GDP) 0.6 0.7 0.6 0.8 0.8 0.8 Fiscal balance (percent of GDP) -3.7 -3.4 -4.7 -4.4 -4.8 -4.7 Debt (percent of GDP) 31.5 30.8 31.9 32.8 33.8 34.8 Primary balance (percent of GDP) -1.8 -1.6 -2.8 -2.5 -2.6 -2.4 International poverty rate (USD 1.9 in 2011 PPP)a,b 14.8 14.1 13.5 12.8 12.3 11.7 Lower middle-income poverty rate (USD 3.2 in 2011 PPP)a,b 52.9 51.7 50.6 49.3 48.3 47.2 Source: World Bank. Notes: (e) = estimate; (f) = forecast. (a) Calculations based on SAR-POV harmonization, using 2010-HIES and 2016-HIES. Actual data: 2016. Nowcast: 2017-2018. Forecast are from 2019 to 2021. (b) Projection using point-to-point elasticity (2010-2016) with pass-through = 1 based on GDP per capita in constant LCU. 65 South Asia Economic Focus Fall 2019 Bhutan South Asia country briefs Growth in Bhutan is forecast to accelerate to 7.4 per- 2018 cent in FY2020 with the commissioning of a new pow- Population, million 0.7 er plant and the completion of maintenance work in GDP, current USD billion 2.5 another. Tourism is emerging as a secondary driver of growth and, together with electricity exports, has con- GDP per capita, current USD 3397.8 tributed to a narrowing of the current account deficit. International poverty rate (USD 1.9)a 1.5 The primary risks to fiscal sustainability and growth Lower middle-income poverty rate (USD 3.2)a 12.0 stem from delays in completing hydro-megaprojects. Upper middle-income poverty rate (USD 5.5)a 38.6 Poverty measured using the USD 3.20 poverty line is projected to fall from 11.2 percent of the population in Gini indexa 37.4 2018 to 9.9 percent in 2019. School enrollment, primary (percent gross)b 92.6 Life expectancy at birth, yearsb 70.6 Sources: WDI, World Bank, and official data. Notes: (a) Most recent value (2017), 2011 PPPs; (b) Most recent WDI value (2017). Figure 1: Real GDP growth and non-hydro export growth. Percent Percent 12 30 10 20 8 10 6 0 4 2 -10 0 -20 2010/11 2011/12 2012/13 2013/14 2014/15 2015/16 2016/17 2017/18 2018/19 (e) 2019/20 (f) 2020/21 (f) Real GDP growth Real hydro export growth (RHS) Sources: Government of Bhutan and staff calculations. Notes: (e) = estimate; (f) = forecast. Figure 2: Actual and projected poverty rates and real GDP per capita. Poverty rate Real GDP per capita Percent Constant LCU 80 120000 100000 60 80000 40 60000 40000 20 20000 66 0 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 (e) 2020 (f) 2021 (f) 0 South Asia International poverty rate Lower middle-income poverty rate Upper middle-income poverty rate Real GDP per capita Economic Focus Source: World Bank. Fall 2019 Notes: (e) = estimate; (f) = forecast. Recent economic developments Overall unemployment is low, but high youth unem- South Asia ployment represents Bhutan’s challenge to create more country briefs and better jobs. Bhutan’s economy continues to be dominated by hy- dropower and its economic relationship with India. Growth is estimated to have rebounded to 5 percent in FY2019 after a deceleration to 4.6 percent in FY2018 Outlook due to maintenance and on-boarding delays on two major power plants (Tala and Mangdechhu). On the Growth is expected to edge up to 7.4 percent in FY2020 demand side, growth was underpinned by exports on the back of increased hydropower exports from the and consumption, reflecting the progress in hydro- newly on-boarded Mangdechhu plant. In the medium power maintenance and a procyclical public demand term, growth is expected to stabilize between 5 and 6 response to the associated income inflows. Services percent, supported by the initiation of new public in- remain the main driver of growth on the supply side, vestment projects under the 12th Five Year Plan and the where wholesale and retail trade has emerged as the construction of the Punatsangchhu I and II hydro plants. key contributing sub-sector. Inflation is expected to increase moderately in the near Approximately 80 percent of Bhutan’s imports are from term, following a strong growth outlook in India and India, and the Bhutanese Ngultrum is pegged to the firming food and fuel prices. Increased hydropower ex- Indian Rupee. Thus, inflation between the countries is ports are expected to narrow the current account defi- closely linked. FY2019 began with a slowdown in infla- cit further, to 10 and 7 percent of GDP in FY2020 and tion, hitting a low of just above 2 percent in July 2018 FY2021, respectively. before gradually increasing to 3.1 percent in April 2019 due to food and oil price dynamics. The exchange rate The fiscal outlook for Bhutan foresees a joint expan- has followed the appreciation of the Indian Rupee in sion of revenue and expenditure that will temporarily recent months, decreasing from 70.78 BTN per USD to widen the fiscal deficit to 4.9 percent of GDP by FY2021. about 69 BTN per USD in August 2019. Revenue growth is expected due to increased hydro revenue from the Mangdechhu power plant and the in- Developments in the hydropower sector also contrib- troduction of a green tax on fuel and a goods and ser- uted to a narrowing of the current account deficit to vices tax (GST). Expenditure growth in the short-term is 16.3 percent of GDP in FY2019, as exports from the expected to outpace revenue growth due to increased Tala plant expanded after the conclusion of main- investments in non-hydro projects as part of the 12th tenance work. At the same time, tourist arrivals in- Five-Year-Plan and public sector wage increases. creased by 8 percent in FY2019, contributing to the growth of service exports. The current account deficit Poverty reduction will continue at a modest pace in was primarily financed through capital inflows from the near term. The USD 3.20 poverty rate is expected to India. Foreign exchange reserves cover nearly 10.1 decline to 9.9 percent in 2019 and 8.7 percent in 2020. months of imports. Diversification into non-hydropower sectors remains the key challenge to accelerating job creation in non- Bhutan has maintained its course of sustainable public fi- farm sectors. Maximizing the growth potential of the nances, with the fiscal deficit estimated to have reduced to tourism sector could significantly contribute to jobs 2.1 percent of GDP in FY2019, from 4.6 percent two years and income growth, especially among the rural poor prior. This reflects revenue growth from reforms that in- and low-skilled. As poverty is almost exclusively rural, creased the corporate income tax base and changed the efforts to develop agribusinesses and increase agricul- valuation rules for sales tax on vehicles in FY2018, and a tural productivity will need to continue, by investing in slowdown in the initiation of new capital projects which the downstream value chain. limited spending growth. Government debt is estimated at 109.3 percent of GDP and is considered sustainable due to a special financing arrangement with India, which covers construction risks of hydropower plants and guar- Risks and challenges antees a return on surplus power purchases. Given its dependence on the hydropower sector, the Little progress was observed in the labor market as la- main risks to Bhutan’s growth and fiscal situation re- bor force participation fell from 65.7 percent in 2017 late to delays and maintenance outages of planned to 62.6 percent in 2018. Agriculture contributes only 10 and existing powerplants, which would delay the large percent to GDP but accounts for 54 percent of employ- expected export and revenue payoffs. Mitigating these ment. Working in agriculture is highly correlated with risks requires reforms that support the development of being poor: about 66 percent of poor household heads a private sector that can complement growth from hy- work in agriculture. Increases in agricultural exports dro, a diversification of public revenue sources towards and productivity helped reduce poverty in recent years. Extreme poverty at USD 1.90 per day is almost elimi- domestic non-hydro tax bases, and a smoothing of the trajectory of public spending to reduce its responsive- 67 nated, a laudable achievement. The USD 3.20 poverty ness to contemporaneous revenue increases. The labor rate (in 2011 PPP) is estimated to have declined from 12 force is young and increasingly educated, presenting a South Asia percent in 2017 to 11.2 percent in 2018. Hydropower is unique opportunity to reap the demographic dividends, Economic Focus capital intensive and contributes little to job creation. provided that job creation can keep pace. Fall 2019 South Asia Table 2: Macro poverty outlook indicators (annual percent change unless indicated otherwise). country briefs 2016 2017 2018 2019 (e) 2020 (f) 2021 (f) Real GDP growth, at constant market prices 7.4 6.3 4.6 5.0 7.4 5.9 Private consumption 3.0 -0.4 1.0 8.7 7.4 4.2 Government consumption 7.3 4.3 5.2 6.8 7.4 5.0 Gross fixed capital investment 11.5 5.5 -3.6 -3.9 -2.5 3.1 Exports, goods and services -1.4 6.8 -2.0 10.9 19.3 0.3 Imports, goods and services 12.3 -7.1 -7.1 6.9 -6.9 -4.6 Real GDP growth, at constant factor prices 7.8 6.0 4.2 5.9 8.8 6.2 Agriculture 4.4 3.6 4.1 4.4 4.7 4.4 Industry 7.6 4.6 1.1 3.9 8.8 3.8 Services 9.2 8.2 7.4 8.4 9.9 9.0 Inflation (consumer price index) 4.4 4.1 2.7 5.0 5.0 5.0 Current account balance (percent of GDP) -30.5 -22.9 -19.6 -16.3 -9.6 -7.3 Fiscal balance (percent of GDP) -1.9 -4.6 -3.1 -2.1 -2.7 -4.9 Debt (percent of GDP) 109.4 108.2 104.0 109.3 105.6 100.6 Primary balance (percent of GDP) -0.4 -3.4 -1.9 -1.2 -1.8 -3.6 International poverty rate (USD 1.9 in 2011 PPP)a,b .. 1.5 1.3 1.1 0.9 0.7 Lower middle-income poverty rate (USD 3.2 in 2011 PPP)a,b .. 12.0 11.2 10.0 8.7 7.8 Upper middle-income poverty rate (USD 5.5 in 2011 PPP)a,b .. 38.6 37.2 35.7 33.0 31.1 Source: World Bank. Notes: (e) = estimate; (f) = forecast. (a) Calculations based on SAR-POV harmonization, using 2017-BLSS. Actual data: 2017. Nowcast: 2018. Forecast are from 2019 to 2021. (b) Projection using neutral distribution (2017) with pass-through = 0.7 based on GDP per capita in constant LCU. 68 South Asia Economic Focus Fall 2019 India South Asia country briefs Growth is estimated to have moderated to 6.8 per- 2018 cent in FY18/19. A further deceleration is expected in Population, million 1356.6 FY19/20 and a gradual acceleration thereafter, subject GDP, current USD billion 2716.6 to significant risks related to external developments and the possibility of renewed fiscal and financial sec- GDP per capita, current USD 2002 tor stress. Poverty declined to an estimated 13.4 percent International poverty rate (USD1.9)a 21.6 in FY15/16 at the USD 1.9 international poverty line. Lower middle-income poverty rate (USD3.2)a 61.1 However, implementation challenges of indirect tax Upper middle-income poverty rate (USD5.5)a 87.3 reforms, weaknesses in the rural economy and a high youth unemployment rate in urban areas may have Gini indexa 35.4 moderated the pace of poverty reduction. Life expectancy at birth, yearsb 68.8 Sources: WDI, World Bank and official data. Notes: (a) Most recent value (2011), 2011 PPPs; (b) Most recent WDI value (2017). Figure 1: Real GDP growth and contributions to real GDP growth. Percent, percentage points 12 10 8 6 4 2 0 -2 -4 2011 2012 2013 2014 2015 2016 2017 2018 2019 (f) Final consumption Gross xed capital formation Net exports Other Real GDP growth Sources: Central Statistics Office and staff calculations. Note: (f) = forecast. Figure 2: Actual and projected poverty rates and real GDP per capita. Poverty rate Real GDP per capita Percent Constant LCU 100 140000 120000 80 100000 60 80000 40 60000 40000 20 20000 0 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 (f) 2020 (f) 2021 (f) 0 69 International poverty rate Lower middle-income poverty rate Upper middle-income poverty rate Real GDP per capita Source: World Bank. South Asia Notes: (f) = forecast. Economic Focus Fall 2019 South Asia Recent economic developments demand (as reflected in a sharp drop in sales of automo- country briefs biles), and credit from non-banking financial companies (NBFCs). Investment will benefit from the recent cut in Economic growth decelerated for the second consecu- the effective corporate tax rate for domestic companies tive year in FY18/19, to 6.8 percent (from 7.2 percent in the medium term, but also will continue to reflect in FY17/18). While industrial output growth increased financial sector weaknesses. Growth is expected to grad- to 6.9 percent, driven by a pick-up in manufacturing ually recover to 6.9 percent in FY20/21 and 7.2 percent and construction activity, growth in agriculture and ser- in FY21/22, as the cycle bottoms-out, rural demand vices moderated to 2.9 and 7.5 percent, respectively. In benefits from the effects of income support schemes, in- the first quarter of FY19/20, the economy experienced vestment responds to tax incentives and credit growth a significant and broad-based growth deceleration (to resumes. However, exports growth is expected to remain a 25-quarter low of 5 percent y-o-y) with a particularly modest, as trade wars and slow global growth depresses sharp deceleration in private consumption on the de- external demand. mand side, and a weakening of growth in both industry and services on the supply side. Going forward, monetary policy is expected to remain a key channel of countercyclical intervention, given be- Reflecting the below-trend economic momentum and nign price dynamics. While the authorities have shown persistently low food prices, headline inflation aver- steadfast commitment to fiscal prudence, the significant aged 3.4 percent in FY18/19 and remained well below growth deceleration as well as the corporate tax cuts un- the RBI’s mid-range target of 4 percent in the first half dertaken to counter it come with heightened risks of fis- FY19/20. This allowed the RBI to ease monetary policy cal slippage. If the center and the states hold on to their via a cumulative 135 basis point cut in the repo rate consolidation objectives, the general government deficit since January 2019 and a shift in the policy stance from is expected to reach 5.8 percent in FY20/21, but reaching “neutral” to “accommodative”. this target may prove challenging. The current account deficit widened to 2.1 percent of External balances are projected to remain stable and GDP in FY18/19 (from 1.8 percent in the previous year), the current account deficit to decline marginally to 2.0 mostly reflecting a deteriorating trade balance. On the percent of GDP in FY19/20, as import demand weakens financing side, significant capital outflows in the first and oil prices remain benign. half of the year (triggered by the tightening in global financial conditions) were followed by a sharp reversal If growth continues to be distribution-neutral, poverty from October 2018 onwards and a build-up of interna- measured at the USD 1.9 PPP/day line, is projected to tional reserves to USD 411.9 billion at the end of the fis- decline to 4.6 percent by FY21/22. Poverty measured at cal year (equivalent to 10 months of imports). Likewise, the USD 3.2 PPP/day line, is expected to fall from 50.4 while the rupee initially lost ground relative to the USD percent in FY14/15 to 31.2 percent in FY21/22. However, (12.1 percent depreciation between March and October the recent contraction in several high-frequency con- 2018), it appreciated by about 7 percent up to March sumption indicators, suggests that vulnerabilities for 2019. poorer households may have increased. By contrast, pol- icy measures to reduce stress in the rural economy, such The general government deficit is estimated to have as the government’s flagship income transfer scheme is widened by 0.2 percentage points, to 5.9 percent of GDP likely to have a positive impact on reducing poverty. in FY18/19. This is despite the central government im- proving its balance by 0.2 percentage points over the previous year. General government debt remained sta- ble and sustainable – being largely domestic and long Risks and challenges term—at around 67 percent of GDP. The main policy challenge is to address the sources of softening private consumption and the structural fac- Poverty has continued to decline, albeit possibly at a tors behind weak investment. This will require restor- slower pace than earlier. Between FY11/12 and FY15/16, ing the health of the financial sector through reforms of the poverty rate declined from 21.6 to 13.4 percent (USD public sector banks’ governance and a gradual strength- 1.9 PPP/day). Subsequently, disruptions brought about ening of the regulatory framework for NBFCs, while by the introduction of GST and demonetization, com- ensuring that solvent NBFCs retain access to adequate bined with stress in the rural economy and a high youth liquidity. It will also require efforts to contain fiscal slip- unemployment rate in urban areas, may have height- pages, as higher-than-expected public borrowings could ened the risks for the poorest households. put upward pressure on interest rates and potentially crowd-out the private sector. Outlook The main sources of risk include external shocks that 70 The significant slowdown in the first quarter of the result in tighter global financing conditions, and new NBFC defaults triggering a fresh round of financial sec- fiscal year and high frequency indicators thereafter tor stress. To mitigate these risks, the authorities will South Asia suggest that output growth will not exceed 6.0 percent need to ensure that there is adequate liquidity in the Economic Focus for the full fiscal year. Consumption is likely to remain financial system, while strengthening the regulatory Fall 2019 depressed by slow growth in rural incomes, domestic framework for NBFCs. Broad-based poverty reduction remains a major chal- force participation rate and high youth unemployment South Asia lenge, in particular with respect to (i) presently ex- present risks to sustaining the current rate of poverty country briefs cluded groups (such as women and scheduled tribes), reduction. Furthermore, outdated information on indi- and (ii) extending gains to a broader range of human cators of poverty and employment limit the scope of re- development outcomes related to health, nutrition, ed- liably correlating growth forecasts with projected rates ucation and gender. The persistently low female labor of poverty reduction. Table 2: Macro poverty outlook indicators (annual percent change unless indicated otherwise). 2016 2017 2018 2019 (f) 2020 (f) 2021 (f) Real GDP growth, at constant market prices 8.2 7.2 6.8 6.0 6.9 7.2 Private consumption 8.2 7.4 8.1 5.0 6.5 8.0 Government consumption 5.8 15.0 9.2 8.6 9.5 7.2 Gross fixed capital investment 8.3 9.3 10.0 8.5 8.2 8.5 Exports, goods and services 5.1 4.7 12.5 6.0 6.1 6.3 Imports, goods and services 4.4 17.6 15.4 5.9 8.1 8.9 Real GDP growth, at constant factor prices 7.9 6.9 6.6 5.8 6.7 7.0 Agriculture 6.3 5.0 2.9 2.5 3.3 3.3 Industry 7.7 5.9 6.9 5.2 6.7 7.3 Services 8.4 8.1 7.5 7.0 7.6 7.7 Inflation (consumer price index) 4.5 3.6 3.4 3.5 4.0 4.0 Current account balance (percent of GDP) -0.6 -1.8 -2.1 -2.0 -2.0 -2.3 Net foreign direct investment (percent of GDP) 1.6 1.1 1.1 1.2 1.4 1.6 Fiscal balance (percent of GDP) -6.9 -5.7 -5.9 -6.0 -5.8 -5.6 Debt (percent of GDP) 67.5 67.6 67.3 66.9 66.5 66.1 Primary balance (percent of GDP) -2.2 -0.9 -1.3 -1.3 -1.2 -0.9 International poverty rate (USD 1.9 in 2011 PPP)a,b 11.3 9.6 8.2 7.0 5.7 4.6 Lower middle-income poverty rate (USD 3.2 in 2011 PPP)a,b 46.6 43.6 40.7 38.0 34.7 31.2 Upper middle-income poverty rate (USD 5.5 in 2011 PPP)a,b 80.4 78.7 76.9 75.3 73.1 70.7 Source: World Bank. Notes: (f) = forecast. Fiscal projections do not take into account the possible effects of changes to corporate taxes announced on September 20, 2019. (a) Calculations based on SAR-POV harmonization, using 2011-NSS-SCH1. Actual data: 2011. Nowcast: 2012-2016. Forecast are from 2017 to 2021; (b) Projection using neutral distribution (2011) with pass-through = 0.733 (rural) and 0.559 (urban) based on HFCE in constant LCU for nowcast and GDP per capita in constant LCU for forecast. 71 South Asia Economic Focus Fall 2019 Maldives South Asia country briefs Real GDP growth is expected to decelerate from 6.7 per- 2018 cent in 2018 to 5.2 percent in 2019, due to a temporary Population, million 0.4 slowdown in construction. Over the medium-term, in- GDP, current USD billion 5.3 frastructure investment and expansion in the tourism sector will support growth. Efforts to rationalize recur- GDP per capita, current USD 11890 rent spending are needed, especially related to subsidies School enrollment, primary (percent gross)a 101.5 and health spending. Poverty is low, but youths continue Life expectancy at birth, yearsa 77.6 to face labor market challenges. Public debt is projected Sources: WDI, World Bank, and official data. to rise further, and international reserves are low. Note: (a) Most recent WDI value (2017). Figure 1: Number of beds in operation. Beds in operation 100 80 60 40 20 0 2012 2013 2014 2015 2016 2017 2018 Resorts (including Marinas) Hotels Guest houses Safari vessels Source: Ministry of Tourism. Figure 2: Poverty headcount ratio. Percent 14 12 10 8 6 4 2 0 Total Male' Atolls Source: National Bureau of Statistics. 72 South Asia Economic Focus Fall 2019 Recent economic developments international poverty line for upper middle-income South Asia countries of USD 5.50 a day. Large disparities in welfare country briefs and other socio-economic outcomes persist, with over Real GDP grew by 6.7 percent in 2018, on the back 90 percent of poor Maldivians living in Atolls. of strong performance in tourism, construction, and trade. Preliminary estimates indicate that growth slowed down to 3.1 percent y-o-y in Q1 2019, com- Labor market challenges prevail for youth and wom- pared to 13.1 percent in Q1 2018. Tourism contribut- en. Over a quarter of women are either unemployed ed approximately 2.4 percentage points, followed by or not looking for a job. Youth unemployment is high transportation and communication (0.7 percentage at 15.3 percent. The government is the top employ- points) and financial services (0.4 percentage points). er among Maldivians. Tourism and construction, the Construction declined by 23.1 percent. The opening main drivers of growth, rely mostly on foreign labor and of new resorts and infrastructure (the international male employment. About two thirds of Maldivians are airport and a connecting bridge) and housing proj- employed in jobs not related to tourism, suggesting a ects have fueled the construction sector over the misalignment between the drivers of growth and aspi- last 5 years, and the decline in construction reflects rations of jobseekers. Measures that foster private sec- the completion of some large infrastructure projects. tor job creation can help reduce pressure on the public Tourist arrivals and bed nights increased by 18.7 per- sector to create jobs. Public-private wage differentials cent and 14.7 percent y-o-y in H1 2019 respectively, and other benefits associated with public employment supported by these investments and the expansion in dis-incentivize young jobseekers from taking up private the guest houses’ sector. sector opportunities. The projected expansion in the young labor force means that private sector alternatives The overall CPI remained unchanged in H1 2019. This will be required. was mainly driven by policy changes that led to a de- crease in prices of staple food items and electricity. Price declines were more pronounced in the atolls, with the Outlook CPI falling on average by 1 percent in H1 2019, whereas the CPI for Male increased by 1.2 percent. Two major Real GDP growth is expected to slow down to 5.2 per- contributors to the decline in the Atolls were food and cent in 2019, with the completion of large infrastructure non-alcoholic beverages and housing and utilities. projects and the slow transition to new ones. Tourism is expected to continue to be the main driver of growth, The current account deficit is estimated to have widened with recent infrastructure investment helping relieve to 25.3 percent of GDP in 2018, from 21.7 percent in 2017. supply bottlenecks in the sector. Growth is expected to The trade deficit increased with the rapid expansion of rebound to 5.5 percent in 2020 as these new projects goods imports (24.0 percent y-o-y) linked to imports of pick up pace. The current account deficit is projected building materials. Services exports performed strongly to narrow over the medium-term, as investment-related (11.4 percent y-o-y growth) thanks to strong tourism re- imports gradually subside. ceipts. The current account deficit was financed mainly through debt flows and direct investment. The nominal The outlook assumes an increase in recurrent spend- exchange rate has remained at around 15.4 MVR/USD ing related to the new administration’s campaign under a de facto stabilized arrangement, whereas the pledges, including an extension of Aasandha coverage overnight deposit rate has been kept at 1.5 percent since for Maldivians living abroad, additional subsidies, and 2014. Gross official reserves decreased to USD 677.4 mil- an expansion of the university scholarship program. lion in June 2019 (2.9 months of imports), from USD Interest costs are also expected to gradually rise over the 726.4 million a year before. Usable reserves stood at USD medium-term, reflecting the country’s increased uptake 276 million. of non-concessional borrowing. The overall fiscal deficit is projected to increase initially before declining over Preliminary estimates suggest that the fiscal deficit the forecast period. Public debt is projected to rise over widened to 4.7 percent of GDP in 2018, from 3 percent the forecast period and peak soon after. in 2017. Public debt is estimated to have reached 59 percent of GDP in 2018. According to the World Bank- IMF debt sustainability analysis, Maldives remains at high risk of debt distress. The overall deficit narrowed Risks and challenges y-o-y in H1 2019. Revenue and grants increased by 5.3 percent y-o-y in H1 2019, with tax revenues growing Risks to the outlook are tilted to the downside. A down- by 6.3 percent y-o-y, non-tax revenues falling by 8.3 turn in the global economy or in China could impact percent y-o-y (driven partly by a decline in SOE div- Maldives’ tourism industry. An increase in oil prices idends) and grants more than doubling. Total spend- could affect the external account, given the country’s ing contracted by 8.0 percent y-o-y in H1 2019, largely heavy reliance on diesel imports. due to under-execution of the public sector investment program (-61.2 percent y-o-y). Recurrent expenditures One key challenge for the Maldives is to strike an ap- 73 increased by 8.8 percent y-o-y. propriate balance between making large investments needed to close existing infrastructure gaps –potentially South Asia The poverty headcount rate is low, at 8.2 percent us- allowing to boost tourism, increase resilience to climate Economic Focus ing the national poverty line and 6.6 percent using the change and ease constraints in service delivery— and Fall 2019 South Asia managing the rapid accumulation of public debt. Large disparities in welfare and other socio-econom- country briefs Containing recurrent spending and improving the ef- ic outcomes across regions are a cause for concern. ficiency of social spending are key areas that require Poverty rates vary widely across geographic areas, attention. The overall level of indebtedness is high and and Maldivians in the Southern Atoll are particular- reserves coverage is low. The large volume of external ly affected by poverty, with almost 1 in 5 being poor. loans and guarantees on non-concessional terms to These spatial disparities require a renewed effort in finance infrastructure projects represents significant economic and social inclusion of all regions across the risks. Maldives. Table 2: Macro poverty outlook indicators (annual percent change unless indicated otherwise). 2016 2017 2018 2019 (f ) 2020 (f) 2021 (f) Real GDP growth, at constant market prices 7.3 6.9 6.7 5.2 5.5 5.6 Real GDP growth, at constant factor prices 7.0 6.0 6.5 5.2 5.5 5.6 Agriculture 1.5 8.3 2.6 3.3 3.4 3.1 Industry 12.3 10.9 11.7 8.0 10.1 8.5 Services 6.7 5.2 6.0 5.0 5.0 5.3 Inflation (consumer price index) 0.5 2.8 -0.1 0.5 0.9 1.2 Current account balance (percent of GDP) -23.4 -21.7 -25.3 -18.5 -15.8 -11.7 Fiscal balance (percent of GDP) -9.9 -3.0 -4.7 -2.9 -4.9 -4.8 Debt (percent of GDP) 56.2 58.5 59.2 59.0 60.2 61.1 Primary balance (percent of GDP) -8.1 -1.5 -3.0 -1.2 -3.1 -3.0 Source: World Bank. Note: (f) = forecast. 74 South Asia Economic Focus Fall 2019 Nepal South Asia country briefs Growth in FY2019 is estimated to have increased to 2018 7.1 percent, driven by private investment and consump- Population, million 29.6 tion. With significant underspending of the budget, GDP, current USD billion 29.0 the fiscal deficit declined in FY2019 while the external deficit remained close to last year’s level. Poverty is de- GDP per capita, current USD 980 clining, but vulnerabilities remain. The medium-term International poverty rate (USD 1.9)a 15.0 outlook is supported by government consumption and Lower middle-income poverty rate (USD 3.2)a 50.8 investment. Risks to the outlook primarily arise from Upper middle-income poverty rate (USD 5.5)a 83.0 capacity constraints, especially at the subnational level, delays in reform implementation and vulnerability to Gini indexa 32.8 natural disasters. School enrollment, primary (percent gross)b 134.1 Life expectancy at birth, yearsb 70.6 Sources: WDI, World Bank, and official data. Notes: (a) Most recent value (2010), 2011 PPPs.; (b) Most recent WDI value (2017). Figure 1: Investments continue to drive real GDP growth. Percent, percentage points Percent 14 9 12 8 10 7 8 6 6 5 4 4 2 3 0 -2 2 -4 1 -6 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 (e) 0 Private consumption Government investment Private investment Real GDP growth (RHS) Sources: Central Bureau of Statistics and staff calculations. Notes: (e) = estimate. Nepal’s fiscal year runs from July 16th to July 15th. Figure 2: Actual and projected poverty rates and real GDP per capita. Poverty rate Real GDP per capita Percent Constant LCU 100 40000 35000 80 30000 60 25000 20000 40 15000 10000 20 0 5000 0 75 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 (e) 2020 (f) 2021 (f) International poverty rate Lower middle-income poverty rate Upper middle-income poverty rate Real GDP per capita South Asia Source: World Bank. Economic Focus Notes: (e) = estimate; (f) = forecast. Fall 2019 South Asia Recent economic developments Outlook country briefs GDP growth is estimated at 7.1 percent in FY2019, GDP growth is projected to average 6.5 percent a year marking three consecutive years of over six percent over the medium term. On the supply side, growth will growth. On the supply side, growth was driven by the be driven by services, underpinned by steady remit- services sector, particularly, retail, hotel, and restaurant tance inflows and high tourist arrivals. The latter will subsectors, which received a boost from an uptick in be supported by the Visit Nepal 2020 program, the com- tourist arrivals and remittance-fueled private con- pletion of the second international airport, and the con- sumption. The agricultural sector contributed an ad- struction of big hotels. On the demand side, investment ditional 1.6 percentage points to growth, supported by and government consumption are expected to be the good monsoons and increased commercialization and main drivers of growth. Public consumption will be sup- availability of agricultural inputs. On the demand side, ported by increased spending on salary and goods and private investment and consumption were the main services. In addition, efforts aimed at building capaci- drivers, contributing 4.9 percentage points each to over- ty at the subnational levels and the implementation of all growth. performance-based contracts at the central level will also result in higher overall public spending. The imple- Given good agricultural production and the peg to the mentation of the 2019 national work plan to minimize Indian rupee, inflation in FY2019 (4.5 percent) remained the trade deficit along with investment-related initia- below target (5.5 percent). Credit growth reached 19.3 tives, such as establishing a one-stop service center, will percent, exceeding deposits growth, and led to a rise in support private investment. However, increased import the banking sector’s credit-to-core capital plus deposit tariffs on selected agricultural products and consum- ratio (which at 75.2 percent remained just below the 80 er goods are expected to lower the growth of private percent regulatory limit). consumption. The current account deficit remained high at 7.7 per- Inflation is expected to increase marginally with higher cent of GDP in FY2019, driven by a persistent trade public sector wages, increases in import duties on agri- deficit. The latter narrowed marginally from 37.5 per- cultural and industrial goods, and the removal of val- cent to 37.1 percent of GDP as goods import growth ue-added-tax exemptions on some intermediate goods slowed down, reflecting lower import demand for and services. On the other hand, the regular supply of industrial supplies (such as cement clinker) and cap- electricity at low cost and low inflation in India will off- ital goods. A part of the trade deficit was financed by set some of the increase. remittances, which were sustained at the same level as last year (25 percent of GDP). Remittance inflows The current account deficit is likely to moderate to 5.9 in FY2019 were supported by the depreciation of the percent of GDP by FY2021, as spending on federalism-re- Nepali rupee against the USD and the increased use lated infrastructure and post-earthquake reconstruction of formal remittance channels. The remaining exter- tapers down and the government starts implementing nal deficit was financed through borrowings and by a work plan for encouraging export-oriented and im- drawing down foreign exchange reserves, which fell to port-substituting industries. Some increase in exports, USD 9.5 billion in July 2019, equivalent to 7.8 months particularly of hydroelectricity, is anticipated in the next of imports. few years, but broader growth in exports will happen only in the longer term as structural reforms start yield- Delays in the enactment of Federal, Provincial, and ing results. Remittances as a share of GDP are expected Local Civil Service Acts and in the establishment of pro- to stabilize at 25 percent over the medium term. vincial civil service commissions adversely impacted the hiring of new staff at the subnational levels. These Government spending is expected to increase to 29.7 delays together with the limited technical capacity of percent of GDP by FY2021 due to salary increases, high- existing staff led to significant underspending of the er social security spending, and a pick-up in capital budget, reducing the fiscal deficit from 6.7 percent of investments. Revenues are also likely to increase with GDP in FY2018 to 1.9 percent in FY2019. With public higher import duties but will lag expenditure growth. debt at 30 percent of GDP, Nepal remains at low risk of As a result, the fiscal deficit is projected to increase to debt distress. 3.3 percent of GDP and public debt is likely to reach 31 percent of GDP by FY2021. The estimated poverty headcount ratio (at the USD 1.90 per person per day international poverty line) was 9.3 percent in 2018, down from 15 percent in 2010. At a higher line of USD 3.20 a day for Nepal, 41 per- Risks and challenges cent of the population was poor in 2018, a 10 percent- The outlook is subject to significant risks and challenges. age-point decrease from 2010. Despite the declining Persistently high trade deficits would raise the risks to 76 poverty trend, vulnerability remains high in Nepal. Almost 10 million people, or close to 32 percent of the population, are estimated to live on incomes between the external sector, especially if remittance growth slows down due to geopolitical tensions in migrant receiving countries and uncertain oil prices. Lower remittances South Asia USD 1.90 and USD 3.20 a day (2018). Climate related could also impact the liquidity in the financial system. Economic Focus shocks, such as floods and earthquakes, further in- Climate-related natural disasters could adversely impact Fall 2019 crease vulnerability. agricultural production and infrastructure, potentially trade; (b) enforce the implementation of policies and South Asia reversing recent gains in poverty reduction. And capaci- regulations related to climate and disaster risk reduc- country briefs ty challenges, particularly at the subnational level, could tion and management; (c) strengthen the preparation of manifest through the under execution of provincial and the medium-term expenditure framework at the subna- local budgets, impacting service delivery. Therefore, ef- tional level; and (d) adopt a legal framework for hiring forts are needed to: (a) raise export competitiveness and staff at the subnational level and build capacity of all establish a legal framework for cross-border electricity staff. Table 2: Macro poverty outlook indicators (annual percent change unless indicated otherwise). 2016 2017 2018 2019 (e) 2020 (f) 2021 (f) Real GDP growth, at constant market prices 0.6 8.2 6.7 7.1 6.4 6.5 Private consumption -0.7 2.6 2.5 6.5 0.9 1.5 Government consumption -0.4 10.5 13.4 8.3 44.4 7.2 Gross fixed capital investment -12.3 44.3 18.1 14.3 7.2 11.4 Exports, goods and services -13.7 11.3 7.8 7.9 8.4 15.0 Imports, goods and services 2.8 27.2 19.0 17.9 7.7 5.5 Real GDP growth, at constant factor prices 0.3 7.7 6.1 6.7 6.4 6.5 Agriculture 0.2 5.2 2.8 5.0 4.0 4.0 Industry -6.4 12.4 9.6 8.1 8.3 8.8 Services 2.4 8.1 7.2 7.3 7.2 7.2 Inflation (consumer price index) 9.9 4.5 4.2 4.5 5.0 5.0 Current account balance (percent of GDP) 6.2 -0.4 -8.2 -7.7 -6.8 -5.9 Fiscal balance (percent of GDP) 1.4 -3.1 -6.7 -1.9 -2.6 -3.3 Debt (percent of GDP) 27.9 26.1 30.2 30.1 30.3 31.4 Primary balance (percent of GDP) 1.8 -2.7 -6.1 -1.3 -1.8 -2.5 International poverty rate (USD 1.9 in 2011 PPP)a,b 11.7 9.9 9.3 8.0 7.0 6.6 Lower middle-income poverty rate (USD 3.2 in 2011 PPP)a,b 45.2 43.1 41.0 39.2 37.4 36.0 Upper middle-income poverty rate (USD 5.5 in 2011 PPP)a,b 79.9 78.4 77.1 75.5 74.1 72.7 Source: World Bank. Notes: (e) = estimate; (f) = forecast. (a) Calculations based on SAR-POV harmonization, using 2010-LSS-III. Actual data: 2010. Nowcast: 2011-2018. Forecast are from 2019 to 2021; (b) Projection using neutral distribution (2010) with pass-through = 0.5 based on GDP per capita in constant LCU. 77 South Asia Economic Focus Fall 2019 Pakistan  South Asia country briefs Pakistan’s economy is slowing as the country passes 2018 through yet another macroeconomic crisis with high Population, million 201.0 twin deficits and low international reserves. With an IMF GDP, current USD billion 314.6 Extended Fund Facility supported stabilization program in place, growth is expected to remain low in the near- GDP per capita, current USD 1565 term. The medium-term growth outlook hinges upon International poverty rate (USD 1.9)a 3.9 the country’s ability to implement necessary structural Lower middle-income poverty rate (USD 3.2)a 34.7 reforms to boost competitiveness and achieve sustained Upper middle-income poverty rate (USD 5.5)a 75.4 growth. Progress in poverty reduction is expected to be limited during the macroeconomic adjustment period. Gini indexa 33.5 School enrollment, primary (percent gross)b 95.9 Life expectancy at birth, yearsb 66.629 Sources: WDI, World Bank, and official data. Notes: (a) Most recent value (2015), 2011 PPPs.; (b) Most recent WDI value (2017). Figure 1: Fiscal deficit and public debt. Percent of GDP Percent of GDP 10 100 8 75 5 50 3 25 0 0 2012 2013 2014 2015 2016 2017 2018 2019 (e) Public debt (RHS) External public debt (RHS) Fiscal de cit Sources: Ministry of Finance and State Bank of Pakistan. Notes: (e)=estimate. Pakistan reports data on fiscal year (FY) basis. The fiscal year runs from July 1 through June 30 in Pakistan. Figure 2: Actual and projected poverty rates and real GDP per capita. Poverty rate Real GDP per capita Percent Constant LCU 100 80000 70000 80 60000 50000 60 40000 40 30000 20000 78 20 10000 0 0 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 (e) 2020 (f) 2021 (f) South Asia International poverty rate Lower middle-income poverty rate Upper middle-income poverty rate Real GDP per capita Economic Focus Source: World Bank. Fall 2019 Notes: (e) = estimate; (f) = forecast. Recent economic developments coordination between federating units, an upward South Asia adjustment in energy prices and tighter monetary country briefs policy. GDP growth (at factor cost) decelerated to 3.3 percent in FY19—2.2 percentage points lower than FY18—as gradual policy adjustments to tackle macroeconomic imbalances started to take effect. These adjustments Outlook included a tightened monetary stance, cuts in public sector development expenditures, and enhanced fo- Growth is projected to decelerate to 2.4 percent in FY20 cus on higher tax collections. As a result, large scale with continued fiscal consolidation and a tight mone- manufacturing, which accounts for half of overall in- tary policy stance. The IMF adjustment program en- dustrial output, contracted by 3.6 percent in FY19. The tails a rebalancing from domestic to external demand. services sector, which contributes over 60 percent to Growth is expected to recover slowly, to 3.0 percent in total output, decelerated to 4.7 percent in FY19 com- FY21, as macroeconomic conditions improve and exter- pared to 6.2 percent last year. In agriculture, major nal demand picks up on the back of structural reforms crops registered a 6.6 percent decline in production and increased competitiveness. This recovery is condi- due to adverse weather conditions. On the demand tional on relatively stable global markets, a decline in in- side, policy adjustments slowed private consump- ternational oil prices and reduced political and security tion growth from 6.8 percent in FY18 to 4.1 percent risks. Inflation is expected to rise in FY20 to 13.0 percent, in FY19, while investment contracted by 8.9 percent and afterwards start declining gradually. The increase in in FY19 as compared to a growth of 7.1 percent in prices will be driven by the second-round impact of ex- FY18. Headline inflation increased to 7.3 percent in change-rate pass-through to domestic prices. Pakistan’s FY19 compared with 3.9 percent in FY18, primarily commercial banks remain well capitalized. However, because of a cumulative depreciation of 25.5 percent rising public sector demand for credit (mainly central of the PKR against the USD during the fiscal year. To government borrowing) and rising interest rates are ex- check these inflationary pressures, the State Bank of pected to crowd out private credit in the near-term. Pakistan (SBP) gradually raised its policy rate by 675 bps to 13.25 by July 2019. The current account deficit is expected to decline to 2.6 percent of GDP in FY20 and further to 2.2 percent of The current account deficit (CAD) narrowed to USD 13.5 GDP in FY21, as increased exchange-rate flexibility will billion (4.8 percent of GDP) in FY19 compared to USD support a modest recovery in exports and rationaliza- 19.9 billion (6.3 percent of GDP) in FY18. The decline in tion of imports. The consolidated fiscal deficit (includ- the CAD was primarily driven by a fall in imports (goods ing grants) is projected to reach 7.5 percent of GDP in imports declined by 7.4 percent while services imports FY20 and remain elevated at 6.2 percent of GDP in FY21. fell by 14.9 percent). Exports, however, did not respond The public debt-to-GDP ratio is expected to remain high to the depreciation in FY19, as regaining competitive- in FY21 at 80.8 percent, increasing Pakistan’s exposure ness after extended periods of an overvalued exchange to debt-related shocks. Fiscal consolidation across the rate will take time. Financial flows increased due to federation will be needed for the public debt to decline, substantial liquidity injections from Saudi Arabia, UAE but the debt-to GDP ratio is not expected to fall below and China. However, these injections did not stem the 70 percent of GDP - the debt burden benchmark for declining (net) foreign reserves, which fell from USD 9.8 high risk emerging markets – over the medium term. billion (1.7 months of import cover) in end-June 2018 to Pakistan’s debt vulnerabilities will remain high due to USD 7.3 billion in end-June 2019 (1.5 months of import large foreign currency debt amortizations and sizeable cover). refinancing of short-term domestic debt. Fiscal performance in FY19 deteriorated substantially Progress in poverty reduction, which was uninterrupt- due to revenue underperformance and higher inter- ed since 2001, is expected to stall during the macroeco- est payments. The consolidated fiscal deficit (including nomic adjustment period, due to decelerating growth grants) stood at 8.8 percent of GDP—2.4 percentage and higher inflation rates. The poverty headcount, mea- points higher than FY18. Tax revenues almost stagnated sured using the USD 1.90 per person per day interna- at last year’s level and non-tax revenues declined by 44 tional poverty line, is projected to remain at the FY19 percent as the exchange rate depreciation reduced SBP level (3.1 percent). Poverty measured using the USD 3.2 profits, resulting in lower transfers to the government. line is expected to decline from 31.4 percent last year to As a result, overall revenues contracted by 6.3 percent. 31.2 percent in FY20, while poverty measured using the With fiscal slippages and the large exchange rate depre- USD 5.5 poverty line is projected at 72.5 percent in FY20, ciation, Pakistan’s public debt rose sharply to 86.5 per- compared to 72.6 in FY19. cent of GDP by end-June 2019 compared to 73.0 percent in FY18. Risks and challenges To restore macroeconomic stability, the government signed a 39-month USD 6 billion Extended Fund 79 Facility (EFF) program with the IMF in July 2019. Key Economic policies over the past few years have result- steps initiated under the EFF include a shift towards ed in increased debt levels and erosion of fiscal and South Asia a market-based exchange rate regime, expenditure external buffers, limiting the economy’s ability to ab- Economic Focus consolidation, increased revenue collections, stronger sorb shocks. The country needs to restore these buffers, Fall 2019 South Asia especially since (i) turbulence in global financial mar- structural reforms. Vulnerable households’ ability to country briefs kets could affect Pakistan’s access to private external weather the economic impact of the crisis will depend financing; and (ii) the weakening global economy and critically on the inclusiveness of growth, food and non- rising trade tensions could dampen external demand. food inflation, and the resilience of sectors relevant for The main domestic risk emerges from potential diffi- their employment (agriculture, construction and whole- culties in implementing the necessary adjustments and sale/retail trade). Table 2: Macro poverty outlook indicators (annual percent change unless indicated otherwise). 2016 2017 2018 2019 (e) 2020 (f) 2021 (f) Real GDP growth, at constant market prices 5.5 5.6 5.8 3.3 2.4 3.0 Private consumption 7.6 8.5 6.8 4.1 0.7 1.5 Government consumption 8.2 5.3 8.6 10.0 1.5 2.7 Gross fixed capital investment 7.5 10.3 7.1 -8.9 1.7 5.9 Exports, goods and services -1.6 -0.6 10.4 13.2 10.8 8.7 Imports, goods and services 16.0 21.2 15.8 5.8 -1.2 1.5 Real GDP growth, at constant factor prices 4.6 5.2 5.5 3.3 2.4 3.0 Agriculture 0.2 2.2 3.9 0.8 1.8 2.0 Industry 5.7 4.6 4.9 1.4 1.2 2.1 Services 5.7 6.5 6.2 4.7 3.0 3.7 Inflation (consumer price index) 2.9 4.2 3.9 7.3 13.0 8.3 Current account balance (percent of GDP) -1.7 -4.1 -6.3 -4.8 -2.6 -2.2 Net Foreign Direct Investment (percent of GDP) 0.8 0.9 1.1 0.6 0.8 1.0 Fiscal balance (percent of GDP) -4.5 -5.8 -6.4 -8.8 -7.5 -6.2 Debt (percent of GDP) 68.7 68.0 73.0 86.5 82.9 80.8 Primary balance (percent of GDP) -0.2 -1.5 -2.1 -3.4 -0.8 0.1 International poverty rate (USD 1.9 in 2011 PPP)a,b 3.7 3.5 3.2 3.1 3.1 3.0 Lower middle-income poverty rate (USD 3.2 in 2011 PPP)a,b 33.7 32.8 31.8 31.4 31.2 30.9 Upper middle-income poverty rate (USD 5.5 in 2011 PPP)a,b 74.6 73.8 72.9 72.6 72.5 72.2 Source: World Bank. Notes: (e) = estimate; (f) = forecast. (a) Calculations based on SAR-POV harmonization, using 2010-PSLM and 2015-PSLM. Actual data: 2015. Nowcast: 2016-2018. Forecast are from 2019 to 2021; (b) Projection using annualized elasticity (2010-2015) with pass-through = 0.25 based on GDP per capita in constant LCU. 80 South Asia Economic Focus Fall 2019 Sri Lanka South Asia country briefs The April terrorist attacks heightened macroeconomic 2018 challenges. Growth is expected to decelerate to 2.7 per- Population, million 21.7 cent in 2019. Reduced tourism receipts will exert pres- GDP, current USD billion 88.9 sure on external accounts, despite reduced import de- mand. Fiscal balances will deteriorate amid contracting GDP per capita, current USD 4104 revenues. Large refinancing needs, weak fiscal buffers International poverty rate (USD 1.9)a 0.8 and high debt make the country vulnerable to rollover Lower middle-income poverty rate (USD 3.2)a 10.1 risks. A slowdown in economic activity will constrain Upper middle-income poverty rate (USD 5.5)a 40.4 job creation and income growth, and the pace of pover- ty reduction. The USD 5.50 poverty rate is projected at Gini indexa 39.8 36.1 percent in 2019. School enrollment, primary (percent gross)b 101.9 Life expectancy at birth, yearsb 75.5 Sources: WDI, Macro Poverty Outlook, and official data. Notes: (a) Most recent value (2016), 2011 PPPs.; (b) Most recent WDI value (2017). Figure 1: Real GDP growth and contributions. Percent, y-o-y 10 8 6 4 2 0 2011 2012 2013 2014 2015 2016 2017 2018 2019 (f) -2 Agriculture Industry Services Net taxes Real GDP growth Source: Department of Census and Statistics and staff calculations. Note: (f) = forecast Figure 2: Actual and projected poverty rates and real GDP per capita. Poverty rate Real GDP per capita Percent Constant LCU 80 600000 500000 60 400000 40 300000 200000 20 81 100000 0 0 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 (f) 2020 (f) 2021 (f) International poverty rate Lower middle-income poverty rate Upper middle-income pov. rate Real GDP per capita South Asia Source: World Bank. Economic Focus Note: (f) = forecast. Fall 2019 South Asia Recent economic developments to poverty reduction. However, low female labor force country briefs participation and the low quality of private sector jobs, including high rates of informal employment, remain Sri Lanka’s economic growth continues to be adversely persistent issues. While overall unemployment was rela- affected by shocks. An unprecedented political contro- tively low at 4.4 percent of the labor force in 2018, youth versy last year led to a subdued growth of 3.2 percent unemployment recorded 21.4 percent, indicating diffi- for 2018. Growth rebounded in the first quarter of 2019 culties with entering the labor market. to 3.7 percent (year-on-year) thanks to relatively benign weather that revived agriculture and related industry sectors. However, the aftermath of the April terrorist at- tacks that killed over 250 people, including tourists, is Outlook expected to take a toll on the economy in the rest of 2019 with decelerating private consumption and investment. The medium-term outlook is subject to the country’s Annual average inflation remained low at 4.2 percent ability to ensure political stability and a return to by end-June 2019, with the moderation of food prices normalcy. Growth for 2019 is expected at 2.7 percent, owing to improved weather, despite the passthrough of as many important economic sectors show relatively currency depreciation in 2018. Citing sluggish growth weak performance. In the medium-term, the economy and favorable inflation expectations, the Central Bank is expected to recover from the disruptions in 2019, and reduced the policy rate corridor by 100 basis points growth is expected to accelerate towards 4.0 percent, in the first eight months of 2019. Nevertheless, the de- gradually closing the output gap. The drivers of the re- mand for private credit has shown limited responsive- covery are anticipated to be investment and exports, as ness amid sluggish economic activity. performance in the tourism sector improves and uncer- tainty is resolved after the elections are held. On the external front, the trade deficit narrowed in the first half of 2019 due to a contraction of motor vehicles The current account deficit is expected to narrow mar- and gold imports, which experienced an import surge ginally in 2019 compared to 2018, thanks to weak im- in the corresponding period of 2018, and increased port demand. Significantly large debt creating flows garments exports. This is expected to have reduced will be required to close the external financing gap the current account deficit for the first half of 2019 when the impact of past one-off FDIs wane. Gross of- despite decelerating tourism receipts and remittances. ficial reserves are expected to remain relatively low, as Notwithstanding historically high debt repayments, the country faces large debt repayments. Provided that official reserves increased thanks to the issuance of the revenue-led fiscal consolidation continues, prima- USD 4.4 billion in sovereign bonds. While reserve cover ry surpluses will return in 2020, which will help bring reached a 12-month high at 5.2 months of merchan- debt to a sustainable path. In the absence of currency dise imports by end-June 2019, external vulnerability depreciation, the debt-to-GDP ratio will stabilize in 2019. remains high with relatively high short-term liabili- The successful completion of the IMF supported reform ties. The rupee remained broadly stable against the program and the continuation of the reform agenda be- USD, with a year-to-date appreciation of 3.4 percent by yond mid-2020 will be critical for macroeconomic sta- end-June. bility and sustainability. The slowdown in economic ac- tivities, especially tourism, trade, transport, construction Fiscal balances deteriorated in the first half of 2019. and other SME businesses, is expected to constrain jobs Tax revenues fell short of expectations due to weak and wage growth in the near-term. The decrease in re- collection of excise taxes from motor vehicles, petro- mittances will also lead to lower contributions to house- leum products and import taxes, amid slow growth. hold income. As a result, the pace of poverty reduction Expenditures increased due to relief packages after the is expected to slow down, with poverty measured using April attacks, settling of arrears and the implementa- the USD 5.50 poverty line projected at 36.1 percent in tion of budget proposals that increase recurrent expen- 2019. ditures in an election year. As a result, the annualized budget deficit is expected to have increased above 6.0 percent of GDP for the first half of 2019. Sri Lanka’s debt to GDP ratio of 82.9 percent and annual gross financ- Risks and challenges ing needs of approximately 18 percent of GDP remain among the highest among middle-income countries. In Risks are tilted to the downside. On the domestic front, addition, the debt portfolio is subject to important risks, a challenging political environment, delays or reversals with over 54 percent of the debt denominated in for- in efforts to strengthen revenues, and a slower than eign exchange. expected recovery of some key economic sectors repre- sent important risks. Mitigating these risks will be key The USD 3.20 poverty rate (used for lower middle-in- to creating private sector jobs and accelerating poverty come countries) declined from 9.5 percent in 2017 to 8.7 reduction. Externally, while Sri Lanka has raised enough 82 percent in 2018. The USD 5.50 poverty rate (used for up- per middle-income countries) declined from 39 percent foreign currency funds to manage immediate debt re- payments, continued large refinancing requirements, to 37.3 percent during the same period. Tourism and weak fiscal buffers and high indebtedness make the South Asia related service sectors have increasingly become a via- economy vulnerable to uncertain global financial con- Economic Focus ble source of employment and income for the poor and ditions. Priority reforms include: (a) continuing fiscal Fall 2019 low-skilled in rural areas in recent years, contributing consolidation by broadening the tax base and aligning spending with priorities; (b) shifting to a private invest- increasing labor force participation, encouraging longer South Asia ment-tradable sector-led growth model by improving working lives and investing in skills to improve produc- country briefs trade, investment, innovation and the business environ- tivity; and (e) mitigating the impact of reforms on the ment; (c) improving governance and SOE performance; poor and vulnerable with well-targeted social protec- (d) addressing the impact of an aging workforce by tion spending. Table 2: Macro poverty outlook indicators (annual percent change unless indicated otherwise). 2016 2017 2018 2019 (f) 2020 (f) 2021 (f) Real GDP growth, at constant market prices 4.5 3.4 3.2 2.7 3.3 3.7 Private consumption 7.4 2.5 2.3 2.6 3.4 3.8 Government consumption 2.3 -5.4 -5.5 4.8 3.5 4.5 Gross fixed capital investment 7.8 5.9 -1.3 -2.4 3.3 4.9 Exports, goods and services -0.7 7.6 0.5 1.1 7.7 4.8 Imports, goods and services 7.9 7.1 1.8 -3.3 4.9 4.5 Real GDP growth, at constant factor prices 4.3 3.4 3.6 2.7 3.3 3.7 Agriculture -3.7 -0.4 4.8 3.4 3.4 3.4 Industry 5.7 4.1 0.9 2.6 3.2 3.5 Services 4.8 3.6 4.7 2.6 3.4 3.8 Inflation (consumer price index) 4.0 6.6 4.3 4.5 4.5 4.5 Current account balance (percent of GDP) -2.1 -2.6 -3.2 -2.7 -2.6 -2.7 Net Foreign Direct Investment (percent of GDP) 0.8 1.5 1.7 1.1 1.3 1.3 Fiscal balance (percent of GDP) -5.3 -5.5 -5.3 -6.5 -5.2 -4.7 Debt (percent of GDP) 78.3 76.9 82.9 82.7 83.6 83.5 Primary balance (percent of GDP) -0.3 0.0 0.6 -0.5 0.7 1.3 International poverty rate (USD 1.9 in 2011 PPP)a,b 0.8 0.7 0.6 0.5 0.4 0.3 Lower middle-income poverty rate (USD 3.2 in 2011 PPP)a,b 10.1 9.5 8.7 8.2 7.4 6.5 Upper middle-income poverty rate (USD 5.5 in 2011 PPP)a,b 40.4 39.0 37.3 36.1 34.4 32.3 Source: World Bank. Notes: (f) = forecast. (a) Calculations based on SAR-POV harmonization, using 2016-HIES. Actual data: 2016. Nowcast: 2017-2018. Forecast are from 2019 to 2021; (b) Projection using neutral distribution (2016) with pass-through = 0.87 based on GDP per capita in constant LCU. 83 South Asia Economic Focus Fall 2019 South Asia at a glance South Afghanistan Bangladesh Bhutan India Maldives Nepal Pakistan Sri Lanka Asia Real GDP 2016 2.4 7.1 7.4 8.2 7.3 0.6 4.6 4.5 7.7 Growth 2017 2.7 7.3 6.3 7.2 6.9 8.2 5.2 3.4 7.2 2018 1.8 7.9 4.6 6.8 6.7 6.7 5.5 3.2 6.6 2019 (f) 2.5 8.1 5.0 6.0 5.2 7.1 3.3 2.7 5.9 2019 Q1 (CY) .. .. .. 5.8 .. .. .. 3.8 .. 2019 Q2 (CY) .. .. .. 5.0 .. .. .. 1.6 .. Inflation 2016 4.3 5.9 4.4 4.5 0.5 9.9 2.9 4.0 4.2 (Consumer OUTPUT and PRICES Price Index) 2017 4.7 5.4 4.1 3.6 2.8 4.5 4.2 6.6 5.1 2018 0.6 5.8 2.7 3.4 -0.1 4.2 3.9 4.3 3.9 2019 (f) 3.1 5.5 5.0 3.5 0.5 4.5 7.3 4.5 4.0 2019 June 3.1 5.5 2.7 3.2 1.7 6.1 8.9 2.1 .. 2019 July 1.8 5.6 .. 3.1 0.5 6.3 10.4 2.2 .. REER (CY) 2016 .. .. .. 105.2 .. .. 110.3 .. 105.7 2017 .. .. .. 110.0 .. .. 106.9 .. 109.7 2018 .. .. .. 105.0 .. .. 103.0 .. 104.8 2019 (f) .. .. .. 106.8 .. .. 103.8 .. 106.5 2019 August .. .. .. 108.0 .. .. 104.4 .. 107.6 2019 Sept. .. .. .. 108.7 .. .. 104.8 .. 108.3 Current 2016 5.6 1.9 -30.5 -0.6 -23.4 -0.6 -1.7 -2.1 -0.7 Account Balance 2017 1.0 -0.5 -22.9 -1.8 -21.7 -1.8 -4.1 -2.6 -2.0 (percent of 2018 0.6 -3.5 -19.6 -2.1 -25.3 -2.1 -6.3 -3.2 -2.5 GDP) 2019 (f) -1.6 -1.7 -16.3 -2.0 -18.5 -2.0 -4.8 -2.7 -2.3 Trade Balance 2016 -30.8 -32.2 -31.1 -1.2 .. -40.7 -14.9 -7.5 -5.0 (percent of GDP) 2017 -32.4 -42.6 -24.0 -2.3 .. -49.2 -18.7 -7.6 -6.9 BALANCE of PAYMENTS 2018 -30.7 -53.7 -19.8 -3.0 .. -57.1 -19.6 -7.7 -8.3 2019 (f) -30.0 -52.8 -19.6 -2.8 .. -59.8 -15.8 -5.6 -7.8 Import 2016 25.8 -7.1 12.3 4.4 .. 2.8 16.0 7.9 4.7 Growth (percent, 2017 8.0 2.9 -7.1 17.6 .. 27.2 21.2 7.1 16.6 y-o-y) 2018 1.0 27.0 -7.1 15.4 .. 19.0 15.8 1.8 15.3 2019 (f) 3.6 4.8 6.9 5.9 .. 17.9 5.8 -3.3 5.7 Export 2016 -0.3 2.2 -1.4 5.1 .. -13.7 -1.6 -0.7 4.4 Growth 2017 7.0 -2.3 6.8 4.7 .. 11.3 -0.6 7.6 4.2 84 (percent, y-o-y) 2018 5.0 8.1 -2.0 12.5 .. 7.8 10.4 0.5 11.9 2019 (f) 10.0 14.9 10.9 6.0 .. 7.9 13.2 1.1 6.5 South Asia Economic Focus Fall 2019 South Afghanistan Bangladesh Bhutan India Maldives Nepal Pakistan Sri Lanka Asia Foreign 2016 .. 8.8 .. 12.1 3.4 .. 5.4 3.8 11.0 Reserves (months of 2017 .. 8.4 .. 10.3 2.8 .. 4.0 3.8 9.5 goods import 2018 .. 7.1 .. 9.5 2.8 .. 2.8 4.4 8.7 cover, CY) BALANCE of PAYMENTS 2019 June .. 7.0 .. 10.7 .. .. 2.9 6.2 9.8 2019 July .. 6.9 .. 10.5 .. .. 3.0 4.6 9.5 Personal 2016 652 13,574 34 62,744 3.8 6,612 19,808 7,262 110,691 Remittances Received 2017 734 13,502 43 68,967 4.0 6,928 19,807 7,190 117,176 (USD million, 2018 385 15,562 58 78,790 4.2 8,064 21,194 7,043 131,101 CY) 2019 June .. 4,374 .. 12,617 .. .. 5,069 1,617 .. 2019 July .. 4,550 .. 14,642 .. .. 5,747 .. Fiscal Balance 2016 0.1 -3.7 -1.9 -6.9 -9.9 1.4 -4.5 -5.3 -6.3 (percent of GDP, FY) 2017 -0.5 -3.4 -4.6 -5.7 -3.0 -3.1 -5.8 -5.5 -5.4 2018 0.6 -4.7 -3.1 -5.9 -4.7 -6.7 -6.4 -5.3 -5.8 GOVERNMENT FINANCES 2019 (f) -0.9 -4.4 -2.1 -6.0 -2.9 -1.9 -8.8 -6.5 -6.0 2020 (f) -0.6 -4.8 -2.7 -5.8 -4.9 -2.6 -7.5 -5.2 -5.8 Public Debt 2016 6.1 31.5 109.4 67.5 56.2 27.9 68.7 78.3 64.5 (percent of GDP, FY) 2017 5.9 30.8 108.2 67.6 58.5 26.1 68.0 76.9 64.5 2018 6.9 31.9 104.0 67.3 59.2 30.2 73.0 82.9 64.8 2019 (f) 7.1 32.8 109.3 66.9 59.0 30.1 86.5 82.7 65.4 2020 (f) 7.1 33.8 105.6 66.5 60.2 30.3 82.9 83.6 64.8 Private 2016 -0.2 5.2 3.0 8.1 .. 0.9 8.2 7.4 7.7 Consumption Growth 2017 4.3 9.2 -0.4 7.6 .. 2.8 7.5 2.5 7.5 (percent, 2018 1.2 8.9 1.0 7.9 .. 3.9 4.6 2.3 7.3 y-o-y) 2019 (f) 2.0 5.8 8.7 5.8 .. 3.7 2.4 2.6 5.2 2020 (f) 2.3 6.3 7.4 6.1 .. 1.2 1.1 3.4 5.4 Gross Fixed 2016 -6.0 9.5 11.5 7.9 .. 16.7 8.7 2.5 8.0 Capital Investment 2017 6.4 10.3 5.5 9.1 .. 31.2 9.1 7.8 9.2 Growth CONSUMPTION and INVESTMENT 2018 0.5 10.7 -3.6 9.8 .. 15.9 2.0 5.0 9.3 (percent, y-o-y) 2019 (f) 3.1 8.6 -3.9 8.9 .. 10.7 -3.6 5.8 8.1 2020 (f) 2.8 8.9 -2.5 8.3 .. 9.3 3.8 5.5 8.1 Net Foreign 2016 -0.1 0.6 0.5 1.6 10.3 0.5 0.8 0.8 1.4 Direct Investment 2017 -0.1 0.7 -0.7 1.1 10.1 0.8 0.9 1.5 1.1 (percent of 2018 -0.1 0.6 0.1 1.1 10.5 0.6 1.1 1.7 1.1 GDP, FY) 2019 (f) -0.1 0.8 .. 1.2 .. .. 0.6 1.1 1.1 2020 (f) 0.1 0.8 .. 1.4 .. 0.8 1.3 1.3 Net Foreign 2016 98.7 -42.1 .. 4,725 132.3 .. -153 -993 .. Portfolio Investment 2017 -29.1 178.8 .. -30,638 -328.7 .. -1,198 -1,772 .. (USD million) 2018 2019 Q1 (CY) .. .. 550.4 253.2 169.5 .. 9,598 13,502 -98.2 .. .. .. 288 -1 -129 1,418 .. .. 85 2019 Q2 (CY) .. 253.2 .. 13,502 .. .. -1,032 1,418 .. South Asia Economic Focus Fall 2019 Notes: (f) Forecast CY Series for Calendar Year FY Series for Fiscal Year Afghanistan’s fiscal year is the calendar year. Bangladesh’s fiscal year runs from July 1st to June 30th. Bhutan’s fiscal year runs from July 1st to June 30th. India’s fiscal year runs from April 1st to March 31st. Maldives’s fiscal year is the calendar year. Nepal’s fiscal year runs from July 16th to July 15th. Pakistan’s fiscal year runs from July 1st to June 30th. Sri Lanka’s fiscal year is the calendar year. Real GDP Growth Source: Central Statistics Office of India, Sri Lanka Department of Census and Survey, and World Bank MTI. Note: Real GDP growth rates (percent change, y-o-y) at Market Prices; Pakistan is in factor prices. Inflation (Consumer Price Index) Source: World Bank DEC GEM and World Bank MTI. Note: Annual percent change in CPI inflation. REER (CY) Source: World Bank DEC GEM. Note: Real effective exchange rate is the nominal effective exchange rate (a measure of the value of a currency against a weighted average of several foreign currencies) divided by a price deflator or index of costs. An increase in REER implies that exports become more expensive and imports become cheaper. Current Account Balance (percent of GDP) Source: World Bank MTI. Note: Does not include grants unless otherwise stated. Trade Balance (percent of GDP) Source: World Bank WDI. Note: Trade balance in goods and services is derived by offsetting value of imports of goods and services against value of exports of goods and services as ratio to GDP. Import Growth (percent, y-o-y) Source: World Bank DEC GEM, World Bank MTI, and staff calculations. Note: Annual trade change is in (respective) fiscal year and covers goods and non-factor services (GNFS) imports. Monthly trade change is in calender year and covers only merchandise. Export Growth (percent, y-o-y) Source: World Bank DEC GEM, World Bank MTI, and staff calculations. Note: Annual trade change is in (respective) fiscal year and covers goods and non-factor services (GNFS) exports. Monthly trade change is in calender year and covers only merchandise. Foreign Reserves (months of goods import cover, CY) Source: World Bank DEC GEM. Note: Annual data for foreign reserves are 12-month averages. Remittances (USD million, CY) Source: Haver Analytics, World Bank WDI, Trading Economics, and staff calculations. Note: Personal remittances including personal transfers and compensation of employees in current USD. Fiscal Balance (percent of GDP, FY) Source: World Bank MTI. Note: Does not include grants unless otherwise stated. Public Debt (percent of GDP, FY) Source: World Bank MTI. Note: Gross public debt stock including domestic and foreign liabilities, end of Period. Private Consumption Growth (percent, y-o-y) Source: World Bank MTI. Note: Annual (respective) fiscal year percent change in gross consumption expenditure. Gross Fixed Capital Investment Growth (percent, y-o-y) Source: World Bank MTI. Note: Annual (respective) fiscal year percent change in gross fixed capital expenditure. Net Foreign Direct Investment (percent of GDP, FY) Source: Haver Analytics and World Bank MTI. Note: Net balance of Foreign Direct Investment assets and liabilities as ratio to GDP. Net Portfolio Investment (USD million) Source: Haver Analytics, IMF BoP database, and staff calculations. Note: Portfolio investment covers transactions in equity securities and debt securities. Balances are calculated as net assets minus net liabilities. Data is in current USD. 86 South Asia Economic Focus Fall 2019 PHOTO BY: SAQIB RIZVI/SHUTTERSTOCK