57499 From Global Collapse to reCovery Economic Adjustment and Growth Prospects in Latin America and the Caribbean THE WORLD BANK From Global Collapse to reCovery Economic Adjustment and Growth Prospects in Latin America and the Caribbean Office of the Chief Economist Latin America and the Caribbean THE WORLD BANK Foreword This semiannual report--a product of the Office of the Chief Econo- mist for the Latin America and the Caribbean Region of the World Bank--analyzes where the Latin America and the Caribbean (LAC) region stands following the global crisis, its growth prospects and main challenges. The first part of the report focuses on macroeco- nomic and financial aspects, emphasizing the outlook going forward. The second part examines some aspects of the adjustment in la- bor markets during this crisis in comparison to previous ones. The preparation of this report was led by Augusto de la Torre, Regional Chief Economist, in close collaboration with César Calderón, Tatiana Didier, Julian Messina, and Sergio Schmukler. Paula Pedro, María Virginia Poggio, and Carlos Felipe Prada provided outstanding re- search assistance. We would like to thank Tito Cordella, Francisco H. G. Ferreira, Samuel Freije-Rodríguez, Gladys López-Azevedo, William Maloney and Lars Christian Moller for their invaluable comments. We also extend our gratitude to M. Ayhan Kose (IMF) for providing us with the results of his research on the decomposition of business cycles into their global, regional, and idiosyncratic components. From Global Collapse to Recovery 3 part I executive summary LAC Recovering The global crisis is now in the rear view mirror and world growth is being restored. In sharp contrast with past episodes of global tur- in a Multi-Polar moil, this time the recovery is led by the periphery--specifically by the larger and more dynamic emerging markets (Brazil, China, India, Global Economy South Korea, Malaysia, Philippines, and Thailand). For this group of emerging markets (EMs), the contraction in economic activity was much smaller than that of rich countries, the recovery started ear- lier, and the rebound has been much steeper. LAC is second among emerging regions, after Asia, in the strength of the recovery. LAC comes out of the crisis with a bruised "income statement", no doubt, but its economic downturn in 2009 was less dramatic than that other of regions and it led to a milder than expected increase in unemployment when compared to past downturns. Moreover, LAC's "balance sheet" was not impaired by the crisis. Due to greatly improved macro and financial policy frameworks in LAC, factors that used to magnify external shocks (i.e., weak currencies, weak fiscal processes, and weak banking systems) this time helped cushion the shock. Several LAC countries were able to conduct countercyclical policy, particularly in the monetary front, for the first time in decades. The effectiveness of countercyclical policies in LAC was complemented and enhanced by multilateral institutions' sizeable, flexible, and timely provision of liquidity and budget-support financing. The current pattern of global recovery has favored LAC so far. Countercyclical policies have supported domestic demand in the larger LAC countries and external demand from fast-growing EMs, especially China, has boosted exports and terms of trade for LAC's net commodity exporters--which are mainly located in South America and account for over 90 percent of the region's population and GDP. Prospects for LAC in the short-term thus look good--regional economic activity is forecast to expand by a solid 4 percent in 2010. Beyond the cyclical rebound, however, a higher trend growth will not be as easy to sustain for LAC because future economic dynamics for the rest of the world are clouded by uncertainty and complexity. It is unclear if rich countries will be able to overcome the growth-impairing effects of high govern- ment indebtedness without a significant increase in inflation fur- ther down the line. And while growth in LAC (especially in South America) can continue on the strength of its ties to emerging Asia, there are doubts about the sustainability of China's invest- From Global Collapse to Recovery 5 Office of Regional Chief Economist ment-reliant/export-based growth model. A smooth LAC's natural resource wealth can broaden the scope shift towards more of a consumption-based growth for seizing the growth opportunity, but only if the asso- model in China would be easier to achieve with the ciated windfall earnings are managed judiciously help of international macroeconomic policy coordi- within a long-term horizon, so as to avoid falling victim nation, which seems unlikely to materialize at this to the so-called "natural resource curse." A clear sign stage. that the downside risks of commodity abundance are being avoided would materialize if commodity export- While central banks in rich countries will have to keep ing countries mange to save (via cyclically adjusted interest rates low to support their sagging economies, primary fiscal surpluses) a substantial fraction of the EMs will have hike them earlier to control inflation ex- commodity-related revenue windfalls. pectations as signs of economic overheating surface. The resulting widening of the interest rate differen- tial will further boost capital flows to LAC, intensifying after the fall, laC stands tall policy tensions vis-à-vis the risks of an overshooting in the appreciation of LAC currencies (with potentially As discussed in our previous semi-annual report of permanent adverse effects on export competitiveness) September 2009, "Update on the Global Crisis: The and excessive credit expansion (which may threaten Worst is Over, LAC Poised to Recover," the 2008­09 financial fragility down the line). The policy debate in crisis reached its darkest phase in the fourth quarter LAC going forward will thus tend to focus on macro- of 2008 and the early months of 2009, when the glob- prudential policies to dampen credit creation as well al credit crunch originated in the US shut down all as on policies to curb undue currency appreciation economic engines of world growth. During that time, (international reserve accumulation, controls on capi- the systemic and global propagation of the downturn tal inflows, fiscal tightening). dominated country-specific strengths and rates of economic growth took a nose dive across countries in The region's major longer-run challenge going forward a highly synchronized manner. A full meltdown of fi- will be to craft a bold productivity agenda. While the nancial intermediation in the rich countries was avert- process of convergence towards rich-country stan- ed and confidence began to stabilize, on the strength dards of living has eluded LAC for more than a cen- of a broad menu of unprecedented risk-absorption tury, hopes that this trend may be changing emerged and stimulus policies, led by the U.S. Federal Reserve before the crisis, when several LAC countries other Bank. In the process, however, the balance sheets of than Chile recorded visible productivity growth. With public sectors in rich countries were weakened by a LAC coming out of this crisis relatively well positioned, major increase in indebtedness, raising the risk of fis- such hopes are rekindling, especially considering that cal crises down the line (see below). the improved macro-financial resiliency of the region gives greater assurance that future gains from growth While the downturn in growth was globally synchro- will not be wiped out by financial crises. In addition, nized, economic performance and policy reactions LAC has been making significant strides in the equity varied across countries, with some coping better than agenda and this could help mobilize consensus in fa- others. Like most countries in the world, LAC came vor of a long overdue growth-oriented reform agenda. out of the crisis with a bruised "income statement" But the jury is out on whether the region will be able to but, unlike rich countries and some of the emerging seize the opportunity to boost long-run growth, espe- economies in Eastern and Southern Europe, LAC's cially considering the large gaps that LAC would need "balance sheet" was not impaired. This remarkable to close in such key areas as saving, human capital fact stands in sharp contrast with LAC's own past ex- accumulation, physical infrastructure, and the ability perience, where domestic currency, banking and/or to adopt and adapt new technologies. debt crises tended to wreak havoc at home following 6 From Global Collapse to Recovery The World Bank a major external shock. In other words, while LAC ex- To be sure, the recession in LAC, while milder than perienced a marked slowdown in growth it did not suf- expected, led to a partial reversal of the robust pov- fer systemic damage, and this has raised the relative erty reduction gains that the region achieved in the attractiveness of many LAC countries as destinations five years prior to the crisis. While 60 million Latin for investment and placed them in a good position to Americans are estimated to have left the poverty resume growth. ranks during 2002­2008, some 9­10 million people joined the poor in 2009, and that number would have In particular, LAC's recession in 2009 (a GDP contrac- been greater had it not been for the fact that--again tion of 2.3 percent) was deeper than that of the East breaking with history--LAC governments were able to Asian Tigers (0.1 percent) but milder than that of high- maintain, and in many cases actually step up, social income countries (3.4 percent) and Eastern Europe assistance programs, including the very effective con- (5.7 percent) (Figure 1.A). Excluding Mexico, more- ditional cash transfers schemes that have become a over, whose output contraction of about 6.5 percent LAC trademark in social policy. was a regional outlier, LAC's GDP hardly contracted in 2009. In addition, LAC's "growth collapse" (the differ- From a cross-country perspective, as analyzed in ence in GDP growth rates between 2009 and 2007) of greater detail in Appendix I.A, growth collapses 6.3 percentage points (pp), although large in absolute around the world tended to be larger in economies terms, was just under that of the industrial countries characterized by greater trade openness, higher trade and East Asian Tigers (6.9 pp and 7.4 pp, respec- dependence on rich country markets, higher share of tively) and much smaller than that of Eastern Europe manufacturing exports, and weaker banking systems. (12.9 pp) (Figure 1.B). Moreover, improvements in LAC's macro-financial FIguRe 1 Real Growth and Growth Forecasts Across Regions A. Growth Across Regions, 2009 B. Growth Collapse Real GDP Growth in 2009 Around the World GDP Growth Collapses Around the World Annual Real GDP Growth Rate Differences Between Growth in 2007 and 2009 10% 0% 8% ­2% 6% ­4% 4% 2% ­6% 0% ­8% ­2% ­10% ­4% ­6% ­12% ­8% ­14% ECA OECD LAC East SSA MENA South China ECA East OECD LAC China SSA MENA South Asian Asia Asian Asia Tigers Tigers Note: Growth collapse is defined as the difference between the GDP growth rate in 2009 vis-à-vis growth in 2007. ECA refers to Eastern Eu- rope and Central Asia countries. East Asian Tigers are Hong Kong,China; Indonesia; Korea,Republic of; Malaysia; Singapore; Taiwan,China; and Thailand. OECD refers to OECD-member countries. LAC refers to countries in Latin America and the Caribbean. SSA refers to Sub- Saharan Africa countries. MENA makes reference to Middle East and North African countries. Data comes from Consensus Forecast as of December 2009 for countries that have not published 2009 growth figures yet. Source: Bloomberg, IMF International Financial Statistics (IFS) and Consensus Forecasts. PART I LAC Recovering in a Multi-Polar Global Economy 7 Office of Regional Chief Economist policy frameworks paid off. They enabled LAC to bet- of LAC's more robust macro-financial and social pol- ter cushion the external shock and undertake coun- icy frameworks to cushion the effects of the external tercyclical policies, especially on the monetary front shock.1 Indeed, the prompt response by multilaterals but also, to a lesser extent, on the fiscal side. While constituted a more prominent feature in the manage- banking system weaknesses did not play a role in ment of this crisis compared to previous ones. In this most LAC countries (with the possible exception of connection, a recent IDB study (Izquierdo and Talvi, some Caribbean countries), the rest of the mentioned 2010) that quantitatively assesses the impact of mul- factors help explain why, for instance, the growth col- tilateral finance during the crisis finds that: (i) devel- lapse was more pronounced in Argentina, Mexico, oping countries with access to multilateral resources Costa Rica, Panama and Peru compared to, say, Bra- outperformed those without it; and (ii) it was the com- zil and Chile, and why the growth collapse was com- bination of multilateral financial support and improved paratively very small in, say, Bolivia. macro-financial fundamentals which enabled LAC to better withstand the crisis. The shock absorption ca- Countercyclical capacity is a new and certainly no- pacity resulting from this combination also helps ex- table feature in LAC's macroeconomic policymaking, plain why labor markets in LAC behaved differently a region where in the past external shocks were in- this time around, with unemployment rising less than stead amplified by weak macroeconomic fundamen- in past crises (controlling for the size of the economic tals. This capacity, furthermore, is a fruit of steady downturn) and the share of informal employment (sur- institution building over the past 25 years involving a prisingly) not increasing. Part II of this report provides virtuous combination of: more robust monetary policy a detailed comparative analysis of the dynamics of la- frameworks (including greater exchange rate flexibility bor markets in this and previous downturns. and credible and professional central banks), more viable fiscal processes (although these tend to remain pro-cyclical in most of LAC, with the salient exception FIguRe 2 LAC: A Safer Integration of Chile), sounder and better regulated banking sys- A Safer Integration in LAC Net Creditor tems, a significant reduction of currency mismatches 10% in debtor balance sheets, and a safer integration into 5% Net Debt Position vis-a-vis Rest of the World international financial markets. The latter reflects a 0% process that begun in this millennium, whereby LAC ­5% ­10% became a net creditor to the rest of the world on the Net Debtor ­15% debt side (that is, the side where systemic vulnerabili- ­20% ties can easily arise) while the rest of the world has be- ­25% come an increasing net claimant on LAC on the equity ­30% Net Equity Position vis-a-vis side (that is, the side where systemic vulnerabilities ­35% Rest of the World are much less likely to emerge) (Figure 2). ­40% 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 Importantly, the flexible, sizeable, and timely provision Source: The net debt position (vis-à-vis ROW) is the sum of debt of liquidity and budget support financing by multilater- assets and reserves minus debt liabilities. In turn, the net equity po- sition (vis-à-vis ROW) is the sum of net FDI assets and net portfolio al financial institutions (including the IMF, World Bank, equity assets. The sample ranges from 1990 to 2007. Source: Lane IDB, and CAF) complemented and boosted the ability and Milesi-Ferretti (2007). 1 The World Bank's commitments to disburse to the LAC region rose from an average of nearly US$ 5 billion per year during FY2003­FY2008 to around US$ 14 billion in FY2009. The average size of IBRD loan to countries in the region increased from US$ 116 million in FY2003­FY2008 to US$ 271 million in FY2009, as a large share of World Bank lending in 2009 took the form of fast-disbursing budget-support finance. 8 From Global Collapse to Recovery The World Bank the global recovery: contrasting set prices, a sharp compression of spreads over U.S. paths for the center and the Treasury securities, and a robust pick-up in commod- periphery ity prices. Moreover, world trade volume, after falling more than 20 percent between April 2008 and May A year has passed since the first signs of a recovery, or 2009, has rebounded significantly, already reaching the so-called "green shoots," started to sprout system- mid-2007 levels and thus attenuating fears of a trade- atically across countries. For the world as a whole, the less recovery. Global liquidity remains abundant and cyclical recovery has so far taken a sharper V-shape a vigorous search for yield seems underway, spurred than was expected a year ago, not least because eco- by low interest rates, and capital flows are returning nomic activity is rebounding from a very low base. to pre-crisis paths while increasingly being directed There has been a strong and steady comeback in as- towards emerging markets (Figure 3). FIguRe 3 Recovery in Financial Markets: A. Stock Prices B. World Trade Volumes Stock Prices Around the World World Trade Volumes Indexes: Jan-06=100 Stock Prices Around the World Seasonally Adjusted, Index 2000=100 World Trade Volumes 170 230 Indexes: Jan-06=100 Seasonally Adjusted, Index 2000=100 EAP 165 170 230 160 165 EAP Exports 180 LAC 155 160 Exports 180 LAC 150 155 UK Imports 130 ECA 145 150 UK Imports 130 ECA 140 145 80 Japan 135 140 Japan 130 135 80 125 130 30 S&P 120 125 Jan-06 Mar-06 May-06 Jul-06 Sep-06 Nov-06 Jan-07 Mar-07 May-07 Jul-07 Sep-07 Nov-07 Jan-08 Mar-08 May-08 Jul-08 Sep-08 Nov-08 Jan-09 Mar-09 May-09 Jul-09 Sep-09 Nov-09 Jan-10 Mar-10 30 S&P Jan-06 Mar-06 May-06 Jul-06 Sep-06 Nov-06 Jan-07 Mar-07 May-07 Jul-07 Sep-07 Nov-07 Jan-08 Mar-08 May-08 Jul-08 Sep-08 Nov-08 Jan-09 Mar-09 May-09 Jul-09 Sep-09 Nov-09 Jan-10 Mar-10 120 Jan-06 Mar-06 May-06 Jul-06 Sep-06 Nov-06 Jan-07 Mar-07 May-07 Jul-07 Sep-07 Nov-07 Jan-08 Mar-08 May-08 Jul-08 Sep-08 Nov-08 Jan-09 Mar-09 May-09 Jul-09 Sep-09 Nov-09 Jan-10 Mar-10 Jan-06 Mar-06 May-06 Jul-06 Sep-06 Nov-06 Jan-07 Mar-07 May-07 Jul-07 Sep-07 Nov-07 Jan-08 Mar-08 May-08 Jul-08 Sep-08 Nov-08 Jan-09 Mar-09 May-09 Jul-09 Sep-09 Nov-09 Jan-10 Mar-10 C. Capital Flows D. Commodity Prices Net Private Capital Flows to Emerging Economies Commodity Prices Net Private CapitalUS$ Billion Flows to Emerging Economies Commodity Index Wheat, Copper and Soybean: Prices Jan-01-05=100 Wheat, Copper and and Soybean, 01-Jan-05=100 1,200 US$ Billion Wheat, CopperOil WTI in Current US$ and Soybean: Index Jan-01-05=100 150 Wheat, Copper Soybean, 01-Jan-05=100 1,200 Oil WTI in Current US$ 1,000 350 Private Flows Wheat 150 1,000 130 800 350 300 Oil WTI,WTI, Current US$ Private Flows Wheat 130 800 300 Copper 110 Oil Current US$ 600 250 Soybean 110 Copper 90 600 250 200 400 Soybean 90 400 FDI 200 150 70 200 FDI 150 70 200 Debt Investment 100 50 0 Debt Investment Portfolio Equity Investment 100 Oil (rhs) 50 0 50 30 ­200 Portfolio Equity Jan-06 Mar-06 May-06 Jul-06 Sep-06 Nov-06 Jan-07 Mar-07 May-07 Jul-07 Sep-07 Nov-07 Jan-08 Mar-08 May-08 Jul-08 Sep-08 Nov-08 Jan-09 Mar-09 May-09 Jul-09 Sep-09 Nov-09 Jan-10 Mar-10 Oil (rhs) 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009f 2010f Investment 50 30 ­200 Jan-06 Mar-06 May-06 Jul-06 Sep-06 Nov-06 Jan-07 Mar-07 May-07 Jul-07 Sep-07 Nov-07 Jan-08 Mar-08 May-08 Jul-08 Sep-08 Nov-08 Jan-09 Mar-09 May-09 Jul-09 Sep-09 Nov-09 Jan-10 Mar-10 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009f 2010f Note: EAP represents East Asia and the Pacific region. Note that the 2009 and 2010 figures on capital flows are forecasts from the Institute of International Finance. Source: Bloomberg, CPB (Netherlands Bureau for Economic Policy Analysis), and Institute of International Finance. PART I LAC Recovering in a Multi-Polar Global Economy 9 Office of Regional Chief Economist While global economic growth is being restored over- FIguRe 4 World Industrial Production all, there is great heterogeneity in the strength of the post-crisis recovery across regions and within regions. World Industrial Production Growth engines are being reignited at different paces Index Apr-08 = 100 105 and intensities across nations. In particular and in sharp contrast with past episodes of global economic 100 Developed Countries and financial turmoil, the recovery this time around 95 has been led not by the center but by the periphery --specifically by the larger and more dynamic emerg- 90 ing markets (namely, Brazil, China, India, South Ko- Emerging Economies rea, Malaysia, Philippines, and Thailand, which jointly 85 account for nearly 55 percent of emerging markets 80 GDP). For this group of emerging markets (dynamic Jan-06 Apr-06 Jul-06 Oct-06 Jan-07 Apr-07 Jul-07 Oct-07 Jan-08 Apr-08 Jul-08 Oct-08 Jan-09 Apr-09 Jul-09 Oct-09 EMs), not only was the fall from peak to through in average industrial production smaller than that of the Note: The group of developed countries refers to OECD countries advanced economies (mainly the U.S., Europe and excluding Turkey, Mexico, Republic of Korea, and Central European Japan) but the through was reached three months countries. Source: CPB (Netherlands Bureau for Economic Policy Analysis). earlier and the rebound since then has been much steeper. Consistent with the heterogeneity of the recovery de- Thus, industrial production for the dynamic EMs fell scribed above, GDP growth forecasts range widely by 13.4 percent from the peak, touched bottom in across regions. The most recent forecasts (i.e., Con- February 2009, and has been increasing sharply sensus forecasts as well as IMF and World Bank pro- since then, reaching pre-crisis levels by the time of jections) put growth in the rich countries at 2.1 percent this writing (Figure 4). When emerging regions are for 2010 and 2.25 percent for 2011. The U.S. is en- compared, emerging Asia and LAC have dominated visaged to grow slightly above that average--flattening the recovery, posting posted cumulative gains of at 3.1 percent per annum during 2010­2011--and 18.7 and 9.1 percent in industrial production since Western Europe somewhat below--at 1.2 percent in their troughs, respectively. By contrast, average 2010 and 1.7 percent in 2011. By contrast, a much industrial production for the rich countries fell by stronger rebound is forecast for the dynamic EMs. 16.9 percent from its peak, reached bottom in May China is expected to post the highest growth rates, at 2009, and has been rising sluggishly since then 9.9 percent in 2010 and 9.1 percent in 2011. East (a cumulative increase of only 3.5 percent in the Asia's GDP is forecast to grow by 5.5 percent in 2010 May 2009-December 2009 period), thus remain- (5.1 percent in 2011) and LAC's economic activity is ing substantially below pre-crisis levels by the time expected not to lag much behind, growing at a solid 4 of this writing. Moreover, capacity utilization in the percent per year during 2010­2011, with Brazil, Peru, rich countries is still below the minimum of previ- Chile and Panama leading the pack (see below). The ous downturns and unemployment rates remain at GDP growth forecast for countries in the Middle East stubbornly high levels (slightly below 10 percent in and North Africa is 4.4 percent per year in 2010­11 the U.S., more than 8 percent in Germany, almost whereas growth in Eastern Europe is the lowest among 20 percent in Spain, and still rising in France and emerging market regions (3.5 percent) (Figure 5). Ireland). By contrast, available data suggest that capacity utilization and employment are at around The considerable variation in the recovery across pre-crisis levels for the dynamic EMs as well as for countries is not surprising given that the idiosyncratic many countries in LAC. strengths of individual economies play a greater role as 10 From Global Collapse to Recovery The World Bank index was back to its pre-crisis peak levels already by FIguRe 5 Regional Growth Forecasts for end-March 2010. Analogous behavior is observed in 2010 and 2011 stock markets in Chile, Colombia, Mexico, and Peru. Real GDP Growth Forecasts for 2010­2011 Annual Real GDP Growth Rate While LAC's is forecast to grow by 4 percent in 2010 12 2010 2011 on a weighted average basis, there is significant het- 10 erogeneity in projected growth rates within the region. 8 South America is envisaged to have a stronger recov- ery (4.7 percent over 2010­11) than Central America 6 (4 percent) and the Caribbean (3.2 percent). Coun- 4 tries in LAC that are experiencing stronger rebounds 2 tend to be characterized by: (i) vigorous expansion in domestic demand; (ii) extensive use of counter- 0 Western Japan Eastern US LAC MENA East China cyclical policies; (iii) economic complementarity to Europe Europe Asian Tigers Asia (especially China); and (iv) commodity abun- dance. The leader is Brazil, where industrial produc- Note: Western Europe comprises Euro Zone countries, Denmark, Sweden, UK, Norway, and Switzerland. Source: Consensus Fore- tion has increased by almost 20 percent from the casts (December 2009 and March 2010), and Bloomberg. trough and its 2010 GDP growth forecast is at 5.5 percent. Peru, Chile, Panama and Mexico are also the effect of the global systemic shock fades. A more expected to record relatively faster GDP growth within systematic analysis of cross-country data suggests the region--between 4 and 5 percent in 2010. Growth that this is indeed the case (Appendix I.B provides rates in the 3­4 percent range for 2010 are expected a more detailed analysis, including a comparison of for Argentina, Bolivia, Colombia, Costa Rica, Domini- the collapse-rebound patterns around the current and can Republic, Paraguay, and Uruguay. Lagging be- previous crises). In effect, countries that were able to hind but rebounding nevertheless are most countries conduct counter-cyclical policies are experiencing a in Central America and the Caribbean. Lastly, Jamaica faster recovery and so are countries that are experi- and Venezuela are expected to grow very little or even encing a rebound in their net exports after the crisis. contract in 2010 (Figure 6). A purely arithmetic bounce-back effect also seems at play, as countries that recorded greater growth col- A few countries in LAC may begin to face the risk of lapses also tend to register faster recoveries. economic overheating, with inflationary pressures ex- pected to increase over the next months. Brazil is the most visible case in point--economic activity is begin- the recovery in laC: things are ning to expand at a pace above the pre-crisis trend looking good overall growth and inflation forecasts are already above the center of the targeted inflation. Moreover, the fiscal As noted, despite being one of the most financially and quasi-fiscal stimulus measures taken in Brazil, globalized regions in the world, LAC is coming out of which are estimated to have totaled around 4 percent this crisis without systemic damage and is experienc- of GDP, will likely not be undone in the near future ing a relatively strong recovery in economic activity due mostly to the electoral cycle. Hence, markets on average. This recovery was preceded by a strong expect the initiation of a monetary policy tightening V-shaped rebound in stock prices and a sharp com- in the near future. For many countries in the region, pression of (sovereign and corporate) spreads. After however, output is expected to remain below potential reaching a trough in October 2008, stock prices in (Mexico and Colombia) and/or inflationary pressures LAC recovered vigorously. The Brazilian stock price are not yet mounting to significant levels (Peru and PART I LAC Recovering in a Multi-Polar Global Economy 11 Office of Regional Chief Economist FIguRe 6 Growth Forecasts for 2010 and 2011 among LAC Countries A. 2010 Growth Forecasts B. 2011 Growth Forecasts Real GDP Growth Forecasts for 2010 Real GDP Growth Forecasts for 2011 LAC countries LAC countries 6.0% 6.0% 5.0% 5.0% 4.0% 4.0% 3.0% 3.0% 2.0% 2.0% 1.0% 1.0% 0.0% 0.0% ­1.0% ­1.0% ­2.0% ­2.0% ­3.0% ­3.0% St. Vc. & Grs. Costa Rica Barbados Venezuela Ant. & Barb. Bahamas St. Kts. & Nv. Jamaica St. Vc. & Grs. Dominica St. Lucia Nicaragua Belize Guatemala Haiti Tri. & Tob. Honduras Ecuador Guyana Colombia Costa Rica Dom. Rep. Paraguay Suriname Bolivia Argentina Uruguay LAC Mexico Panama Chile Peru Brazil Dom. Rep. St. Lucia Paraguay Peru Uruguay Belize Panama Argentina Mexico LAC Brazil Venezuela Haiti Nicaragua Bolivia Guatemala Honduras Ecuador Dominica Chile Guyana Jamaica Colombia Source: Latin American Consensus Forecasts as of March 2010, IMF's World Economic Outlook, IMF's Regional Economic Outlook. Chile). After engineering a massive cut in policy rates For the small, open economies in Central America and (Figure 7), Latin central banks are thus expected to the Caribbean, where the scope for independent mon- start a gradual normalization of interest rates, but the etary policy is very narrow or non-existent, domestic timing and extent of interest rates increases will likely price increases will tend to reflect developments in vary significantly across countries depending on the import prices (including foods and fuels) as well as output gap and inflationary expectations (Figure 8). supply conditions in the agricultural sector. Be it as it FIguRe 7 Output Gap and Inflation A. 2010 B. 2011 2010 Output and Inflation Gaps 2011 Output and Inflation Gaps 0.001 0.0006 URY URY 0.0004 BRA 0.0005 0.0002 PRY 0 PER BRA ARG CHL BOL Output Gap Output Gap 0 BOL PRY ­0.0002 PER MEX PAN ­0.0004 COL COL PAN ARG ­0.0005 CHL ECU ­0.0006 SLV GUA DOM ­0.0008 GUA SLV CRI CRI DOM ­0.001 MEX HND ­0.001 ECU NIC NIC ­0.0012 HND ­0.0015 ­0.0014 ­4 ­2 0 2 4 6 ­2 0 2 4 6 Inflation Pressure in percentage points Inflation Pressure in percentage points Note: Inflation pressures are calculated as the difference between the 2010 inflation rate forecast and an estimated target of 4% (assumed to be the target for most countries in the region). The output gap calculated as the difference between the (log of) actual and potential GDP, with the latter being calculated using the Hodrick-Prescott filter. Source: LCRCE Staff calculations based on Consensus Forecasts as of March 2010. 12 From Global Collapse to Recovery The World Bank very short-run, however, is plagued by great complex- FIguRe 8 Monetary Policy Rates in LAC ity and uncertainty. Some of the regional differences countries and relevant factors affecting growth prospects out- Monetary Policy Rates side LAC are briefly and selectively discussed in the Inflation-Targeting Latin American Countries and the US 15.0% rest of this section. Brazil 13.0% 11.0% One important point to keep in mind in this regard is 9.0% Colombia that, as the region diversifies its linkages to the rest of Mexico the world, the sources and nature of external shocks 7.0% Chile Peru will also diversify. In particular, shocks coming from 5.0% other emerging countries may become increasingly 3.0% US more relevant for LAC in the future. The volatility of 1.0% such shocks may be higher than those coming from ­1.0% international financial markets and rich country de- Jun-07 Aug-07 Oct-07 Dec-07 Feb-08 Apr-08 Jun-08 Aug-08 Oct-08 Dec-08 Mar-09 May-09 Jul-09 Sep-09 Nov-09 Jan-10 Mar-10 mand for LAC exports. In the past, LAC's trade was on average mostly linked to the U.S. and, to a lesser Source: Bloomberg. extent, Europe. The region now relies much more on external demand from other emerging markets, par- may, as interest rates increase in most of LAC while re- ticularly China and the South East Asian countries maining low in rich countries, capital inflows are likely (Figure 9). Kose, Otrok and Prasad (2008) find that to surge, creating complex challenges for LAC central business cycles among LAC-7 countries are increas- bankers, as discussed in more detail below. ingly accounted for by "regional" factors (that is, fac- tors affecting individual regions rather than the whole world) and, although there is no one-to-one mapping, Global growth prospects: it is likely that those regional factors are mainly driven obscured by clouds by regional shocks. As a result, the share of output variation for LAC-7 countries accounted for by regional Whereas the prospects for growth in LAC countries in the near future are promising, the region is not isolated from the world economy and a significant FIguRe 9 Variation in Export Market fraction of the downside risks to growth lies in the ex- Shares in LAC ternal environment. To be sure, due to the mentioned Variation in Export Market Shares for LAC Countries substantial improvements in macro-financial policy 2008 vs. 1990, in percentage points frameworks LAC is much less vulnerable to shocks 20% Euro Zone US China than it used to be. Hence, domestic macro-financial 15% crisis are arguably less likely than in the past to erase 10% LAC's gains from growth going forward. But the sad 5% fact remains that LAC has not managed to sustainably grow above the world average during the last hun- 0% dred years! In general, LAC's growth has tended to ­5% track closely global growth. Full decoupling between ­10% LAC's growth and world growth is thus a chimera, im- plying that the growth prospects for LAC can only be ­15% ARG BRA CHL COL ECU MEX PER PRY URY VEN assessed in the context of the growth prospects for the world. Growth in the rest of the world beyond the Source: IMF's Direction of Trade Statistics (DOTS). PART I LAC Recovering in a Multi-Polar Global Economy 13 Office of Regional Chief Economist factors increased from 1.3 percent to 7.8 percent on lus injections. So, prospects will largely depend on average between 1960­1984 and 1985­2008, reach- domestic private demand, which is still relatively ing almost 16 percent in Peru and 10 percent in Bra- weak and shows no clear signs of a solid recovery zil. Global (rather than "regional") factors (i.e. related yet. Moreover, the management of the crisis implied to fluctuations affecting the entire world) are nonethe- that balance-sheet problems shifted from the private less still important for LAC, explaining more than 17 to the public sector, and this is likely to pose down- percent of output variations in Brazil and around 9 side risks to a sustainable growth path in the medium percent for non-LAC7 countries (e.g., mainly Central term. The extraordinarily high level of government American and the Caribbean countries) (Figure 10).2 indebtedness significantly limits the scope for fiscal expansionary policies and this shifts the countercycli- Consider now the growth prospects for the rich coun- cal burden to monetary policy. While the U.S. Fed is tries. The recovery in the United States has so far been already unwinding some credit support facilities that driven mostly by a restocking of inventories, massive help stabilized the financial system, interest rates in stimulus policies, and a revival of net exports. There the U.S. are likely to remain at near-zero for a while in is uncertainty about the sources of future growth as order to continue providing an expansionary impulse the process of inventory restocking ends and the ef- to the economy. But this, in the absence of vigorous fects of the stimulus fizzle away--towards the second and well-designed regulatory reform, may foster a half of this year in the absence of additional stimu- buildup of financial excesses similar to those that led FIguRe 10 The Increased Importance of Regional Factors and Decreased Global Ones A. Global Factors B. Regional Factors Percentage of Output Fluctuations Percentage of Output Fluctuations Explained by Global Factors Explained by Regional Factors 50.0% 22.0% 1960­1984 1960­1984 45.0% 1985­2008 1985­2008 40.0% 17.0% 35.0% 30.0% 12.0% 25.0% 20.0% 7.0% 15.0% 10.0% 2.0% 5.0% 0.0% ­3.0% Colombia Venezuela, Brazil Mexico Peru Chile Argentina Venezuela, Mexico Brazil Chile Peru Colombia Argentina RB RB Note: Global factors encompass all fluctuations that are common among industrial countries, emerging markets and other developing coun- tries. Regional factors capture fluctuations that are common only within a particular group of countries (say, emerging markets). Source: Kose, Otrok, and Prasad (2008). 2 Kose, Otrok and Prasad (2008) use dynamic factor models to decompose the fluctuations in output, consumption and investment for a sample of 106 countries over the period 1960­2008 into global factors, regional factors, country factors, and idiosyncratic factors. The global factor includes all fluctuations that are common to industrial countries, emerging markets and other developing countries, and across all variables. Regional factors, on the other hand, are those that capture fluctuations that are common only to a particular group of countries, and across all variables. For instance, the regional factors for LAC-7 countries involve fluctuations affecting only theirs and other emerging markets' output, consumption and investment. 14 From Global Collapse to Recovery The World Bank to the crisis. Moreover, there are strong doubts that Japan has experienced a slow export-based recovery the U.S. will be able to grow at the very high rates, or so far. But it has still not seen a recovery in domestic undertake the extent of fiscal adjustment, required to (consumption and investment) demand and unem- restore public sector debt to robust viability. This, in ployment remains high. Moreover, the export-led re- turn, raises the inflation specter--i.e., the specter that bound might slow down as inventory restocking ends. debt might have to be inflated away down the road. The country remains caught in a long-lasting liquidity trap, with deflationary pressures ensuring that interest In Western Europe things look even worse and overall rates will stay at near zero levels. The lack of room on growth prospects are grim, although there is a wide the monetary policy front is putting an excess burden variation in macro-financial circumstances among its on fiscal policies. However, the country does not have countries. The forces of Western European growth re- much space on this front due to its large (gross) debt covery have been the same temporary ones at work in position of almost 200 percent of GDP. the U.S.--inventory restocking and stimulus packages, along with net exports. Domestic demand is also at de- Turning now to emerging markets, China is playing an pressed levels and investment is not expected to pick important role in the world's recovery. During the cri- up significantly in the near future. But the challenges sis, its main engine of growth, exports, collapsed. To for growth in Western Europe are more daunting. Un- sustain economic activity, the government launched a employment is extremely high and unyielding, not least massive and unprecedented stimulus focused on in- because of the well-known rigidities in European labor frastructure investments at the local government level, markets which are politically difficult to alter. Perhaps which were spurred by a gigantic increase in domestic more importantly, the European Union (EJU) does not credit. The resulting investment boom in China (invest- have the flexibility to tailor monetary policy to the differ- ment is estimated to have risen to nearly 60 percent of ent needs of different EU countries. In the absence of GDP) led to a sharp rise in imports, particularly of in- differentiated exchange rate adjustment mechanisms dustrial metals and minerals, which contributed to the and given the severe restrictions on the ability to trans- rebound in commodity prices and helped with the re- fer fiscal resources among Euro members, recession- covery in other emerging markets, including of course ary forces are likely to be stronger and more persistent. the net commodity exporting countries of South Amer- This is particularly troublesome for EU countries with ica. However, household consumption in China seems high debt levels and/or weak fiscal stances such as to be growing at the same pace as real GDP (even if Greece, Ireland, Italy, Portugal, and Spain. The prob- real wages seem to be rising a bit faster), suggesting lems with Greece's debt have already led to increased that the high domestic saving rate has remained es- fears of a double-dip recovery path in the EU, as the sentially unchanged. While China is expected to con- contractionary effects of fiscal austerity measures may tinue to grow strongly in the short-run, as its stimulus spread.3 Given the sluggish recovery and the challeng- dwindles (a few measures have already been taken es faced on the fiscal side, an exit from the monetary towards a normalization of conditions) uncertainties stimulus is unlikely to occur soon, and interest rates in rise about the durability and sustainability of China's the EU will also remain at very low levels for a while. investment-based growth pull on the rest of the world.4 3 A resolution to Greece's debt sustainability problems seems to be underway with both the IMF and Euro zone members involved in a rescue package, but this will have to be accompanied with painful fiscal austerity measures. 4 China is capable of designing the appropriate mix of policies to stimulate domestic demand, and consumption in particular, if external demand indeed remains weak. This would entail a shift towards a domestically-oriented growth model. Another way to stimulate domestic (private) consumption is the real appreciation of the Chinese Yuan. Real appreciation can be achieved through either domestic price inflation or a change in the current exchange rate policy. This is more likely to happen through a slow and gradual change in the nominal exchange rate in the event of high inflation. In the event of mounting inflationary pressures, Chinese authorities may allow the exchange rate to gradually adjust, thus giving a boost to domestic consumption. PART I LAC Recovering in a Multi-Polar Global Economy 15 Office of Regional Chief Economist The significant growth recovery in other Asian coun- ever, capital flows to EMs are likely to pick up further tries has been driven by their position in the global as the interest rate differentials rise--that, is, as EMs production chain, including in particular the integrat- increase interest rates to control inflation expectations ed distribution of manufacturing production stages in the midst of a possible economic overheating while within Asia. Together with China and India, East Asian rich countries keep interest rates low to stimulate their countries, have been the main drivers of global growth sagging economies. since January 2009 and will likely continue to be so in the near future. At the same time, however, with actu- al GDP growth rapidly approaching potential growth, FIguRe 11 Gross nominal liabilities for inflationary pressures are surfacing, and this is induc- developed countries and LAC-7 ing Asian EMs to normalize monetary conditions and even initiate the cycle of monetary policy tightening Gross Nominal Liabilities % of GDP that would cool down growth. Malaysia already raised 200 the policy interest rate and others are expected to fol- 200 2007 2010 low soon. 167 160 In sum, the prospects for global growth are clouded 120 by great uncertainty and complexity regarding at least 98 100 89 89 two main questions. The first question is whether rich 80 71 74 63 countries will be able to overcome the growth-impairing 47 effects of their damaged balance sheets (Figure 11). 40 33 33 The difficulty involved in meeting this challenge is clear when one considers the large size of primary fis- 0 US Euro Japan UK OECD LAC-7 cal surpluses that the U.S., Europe and Japan would have to generate in order to bring down their debts to Source: EIU and Consensus Forecasts (various issues). sustainable levels without inflation (Figure 12). The second question is whether Asian EMs will be able to FIguRe 12 Required Fiscal Adjustments sustain in the medium-term their current growth pat- tern. This is equally uncertain considering that their Required Fiscal Adjustment Between 2010 and 2020 To Reach Debt Levels of 60% current growth model remains frankly reliant on ex- in % of GDP ports. While the domestic market has played an im- 20 Developed Countries portant role in the recovery for some of the dynamic 15 EMs (including Brazil), world consumption demand needs to be recomposed to put world export growth 10 on a high and sustainable path over the medium- Other Emerging Economies term. Ideally, more consumption demand will have 5 Latin America to come from the surplus EMs, Germany and Japan (and less from the U.S.) and this is difficult to envis- 0 age in the absence of stronger currencies in Asia, par- ticularly China. But it is utterly unclear whether and ­5 Colombia Brazil Chile Mexico Peru Argentina Hungary Bulgaria Ukraine Turkey Philippines Indonesia Russia Romania South Africa Germany Italy France Portugal Spain UK US Japan Ireland Greece how this might happen in the absence of effective in- ternational macroeconomic policy coordination involv- ing not just the rich countries but also the dynamic Source: IMF Fiscal Affairs Deparment, 2010. "Strategies for Fis- EMs. That type of coordination unfortunately appears cal Consolidation in the Post-Crisis World." Washington, DC: IMF, unfeasible in the near future. In the meantime, how- February. 16 From Global Collapse to Recovery The World Bank Challenges ahead for laC FIguRe 13 Terms of Trade So far for LAC, the pattern of global recovery discussed Cumulative Change in Terms of Trade above has brought benefits. Countercyclical policies Bolivia Ecuador have supported domestic demand in the larger LAC Chile countries and external demand from fast-growing Paraguay Trin. & Tob. EMs, especially China, has boosted exports and terms Colombia Peru of trade for LAC's net commodity exporters--which Argentina are mainly located in South America and account for Mexico Uruguay over 90 percent of the region's population and GDP Brazil Panama (Figure 13). Until now, moreover, the rebound in com- Guatemala Costa Rica modity prices and relatively fast pace of recovery in Nicaragua 2008q4­2009q4 LAC has not awaken major inflationary pressures. This Dominica 2001q4­2008q2 Dom. Rep. has hitherto allowed LAC central bankers to keep in- Honduras terest rates low and this has, in turn, helped dampen ­40% ­20% 0% 20% 40% 60% 80% 100% 120%140% the pressures for LAC currencies to strengthen rela- Note: The cumulative variation in the terms of trade index is calcu- tive to other currencies. Things will not be as easy for lated using quarterly data. The blue bars represent the cumulative percentage change during the recent commodity price boom up LAC going forward, however, given the prospects and to the peak in 2008q2. The red bars capture the cumulative per- policy complications outside LAC as discussed in the centage change in terms of trade from its trough in 2008q4 to the most recently available quarter (2009q4). Source: WDI, DECPG, previous section. This section briefly discusses some and Haver Analytics. of the challenges ahead for LAC. Living with capital inflows and exchange wards monetary policy. LAC policy makers will thus rate appreciations likely face increasingly difficult policy tensions vis-à- vis the risks of overshooting in currency appreciations As noted, interest rate differentials (already at non- (which could inflict adverse and permanent effects on trivial levels given the near zero rates in developed export competitiveness) and capital inflows-induced countries) will widen further, likely leading to greater excessive credit expansion (which could weaken the capital flows to LAC. And this will happen in a context financial system down the line). There is of course where foreign capital, especially from the equity side, no easy way out of these tensions. The option to step is already flowing to LAC in considerable amounts, not up exchange rate intervention--aimed at dampening least because of many LAC countries withstood the appreciation pressures--will naturally be considered global crisis well and came out of it as relatively more despite its costs--for it calls for a significant degree attractive destinations for investment. The pressures of sterilization, via the issuance of relatively expensive for LAC currencies to appreciate will be further en- local-currency debt securities, of the monetary impact hanced in the case of commodity exporters--where a of such intervention. The ongoing strengthening of commodity export boom can easily materialize. LAC currencies and/or the accumulation of interna- tional reserves in several LAC countries (e.g., Brazil, Consequently, countries in the LAC region will con- Chile, Colombia, Mexico, and Peru) already reflects tinue to face significant challenges in the monetary/ the interventions in the exchange rate market moti- exchange rate front. Having moved to more robust vated by the need to find a reasonable balance in the monetary policy frameworks that rely on greater ex- face of the tensions (Figure 14). change rate flexibility, and given a generally limited scope for fiscal policy adjustment, the feasible fiscal- Not surprisingly, therefore, alternative options to "lean monetary policy mix will likely shift the burden to- against the wind" in exchange rate markets might PART I LAC Recovering in a Multi-Polar Global Economy 17 Office of Regional Chief Economist FIguRe 14 Exchange Market Pressures A. Brazil B. Chile 8 6 8 3 2 3 4 6 1 2 2 4 0 1 0 2 ­1 0 Reserve Accumulation Reserve Accumulation ­2 0 Appreciation Pressures ­2 Appreciation Pressures ­1 Exchange Market Pressure Exchange Market Pressure Reserve Accumulation Reserve Accumulation ­4 ­2 Appreciation Pressures ­3 ­2 Appreciation Pressures Jan-03 May-03 Sep-03 Jan-04 May-04 Sep-04 Jan-05 May-05 Sep-05 Jan-06 May-06 Sep-06 Jan-07 May-07 Jan-08 May-08 Sep-08 Jan-09 May-09 Sep-09 Jan-10 Jan-03 May-03 Sep-03 Jan-04 May-04 Sep-04 Jan-05 May-05 Sep-05 Jan-06 May-06 Sep-06 Jan-07 May-07 Jan-08 May-08 Sep-08 Jan-09 May-09 Sep-09 Jan-10 Sep-07 Sep-07 Exchange Market Pressure Exchange Market Pressure ­4 ­3 Jan-03 May-03 Sep-03 Jan-04 May-04 Sep-04 Jan-05 May-05 Sep-05 Jan-06 May-06 Sep-06 Jan-07 May-07 Jan-08 May-08 Sep-08 Jan-09 May-09 Sep-09 Jan-10 Jan-03 May-03 Sep-03 Jan-04 May-04 Sep-04 Jan-05 May-05 Sep-05 Jan-06 May-06 Sep-06 Jan-07 May-07 Jan-08 May-08 Sep-08 Jan-09 May-09 Sep-09 Jan-10 Sep-07 Sep-07 8 C. Colombia 5 D. Peru 4 6 8 5 3 4 4 6 2 3 2 4 1 2 0 0 2 1 Reserve Accumulation ­1 Reserve Accumulation 0 ­2 0 Appreciation Pressures Appreciation Pressures ­2 Exchange Market Pressure Reserve Accumulation ­1 Exchange Market Pressure Reserve Accumulation ­4 ­2 Appreciation Pressures ­3 Appreciation Pressures ­2 Jan-03 May-03 Sep-03 Jan-04 May-04 Sep-04 Jan-05 May-05 Sep-05 Jan-06 May-06 Sep-06 Jan-07 May-07 Jan-08 May-08 Sep-08 Jan-09 May-09 Sep-09 Jan-10 Jan-03 May-03 Sep-03 Jan-04 May-04 Sep-04 Jan-05 May-05 Sep-05 Jan-06 May-06 Sep-06 Jan-07 May-07 Jan-08 May-08 Sep-08 Jan-09 May-09 Sep-09 Jan-10 Sep-07 Sep-07 Exchange Market Pressure Exchange Market Pressure ­4 ­3 Jan-03 May-03 Sep-03 Jan-04 May-04 Sep-04 Jan-05 May-05 Sep-05 Jan-06 May-06 Sep-06 Jan-07 May-07 Jan-08 May-08 Sep-08 Jan-09 May-09 Sep-09 Jan-10 Jan-03 May-03 Sep-03 Jan-04 May-04 Sep-04 Jan-05 May-05 Sep-05 Jan-06 May-06 Sep-06 Jan-07 May-07 Jan-08 May-08 Sep-08 Jan-09 May-09 Sep-09 Jan-10 Sep-07 Sep-07 Note: The Exchange Market Pressure Index is the weighted average of year-on-year percentage changes in: (a) the nominal exchange rate of the local currency vis-à-vis the US dollar (such that an increase represents an appreciation of the LAC currency), and (b) the level of international reserves. The weights are given by the inverse of the annual standard deviation of the changes in the nominal exchange rate and the standard deviation of the changes in reserves. An increase in the Exchange Market Pressure index signals appreciation pressures and/or accumulation of reserves. Source: LCRCE Staff calculations based on IMF's IFS. become an important item in the monetary policy de- tions, although these controls might bring distortions to bate in LAC going forward. To the extent that currency financial systems that would be hard to reverse in the appreciation in LAC would be lower if China's currency future. Furthermore, such controls tend to be easy to were allowed to strengthen, LAC countries might also circumvent and thus become less and less ineffective interpret their intervention in exchange rate markets as over time. Alternatively, LAC authorities might focus on a way to offset a global distortion. The actual mix of averting the excessive credit expansion that could result policies will, however, depend on whether flows and from the surge in capital inflows. In this case, macro- associated appreciation pressures are perceived to be prudential policy tools could be considered, including, permanent or temporary. Controls on capital inflows, as for instance, unremunerated reserve requirements on pointed out in a recent IMF Staff Position Note (Ostry financial intermediaries and/or counter-cyclical pru- et al, 2010), is likely also to be part of the menu of op- dential (capital or provisioning) norms. Lastly, although 18 From Global Collapse to Recovery The World Bank harder to implement, fiscal tightening will also have FIguRe 15 Trends in GDP per capita of to be considered, as it could allow interest rates to be Latin America and East Asian Tigers vis-à- lower, and the real exchange rate to appreciate less, vis the United States than otherwise. In the end, with no simple solution in sight, countries in LAC are likely to have to learn to live Relative GDP Per Capita of Selected Regions relative to the US with more appreciated real exchange rates, for which 80.0% the only lasting solution is productivity growth. 70.0% Interwar Alliance Lost Washington Period for Decade Dissensus 60.0% Progress Gold Standard US Washington Assembling a productivity agenda while Period Recovers Imports Consensus 50.0% Substitution averting the downside risks of commodity 40.0% wealth LAC/US 30.0% 20.0% With the possible exception of Chile, LAC has failed to engineer sustained high growth. The process of con- 10.0% Asian Tigers/US vergence towards rich-country standards of living has 0.0% 1900 1906 1912 1918 1924 1930 1936 1942 1948 1954 1960 1966 1972 1978 1984 1990 1996 2002 2008 in fact eluded LAC for more than a century, in sharp contrast with the East Asian tigers where convergence Note. The group of East Asian tigers includes Hong Kong (China), has been taking place at a rapid rate since the 1970s Indonesia, Malaysia, Republic of Korea, Singapore, Thailand, and (Figure 15). Some hope that this trend may be chang- Taiwan (China). Source: LCRCE Staff calculations based on Mad- dison (2007, 2009), WDI and DECPG. ing emerged before the crisis, during the 2002­2008 period, where several LAC countries other than Chile (chiefly Peru and Panama, but also Brazil and Colom- numerous or unduly complex, policies to ignite growth bia) recorded significantly higher growth rates than in that are adequately tailored to the circumstances of rich countries on the strength not just of favorable ex- individual countries. Once ignited, growth would have ternal conditions (low interest rates, abundant liquidi- to be sustained through perseverant and major re- ty, and high commodity prices) but also of productivity form efforts aimed at eliminating well-know obstacles growth. With LAC coming out of this crisis relatively that undercut efficient resource allocation--such that well positioned, those hopes appear to be rekindling. competitive market forces are honed and relative pric- Moreover, as noted, the improved macro-financial es better reflect relative scarcities. Also, the large gaps resiliency of the region gives greater assurance that that LAC has in education, physical infrastructure, whatever gains from growth that LAC could achieve in and the ability to adopt and adapt new technologies the future will be less likely to be wiped out by finan- relative to, say, the Asian Tigers would have to be sys- cial crises, as tended to be the case in the past. In ad- tematically closed. The associated need for high in- dition, LAC has been making significant strides in the vestment levels would have to be supported via higher equity agenda--through noticeable improvement in national savings (and the willingness to keep savings poverty-reducing social policies--and this could help in the country) as well as prudent access to foreign mobilize consensus in favor of a long overdue reform savings, particularly in the form of FDI. That would agenda aimed at raising productivity growth. in turn require a leap in the quality of the investment climate, including through a better contractual and Some of the key conditions for LAC to raise its growth informational environment, much reduced corruption rate sustainably above the world's average are, there- levels, sensible regulation, a major reduction in crime fore, in place. But the jury is still out on whether the and violence, a major thinning of red tape, etc. region will be able to seize the opportunity, as LAC fac- es a very tall order in this regard. Seizing the oppor- The scope for seizing the opportunity to move to a high- tunity will require well-designed, but not necessarily er growth path can be greatly enhanced in commodity PART I LAC Recovering in a Multi-Polar Global Economy 19 Office of Regional Chief Economist rich countries. But this can become a reality only if the tions and work effort that could result from the abil- associated windfall earnings are managed judiciously ity of special interests to capture the natural resource within a long-term horizon, so as to avoid falling victim rents to boost wasteful current spending). A clear sign to the so-called "natural resource curse." This would that these bad side effects are being avoided would require that policies and institutions are in place to be given by a demonstrated ability of governments to dampen the possible welfare-reducing "Dutch Dis- save a substantial fraction of the commodity-related ease effects" (via an overvalued currency that unduly revenue windfall--that is, by the materialization of hinders non-commodity export diversification) and significant and continuous cyclically-adjusted primary "rent-seeking effects" (i.e., the weakening of institu- fiscal surpluses going forward. 20 From Global Collapse to Recovery part II executive summary Labor Markets in In a refreshing break from the past, economic activity in the Latin American and Caribbean (LAC) region was not hit as hard by the LAC during the global financial crisis as in previous recessions, or as compared to other regions. In addition, controlling for the size of the eco- 2009 Downturn: nomic downturn, the region experienced significantly milder labor market adjustments than in previous crises. Unemployment did increase (about 3.5 million joined the ranks of the unemployed) Common Patterns, but at a slower pace than in the recessions of the late 1990s. Per- haps even more surprisingly, the share of informal employment Surprises & Puzzles --a commonly used indicator of deterioration of the quality of employment--did not rise. These mild quantity adjustments in the labor market are especially puzzling in light of evidence that real wages in LAC did not decline during the 2009 recession, re- flecting the combination of low inflation and downward rigidity in nominal wages. The overall picture of labor market adjustments during this crisis therefore stands in sharp contrast with past cri- sis episodes, during which significant increases in unemployment were generally combined with increasing informality and a sharp fall in real wages. Behind these general patterns, substantial differences across countries remain. Unemployment rose relatively rapidly in some countries, such as Chile and Mexico, while in Brazil and Peru it returned to pre-crisis levels rather quickly. Another common trend is that the crisis had a differential impact across genders, hitting males harder than females. While labor force participa- tion remained relatively stable overall, the share of female workers increased relative to males. Hence, the crisis did not reverse the trend towards closing the gap between male and female participa- tion rates. In the countries studied, with the exception of Colombia and Mexico, men became unemployed proportionally more than women. While the reasons behind the better labor market performance remain to be studied, several factors may have played a role. The recent shock was transmitted mainly through the trade channel, thus affecting primarily the relatively land- and capital-intensive tradable sectors. In addition, the domestic policy environment has also changed: Sounder macro-financial policy frameworks and timely finance from multilateral institutions allowed the imple- mentation of counter-cyclical fiscal policies that helped dampen the unemployment response. Finally, specific labor market poli- cies, such as temporary employment programs and employment subsidies, may have also contributed to mitigating the impact of the external shock. From Global Collapse to Recovery 21 Office of Regional Chief Economist adjustments in employment, gentina and Brazil experienced a decline in employ- unemployment, and labor force ment rates of around 1pp but quickly bounced back participation5 in 2009Q3 to pre-recession levels. Chile also experi- enced a sharp decline in employment but it had not In the last quarter of 2008, the world economy en- rebounded by 2009Q3. In other LAC countries the la- tered into the worst recession since the Great Depres- bor market response actually began earlier. This was, sion. As discussed in Appendix I.A of the Part I of this for example, the case of Mexico and Peru. Whereas report, the Latin America and the Caribbean (LAC) in Mexico the employment rate fell by 2.1 pp from region's economic performance during this crisis was 2008Q1 to 2009Q1, employment declined by 1.6 pp better than in most of its own past downturns, and in Peru over the same period.6 than in other emerging regions. Nonetheless, some of the pains of a recession were unavoidable. The unemployment rate is affected both by labor supply and labor demand forces. During downturns, While the consequences of the crisis on labor markets labor demand typically falls while the effect of a reces- around the world are still unfolding, it is clear that the sion on the supply for labor is in principle ambiguous. impact is quite large. In Eastern Europe, the unemploy- Two opposite forces operate on the supply of labor ment rate increased by 2 percentage points (pp) dur- and, hence, on the degree of labor force participation ing 2009 while it grew by 2.4 pp on average in the during recessions. On the one hand, the prospects developed world, with the largest increases in countries of finding a job shrink as the unemployment rate in- such as Ireland (10pp since 2008Q1) and Spain (9pp creases, hence reducing the incentives for actively within the same period). In LAC, the unemployment looking for a job. On the other hand, the reduction rate rose by 1.2 pp, from 7 percent in 2008 to 8.2 per- of employment, hours worked, or real wages among cent in 2009, according to the latest estimates, imply- the breadwinners of the family might provide an in- ing that the number of the unemployed increased by centive for the females and youth in the family (who 3.5 million to a total of 22.5 million by the end of the are traditionally less attached to the labor market) 2009 (ILO, 2010). In spite of this general pattern, both to search for a job.7 The final impact on overall la- the level and the change in unemployment rates varied bor force participation rates depends on the relative widely across countries. Encouragingly, unemployment strengths of these forces, and of the generosity and started to fall in some countries starting in 2009Q2, but coverage of safety nets in the economy, which might it is still too early to tell if this favorable development alleviate the welfare consequences of unemployment constitutes a change in trend. It might instead simply among some of the members of the household. In reflect a seasonal pattern, as employment in the region LAC, the relative paucity of functional unemployment typically rises in the second half of the year. insurance schemes tends to strengthen incentives to remain in the job search, including in the informal Employment in most LAC countries started to contract sector. What was the actual net effect of these forces after the fourth quarter of 2008 (2008Q4). Thus, Ar- in LAC during the crisis? 5 This section draws on data from Argentina, Brazil, Chile, Colombia, Mexico, and Peru, for which aggregate labor market data are available up to the third quarter of 2009. 6 For an early assessment of the impact of the crisis on labor markets in the LAC region, see Freije-Rodríguez and Murrugarra (2009) 7 Labor force participation varies substantially by gender among LAC countries. On average, female participation rates are lower than those of males, although these differences are quite heterogeneous across LAC countries. Participation rates for females in the region are lowest in Chile and Mexico (slightly above 40 percent) and highest in Peru (around 60 percent). In Colombia, Argentina and Brazil they are rapidly approaching 50 percent. There is also quite a lot of heterogeneity across countries in the participation of men, with a relatively low rate in Brazil, around 66 percent, that is, more than 10 pp below the rates in the countries with the highest rates in our sample, Mexico and Peru. 22 From Global Collapse to Recovery The World Bank The overall participation rate remained relatively it increased mildly in Brazil and Colombia (around 0.5 constant in most countries, but female participation pp). The increase in female unemployment rates, by increased relative to males. During the global down- contrast, has been weaker. Again, Chile experienced turn, overall participation rates have barely changed a relatively large increase in female unemployment in in Brazil, Chile, Mexico and Peru. The largest excep- 2009Q3 (1.6 pp yoy)--which is however only about tion to this pattern was observed in Colombia, which half the increase in the rate of male unemployment. experienced a net increase of 2.5 pp during 2008Q3­ Female unemployment declined mildly in Brazil, and 2009Q3, followed by Argentina (a 1 pp increase dur- more sharply in Peru (by 1.7 pp yoy) to reach 9.5 ing the same period). While the actual response of percent (the lowest in the last 5 years). The outlier male and female participation rates to the recession was Mexico, where female unemployment rates rose varied across LAC countries, the pre-existing trend by 2.4 pp. toward the narrowing of gender gap in labor force par- ticipation was either maintained or even accentuated The effect of the crisis on the gender composition of during the crisis. This was because, similarly to previ- employment and unemployment is likely related to ous crisis, male participation rates tended to decline its differential effect across the manufacturing and while female participation rates tended to increase; service sectors. As shown in Appendix I.A of Part 1, or in the countries where both increased in tandem, countries with higher trade exposure and specialized female participation tended to increase more (see in manufacturing exports experienced larger growth Figure 16). collapses. Hence, the external shock was transmitted to the region through a fall in the demand for trad- Even with rising female participation relative to males, able goods. It is well known that manufacturing traded the unemployment rate of males rose relative to fe- sectors tend to employ males in a larger proportion males in the LAC-6 (Figure 17). Male unemployment than females. Hence, a shock hitting those sectors is rates increased the most in Chile, by 2.9 pp year- likely to have a larger impact on males. Figure 18 illus- on-year (yoy) in 2009Q3, followed by Argentina and trates that the reduction in employment was strongest Mexico (2.2 and 1.4 pp, respectively). Male unem- in manufacturing sectors during the recession. It de- ployment remained relatively stable in Peru whereas picts the variation in the share of workers employed in FIguRe 16 Participation Rate by Gender in Selected LAC Countries A. Male Population B. Female Population 85% 85% Peru Mexico 75% Argentina 75% Chile Colombia 65% Brazil 65% Peru 55% 55% Argentina Brazil 45% 45% Colombia Mexico Chile 35% 35% Mar-05 Mar-07 Mar-08 Mar-09 Mar-05 Mar-07 Mar-08 Mar-09 Sep-05 Mar-06 Sep-06 Sep-07 Sep-08 Sep-09 Sep-05 Mar-06 Sep-06 Sep-07 Sep-08 Sep-09 Source: International Labor Organization PART II Labor Markets in LAC During the 2009 Downturn 23 Office of Regional Chief Economist FIguRe 17 Unemployment Rate by Gender in Selected LAC Countries A. A. Male Population B. B. Female Population 20% 20% 18% 18% Colombia 16% 16% 14% 14% Argentina 12% 12% Brazil Argentina 10% Colombia 10% Peru 8% 8% Chile 6% Peru 6% Chile Brazil Mexico 4% 4% 2% Mexico 2% 0% 0% Mar-05 Sep-05 Mar-06 Sep-06 Mar-07 Sep-07 Mar-08 Sep-08 Mar-09 Sep-09 Mar-05 Sep-05 Mar-06 Sep-06 Mar-07 Sep-07 Mar-08 Sep-08 Mar-09 Sep-09 FIguRe 18 Change in Employment Levels by Activity A. Industry B. Services in Percentage Points in Percentage Points 2.5% 2.5% 2.0% Female Male 2.0% 1.5% 1.5% 1.0% 1.0% 0.5% 0.5% 0.0% 0.0% ­0.5% ­0.5% ­1.0% ­1.0% ­1.5% ­1.5% Female Male ­2.0% ­2.0% ­2.5% ­2.5% Argentina Brazil Chile Mexico Argentina Brazil Chile Mexico Note: Differences in the change of the share of manufacturing and services employment share between June 2009 to June 2008 and June 2008 to June 2007. Source: Cho and Newhouse (2010). manufacturing and services in the 2008Q2­2009Q2 major countries, except for Argentina. Hence, in con- period compared to the preceding year (2007Q2­ trast with previous recessions in the region, when the 2008Q2) in Argentina, Brazil, Chile and Mexico.8 We external sector actually helped dampen the labor mar- observe that the share of employment in services has ket response to the shock, these numbers suggests an increased faster relative to manufacturing during the asymmetric effect of the downturn in global demand recession than during the last year of expansion in all for tradable versus non tradable goods in LAC. 8 The double difference is intended to isolate the change in shares associated with the recession from the secular structural change towards the service sector that is taking place in every country. 24 From Global Collapse to Recovery The World Bank No clear signs of a crisis- available so far, however, do not suggest a reversal induced expansion in informal of the rising trend in the formal employment share, employment although the rate of increase in formality did appear to decelerate during the crisis.9 The evolution of formal Not only were unemployment increases relatively mild employment during this global crisis can be gauged in most countries, but rises in informality--a typical by looking at the share of workers with a labor contract byproduct of economic downturns in Latin America-- and that of workers with health insurance. The share were conspicuous only by their absence. The deterio- of formal employment in total employment (using hav- ration in economic activity in LAC labor markets has ing a labor contract as a measure of formality) kept on traditionally led to an increase in informal employ- rising during the crisis in all countries but Colombia: ment, or in hidden unemployment. Although informal between 2008Q2 and 2009Q2 in Brazil (by 3.4 pp), employment had progressively lost ground relative Mexico (1.4 pp), and Peru (3.1 pp) (Figure 19). When to formal employment in the years preceding the we use workers with access to health insurance as an global crisis, one might have expected the crisis to indicator of formality the message is less clear. Formal have reversed this trend, especially considering the employment declined considerably in Mexico (by 1.8 countercyclical nature of informal unemployment and pp) and Colombia (by 1.4 pp), but increased sharply the dearth of formal hiring observed in previous re- in Argentina (by 3.5 pp) and Peru (by 6.6 pp). cessions (Bosch and Maloney, 2008). The indicators Zooming in on the dynamics FIguRe 19 Type of Labor Contracts in of labor markets: the cases of Selected LAC Countries argentina and brazil Share of Workers Relative to the Total Number of Employees We now examine the behavior of the different segments Change from 2008 to 2009, in Percentage Points 7 of the active labor force--formal employment, informal With Contract With Health Insurance 6 employment, unemployment and self-employment-- 5 during the business cycle in Argentina and Brazil, us- 4 ing quarterly information for the period 1991­2009.10 3 Figure 20 depicts their evolution in Argentina and Bra- 2 zil. The shaded areas in the figures represent reces- 1 sions--as defined in Appendix II.A. The message from 0 ­1 the previous section is reinforced. Formal employment ­2 stagnated somewhat with respect to the expansion pe- ­3 riod, but it did not lose relative weight in favor of infor- Argentina Brazil Colombia Ecuador Mexico Peru mality as it did during previous recessions. Note: For Argentina and Brazil, LCRCE Staff calculations based on Households Surveys. When the bars are not shown in the graph is because the information for that specific country is not available. In Argentina, the unemployment rate increased rap- Source: International Labor Organization. idly in the previous downturn (1998Q2­2002Q1), 9 There is no consensus on the best definition of informality (Perry et al., 2007) and data availability limitations preclude us from presenting a full picture of the behavior of the informal segment of the labor market (under alternative definitions) during this crisis. 10 The quarterly data on the different segments of the labor market spans from 1995Q2 to 2009Q3 in Argentina, and from 1991Q1 to 2009Q4 in Brazil. Formal employment in Argentina is defined as the share of workers having access to both health insurance and a pension benefit through their jobs. In the case of Brazil, formal employees are defined as those with "carteira de trabalho assinada," a signed labor card that gives access to social insurance or social protection. Lastly, we consider a residual employment condition, which we call "other employees" and comprises owners of firms and unpaid employees. PART II Labor Markets in LAC During the 2009 Downturn 25 Office of Regional Chief Economist reaching a peak close to 21 percent in 2002Q3 be- Brazil exhibits an upward trend in formal employment fore declining sharply (see Figure 20.A). The initial since the second half of 2004, which led to a net cu- absorption of this large number of unemployed work- mulative gain of the share of formal workers of 8 pp by ers coincided with a massive increase of the share of 2009Q4 (Figure 20.B). This steady increase in formal informal employment, which rose by 9 pp between employment was not reversed by the crisis, although 2002Q3 and 2004Q3. The formal sector remained the share of informality declined at a slower pace in stubbornly closed to new hires with the result that the second half of the 2000s (falling by only 0.7 pp the share of formality continued a downward trend in 2009). The share of self-employed increased by for more than a year after the recession. When the nearly 0.4 pp during 2009. recovery strengthened, starting in the second half of 2003, the formal sector reactivated and started absorbing unemployed and informal workers. At a milder labor market response the start of the current downturn, by contrast, the to crisis this time around upward trend in formal employment in Argentina leveled off somewhat but did not decrease. Interest- In sum, while unemployment did rise in the LAC-6 ingly, the share of informal employment continued countries during the recent recession, these increas- to lose ground during this period (by 0.9 pp) in favor es were relatively muted. Labor force participation of formal employment and the self-employed. This rose, especially among females, and there were no finding reflects some degree of substitution between clear signs of deterioration in the quality of employ- informality and self-employment during the last re- ment, at least in terms of growing informality.11 Next, cession in Argentina. we investigate whether labor market responses during FIguRe 20 Active Labor Force by Status A. Argentina B. Brazil 60% 60% 50% 50% Formal Employee 40% Formal Employee 40% Informal Employee 30% 30% Informal Employee 20% Self-Employed 20% Self-Employed Unemployed Unemployed 10% 10% Employed. Others Employed. Others 0% 0% Jun-95 Sep-96 Jan-98 Apr-99 Jul-00 Oct-01 Jan-03 Apr-04 Jul-05 Oct-06 Jan-08 Apr-09 Mar-91 Jun-92 Sep-93 Jan-95 Apr-96 Jul-97 Oct-98 Jan-00 Apr-01 Jul-02 Oct-03 Jan-05 Apr-06 Jul-07 Oct-08 Note: Shaded areas are recessions, as defined in Appendix II.A. Source: LCRCE calculations based on Households Surveys. For Argentina, Encuesta Permanente de Hogares (EPH) y Encuesta Permanente Continua de Hogares (EPHC), for Brazil, Pesquisa Mensal de Emprego (PME). 11 In an earlier assessment, Ferreira and Schady (2009) have a relatively benign view of the effects of the global crisis on Latin American labor markets. 26 From Global Collapse to Recovery The World Bank the current downturn were similar to those in previous unemployment in previous recessions, with increases recessions.12 of 8.3 pp and 5.6pp, respectively. In contrast, unem- ployment rose only mildly during the current recession The main conclusion of our analysis is that, control- --by 0.4 pp and 2.3 pp in Argentina and Chile, re- ling for the size of the downturn, the rise in unem- spectively. Note also that even in Mexico, which ex- ployment this time around was significantly smaller perienced the sharpest GDP contraction in the region than in previous crisis episodes. This result is even this time around, the unemployment response was more surprising considering that average real wages less than half that observed during the Tequila crisis, (in the economy as a whole) did not fall during this i.e., between 1994Q4 and 1995Q3. recession--given the combination of low inflation and downward rigidity in nominal wages--whereas they The milder unemployment response this time around fell sharply in previous crises (on account of signifi- could be attributed either to softer GDP contractions cant inflation spikes). Let us now consider some of the (than in previous crises) or to a weaker link between details behind this result. the change in economic activity and the change in unemployment (that is, to a lower elasticity of un- Figure 21.B illustrates the fact that the rise in unem- employment to output). First, GDP did contract by ployment during previous recessions substantially ex- less during the current crisis, but not in all countries ceeded that observed in the current one. Argentina (Figure 21.A). Argentina, Colombia and Mexico ex- and Chile experienced the largest surges in the rate of perienced larger contractions in economic activity FIguRe 21 GDP Growth and Change in the Unemployment Rate During Recessions A. GDP Growth (%) B. Change in Unemployment (p.p.) Current Recession Current Recession Mexico Mexico Previuos recession Previous recession Colombia Colombia Chile Chile Brazil Brazil Argentina Argentina ­30 ­25 ­20 ­15 ­10 ­5 0 0 2 4 6 8 10 Note: Previous recession periods are: Argentina (1998.Q4­2002.Q2); Brazil (1997.Q4­1998.Q2); Chile (1998.Q3­1999.Q4); Colombia (1998.Q3­1999.Q4); and Mexico (1995.Q1­1996.Q1). Current recession periods are: Argentina (2008.Q3­2009.Q2); Brazil (2008.Q4­ 2009.Q2); Chile (2008.Q3­2009.Q3); Colombia (2008.Q3­2009.Q2); and Mexico (2008.Q2­2009.Q2). Source: LCRCE Staff calculations based on National Statistical Institutes data. 12 The samples used are constrained by the availability of quarterly unemployment data. We use the Harding and Pagan (2002) adaptation of the Bry-Boschan algorithm (BBQ) to date recessions from quarterly GDP series in Argentina, Brazil, Chile, Colombia, and Mexico. A more detailed description of the algorithm is provided in Appendix II.A. As a result, the measures of GDP contraction, i.e., of decreases in the level of GDP, used in this part of the report do not coincide with those in Part I, where growth collapse is defined simply as the difference in the annual GDP growth rate between 2007 and 2009. PART II Labor Markets in LAC During the 2009 Downturn 27 Office of Regional Chief Economist during their respective previous downturns compared FIguRe 22 Sensitivity of Unemployment to the recent global crisis. In the case of Argentina, the to Economic Activity crisis of 1998­2002 was associated with a dramatic collapse of GDP in real terms (28 percent), while dur- Elasticity of Unemployment With Respect to GDP Growth 2.5 ing the current episode GDP contracted by almost 10 Previous recession percent. Mexican GDP fell in real terms by nearly 15 Current Recession 2.0 percent during the Tequila crisis, while it shrank by 12 percent during 2008­2009. In Colombia differences 1.5 are marked, but the contractions are overall smaller (7 percent during the previous recession vs. 1 percent 1.0 during the global crisis). Brazil and Chile instead suf- fered a similar contraction in GDP during the current 0.5 crisis compared to certain previous episodes. 0.0 Argentina Brazil Chile Colombia Mexico Second, while there is a great deal of heterogeneity in the responses among the 5 LAC countries analyzed, it Note: Previous recession periods are: Argentina (1998.Q4­2002. Q2); Brazil (1997.Q4­1998.Q2); Chile (1998.Q3­1999.Q4); Co- is clear that the rise in the unemployment rate (in per- lombia (1998.Q3­1999.Q4); and Mexico (1995.Q1­1996.Q1). centage points) for each one percent contraction in Current recession periods are: Argentina (2008.Q3­2009.Q2); Brazil (2008.Q4­2009.Q2); Chile (2008.Q3­2009.Q3); Colombia GDP (i.e., the semi-elasticity of unemployment to out- (2008.Q3­2009.Q2); and Mexico (2008.Q2­2009.Q2). Source: put) was significantly larger in past downturns com- LCRCE Staff calculations based on National Statistical Institutes data. pared to the current one (Figure 22). For instance, the semi-elasticity of unemployment to output has halved in Chile, from 1 in the previous recession to tion in Colombia merit further study. Other elements 0.5 during the last episode. A similar reduction is ob- might be at play behind the sharp output elasticity of served in Mexico, although the semi-elasticities there unemployment in Colombia. Colombia has one of the are smaller (0.25 during the Tequila crisis vs. 0.12 most binding minimum wages in the region, and pre- during the current one). In Argentina and Brazil, the vious studies have established their negative impact semi-elasticities of unemployment were 6 and 3 times on employment (Cunningham, 2007). During the re- larger in past downturns, respectively, than in the cur- cession, these negative effects on employment might rent one. have exacerbated. The sole exception is Colombia, where the semi-elas- Aggregate wage behavior fails to explain the relatively ticity of unemployment to output during the previous muted response of unemployment to output contrac- recession was about one third the one observed in tion among LAC countries during the current global the recent downturn. The Colombian result might be crisis. While in previous recessions average real wages related to the massive increase in labor force partici- in the whole economy fell rapidly due to high infla- pation observed during the downturn. During 2008Q3 tion, this time around they exhibited an upward trend, and 2009Q2, overall participation increased by 2.5 pp pointing towards downward rigidity in nominal wages in Colombia while in the other four countries it barely in a new LAC context of low inflation.13 In Argentina, changed. The forces behind this increase in participa- real wages for males (formal and informal workers) 13 Unfortunately, we only had access to quarterly data on average wages for Argentina and Brazil. Instead, most countries publish data on compensation per employee in the industrial sector, which are shown in the case of Mexico. In Figure 2.8 recessions are represented by the shaded areas, and we have set the CPI and wage index to 100 for the first quarter of the recession to highlight the behavior of prices during recessions. 28 From Global Collapse to Recovery The World Bank declined sharply in the previous recession, especially Another conjecture could be that the rising formal em- after the convertibility was abandoned and inflation ployment in the second half of the 2000s might have spiked at the end of 2001. During the current reces- reduced the volatility of LAC labor markets inasmuch sion, real wage behavior has been apparently differ- as it came with increasing job protection. If this is the ent, although the interpretation of the data is clouded case, one might expect a slower job recovery this time by well-know concerns with an underestimation of the around compared to previous recessions. This would CPI. After a moderate decline in the first quarter of the be because firing costs/restrictions can be expected recession, reported real wages returned to their pre-re- to reduce also hiring when conditions turn favorable, cession upward trend (Figure 23.A and Figure 23.B).14 as firms anticipate costs associated with future layoffs In Brazil, after an initial upward spike real wages de- (Bentolila and Bertola, 1990). clined sharply throughout the 1997­1998 recession, while they showed an upward trend during the current Lastly, countercyclical fiscal policies, to the extent that turmoil (Figure 23.C and Figure 23.D). The accelera- they were implemented in LAC, might have supported tion of inflation in Mexico during 1995­1996 led to a employment, for example, through spending on pub- major downward adjustment in real wages (18 pp in lic works. In addition, efforts to prevent labor shedding a year), and while there are signs of wage moderation were an important ingredient of stimulus packages in during the current crisis there was no clear downward some LAC countries. Overall, it is estimated that the adjustment (Figure 23.E and Figure 23.F). region's government had planned to invest an addi- tional US$25 billion in 2009, which would imply a 20 The milder response of unemployment to the decline percent increase in public spending on infrastructure in real economic activity during the global crisis com- above the original targets (Schwartz, Andrés, and Dra- pared to previous recessions constitutes an apparent goiu, 2009). Although it is premature to assess the puzzle that deserves further research, especially since effectiveness of these programs on employment, they we observe, if anything, stronger wage rigidities in the might have helped mitigating the unemployment re- recent past rather than in previous recessions with the sponse. Regarding labor market policies, Appendix available data. What explains the different behavior this II.B provides a description of policy interventions spe- time around? Among the many conjectures, it could be cifically designed to cope with the effects of the inter- argued first that the nature of the recent shock is differ- national financial crisis on labor markets in selected ent. Whereas the current crisis originated outside the LAC countries. region, previous recessions were either home-brewed or externally generated but magnified by internal weak- We conclude this section with a closer look at the evo- nesses. Previous recessions affected most sectors of lution of formal and informal employment and wages in economic activity while the current global shock was Argentina and Brazil. Compared to Figure 20 presented transmitted mainly through the external sector, which above, we focus our discussion here on the compara- is typically (land or) capital intensive. Moreover, to the tive performance of the shares of formal and informal extent that this recession was perceived as a tempo- employment in the active labor force during the reces- rary phenomenon, firms in sectors where training and/ sionary periods associated to the current crisis vis-à-vis or firing are costly (i.e., formal sectors) might have previous crises (Figure 24).15 Formal employment de- opted for labor hoarding rather than layoffs. creased significantly in Argentina and Brazil in previous 14 The corresponding graphs for females present a qualitatively similar pattern. We prefer to show the evolution of male wages since it is well known that compositional changes in the labor force across the business cycle are more pronounced for females than for males. 15 As before, in Argentina the point of comparison is the recessionary period from 1998Q2 to 2002Q2 associated to the financial crisis. In Brazil we identify two previous recessionary periods, 1997Q4 to 1998Q2 and 1999Q4 to 2002Q2. PART II Labor Markets in LAC During the 2009 Downturn 29 Office of Regional Chief Economist FIguRe 23 Evolution of Real Wages and CPI (Males) A. Argentina B. Argentina Previous Crisis Current Crisis Index Jun-98 = 100 Previous Crisis Index Jun-08 = 100 Current Crisis 150 Previous Crisis Index Jun-98 = 100 150 Current Crisis Index Jun-08 = 100 150 140 Index Jun-98 = 100 150 140 Index Jun-08 = 100 150 150 140 130 140 130 140 140 130 120 130 120 130 130 120 110 120 110 120 CPI 120 110 100 110 100 110 CPI 110 CPI 100 90 100 90 CPI CPI 100 100 90 80 90 80 CPI Real Mean Wage 90 Real Mean Wage 90 Real Mean Wage 80 70 80 70 Real Mean Wage Real Mean Wage 80 80 70 60 Real Mean Wage 70 60 70 70 Mar-07 Jun-07 Sep-07 Dec-07 Mar-08 Jun-08 Sep-08 Dec-08 Mar-09 Jun-09 60 60 Jun-97 Sep-97 Dec-97 Mar-98 Jun-98 Sep-98 Dec-98 Mar-99 Jun-99 Sep-99 Dec-99 Mar-00 Jun-00 Sep-00 Dec-00 Mar-01 Jun-01 Sep-01 Dec-01 Mar-02 Jun-02 Sep-02 Dec-02 Mar-03 Jun-03 Sep-03 Dec-03 Mar-04 60 60 Mar-07 Jun-07 Sep-07 Dec-07 Mar-08 Jun-08 Sep-08 Dec-08 Mar-09 Jun-09 Jun-97 Sep-97 Dec-97 Mar-98 Jun-98 Sep-98 Dec-98 Mar-99 Jun-99 Sep-99 Dec-99 Mar-00 Jun-00 Sep-00 Dec-00 Mar-01 Jun-01 Sep-01 Dec-01 Mar-02 Jun-02 Sep-02 Dec-02 Mar-03 Jun-03 Sep-03 Dec-03 Mar-04 Mar-07 Jun-07 Sep-07 Dec-07 Mar-08 Jun-08 Sep-08 Dec-08 Mar-09 Jun-09 Jun-97 Sep-97 Dec-97 Mar-98 Jun-98 Sep-98 Dec-98 Mar-99 Jun-99 Sep-99 Dec-99 Mar-00 Jun-00 Sep-00 Dec-00 Mar-01 Jun-01 Sep-01 Dec-01 Mar-02 Jun-02 Sep-02 Dec-02 Mar-03 Jun-03 Sep-03 Dec-03 Mar-04 C. Brazil D. Brazil Previous Crisis Current Crisis Index Sep-97 = 100 Previous Crisis Index Sep-08 = 100 Current Crisis 115 Previous Crisis Index Sep-97 = 100 115 Current Crisis Index Sep-08 = 100 115 Index Sep-97 = 100 115 Index Sep-08 = 100 110 115 110 115 110 110 105 110 105 110 105 Real Mean Wage 105 100 105 Real Mean Wage 100 105 100 Real Mean Wage 100 Real Mean Wage 95 100 95 CPI 100 Real Mean Wage 95 95 Real Mean CPI Wage 90 95 CPI 90 95 CPI 90 CPI 90 85 CPI 90 85 90 85 85 80 85 80 85 Mar-96 Jun-96 Sep-96 Dec-96 Mar-97 Jun-97 Sep-97 Dec-97 Mar-98 Jun-98 Sep-98 Dec-98 Mar-99 Jun-99 Sep-99 Dec-99 Mar-00 80 Mar-07 Jun-07 Sep-07 Dec-07 Mar-08 Jun-08 Sep-08 Dec-08 Mar-09 Jun-09 Sep-09 Dec-09 80 80 Mar-96 Jun-96 Sep-96 Dec-96 Mar-97 Jun-97 Sep-97 Dec-97 Mar-98 Jun-98 Sep-98 Dec-98 Mar-99 Jun-99 Sep-99 Dec-99 Mar-00 Mar-07 Jun-07 Sep-07 Dec-07 Mar-08 Jun-08 Sep-08 Dec-08 Mar-09 Jun-09 Sep-09 Dec-09 80 Mar-96 Jun-96 Sep-96 Dec-96 Mar-97 Jun-97 Sep-97 Dec-97 Mar-98 Jun-98 Sep-98 Dec-98 Mar-99 Jun-99 Sep-99 Dec-99 Mar-00 Mar-07 Jun-07 Sep-07 Dec-07 Mar-08 Jun-08 Sep-08 Dec-08 Mar-09 Jun-09 Sep-09 Dec-09 e. Mexico Previous Crisis Mexico F.Current Crisis Index Dec-94 = 100 Previous Crisis Index Dec-07 = 100 Current Crisis 180 Previous Crisis Index Dec-94 = 100 180 Current Crisis Index Dec-07 = 100 180 170 Index Dec-94 = 100 180 170 Index Dec-07 = 100 CPI 180 170 160 180 170 160 CPI 170 160 150 170 160 150 CPI 160 150 140 160 150 140 150 140 130 150 140 130 140 130 120 140 130 120 CPI 130 120 110 130 120 110 CPI 120 110 100 120 110 100 Real Compensation per CPI 110 Real Compensation per 110 100 90 Employee -Industry 100 90 Employee-Industry Real Compensation per 100 Real Compensation per 100 90 80 Employee -Industry 90 80 Employee-Industry Real Compensation per Real Compensation per 90 80 70 Employee -Industry 90 80 70 Employee-Industry Mar-94 Jun-94 Sep-94 Dec-94 Mar-95 Jun-95 Sep-95 Dec-95 Mar-96 Jun-96 Jun-07 Sep-07 Dec-07 Mar-08 Jun-08 Sep-08 Dec-08 Mar-09 Jun-09 Sep-09 Dec-09 80 70 80 70 Mar-94 Jun-94 Sep-94 Dec-94 Mar-95 Jun-95 Sep-95 Dec-95 Mar-96 Jun-96 Jun-07 Sep-07 Dec-07 Mar-08 Jun-08 Sep-08 Dec-08 Mar-09 Jun-09 Sep-09 Dec-09 70 70 Mar-94 Jun-94 Sep-94 Dec-94 Mar-95 Jun-95 Sep-95 Dec-95 Mar-96 Jun-96 Jun-07 Sep-07 Dec-07 Mar-08 Jun-08 Sep-08 Dec-08 Mar-09 Jun-09 Sep-09 Dec-09 Note: The graph depicts the evolution of quarterly average male real wages for the whole economy (all sectors) and the CPI index around recession periods for Argentina and Brazil. For Mexico, wage data refers to average compensation per employee in the manufacturing sec- tor. Source: LCRCE calculations for Argentina based on Encuesta Permanente de Hogares (EPH) and Encuesta Permanente Continua de Hogares (EPHC) and for Brazil based on Pesquisa Mensal de Emprego (PME). For Mexico, Banco de México and INEGI. 30 From Global Collapse to Recovery The World Bank FIguRe 24 Changes in Active Labor Force by Status During Recessions (Percentage Point Variation Between Peak and Trough) A. Argentina B. Brazil 1.0 1.0 Previous Recession (98q2­02q1) 0.5 Previous Recession (97q3­98q1) Current Recession (08q3­09q1) Previous Recession (99q3­00q1) 0.0 0.5 Current Recession (08q3­09q1) ­0.5 ­1.0 0.0 ­1.5 ­2.0 ­0.5 ­2.5 ­3.0 ­1.0 ­3.5 ­4.0 ­1.5 Formal Employment Informal Employment Formal Employment Informal Employment Source: LCRCE calculations based on Households Surveys: Pesquisa Mensal de Emprego (PME); Encuesta Permanente de Hogares (EPH) and Encuesta Permanente Continua de Hogares (EPHC). crises while it slightly increased in both countries dur- Final remarks ing the current downturn. In contrast, informality ei- ther increased or remained constant in previous crises Our exploratory assessment of the evolution of labor while it has declined in Argentina and Brazil this time market aggregates in selected LAC countries during around. Counter-cyclical fiscal policies, to the extent the recent crisis shows that labor markets were not hit that they targeted public works and contracts with as hard as in previous recessions. Neither quantities-- firms in the formal sector, might help explaining the unemployment, employment--nor wages suffered differential behavior across the recessions. as much as in the past, while indicators of labor qual- ity such as the share of formal employment continued Real wages in the formal and informal sector in Ar- growing, albeit at a slower pace. While the differen- gentina and Brazil might also help to shed some light tial impact of the crisis with respect to past episodes on the different behavior of LAC labor markets during deserves further study, it may be explained partly by the current recession. In Argentina, the sharp decline the nature of the external shocks. The greater ability of of relative wages of informal with respect to formal the region to withstand shocks (due to sounder macro- workers in previous crises (especially after exiting the financial policy frameworks) may also have played an convertibility program) may help understand the dif- important role. Counter-cyclical fiscal policies were im- ferent behavior of quantities during the recession. In- plemented to one extent or another in most countries stead, the lack of dynamism of informal employment studied here, which may have dampened the initial during the current downturn might be related to wage unemployment response. It is also possible that labor setting: wages of informal workers have grown as fast market policy responses may have helped mitigate the as the wages of formal employees (Figure 25.A and impact of the shocks on labor markets. A thorough Figure 25.B). In contrast, we do not find a differential evaluation of the effectiveness of these policies and wage behavior across crises episodes in Brazil (Figure their role in dampening the labor market response in 25.C and Figure 25.D). During the two recessions, the region constitutes an important issue that warrants wages of formal employees decline relative to wages additional research. of informal workers. PART II Labor Markets in LAC During the 2009 Downturn 31 Office of Regional Chief Economist FIguRe 25 Evolution of Real Wages and CPI (males) by Status A. Argentina B. Argentina Previous Crisis Current Crisis Index Jun-98 = 100 Index Jun-08 = 100 160 Previous Crisis 160 Current Crisis Index Jun-98 = 100 Index Jun-08 = 100 160 CPI 160 140 140 CPI Real Wages - 140 120 140 120 Formal Employee Real Wages - 120 100 Real Wages - 120 100 CPI Formal Employee Formal Employee 100 80 Real Wages - 100 80 Real CPI Wages - Formal Employee Informal Employee Real Wages - Real Wages - 80 60 80 60 Informal Employee Informal Employee Real Wages - 60 40 Informal Employee 60 40 Jun-97 Sep-97 Dec-97 Mar-98 Jun-98 Sep-98 Dec-98 Mar-99 Jun-99 Sep-99 Dec-99 Mar-00 Jun-00 Sep-00 Dec-00 Mar-01 Jun-01 Sep-01 Dec-01 Mar-02 Jun-02 Sep-02 Dec-02 Mar-03 Jun-03 Sep-03 Dec-03 Mar-04 Mar-07 Jun-07 Sep-07 Dec-07 Mar-08 Jun-08 Sep-08 Dec-08 Mar-09 Jun-09 40 40 Jun-97 Sep-97 Dec-97 Mar-98 Jun-98 Sep-98 Dec-98 Mar-99 Jun-99 Sep-99 Dec-99 Mar-00 Jun-00 Sep-00 Dec-00 Mar-01 Jun-01 Sep-01 Dec-01 Mar-02 Jun-02 Sep-02 Dec-02 Mar-03 Jun-03 Sep-03 Dec-03 Mar-04 Mar-07 Jun-07 Sep-07 Dec-07 Mar-08 Jun-08 Sep-08 Dec-08 Mar-09 Jun-09 C. Brazil Previous Crisis D. Brazil Current Crisis Index Sep-97 = 100 Index Sep-08 = 100 115 Previous Crisis 115 Current Crisis Index Sep-97 = 100 Index Sep-08 = 100 110 115 110 115 Real Wages - Real Wages - 105 110 105 110 Informal Employee Informal Employee Real Wages - Real Wages - Real Wages - 100 105 100 105 Formal Employee Informal Employee Informal Employee Real Wages - 95 100 95 100 Formal Employee Real Wages - CPI CPI 90 95 Formal Employee 90 95 Real Wages - CPI CPI 85 90 Formal Employee 85 90 80 85 80 85 Mar-07 Jun-07 Sep-07 Dec-07 Mar-08 Jun-08 Sep-08 Dec-08 Mar-09 Jun-09 Sep-09 Dec-09 Mar-96 Jun-96 Sep-96 Dec-96 Mar-97 Jun-97 Sep-97 Dec-97 Mar-98 Jun-98 Sep-98 Dec-98 Mar-99 Jun-99 Sep-99 Dec-99 Mar-00 80 80 Mar-07 Jun-07 Sep-07 Dec-07 Mar-08 Jun-08 Sep-08 Dec-08 Mar-09 Jun-09 Sep-09 Dec-09 Mar-96 Jun-96 Sep-96 Dec-96 Mar-97 Jun-97 Sep-97 Dec-97 Mar-98 Jun-98 Sep-98 Dec-98 Mar-99 Jun-99 Sep-99 Dec-99 Mar-00 Source: LCRCE Staff calculations based on Households Surveys. For Argentina, Encuesta Permanente de Hogares (EPH), and Encuesta Permanente Continua de Hogares (EPHC); for Brazil, Pesquisa Mensal de Emprego (PME). 32 From Global Collapse to Recovery appeNdIx I.a The global financial crisis hit the world economy suddenly, severely, and in a synchronized fashion. No country was left unscathed. How- The Great Growth ever, the growth consequences of the crisis vary considerably across countries. So far several explanations have been put forth and, al- Collapse in 2009 though there is still no consensus on the relative importance of dif- ferent factors in the international transmission of this crisis, it has been argued the cross-country differences in growth response at the onset of the crisis may reflect differences in: (i) the exposure to real and financial manifestations of the global shock, (ii) the institutional and macroeconomic framework in place, and (iii) the response engi- neered by policymakers. In other words, provided that all countries were hit by a global shock (similar in nature across countries), the differences in growth performance may reflect the cross-country vari- ance in terms of resilience (external and macroeconomic vulnerabili- ties as well as other factors that might amplify shocks). This appendix investigates the relationship between the growth collapse at the onset of the crisis and pre-crisis characteristics of the domestic economy and its linkages to the rest of the world. The findings are presented in Table I.A and compared in light of recent empirical work. Our dependent variable is the change in the GDP growth rate in 2009 relative to 2007. Earlier attempts to characterize the crisis focus on measuring the severity of the crisis in its different financial and real manifestations (Rose and Spiegel, 2009a, b). These studies are un- able to link the growth decline since the beginning of the crisis to either pre-existing conditions in the domestic economies (except for sharp rises in equity prices) or international trade and financial linkages. We conduct a regression analysis for a sample of 103 coun- tries and find that output during the crisis was hit the hardest in advanced countries (see Table A). Controlling for income per capita, we estimate the link between the growth collapse in 2008­9 and its potential causes, as classified in the following groups: (a) trade and financial linkages, (b) macroeconomic and financial vulnerabilities, and (c) regulatory quality. Trade and financial linkages. The decline in growth performance was larger in countries with higher trade exposure, although no sig- nificant role for the composition of trade was found. On the other hand, recent evidence suggests that the growth collapse was deeper in countries with high shares of manufacturing exports (Berkmen et al. 2009; Lane and Milesi-Ferretti, 2010), but this effect ceases to be significant when the sample is restricted to emerging markets (Blanchard et al. 2010). From Global Collapse to Recovery 33 Office of Regional Chief Economist We fail to find a robust link between the overall mea- Countries with financial vulnerabilities also tended to sure of financial openness and the growth collapse. suffer a larger growth collapse. Table A shows that However, the structure of foreign assets and liabilities high pre-crisis credit growth and highly-leveraged may have played a significant role. Table I shows that economies (i.e., high loan-to-deposit ratios) faced safer modes of integration (i.e., lower debt-to-equity sharper declines in growth. Existing evidence shows ratios) were associated to milder growth declines. This that this effect was particularly more severe among finding may reflect the insulation properties against EU accession countries (Berkmen et al. 2009). external shocks of this mode of integration (Lane and Milesi-Ferretti, 2010). Finally, Blanchard et al. (2010) Quality of the regulatory framework. Giannone et al. show that countries with external finance vulnerabili- (2010) examines the relation between regulation ties (i.e. higher short-term debt to GDP ratios) have quality and growth performance during the crisis. Al- also suffered larger growth collapses. though countries with deeper financial markets may be better prepared to cope with crisis episodes, the Macroeconomic and financial vulnerabilities. Our evi- authors find that countries that adopted more market- dence suggests that the growth collapse (if any) was friendly policies (i.e., deregulated credit markets) were milder in countries with ability to implement counter- hit the hardest. However, the quality of regulation in cyclical policies and with lower pre-crisis inflation. credit markets ceases to explain the growth collapse However, we fail to find a robust role for the flexible if only emerging markets are analyzed (Blanchard et exchange rate arrangements. al. 2010). 34 From Global Collapse to Recovery The World Bank TABle I.A Characterizing the Growth Collapse Dependent Variable: Growth Collapse between 2007 and 2009 (1) (2) (3) (4) (5) (6) (7) (8) Openness Trade Openness 0.028*** 0.018* 0.022 0.019 0.019** 0.020** 0.022** (Export and Imports, % of GDP) (0.009) (0.011) (0.013) (0.011) (0.009) (0.008) (0.009) Financial Openness ­0.030 0.097 (Foreign Assets and Liabilities, % of GDP) (0.178) (0.092) Domestic Demand Private Consumption ­0.078* (% of GDP) (0.040) Composition of Trade Share of Natural Resource Exports ­0.026 (% of Total Exports) (0.029) Share of Manufacturing Exports 0.011 (% of Total Exports) (0.027) Structure of External Liabilities Equity-related Financial Openness ­1.554 (Assets and Liabilities, % of GDP) (0.937) Debt-related Financial Openness 0.861** (Assets and Liabilities, % of GDP) (0.389) International Reserves 4.851 (% of GDP) (3.404) Resilience of Domestic Financial Systems "Excess" Credit Growth 16.02** (Average 2006­7 vs. 2002­5) (6.179) Loan to Deposit Ratios 3.521** (Ratio, 2007) (1.548) Macroeconomic Policies Fiscal Pro-Cyclicality 4.797*** (Correlation of Govt. Spending and GDP) (1.505) Inflation Rates 0.396*** (Level of YoY CPI) (0.128) Observations 103 85 83 84 50 99 96 99 R-squared 0.106 0.081 0.047 0.101 0.155 0.188 0.173 0.208 Robust standard errors are reported in brackets. All regressions include controls for GDP per capita (PPP measure). ***, **, and * indicate statistical significance at one, five, or ten percent, respectively. APPeNDIX I.A The Great Growth Collapse in 2009 35 appeNdIx I.b This appendix characterizes the main features of the current recov- ery from the global downturn of 2008­9 and contrasts this episode Characterizing with previous recoveries from financial crises and global downturns. An exploratory econometric assessment of the forces driving the re- the Global covery is presented in Table I.B. Recovery: We find robust evidence of a bounce-back effect: countries that suf- fered a larger collapse in GDP growth rates between 2007 and 2009 How fast? What are expected to have a stronger recovery (see regression [1]). How- ever, our estimates suggest that the recovery is not steeper than the is driving it? recession. For example, Mexican real output growth declined by 9.9 percentage points (from 3.3 percent growth in 2007 to a contrac- tion in economic activity of 6.5 percent in 2009), and our estimates suggest a cumulative growth surge of only 5.5 percent for the years 2010­11. It should be highlighted that these figures are based on current forecasts and can be altered significantly if the perspectives for global growth (and the US in particular for the Mexican example) change. In short, our estimations suggest--regardless of the econo- metric specification used--that the rebound in economic activity will be sluggish. This result is consistent with recoveries from financial crises as well as those from global downturns (Terrones et al. 2009). Weak and slow recoveries from these turmoil periods tend to reflect declining asset prices, decreased levels of investment, and weak credit growth. The current rebound in economic activity is only partly engineered by a recovery in trade flows--see regression [2]. The unconditional correlation suggests only a modest economic impact: an increase in total trade of 10 percentage points of GDP would lead to an in- crease in GDP growth of only 0.2 percent. Moreover, the boost from greater trade flows appears to be negligible when controlling for the bounce-back effect, suggesting that the trade channel might also re- flect a bounce-back effect from the collapsed trade volumes in the last quarter of 2008. It should be noticed that the trade effect on the recovery is primarily driven by a rebound in the manufacturing sectors--see regression [3]. In comparison to past episodes, net ex- ports have typically operated as an engine that pulls the economy out of the recession (especially from those caused by financial crises). However, this pull factor loses steam during global downturns, which is the current case. The evidence suggests that export growth is slug- gish during global downturns due to a sharp decline in U.S. imports (Terrones et al. 2009). As noted in the main text, advanced economies and emerging mar- kets quickly implemented policies to counteract the recessionary From Global Collapse to Recovery 37 Office of Regional Chief Economist effects of the global financial crisis. Our estimations cal position. Moreover, Cerra et al. (2009) add that show that the recovery was indeed stronger in coun- monetary policy appears to have no significant effect tries with the ability to run countercyclical policies, on the strength of the recovery among developing while the floating exchange rate regimes did not seem countries and that floating rates have helped develop- to play a significant role. Terrones et al. (2009) find ing countries rebound from previous recessions. that stronger recoveries in industrial economies are also characterized by discretionary increases in gov- The main messages from our regression analysis are ernment spending and reductions in nominal and real summarized in the last four columns of Table B. Con- interest rates. However, the authors suggest that the trolling for the bounce-back effect and openness to stimulus engineered by the government becomes less trade, we analyze the effect of different components effective if it compromises the sustainability of the fis- of aggregate demand separately on the strength of the TABle I.B. Characterizing the Growth Recovery Dependent Variable: Expected Recovery in Growth Rates between 2009 and 2011 (1) (2) (3) (4) (5) (6) (7) (8) Bounce Back Effects Growth Collapse between 2007­2009 0.557*** 0.546*** 0.542*** 0.441*** 0.527*** 0.494*** 0.467*** (Percentages) (0.073) (0.083) (0.075) (0.079) (0.084) (0.083) (0.083) Openness Trade Openness 0.018** 0.000 0.000 0.003 0.002 0.001 0.000 (Export and Imports, % of GDP) (0.008) (0.006) (0.005) (0.005) (0.005) (0.005) (0.006) Financial Openness (Foreign Assets and Liabilities, % of GDP) Composition of Trade Share of Manufacturing Exports 0.022* (% of Total Exports) (0.012) Pull and Push Factors and Macroeconomic Policies Fiscal Pro-Cyclicality -1.461* (Correlation of Govt. Spending and GDP) (0.814) Increase in Government Consumption 87.93*** (2007 vs. 2009, % of GDP) (26.74) Increase in Private Household -2.596 Consumption (7.206) 2007 vs. 2009 (% of GDP) Increase in Investment (Capital -11.39 Formation) (10.62) 2007 vs. 2009 (% of GDP) Increase in Net Exports 13.02** (2007 vs. 2009, % of GDP) (5.917) Observations 81 75 68 74 70 71 71 71 R-squared 0.684 0.104 0.665 0.645 0.711 0.638 0.656 0.669 Robust standard errors are reported in brackets. All regressions include controls for GDP per capita (PPP measure). ***, **, and * indicate statistical significance at one, five, or ten percent, respectively. 38 From Global Collapse to Recovery The World Bank recovery in economic activity. The results suggest that downturn thanks in part to bold actions on the fiscal the recovery from the current global financial crisis front. The ratio of government consumption to GDP is driven by a rebound in domestic demand, in par- moved from a below-trend average before the crisis ticular by higher government consumption, and to a to above trend at the onset. Although not reported lesser extent net exports. here, there is no qualitative difference in the behavior of government consumption between HICs and MICs. Next, we conduct an event study analysis to examine They both have increased on average 1 percent of whether the different components of aggregate de- GDP in 2009 and are expected to remain at this new mand have played a different role this time around in level for the next couple of years. comparison to past recoveries from crises episodes (See Figures I.B.1­I.B.6).16 The decline in GDP growth Second, the decline in investment was not as sharp, rates has been sharper during the current global especially in MICs, as in previous episodes of severe downturn among high-income countries (HICs) and crisis (say, the collapse of investment in the East Asian the expectations of a swift recovery are disappoint- tigers after the 1997­8 crisis). In the current cycle, in- ing. On the other hand, GDP growth in middle-income vestment fell on average less than 2 percentage points countries (MICs) has declined at a similar pace to that of GDP for both HICs and MICs whereas it dropped of HICs (although growth rates were higher in the for- by more than 4 percentage points in MICS in past mer group) but has quickly rebounded. Even though episodes. Lastly, this recovery has also been strength- MICs seem to have a growth engine of their own, the ened to some extent by a rebound in net exports that sustainability of growth in the medium term may still have risen to above trend levels after an initial collapse depend largely on the growth prospects of advanced on the onset of the crisis. This pattern is somewhat countries. Analogously to MICs, growth in LAC was unusual and weaker if compared to previous crises above trend in the run-up to the crisis. Afterwards, it episodes, especially in MICs. The typical behavior of experienced a sharp decline; however, a fast rebound net exports is to simply increase as the crisis unfolded to trend growth is expected during the recovery phase from slightly below trend to significantly above trend (See Figure I.B.3). levels, pulling economies out of the crises. This time, however, we only observe a sharp decline from above When looking at the different components of aggre- trend levels and a mild recovery compared to previ- gate demand, we find some distinctive features that ous severe episodes. In other words, the prospects of characterize the current global downturn vis-à-vis a sluggish growth in trade (and, especially, in imports previous episodes of severe crises (i.e., episodes in from advanced countries) may hinder the likelihood which a country was hit by more than one crisis type: of a stronger and faster recovery based on an external banking, currency, external debt, or domestic debt push. crises). First, countries have emerged from the global 16 The data for GDP and its demand components for 2010 and 2011 are forecasts from the Economist Intelligence Unit. APPeNDIX I.B Characterizing the Global Recovery: How fast? What is driving it? 39 Office of Regional Chief Economist FIguRes I.B.1­I.B.6 A. I.B.1. Real GDP Growth: High-Income B. I.B.2. Real GDP Growth: Middle-Income Real GDP Growth Around Crises Real GDP Growth Around Crises High-Income Countries Middle-Income Countries 4% Real GDP Growth Around Crises 4% Real GDP Growth Around Crises 3% 3% High-Income Countries Middle-Income Countries 2% 4% Real GDP Growth Around Crises 2% 4% Real GDP Growth Around Crises Current Crisis 1% Severe Crises Episodes 1% 3% High-Income Countries 3% Middle-Income Countries 4% 0% 2% 4% 0% 2% Percentage Percentage Points Percentage Percentage Points Current Crisis 3% ­1% 1% Severe Crises Episodes 3% ­1% 1% 2% ­2% 0% 2% ­2% 0% Current Crisis Percentage Points Percentage Points 1% Severe Crises Episodes 1% ­3% ­1% ­3% ­1% 0% Current Crisis 0% ­4% ­2% ­4% ­2% Points Points Severe Crises Episodes ­1% ­5% ­3% ­1% ­5% ­3% ­2% ­6% Current Crisis ­2% ­6% ­4% ­4% ­3% ­3% Severe Crises Episodes ­7% ­5% ­7% ­5% Current Crisis ­4% ­8% ­6% ­4% ­8% ­6% Severe Crises Episodes ­5% ­9% ­7% ­5% ­9% ­7% ­6% ­10% ­8% ­6% ­10% ­8% ­7% ­9% T-3 T-2 T-1 T T+1 T+2 T+3 ­7% ­9% T-3 T-2 T-1 T T+1 T+2 T+3 ­8% ­10% ­8% ­10% ­9% T-3 T-2 T-1 T T+1 T+2 T+3 ­9% T-3 T-2 T-1 T T+1 T+2 T+3 C. ­10% Real GDP Growth Around Countries I.B.3. Real GDP Growth: LACCrises ­10% D. I.B.4. Government Consumption Government Consumption Around Crises T-3 T-2 LAC Countries T-1 T T+1 T+2 T+3 T-3 T-2 T-1 % ofTGDP T+1 T+2 T+3 4% 0.8% Real GDP Growth Around Crises Government Consumption Around Crises 3% LAC Countries % of GDP 2% 4% Around Crises Real GDP GrowthCurrent Crisis 0.6% 0.8% Government Consumption Around Crises 1% 3% LAC Countries Severe Crises EpisodesGDP % of 4% 0% 0.8% 0.4% 0.6% 2% Current Crisis Percentage Percentage Points Percentage Percentage Points 3% ­1% 1% Severe Crises Episodes 2% ­2% 0% Current Crisis 0.6% 0.2% 0.4% Percentage Points Percentage Points 1% ­3% Severe Crises Episodes ­1% 0% ­4% ­2% 0.4% 0.0% 0.2% Points Points ­1% Severe Crises Episodes ­5% ­3% ­2% ­6% ­4% 0.2% ­0.2% 0.0% ­3% ­7% Severe Crises Episodes ­5% ­4% ­8% ­6% 0.0% ­0.4% ­0.2% Current Crisis Severe Crises Episodes ­5% ­9% ­7% ­6% ­10% ­8% ­0.2% ­0.6% ­0.4% Current Crisis ­7% ­9% T-3 T-2 T-1 T T+1 T+2 T+3 T-3 T-2 T-1 T T+1 T+2 T+3 ­8% ­10% ­0.4% ­0.6% Current Crisis ­9% T-3 T-2 T-1 T T+1 T+2 T+3 T-3 T-2 T-1 T T+1 T+2 T+3 ­10% Net Exports Around Crises ­0.6% Investment Around Crises T-3 T-2 T-1 % ofTGDP T+1 T+2 T+3 T-3 T-2 T-1 % ofTGDP T+1 T+2 T+3 4% Exports e. I.B.5. NetAround Crises Net Exports 2% F. I.B.6. Investment Investment Around Crises % of GDP % of GDP 4% 3% Net Exports Around Crises 2% 1% Investment Around Crises % of GDP % of GDP 4% Current 2% 1% Percentage Percentage Points Percentage Percentage Points 3% 2% 0% 3% Current 1% Current Crisis Percentage Points Percentage Points 2% 1% 0% ­1% Current Points Points 2% 0% Current Crisis 1% 0% ­1% ­2% Current Crisis 1% ­1% 0% ­1% ­2% ­3% Severe Crises Episodes ­2% Severe Crises Episodes 0% ­1% ­2% ­3% ­4% Severe Crises Episodes T-3 T-2 T-1 T T+1 T+2 T+3 T-3 T-2 T-1 Severe Crises T Episodes T+1 T+2 T+3 ­1% ­3% ­2% Severe Crises Episodes ­4% T-3 T-2 T-1 T T+1 T+2 T+3 T-3 T-2 Episodes T+1 Severe Crises T T-1 T+2 T+3 ­2% ­4% T-3 T-2 T-1 T T+1 T+2 T+3 T-3 T-2 T-1 T T+1 T+2 T+3 Note: For the current crisis the T-period is 2008. The periods T+1, T+2 and T+3 are forecasts from EIU. Source: LCRCE Staff calculations using WDI and EIU data. 40 From Global Collapse to Recovery appeNdIx II.a Our goal is to compare the behavior of labor markets during the cur- rent economic recession relative to previous ones. Hence, we need Business Cycle to identify the cyclical turning points for each country. We follow the so-called BBQ algorithm for dating recessions, which was developed Dating by Bry and Boschan (1971) and later extended by Harding and Pa- gan (2002) to accommodate quarterly series. The algorithm defines cycles by identifying troughs and peaks--any quarter lower/higher than the preceding and succeeding two quarters--and using various censoring rules about the minimum duration of the overall cycle as well as the phases of expansion and contraction. We define economic cycles based on movements in quarterly real GDP in Argentina, Brazil, Chile, Colombia and Mexico. Sample sizes, reported in Table II.A.1, vary across countries. Overall, we find 28 re- cessions. The duration of a recession (expansion) counts the number of quarters between a peak and trough (between a trough and peak). The average duration of a recession in the five countries studied is slightly above a year, at 5 quarters. Table II.A.2 below provides some summary statistics on the average severity of each recession in each country in our sample. We report the duration of its recession and its amplitude. The amplitude of a recession (expansion) measures the percentage change in GDP between a peak and trough (between trough and peak). TABle II.A.1 Dating Recessions in Selected LAC Countries Recessions Expansions Nº of Nº of Nº of Avg. Nº of Nº of Avg. Country Sample Quarters Episodes Quarters Duration Episodes Quarters Duration Argentina 1990q1­ 79.00 5.00 37.00 7.40 4.00 42.00 10.50 2009q3 Brazil 1991q1­ 75.00 7.00 25.00 3.57 6.00 50.00 8.33 2009q3 Chile 1980q1­ 119.00 5.00 28.00 5.60 3.00 91.00 30.33 2009q3 Colombia 1994q1­ 62.00 3.00 11.00 3.67 2.00 51.00 25.50 2009q2 Mexico 1980q1­ 119.00 8.00 42.00 5.25 7.00 77.00 11.00 2009q3 Source: LCRCE Staff Calculations From Global Collapse to Recovery 41 Office of Regional Chief Economist TABle II.A.2 Dating Recessions in Latin America A. Argentina B. Brazil Quarter Quarter Period Duration Amplitude Amplitude Period Duration Amplitude Amplitude 1992q2­1993q1 4 ­6.7% ­1.7% 1991q3­1992q1 3 ­11.2% ­3.7% 1994q2­1996q1 8 ­8.1% ­1.0% 1993q3­1994q1 3 ­7.5% ­2.5% 1998q2­2002q1 16 ­28.0% ­1.8% 1994q4­1996q1 6 ­8.0% ­1.3% 2008q2­2009q1 4 ­9.9% ­2.5% 1997q3­1998q1 3 ­8.7% ­2.9% 1999q3­2000q1 3 ­2.8% ­0.9% 2001q2­2000q2 4 ­3.7% ­0.9% 2008q3­2009q1 3 ­9.2% ­3.1% C. Chile D. Colombia Quarter Quarter Period Duration Amplitude Amplitude Period Duration Amplitude Amplitude 1981q2­1982q4 7 ­20.5% ­2.9% 1998q2­1999q2 5 ­6.8% ­1.4% 1985q2­1986q3 6 ­2.8% ­0.5% 2002q2­2002q4 3 ­0.7% ­0.2% 1988q1­1988q3 3 ­3.6% ­1.2% 2008q2­2008q4 3 ­1.2% ­0.4% 1998q2­1999q3 6 ­5.5% ­0.9% 2008q2­2009q2 5 ­4.7% ­0.9% e. Mexico Period Duration Amplitude Quarter Amplitude 1981q4­1984q3 12 ­5.9% ­0.5% 1985q4­1986q3 4 ­8.6% ­2.1% 1987q4­1988q3 4 ­6.4% ­1.6% 1992q4­1993q3 4 ­5.2% ­1.3% 1994q4­1995q3 4 ­15.0% ­3.8% 1997q4­1998q3 4 ­3.0% ­0.8% 2000q4­2001q3 4 ­5.6% ­1.4% 2007q4­2009q1 6 ­11.7% ­1.9% Source: LCRCE Staff Calculations 42 From Global Collapse to Recovery appeNdIx II.b This Appendix provides some examples of policies that were either Labor Market introduced in Latin American labor markets or reinforced as a con- Policy Responses sequence of the international crisis. Its aim is not exhaustive, and the country coverage responds to the countries included in this note. to the Crisis17 argentina Argentina responded to the global crisis both reinforcing existing labor market policies and creating new ones. Labor market policies aimed at mitigating the effects of the international crisis can be grouped on three fronts: i) active labor market policies, ii) social protection and policies for unemployment protection, and iii) employment subsidies. Argentina introduced "Jóvenes con más y mejor trabajo," a new ac- tive labor market program in June 2008, providing training and an opportunity to finish regular school to unemployed young workers (between 18 and 24) with unfinished primary or secondary school. It also reinforced an existing program, the "Seguro de capacitación y empleo," which combines elements of non-contributory benefits for the unemployed with active labor market policies in the form of assistance in finding a job and training. While individuals were in principle eligible for the program for a maximum of 24 months, the government decided to extend eligibility beyond this limit up to De- cember 31st 2009 in case the worker remained unemployed. Simi- larly, contributory unemployment benefits were extended. Regarding the protection of jobs, the "Programa de recuperación productiva" (REPRO) provides income support to firms in distress in exchange of a compromise not to cut employment. In principle, any firm engaged in a declining activity in the private sector is eligible to receive a monthly fixed-sum of AR$150 (US$ 38) for each of its employees up to a maximum period of six months. In exceptional circumstances, the benefit can be extended as well as the receiving period, up to no more than 12 months. Participating firms have to show the severity of the economic crisis hitting them, and provide plans to show their future viability. The program was extended as a consequence of the crisis up to December 31st 2009. brazil Brazil articulated the response of its labor market policies to the in- 17 Some parts of this appendix draw heavily ternational crisis on three fronts: strengthening contributory unem- on Freije-Rodríguez and Murrugarra (2009). ployment insurance, extending non-contributory income support From Global Collapse to Recovery 43 Office of Regional Chief Economist programs and raising minimum wages. Unemploy- is included in a package of reforms of the "Seguro de ment insurance in Brazil is quite limited. As a con- cesantía" that was launched before the start of the sequence of the international crisis, the government international crisis, but came into effect around May extended the eligibility of the "Seguro desemprego" 2009 and included special provisions to cope with the for two additional months for the "most affected sec- crisis. Among the latter, an interesting supplementary tors and states." The maximum period of eligibility measure is the program of "Protección al empleo y became 7 months. Non-contributory benefits through capacitación laboral," which allows workers on a vol- the social assistance program "Bolsa familia" were untary basis and as an alternative to a layoff to collect also strengthened. The ceiling of eligibility was raised during up to 5 months benefits from the "Seguro de from R$120 (US$ 68) per capita to R$137 (US$ 77) cesantía" while on training. The benefits are defined per capita, which implies that some 1.3 million new as 50 percent of the average wage of the worker dur- families became eligible by the end of 2009. The ing the previous 6 months. During the period the ben- third income support measure was to increase the efits are collected, the worker keeps his or her labor minimum wage. In February 2009, the minimum contract and all its related benefits (e.g. health insur- wage was raised from R$415 (US$ 235) to R$465 ance, pension contributions and different insurance (US$ 262), which has effects not only on low wage policies). The measure was designed to be in effect workers, since the minimum wage is used to define between July 1st 2009 and July 1st 2010. the benefits obtained through several social security and social assistance transfers. Colombia Chile Colombia articulated its labor market policy response to the crisis on two fronts: i) a series of tax cuts for We identify two set of reforms that tried to mitigate small and medium size enterprises, in an effort to re- the impact of the international financial crisis on the vitalize this sector and ii) enhancing a major training Chilean labor market. The first falls within the broad program, through the "Servicio nacional de aprendiza- umbrella of employment subsidies, while the second je" (SENA) to provide additional training to 250.000 is related to the innovative program of unemployment young workers. Eligible individuals are between 14 benefits "Seguro de cesantía." Within the employ- and 26 years of age, unemployed, who live in condi- ment subsidies, a new youth employment subsidy tions of severe or extreme poverty in rural and urban law was launched on January 2009. This program areas. SENA will provide to these young workers some provides a 30 percent subsidy of the annual income 187 training programs. Up to 70 percent of the new for those individuals aged between 18 and 24, with training positions were posted during 2009. The pro- finished secondary education and working in the for- gram also includes some support to participants in mal sector, whose monthly (annual) incomes is below the form in kind transfers (e.g. covering transporta- Ch$360,000 (approximately US$ 690). Within the tion). Firms interested in hiring trainees from the pro- same set of policies, and during the same month, a gram may not have cut employment in the previous second law was passed that provides temporary in- three months. come tax reductions for individuals and tax credits for firms that carry out training activities with their workers. The law provides special tax reductions to mexico firms employing individuals who are beneficiaries of certain social programs such as "Subsidio familiar," Mexico has in place several active and passive labor "Asignación familiar," "Chile solidario" and "Asig- market policies. During the recent crisis, some of nación maternal." The second set of policy reforms these policies were strengthened or expanded. On the 44 From Global Collapse to Recovery The World Bank side of active labor market policies, the "Programa de up to September in a maximum amount that equals empleo temporal" (PET) which originally was opera- a third of the fall in sales during the corresponding tive only in marginal rural areas was extended to cover period. urban areas. The new program is called "Programa de empleo temporal ampliado" (PETA). In order to make this expansion operative, some 500 million pe- peru sos were added to the program, with a budget total in 2009 of 2,200 million pesos (approximately US$ 179 In 2009, the Peruvian Congress introduced temporary million). Also related to active policies, 1,250 addition- payroll tax holiday measures. Bonuses payable twice al million pesos (about US$ 100 million) were added yearly and equivalent to two monthly wages have been to the budget for intermediation services controlled exonerated from social and pension contributions un- by the Labor secretary. Passive labor market policies til 2010. As a result, workers' income will increase by were also enhanced. Individuals' ability to withdraw 3.7 percent relative to pre-tax incomes, while firms funds from retirement accounts while unemployed will not have to pay social security contributions on was extended, and health insurance coverage of up to these bonuses. The estimated fiscal cost of this mea- six months after dismissal for the unemployed worker sure of US$212 million is initially covered by the so- and his or her family was granted. A new program, cial security administration (EsSalud) and the public the "Programa para la preservación del empleo" was pension system (ONP), although these institutions are launched in January 2009 with the objective of main- expected to seek compensation from the Central Gov- taining employment in firms operating in tradable ernment. The Peruvian Government also launched sectors in manufacturing that have been especially a publically financed retraining program, "Revalora hit by the slowdown in global demand. The program, Peru", for workers who have lost their job since 2008. operating from February to September of the same Training by the relevant technical education institu- year, was funded with 2,000 million Mexican pesos tions is offered in the areas of construction, manufac- (US$ 162 million). Eligible firms who can prove they turing, and tourism. Unemployed workers can register are under temporary distress might receive a subsidy themselves in decentralized Labor Ministry offices to of up to 5,100 pesos per worker (US$ 414) during this enroll in the program. period. In exchange, firms commit to limit the layoffs APPeNDIX II.B Labor Market Policy Responses to the Crisis 45 References Bentolila, S., and G. Bertola, 1990. "Firing costs and labor demand: How bad is Eurosclerosis?" Review of Economic Studies 57, 381­402 Berkmen, P., G. Gelos, R. Rennhack, and J.P. Walsh, 2009. "The global financial crisis: Explaining cross-country differences in the out- put impact." IMF Working Paper WP/09/280, December Blanchard, O.J., H. Faruqee, and M. Das, 2010. "The initial impact of the crisis on emerging market countries." International Monetary Fund, manuscript, March Bosch, M. and W.F. Maloney, 2008. "Cyclical movements in unem- ployment and informality in developing countries." The World Bank Policy Research Working Paper 4648, June Bry, G., and C. Boschan, 1971. "Cyclical analysis of economic time series: Selected procedures and computer programs." NBER Techni- cal Working Paper No. 20 Cerra, V., U. Panizza, and S.C. Saxena, 2009. "International evidence on recovery from recessions." IMF Working Paper WP/09/183, Au- gust Cho, Y., and D. Newhouse, 2010. "Workers' vulnerability to the finan- cial crisis." The World Bank, manuscript Cunningham, W., 2007. "Minimum wages and social policy: Lessons from developing countries". The World Bank. Washington, DC. Ferreira, F.H.G., and N. Schady, 2009. "Social consequences of the global financial crisis in Latin America: Some preliminary, and sur- prisingly optimistic, conjectures." Washington, DC: The World Bank, LCR Crisis Brief Series, November Freije-Rodríguez, S., and E. Murrugarra, 2009. "Labor markets and the crisis in Latin America and the Caribbean: A preliminary review for selected countries." Washington, DC: The World Bank, LCR Crisis Brief Series, June Giannone, D., M. Lenza, and L. Reichlin, 2010. "Market freedom and the global recession." European Central Bank, manuscript, Januar Harding, D., and A. Pagan, 2002. "Dissecting the cycle: A method- ological investigation." Journal of Monetary Economics 49(2), 365­ 381 From Global Collapse to Recovery 47 Office of Regional Chief Economist International Labour Organization, 2010. "Global Em- Perry, G., W. Maloney, O. Arias, P. Fajnzylber, A. ployment Trends." Geneva: ILO, Various issues. Mason and J. Saavedra-Chanduvi, 2007. "Informal- ity: Exit and exclusion." Washington, DC: The World Izquierdo, A., and E. Talvi, 2010. "The aftermath of Bank, LAC Regional Flagship Report the crisis: Policy lessons and challenges ahead for Latin America and the Caribbean." Washington, DC: Rose, A.K., and M.M. Spiegel, 2009a. "Cross-country Inter-American Development Bank, March causes and consequences of the 2008 crisis: Early warning." NBER Working Paper 15358, September Kose, M. A., C. Otrok, and E. Prasad, 2008. "Global business cycles: Convergence or decoupling?" NBER Rose, A.K., and M.M. Spiegel, 2009b. "Cross-country Working Paper 14292, October causes and consequences of the 2008 crisis: Interna- tional linkages and American exposure." NBER Work- Lane, P.R., and G.M. Milesi-Ferretti, 2010. "The ing Paper 15358, September cross-country incidence of the global crisis." Wash- ington, DC: International Monetary Fund, manuscript, Schwartz, J.Z., L.A. Andrés, and G. Dragoiu, 2009. January "Crisis in Latin America: Infrastructure investment, employment and the expectations of stimulus." The Maddison, A., 2007. The world economy: A millennial World Bank Policy Research Working Paper 5009, perspective. Paris, OECD July. Maddison, A., 2009. Historical statistics of the world Terrones, M., A. Scott, and P. Kannan, 2009. "From economy: 1­2008 AD. University of Groningen, Gron- recession to recovery: How soon and how strong?" ingen Growth and Development Centre, available on- IMF World Economic Outlook, April, Chapter 3, pp. line at http://www.ggdc.net/maddison/ 103­138 Ostry, J.D., A.R. Ghosh, K. Habermeier, M. Chamon, The World Bank, 2009. "Update on the global crisis: M.S. Qureshi, and D.B.S. Reinhardt, 2010. "Capital The worst is over, LAC poised to recover." Washing- inflows: The role of controls." IMF Staff Position Note ton, DC: The World Bank, LCRCE Semi-Annual Report SPN/10/04, February for the 2009 Annual Meetings, October 48 From Global Collapse to Recovery THE WORLD BANK Office of the Chief Economist Latin America and the Caribbean 1818 H St. NW Washington, DC 20433 www.worldbank.org/lac