Contents Acknowledgements ..................................................................................................................................................... iii Executive Summary .................................................................................................................................................... iv 1. Progress in Reducing Poverty .................................................................................................................................... 1 2. The Inclusiveness of Economic Progress ......................................................................................................... 10 3. Evolution of Living Conditions and Economic Mobility ................................................................................. 17 4. Inequality of Opportunities ................................................................................................................................... 22 5. Productivity, Market Development and Vulnerability in Agriculture ........................................................... 25 6. Human Capital, Labor Force and Jobs .............................................................................................................. 29 References ..................................................................................................................................................................... 33 FIGURES Figure 1.1. Robust growth has led to a sustained rise in GPD per capita in Mozambique ....................... 1 Figure 1.2. The services and extractive sectors are increasingly supporting GDP growth ....................... 2 Figure 1.3. For most workers their primary job is still in agriculture .............................................................. 3 Figure 1.4. Poverty has been falling since the early 2000s, but the pace accelerated after 2008 ......... 4 Figure 1.5. Welfare levels have not converged between urban and rural areas ........................................ 4 Figure 1.6. Since the early 2000s, nearly 8 out of 10 poor people are in rural areas ............................... 4 Figure 1.7. Niassa, Nampula and Zambezia are the provinces with the highest poverty rates ................ 7 Figure 1.8. The incidence of monetary poverty fell in Mozambique and other countries in the region .. 7 Figure 1.9. In recent years growth has been more poverty reducing in Mozambique ........................... 8 Figure 1.10. Poverty will fall markedly moving forward if growth is strong and more equally shared .. 9 Figure 2.1. Growth after the late 2000s benefitted mostly the non-poor, chiefly in urban areas ........ 10 Figure 2.2. Positive shared prosperity reversed after 2008 signaling weaker inclusiveness .................. 11 Figure 2.3 The distribution of household consumption is highly unequal by regional standards ...... 12 Figure 2.4. Inequality is high and increasing, a trend driven by worse inequality in urban areas ......... 12 Figure 2.5. Higher inequality has offset the contribution of growth to poverty reduction .................... 13 Figure 3.1. Higher school participation is slowly increasing educational attainment .............................. 17 Figure 3.2. Infant and maternal mortality rates have fallen ............................................................................. 18 Figure 3.3. Access to basic services is imprving but are yet far from universal ......................................... 18 Figure 3.4. Location is a strong determinant of access to basic public services ...................................... 18 Figure 3.5. Ownership of traditional and modern assets has increased ..................................................... 19 Figure 3.6. The prevalence of multiple deprivations has declined but mostly in urban areas .............. 20 Figure 3.7. Non-monetary deprivations continue to be larger among the monetary poor .................. 20 Figure 3.8. The chronic poor remains the largest welfare group in the population ................................ 21 Figure 4.1. Human opportunities are more unequally allocated in the poorest provinces ................... 23 Figure 4.2. Location, consumption and parental education drive the inequality of opportunity ..... 24 i Figure 5.1. Average maize yields are lower in Mozambique than in other neighboring countries .............. 25 Figure 5.2. There is low adoption of modern agricultural inputs among farmers in Mozambique ............. 26 Figure 5.3. Poverty rates are higher in provinces with lower maize yields per hectare .................................. 26 Figure 5.4. Modern inputs and market orientation are correlated with higher agricultural productivity .... 27 Figure 5.5. The more isolated a province is from the nearest market the higher is its poverty rate ............ 28 Figure 5.6 Maize yields per hectare are lower for farmers that experienced droughts and/or floods ......... 28 Figure 6.1. Educational attainment in Mozambique is increasing across the board ........................................ 29 Figure 6.2: The risk of dropping out of school is higher for children from poor households ....................... 30 Figure 6.3. Skilled workers in urban areas experience the highest returns to schooling ................................ 31 Figure 6.4. Per capita expenditures are higher in households hwith jobs outside agriculture ....................... 32 TABLES Table 1.1. Poverty headcount ratio for national poverty line and the US$1.9 PPP poverty line ..................... 5 Table 1.2. The total number of poor has increased, mostly in rural areas .......................................................... 5 Table 2.1. The services sector is gradually playing a greater role in the economy ........................................... 13 Table 2.2. Labor productivity growth is the single greatest contributor to growth in GDP per capita ......... 14 Table 2.3. The livelihoods of the poor differ from those of the non-poor in many key aspects .................. 16 Table 4.1. The distribution of opportunities is highly unequal but is slowly improving .................................... 22 BOX Box 1. Measuring poverty in Mozambique ............................................................................................................. 6 ii mozambique poverty assessment | acknolwedgements Acknowledgements The World Bank greatly appreciates the close collaboration with the Government of Mozambique (the Ministry of Economy and Finance and the National Institute of Statistics) in the preparation of this report. The core team preparing this report consisted of Javier E. Baez (Senior Economist, GPV01, World Bank), German Caruso (Economist, GPV04), Chiyu Niu (Consultant, World Bank) and Cara Myers (Consultant, Harvard University). The following people contributed to this report through the preparation and technical review of background papers and analytical work that form the basis for several chapters of this report: Juan Carlos Parra (Senior Economist, GPV04), Lidia Ceriani (World Bank/Georgetown), Nobuo Joshida, (Lead Economist, GPV01), Chuqiao Bi (Consultant, World Bank), Carlos Da Maia (Economist, GPV01), Anna Carlotta Allen Massinge (Research Analyst, GMTA4), Shireen Mahdi (Senior Economist, GMTA4), Peter Anthony Holland (Program Leader, AFCS2), Sara Troiano (Young Professional, GTD11), Ghada Elabed (Agricultural Economist, GFA07), Jan Joost Nijhoff (Senior Agriculture Economist, GFA02), Donald Larsson (Consultant), Ian Walker (Lead Economist, GPSJB), Ulrich Lachler (Consultant) and Paul Christian (Economist, DECIE). The team acknowledges the collaboration and insightful comments from the Directorate of Economic and Financial Studies (DEEF) at the Ministry of Economy and Finance from the Government of Mozambique. The core team received guidance and comments drafts of the report and presentations from Pierella Paci (Practice Manager, GPV01), Mark Lundell (Country Director, AFCS2), Carolin Geginat (Program Leader, AFCS2) and Raymond Bourdeaux (Program Leader, ACFS2). The team also acknowledges comments from other World Bank colleagues from the Mozambique Country Office during two presentations with preliminary findings of the report. Sections 6.3 and 6.4 in Chapter 6 of this report (“Human Capital, Labor Force and Jobs”) draw from the report “Let’s Work Mozambique Country Pilot – Job Diagnostics Report” (World Bank, 2017). The elaboration of this poverty was largely possible to the generous financial support from the Belgian Poverty Reduction Partnership III (BPRP III) Trust Fund. iii executive summary | mozambique poverty assessment Executive Summary Mozambique has experienced strong absolute terms, however, the number of poor and sustained economic growth in the increased –owing largely to rapid growth in last two decades. population, from 11 million in 2002/03 to 12.3 million in 2014/15. Growth of its Gross Domestic Product (GDP) expanded at an annual average rate of 7.2 Over the long term, poverty has fallen percent between 2000 and 2016, making it more slowly than expected considering one of the fastest-growing countries in Sub- the strong growth performance, yet Saharan Africa (SSA). The economic expansion growth has become more poverty has boosted incomes and living standards. reducing in recent years. GDP per capita, for instance, grew annually on average by 4 percent over the same period, The fall in poverty in Mozambique is consistent climbing from $561 to $1,128 (2011 PPP). with the trend seen in many other countries in Growth has been supported by a rebounding the region. Yet, looking at the last two decades, agricultural sector, particularly in the first phase economic growth and poverty reduction of the post-war period, increased productivity are not as strongly linked in Mozambique as in trade, transport and communications and in other countries. Estimates of the growth financial services, sound macroeconomic elasticity of poverty reduction for a group of management, large-scale foreign investments selected countries in Eastern Africa with two projects and significant donor support. More poverty measurements in the last decade recently, however, growth has slowed down due show that the responsiveness of monetary mainly to macroeconomic factors and severe poverty to raising levels of income per capita natural disasters. in Mozambique is moderate. For instance, a one percentage increase in GDP per capita in High and stable growth has led to Uganda is associated with a fall in poverty of poverty reduction, especially after the 0.95 percent. An equivalent change in GDP late 2000s. per capita reduced poverty by 0.3 percent in Mozambique, less than a third than in Uganda. Poverty has been on a declining trend following However, the latest numbers indicate that the sustained strong growth in the 2000s. Poverty country may be gradually becoming better at numbers based on the official methodology leveraging strong growth for poverty reduction. show that the poverty headcount fell from 52.8 The elasticity rose from 0.08 (2002/03-2008/09) percent in 2002/03 to 46 percent in 2014/15. to 0.68 (2008/09-2014/15). This study, which examines the evolution of poverty using a different poverty measurement While household consumption growth methodology, also finds a downward trend. As accelerated after 2008, it became less of 2014/15, the share of Mozambicans living inclusive. beneath the poverty line is 48.4 percent, below the levels of poverty recorded in 2002/03 and Who benefitted the most from economic 2008/09, 60.3 and 58.7 percent, respectively. progress in Mozambique? The answer depends This is equivalent to an average reduction on what period is analyzed. Most of the 2000s in poverty of 1 percentage point per year. In (the period 2002/03-2008/09) recorded a iv mozambique poverty assessment | executive summary small reduction in poverty because of meagre fallen in both rural and urban areas, from 69 consumption growth (0.11 percent). However, percent to 56.0 percent in the former, and from this slow growth was “pro-poor”, namely it 41.1 percent to 32 percent in the latter. However, benefitted disproportionately low-income rural areas continue to lag behind urban areas: households more, amongst all those located since the early 2000s, nearly 8 out of 10 poor in rural areas. Nevertheless, while growth people have been in rural areas. There are also accelerated at the end of the 2000s, its gaps across provinces. Despite the generalized distributional pattern reversed, turning into “pro- decline in poverty, welfare levels remain low rich”. Annual growth in consumption per capita in the Northern and the Center Regions of the picked up, averaging 4.3 percent (2008/09 and country relative to the South. Poverty continues 2014/15). Stronger growth for everyone resulted to be high in Zambezia, Nampula and Niassa, in faster poverty reduction, yet it benefited historically the provinces with the highest poverty chiefly the upper parts of the distribution. Annual levels. In contrast, Maputo Province and Maputo consumption growth for the top quintile was City show the largest decline even though they 7.5 percent, three times faster than the rate of had the lowest poverty levels back in 2002/03. the bottom 40. The “pro-richness” of growth is limiting Mozambique’s progress in achieving The increasing role of services in the shared prosperity and reducing inequality. The economy and favorable macroeconomic Gini coefficient increased from 0.47 to 0.56 conditions contributed to faster between 2008/09 and 2014/15 –largely an urban consumption growth after the late 2000s. phenomenon, placing Mozambique among the most unequal countries in SSA. Mozambique is undergoing a process of structural change whereby the sources of Had growth been more equally shared growth have gradually shifted away from Mozambique would have achieved twice agriculture. The GDP share of agriculture fell as much poverty reduction after 2000. from 38.1 to 25.5 percent between 1996 and 2014. While the emergence of manufacturing The weaker inclusiveness means that many low- is characterized by capital intensive activities income Mozambicans are missing out on the (largely “megaprojects” in extractive, export- benefits of progress. Changes in poverty can be oriented industries) with higher value added decomposed into “growth” and “redistribution” but low job creation, the increasing role of effects. The analysis shows that consumption services in the economy has offered a path to growth (“growth effect”) has been the main force jobs outside agriculture. Between 2008 and behind the fall in poverty. In contrast, the increase 2014, the jobs share of services increased fast, in inequality in the distribution of consumption moving from 15 to 24 percent. The GDP share of (“redistribution effect”) has offset part of the services also increased by almost 6 percentage gains. More specifically, the “growth effect” alone points, reaching 55.7 percent. After 2008, would have reduced poverty by 23.1 percentage labor productivity growth – the main engine points between 2002 and 2014 – bringing the of economic growth in the last two decades – poverty headcount down to 37.2 percent rather has been largely driven by the redeployment of than 48.4 percent – had that growth been more labor away from agriculture and into sectors with inclusive. Instead, inequality in the distribution of higher productivity growth, chiefly in services, consumption growth increased poverty by 11.2 where productivity is over six times larger percentage points. despite high levels of informality. Moreover, the macroeconomic framework (fiscal expansion, Faster poverty reduction in some of strong credit growth, large influx of foreign direct the areas of the country where poverty investments) provided the conditions for faster was lowest a decade and half ago has private consumption growth. limited the convergence in welfare levels between regions. Economic progress also translated into improvements in non-monetary The evolution of poverty displays regional dimensions of well-being … differences. The share of poor households has v executive summary | mozambique poverty assessment The average household has better standards is likely to continue trapped into chronic of living today than at the turn of the century. poverty unless they break the cycle of physical Progress in closing consumption deficits, albeit deprivation and accumulate human, physical at a moderate pace, has been accompanied and financial capital to enter a path of stable by improvements in other dimensions of well- income growth. Another 25 percent of the being. School enrollment and attendance show population is not monetarily poor but faces a continued improvement since the early 2000s. high risk of sliding back into poverty because of Individuals ages 20 to 65 have on average 5.1 the high economic insecurity brought about by years of schooling, compared to 2.4 in 2002/03. its multiple non-monetary deprivations. Mozambicans are living longer. Life expectancy increased by nearly 9 years since 2001, from Is Mozambique on a path to end extreme 48.8 to 57.6. Infant mortality, expressed as the poverty by 2030? It is unlikely, but number deaths per thousand live births, fell from poverty will fall significantly if growth is 99.1 in 2003 to 68.1 in 2011. Other key health high, stable and more broadly shared. indicators as maternal mortality and morbidity are also moving in the right direction. These Projections under an optimistic (high growth), changes are coupled with improvements in the pro-poor (inclusive) scenario show that poverty quality of housing and increased ownership of is unlikely to be eradicated by 2030 but it can traditional and modern assets. be reduced to 21.8 percent, a remarkable achievement. However, if growth remains strong But large inequalities of opportunities but pro-rich, as in recent years, the projections remain across the population, indicate that poverty will fall at most to around limiting the degree in which the poor 32 percent by 2030. If consumption growth is participate in the growth process and equally distributed across the population but share in its proceeds. below past performance, reflecting the slower economic growth experienced in recent years, While multidimensional poverty has fallen, around 36 percent of the Mozambicans will still it remains high. Improvement in several be poor by 2030. These simulations underscore dimensions occurred from low levels, which that achieving robust, inclusive growth is the means that the remaining gaps are still large. right mix to maximize poverty reduction Indicators such as access to electricity, food moving forward. security and stunting, among others, showed little or no improvement during the period with Strengthening the linkages between the strongest economic growth on record. growth and poverty reduction requires a Progress has not been even neither across mix of policies aimed at achieving three income groups nor across areas. The Human overarching objectives. Opportunity Index, a measure that summarizes the level of basic opportunities in a society and The first objective is to continue expanding the how equitable they are distributed, reveals that availability of basic services while addressing the chances of Mozambican children later in the remaining large inequalities in terms of life are largely influenced by their location and access and quality to improve and equalize family background, chiefly household income opportunities for all citizens. Enabling the and school attainment of the household head. poor with the skills and assess to participate in the growth process and share in its proceeds Nearly one in two Mozambicans are will bolster growth and economic progress. trapped in chronic poverty and close to The second objective is to foster economic 25 percent of the population is highly diversification, job growth in productive, labor- vulnerable to fall into poverty. intensive production, and agglomeration of firms and markets. The private sector is typically the Almost half of the population (46.3 percent) main engine for broad-based growth through continues to be poor in monetary and non- job creation. The government can play a critical monetary sense, most of whom (84.9 percent) role by implementing policies and regulations are in rural areas. This segment of the population aimed at promoting an environment conducive vi mozambique poverty assessment | executive summary to achieving high private investment rates risk. Cutting across these three overarching and strong firm growth. The third objective is objectives is the need to protect the significant to raise productivity in agriculture – a sector gains in poverty reduction achieved so far to that still supports the livelihoods of most rural avoid letting the one in four Mozambicans that households and the poor – by addressing the have high economic insecurity slide back into limited linkages of farmers with input and output poverty and deprivation. markets, and the weak resilience to weather vii 1. progress in reducing poverty | mozambique poverty assessment 1 Progress in Reducing Poverty Mozambique has experienced strong stability provided the foundation for robust and sustained economic growth in the growth. The economic expansion has been last two decades. supported by a rebounding agricultural sector, increased productivity in trade, transport 1. Mozambique experienced strong and and communications and financial services, sustained economic growth in the last two structural reforms and sound macroeconomic decades. Growth of its Gross Domestic Product management, large-scale foreign investments (GDP) picked up following the end of the war projects and significant donor support. The in 1992, expanding at an annual average rate of economic expansion boosted incomes and 7.2 percent between 2000 and 2016, making living standards. GDP per capita grew annually it one of the fastest-growing countries in Sub- on average by 4 percent, climbing from $561 to Saharan Africa. Political and macroeconomic $1,128 (2011 PPP) (Figure 1.1). Figure 1.1. Robust growth has led to a sustained rise in GPD per capita in Mozambique $1,300 14% $1,200 12% $1,100 10% $1,000 $900 8% $800 6% $700 4% $600 $500 2% $400 0% 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 GDP per Capita, PPP 2011 (LHS) Real GDP growth (RHS) Source: National Institute of Statistics of Mozambique (INE) and World Bank using WDI 1 mozambique poverty assessment | 1. progress in reducing poverty The services sector and investments GDP growth in the late 1990s and early 2000s in mega-projects have driven growth (at an average 3.3 percentage points). since the early 2000s, partly shifting the sectoral composition of employment. 3. Changes in the sectoral composition of employment reflect the ongoing structural 2. An expansion of the services sector and transition of the economy. Notwithstanding investments in megaprojects contributed the falling share of agriculture in total to rapid accumulation of physical capital employment, most people continue to work and an increase in total factor productivity. in this sector. Almost 3 in 4 workers are Investments in reconstruction and the mostly engaged in agriculture. Led by the incorporation of new workers into the labor modest structural transition of the economy, a force, particularly in agriculture, led growth in growing proportion of workers is employed in the early postwar period. These trends began to the service sector – the share increased from change at the end of the 1990s. The contribution 9 percent in 1997 to 24 percent in 2015. In of agriculture to GDP growth fell from an contrast, the contribution of the industry sector average of 6 percentage points in the middle of to employment is rather limited, oscillating the 1990s to 1.1 percentage points in the early between 3.4 percent and 4.9 percent in the 2000s (Figure 1.2). At the same time, the services last two decades despite its larger impact on sector increased its role in the economy, from growth (Figure 1.3). This is largely explained contributing an average 0.9 to 2.8 percentage by the concentration of investments in points to overall growth between the middle large-scale capital-intensive projects that are of the 1990s and the middle of the 2010s. The characterized by weak backward and forward manufacturing sector was a large contributor to linkages with other parts of the economy. Figure 1.2. The services and extractive sectors are increasingly supporting GDP growth 10% 9.0% 8.3% 7.8% 8% 6.9% 6.4% 6% 4% 2% 0% 1992-6 1997-2001 2002-2006 2007-2011 2012-2016 Agriculture Extractives Manufacturing Services GDP growth Source: National Institute of Statistics of Mozambique (INE) 4. Recent economic developments have shifted undisclosed commercial loans. Together, these Mozambique to a slower growth trajectory. events contributed to a sharp pace of currency The economy has been growing at a reduced depreciation and soaring inflation. Confidence pace since 2015, largely driven by an ongoing in the economy also faltered as the debt crisis economic downturn, bouts of low commodity continues to be transmitted to the real sectors prices, the occurrence of natural disasters and of the economy, derailing Mozambique’s track the revelation of USD 1.4 billion in previously record for high growth and economic stability. 2 1. progress in reducing poverty | mozambique poverty assessment Figure 1.3. For most workers their primary job is still in agriculture (Employment by economic sectors, selected years) 100% 9.0% 4.4% 16.1% 15.0% 24.0% 80% 3.4% 4.7% 4.9% 60% 40% 86.6% 80.5% 80.4% 71.0% 20% 0% 1997 2003 2009 2015 Agriculture Industry Services Source: National Institute of Statistics of Mozambique (INE) High and stable growth has led to 6. Poverty has been on a declining trend poverty reduction, especially after the following strong economic growth in the late 2000s. 2000s. Based on data from the IOF-2014/15, 48.4 percent of Mozambicans live beneath the poverty 5. The measurement of poverty in Mozambique line, lower than the levels of poverty recorded in is based on the value of a “minimum” level of 2002/03 and 2008/09, 60.3 and 58.7 percent, consumption necessary for normal short- and respectively (Figure 1.4). This corresponds to an long-term human well-being, which is estimated annual reduction in poverty, on average, of 1 from household surveys collected nearly every percentage point. Yet, poverty fell markedly faster 5 or 6 years. Under this method, households in the period 2008/09-2014/15 (on average 1.8 not deemed poor have consumption levels that percentage points annually) than in the period are enough to meet their basic food needs and 2002/02-2008/09, where the poverty rate barely other non-food essential expenditures.¹ The dropped (on average 0.26 percentage points household budget survey used in this study, annually). The official numbers, reported in the known as Inquérito aos Orçamentos Familiares Fourth National Poverty Assessment conducted (IOF) (household survey of living conditions), by the Government of Mozambique (2016), also is collected by the National Statistics Office of reflect a downward trend in poverty –from 52.8 Mozambique (Instituto Nacional de Estatistica, percent in 2002/03 to 46.1 percent in 2014/15– INE).² The surveys are representative at the and faster reduction in recent years. national, rural-urban, and provincial levels.³ 1 The official methodology defines this “minimum” level using 13 different values (poverty lines) for an equal number of regions. The methodology employed in this study also follows the “basic needs” approach but defines only one absolute minimum level of necessary resources for the entire country. The methodology of this study also adjusts household consumption to reflect regional differences in prices and temporal differences in prices over the course of the data fieldwork. See Box 1 and the full report for more details about the methodological differences. 2 The first survey was collected by INE between February 1996 and April 1997. Due to better comparability of the data, the analysis is based on the last three waves of the household budget survey (2002/03, 2008/09 and 2014/15). Contrary to the previous rounds of the IOF, the survey collected in 2014/15 was implemented as a panel. However, for poverty measurement purposes, the three quarters of data collected as part of the IOF-214/15 have been appended as a pooled cross section. 3 This decision was made by the Mozambican Ministry of Economy and Finance (MEF) to capture the seasonality of consumption and avoid using the longitudinal nature of the data that was affected by high attrition rates. More details about these issues are discussed in the full poverty assessment report. 3 mozambique poverty assessment | 1. progress in reducing poverty Figure 1.4. Poverty has been falling since the early 2000s, but the pace accelerated after 2008 (World Bank methodology) (Official methodology) 70 70 65 60.3 65 Poverty rate, % Poverty rate, % 58.7 60 60 55 55 52.8 48.4 51.7 50 50 46.1 45 45 40 40 2002/2003 2008/2009 2014/2015 2002/2003 2008/2009 2014/2015 Source: MEF (2016) and World Bank using IOF-2002/03, IOF-2008/09 and IOF-2014/15 Rural areas continue to lag behind. dropping from 41.1 percent to 32 percent (Figure 1.5). In relative terms, poverty declined faster in 7. The share of households living in poverty urban centers (23.2 percent) that in rural areas has fallen in rural and urban areas, but poverty (18.8 percent). Rural households have been remains significantly higher in the former. Poverty concentrated in the bottom part of the distribution. in rural areas declined from 69 percent in 2002/03 Nearly 8 out of 10 poor people are in rural areas, to 66.4 percent in 2008/09 and to 56.0 percent fairly similar to the urban-rural composition in 2014/15. Urban poverty shows a similar trend, observed in the early 2000s (Figure 1.6).⁴ Figure 1.5. Welfare levels have not converged Figure 1.6. Since the early 2000s, nearly 8 out between urban and rural areas of 10 poor people are in rural areas (poverty rates by area) (Rural and urban poverty shares) 80 (2002/2003) (2014/2015) 70 69.0 66.4 22.2% 21.0% Poverty rate, % 60 Rural 56.0 50 41.1 41.7 40 32.0 Urban 30 77.8% 79.0% 20 Rural Urban 2002/2003 2008/2009 2014/2015 Source: World Bank using IOF-2002/03, IOF-2008/09 and IOF-2014/15 Source: World Bank using IOF-2002/03, IOF-2008/09 and IOF-2014/15 8. The downward trend in poverty is also PPP can also be used to examine the level and observed when measured using an international evolution of poverty in Mozambique. This line poverty line. The global poverty line of US$1.90 is not a substitute of the official poverty line but 4 The rural and urban divide in poverty is also evident in the official numbers (Government of Mozambique, 2016). However, the composition of poverty across areas is different. Urban poverty is higher whereas rural poverty is lower in the official estimates relative to the methodology followed in this study. 4 1. progress in reducing poverty | mozambique poverty assessment rather an international threshold that is used to in 2008/09 and to 62.9 percent in 2014/15 measure and track poverty trends worldwide when this poverty line is used as the threshold (see Box 1). The poverty headcount ratio fell of reference (Table 1.1). from 78.5 percent in 2002/03 to 67.9 percent Table 1.1. Poverty headcount ratio for national poverty line and the US$1.9 PPP poverty line 2002/03 2008/09 20014/15 National 60.3% 58.7% 48.4% Urban 41.7% 41.1% 32.0% Rural 69.0% 66.4% 56.0% US $1.9 PPP Poverty Line 78.5% 67.9% 62.9% Source: World Bank using IOF-2002/03, IOF-2008/09 and IOF-2014/15 and Povcalnet 9. Owing largely to the rapid growth in nearly 3 percent. The average total fertility rate population, the absolute number of poor is estimated at 5.9 children per woman, nearly people in Mozambique has increased over one child more than the average for countries time despite the decline in the overall in the region. The rapid increase in population poverty rate. Long-term demographic trends, is making it more difficult to reduce the number particularly high and stagnant fertility rates, have of poor people even though the poverty rates slowed down the pace of poverty reduction. have been falling. Indeed, as of 2014/15, the The population of Mozambique increased from country has 12.3 million people living below the 18 million in 2000 to a projected 28.8 million in poverty line, 1.3 million more than in 2002/03 2017, equivalent to an average growth rate of (Table 1.2). Table 1.2. The total number of poor has increased, mostly in rural areas (in thousands) National Rural Urban 2002/03 11,032 8,582 2,450 2008/09 12,647 9,959 2,688 2014/15 12,336 9,752 2,584 Source: World Bank using IOF-2002/03, IOF-2008/09 and IOF-2014/15 Poverty reduction has also been uneven and Maputo City, which recorded the largest across regions decline. By 2014/15, poverty rates in these three provinces are well above the national average 10. Welfare levels remain low in the Northern (48.4 percent): Niassa (67 percent), Nampula and the Center Regions relative to the South. (65 percent) and Zambezia (62 percent). Back There are large spatial differences in poverty in 2002/03 the provincial rankings differed levels and changes across provinces.5 Poverty strongly, with Tete, Gaza and Inhambane continues to be high in Zambezia, Nampula exhibiting poverty rates above 70 percent. and Niassa, in contrast to Maputo Province Since then poverty reduction has been faster in 5 Mozambique is administratively divided into 10 provinces and one capital city (Maputo) with provincial status. 5 mozambique poverty assessment | 1. progress in reducing poverty these provinces, falling by around 40-50 percent. indicators, with a decline of 70 percent, even Likewise, Maputo Province and Maputo City though they had significantly lower poverty than recorded the largest improvements in poverty other areas of the country in 2002/03 (Figure 1.7). Box 1. Measuring poverty in Mozambique Official poverty measurement methodology The official methodology to estimate poverty in Mozambique was developed by the Mozambican Ministry of Economy and Finance with technical assistance from UN-Wider. As most countries in Sub-Saharan Africa (SSA), the poverty estimates are based on aggregate household consumption as the key welfare indicator. The consumption aggregate comprises food consumption, including food produced by households themselves, as well as expenditures on a range of nonfood goods (including durables such as car, TVs, computers, etc.) and services (e.g., housing, clothing, utilities, transportation, communication, health, education, etc.). Price deflators are used to adjust the consumption aggregate for differences in prices across geographic areas as well as differences across time over the course of the IOF fieldwork. The poverty lines are based on the cost-of-basic-needs (CBN) approach. The methodology defines food poverty lines for 13 geographic regions anchored in the cost of region- specific food baskets that provide 2,150 calories per person per day. These lines are augmented to include an allowance for basic non-food needs. The regional poverty lines are re-estimated every time there is a new household budget survey. The poverty rate measures the proportion of people whose monthly price-adjusted total household consumption per capita is below the poverty line in the corresponding year and region. The values of the poverty lines used in the estimation of poverty with the IOF-2014/15 are found in Annex 2. Poverty measurement methodology followed in this study This Poverty Assessment followed a methodology that is close to the methods used in most countries within and outside SSA. This methodology is also based on the CBN approach and, for that reason, most of the concepts underlying it are analogous to those underpinning the official methodology. There are, however, some important differences. Regarding the consumption aggregate, the difference lies chiefly in the assumptions adopted to impute the value of services delivered by durable goods. Another difference is the use of a single poverty line. The food poverty line was calculated using the average food basket and price per calorie of households between the percentiles 40th and 60th in the IOF 2014/15. The reference food basket obtained corresponds to 1,460 calories. This value is below the calorie requirement of a typical Mozambican for adequate nutrition – usually around 2,100 calories per person per day. Yet, a decision was made to not scale up the value of the food basket to ensure consistency with the issue of systematic underestimation of calorie consumption in the IAF and IOF surveys and to reflect the behavior of households as depicted in the actual data. Like the official methodology, the non-food poverty line adds the cost of other essentials observed in the reference group. The 2014/15 poverty line was deflated to 2002/03 and 2008/09 values using the official Consumer Price Index (CPI). Finally, the consumption aggregate was adjusted to account for geographic food price variation using a Paasche index for each household. Annex 1 provides more details. The World Bank $1.9 International Poverty Line The World Bank uses a global poverty line set at $1.9 per person per day using 2011 prices to measure and track the evolution of poverty line worldwide. This line is not a substitute of the official poverty line, which is defined based on each country’s specific economic and social circumstances. The value is derived from the national poverty lines of the 15 countries (including Mozambique) with the highest levels of poverty in 2005. To ensure that the same quantity of goods and services are priced equivalently across countries, the 15 poverty lines are converted to a common currency using 2011 purchasing 6 1. progress in reducing poverty | mozambique poverty assessment power parity (PPP) exchange rates. The average of these 15 lines in PPP terms was $1.9 per person per day. The international line is above the average of the 13 poverty official lines in Mozambique for 2014/15 ($1.54 2011 PPP) and the WB poverty line estimated for this report ($1.49 2011 PPP). Figure 1.7. Niassa, Nampula and Zambezia are the provinces with the highest poverty rates (poverty rates across provinces) 100 79 75 74 90 70 67 65 69 63 62 63 62 65 61 65 62 Percentage 55 54 57 60 44 48 48 50 50 50 49 42 37 38 35 40 20 12 13 14 4 0 2002/03 2008/09 20014/15 2002/03 2008/09 20014/15 2002/03 2008/09 20014/15 2002/03 2008/09 20014/15 2002/03 2008/09 20014/15 2002/03 2008/09 20014/15 2002/03 2008/09 20014/15 2002/03 2008/09 20014/15 2002/03 2008/09 20014/15 2002/03 2008/09 20014/15 2002/03 2008/09 20014/15 Tete Inhambane Gaza Zambezia Niassa Manica Nampula Sofala Cabo Maputo Maputo Delgado Province City Source: World Bank using IOF-2002/03, IOF-2008/09 and IOF-2014/15 Economic growth and poverty have circa the middle of the 2000s and early/middle become more closely linked in recent years. 2010s suggests that the direction and pace of poverty reduction in Mozambique is within the 11. The fall in poverty is consistent with the range seen for most countries in the region. trend seen in other countries in the region in For illustration, poverty has fallen by nearly 8 the last decade or so. Poverty has been falling in percentage points in Rwanda between 2005 and most countries in sub-Saharan Africa in the last 2010, by almost 7 percentage points in Zambia 10-15 years. Comparisons of poverty levels and between 2010 and 2015 and by close to 4 changes across countries are difficult because percentage points in Tanzania – in Mozambique the years of the surveys vary from country to it fell by 5 percentage points in the years between country. Yet, looking at the evolution of numbers 2008/09 and 2014/15 (Figure 1.8). Figure 1.8. The incidence of monetary poverty fell in Mozambique and other countries in the region (Percentage of the population below the US$1.9 PPP poverty line) 90 78.5 81.8 80 73.6 70.9 67.9 68.7 68.0 64.4 70 62.9 60.3 62.2 57.5 60 52.7 48.9 50 41.5 40 32.5 30.1 33.6 30 20 10 0 2002 2008 2014 2001 2010 2004 2010 2005 2010 2002 2009 2010 2015 2007 2011 2000 2008 2005 Mozambique Madagascar Malawi Rwanda Uganda Zambia Tanzania Angola Kenya Source: World Bank using WDI 7 mozambique poverty assessment | 1. progress in reducing poverty 12. The pace of poverty reduction in recent time where GDP per capita grew at a stable years reveals that economic growth and rate (on average around 4 percent per year) but poverty became more closely linked. The GDP poverty evolved at different rates. For most of the per capita growth elasticity of poverty gives an 2000s (2002/03-2008/09), poverty dropped by estimate of how closely (or not) are growth and a total of 1.6 percentage points, which translates poverty linked.⁶ At 0.3 over the period 2002/03- into an elasticity of 0.08. However, the pace 2014/15, the response of poverty to fast and picked up remarkably after 2008/09, with the accelerating economic growth in Mozambique elasticity rising to 0.68 for the period 2008/09- has been relatively modest.7 This relatively low 2014/15 (Figure 1.9). value is the result of averaging two intervals of Figure 1.9. In recent years growth has been more poverty reducing in Mozambique (GDP per capita growth elasticity of poverty, Mozambique) 0.80 0.68 0.60 0.40 0.30 0.20 0.80 0.00 2002/03 - 2014/15 2002/03 - 2008/09 2008/09 - 2014/15 Source: World Bank using WDI, IOF-2002/03, IOF-2008/09 and IOF-2014/15 Mozambique is not yet on a path to end recovering and sustaining strong consumption extreme poverty by 2030 but more broadly growth and making it more inclusive. If the shared growth can bring the country faster pattern of growth is simulated to stay pro-rich, to this target. as in recent years, poverty will fall but at a slow pace, and inequality will worsen even further. 13. Projections of consumption growth can By 2030 poverty would fall from 48.4 percent give an idea as to whether Mozambique is on to 32.1 percent, over 10 percentage points less a path to end extreme poverty –or reduce it than the reduction achieved under a pro-poor substantially– by 2030. Figure 1.10 shows the growth scenario. A third scenario shows that trends for the poverty rates between 2015 and nearly 36 percent of the Mozambicans would 2030 based on three different scenarios.8 An be poor by 2030 if consumption growth is optimistic pro-poor growth scenario could equally distributed population but below past reduce poverty significantly by 2030 but it performance reflecting the slower economic will not eradicate it. Monetary poverty would growth experienced in recent years. fall from 48.4 to 21.8 percent, but this requires 6 It measures the percentage change in poverty with respect to a 1 percentage change in GDP (or consumption per capita). 7 A 1 percentage increase in GDP per capita in Uganda is associated with a fall in poverty of 0.95 percentage points for similar period (elasticity = 0.95). 8 The three scenarios are as follows: 1) neutral but lower growth scenario – assumes annual distribution-neutral consumption growth of 2 percent, below the average annual growth of mean consumption recorded for the period 2002-2014 (2.9 percent) to reflect the lower economic growth of recent years; 2) pro-rich growth scenario – assumes that the speed of consumption growth for the upper half of the distribution is 3.5 percent, more than twice the growth rate of the bottom 50 (1.5 percent) and 3) pro-poor growth scenario – assumes the opposite of the growth levels set in the pro-rich scenario for the bottom and upper halves of the distribution, namely 1.5 percent for the top 50 and 3.5 percent for the bottom 50. 8 1. progress in reducing poverty | mozambique poverty assessment Figure 1.10. Poverty will fall markedly moving forward if growth is strong and more equally shared (poverty headcount projections under different scenarios) 60 Projected poverty rate, % 50 40 35.8 30 32.1 20 21.8 10 0 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 Pro-rich growth Neutral but lower growth Pro-poor growth Source: World Bank using IOF-2014/15 9 mozambique poverty assessment | 2. the inclusiveness of economic progress 2 The Inclusiveness of Economic Progress While household consumption growth “pro-poor”, namely benefited mostly the accelerated after 2008, it became less bottom 50 percent. The broad-based pattern inclusive, limiting the fostering of shared of growth reversed after 2008. The growth prosperity and raising inequality. incidence curve (GIC), which shows the percent change in average consumption for 14. While growth picked up in recent years, each percentile of the distribution, indicates it has been benefiting proportionally more that growth between 2008/09 and 2014/15 the wealthier segments of the population. became stronger (4.34% as measured by Taking a closer look at changes in the the growth rate at the mean) but “pro-rich”, distribution of consumption over time sheds particularly in urban areas (right panel of Figure light on which income groups benefitted the 2.1). As noted below, these two features explain most from economic growth. The period why faster poverty reduction took place 2002/03-2008/09 recorded a small reduction alongside with increasing inequality. Annual in poverty because of slow average growth consumption growth for the top quintile was in consumption (0.11 percent annually). But in the order of 7.5 percent, nearly three times despite being low, consumption growth was faster than the growth rate of the bottom 40. Figure 2.1. Growth after the late 2000s benefitted mostly the non-poor, chiefly in urban areas (Consumption Growth Incidence Curves with 95% confidence intervals nation-wide, urban and rural, 2008/09 – 2014/15) 10 12 Annual growth rate, % Annual growth rate, % 8 10 6 8 4.34 6 4 4 2 2 0 0 0 10 20 30 40 50 60 70 80 90 0 10 20 30 40 50 60 70 80 90 -2 GIC Mean Growth Rate 95% CI Urban GIC Rural GIC Note: Dotted lines show 95% confidence intervals. Source: World Bank using IOF-2008/09 and IOF-2014/15 10 2. the inclusiveness of economic progress | mozambique poverty assessment 15. A less inclusive pattern of growth limits Since consumption growth was higher among Mozambique’s progress in fostering shared the better off after 2008, the shared prosperity prosperity faster. The Shared Prosperity Indicator premium turned negative relative to earlier years captures two key elements, economic growth (-3.5 percent) as the average consumption of the and equity. Strong and stable economic growth top 60 grew faster (6.8 percent on average per is necessary to increase the living standards of year) than the consumption of the bottom 40 the population. But for robust growth to trickle, (3.2 percent) (Figure 2.2). it needs to be inclusive of the less well-off. Figure 2.2. Positive shared prosperity reversed after 2008, signaling weaker inclusiveness (Average annual consumption growth for the bottom 40, top 60 and shared prosperity premium) 8 6.8 Annual growth rate, % 6 4 3.2 2.9 2.3 2.3 2 1.4 0 -2 -0.6 -0.9 -4 -3.5 -6 2002-2008 2008-2014 2002-2014 Top 60 Bottom 40 Premium Note: Shared prosperity premium is the difference between the consumption growth rates of the bottom 40 and the top 60. Source: World Bank using IOF-2002/03, IOF-2008/09 and IOF-2014/15 16. Consequently, inequality is not only high but size of the changes are different (Government rising. Mozambique is among the most unequal of Mozambique, 2016). The Gini coefficient countries in sub-Saharan Africa as measured by hovered around 0.40-0.42 from 1996/97 to the Gini coefficient9 (Figure 2.3). The recent pro- 2008/09 but increased since after reaching 0.47 richness of growth is contributing to raise it even in 2014/15. Additional disaggregation of the data further. The Gini coefficient increased from 0.47 shows that the worsening of income inequality to 0.56 between 2008/09 and 2014/15. Inequality is largely the result of higher concentration in numbers reported with the official methodology urban areas (Figure 2.4). produce the same trend albeit the levels and 9 The Gini Coefficient is the most popular measure of inequality. It is derived from the Lorenz curve, which shows the cumulative proportion of the population on the horizontal axis and the cumulative proportion of consumption or income on the vertical axis, sorted from the poorest to the richest household. The Gini is calculated as the ratio of the area between the Lorenz Curve and the diagonal of perfect equality, namely each household has the same consumption/income share. The Gini coefficient ranges from 0 (perfect equality) to 1 (perfect inequality). 11 mozambique poverty assessment | 2. the inclusiveness of economic progress Figure 2.3. The distribution of household consumption is highly unequal by regional standards (Gini coefficient for selected countries and years) 0.7 0.63 0.61 0.6 0.56 0.56 0.56 0.54 0.52 0.52 0.51 0.49 0.46 0.5 0.47 0.47 0.43 0.40 0.4 0.3 Swaziland Mozambique (2008/09) Mozambique Republic (2008) (2014/15) (2011) Guinea-Bissau South Africa Botswana (2009) Central African Zambia (2010) Lesotho (2010) Rwanda (2005) (2009) (2010) Kenya (2005) Malawi (2010) Madagascar (2012) Tanzania (2007) Mozambique (2002/03) Source: World Bank using WDI Figure 2.4. Inequality is high and increasing, a trend driven by worse inequality in urban areas (Consumption-based Gini coefficient) 0.7 0.62 0.6 0.56 0.55 0.47 0.49 0.47 0.48 0.48 0.5 0.42 0.47 0.42 0.47 0.42 0.43 0.43 0.36 0.37 0.37 0.4 0.3 0.2 0.1 0 2002/03 2008/09 2014/15 2002/03 2008/09 2014/15 2002/03 2008/09 2014/15 National Urban Rural Official WB Source: World Bank using IOF-2002/03, IOF-2008/09 and IOF-2014/15 Had growth been more equally shared effect) is driving the fall in poverty. In contrast, Mozambique would have achieved twice the increase in inequality in the distribution of as much poverty reduction after 2000. consumption (redistribution effect) offsets half of the contribution of the growth effect to poverty 17. The decline in poverty in Mozambique has reduction, increasing the incidence of poverty. been hindered by high and rising inequality. The growth effect alone would have reduced Changes in poverty can be decomposed into poverty by 23.1 percentage points between 2002 “pure growth” and “redistribution” elements to and 2014 – bringing the poverty headcount down shed light on whether the benefits of growth reach to 37.2 percent rather than the actual 48.4 percent the poor.¹⁰ As shown in Figure 2.5, the increase – had the redistribution effect not increased in mean household consumption (growth poverty by 11.2 percentage points. 10 This report follows the method proposed by Datt and Ravallion (1992). This decomposition is based on the idea that that a measure of monetary poverty can be expressed as the product of mean consumption and a parameterized Lorenz curve. Keeping the Lorenz curve constant gives the distribution neutral growth that would drive the average increase in consumption across the population, for instance, raising the levels of consumption of all households by the same rate. The other part is derived from holding the mean consumption constant (a mean-preserving redistribution) to capture the change in the shape of the consumption distribution driven by, for instance a faster growth in the consumption of the poorest relative to the consumption growth of the richest. There is also a third, much lower “price” effect explained by price adjustments made to the poverty line over time. 12 2. the inclusiveness of economic progress | mozambique poverty assessment Figure 2.5. Higher inequality has offset the contribution of growth to poverty reduction (Growth, redistribution effects on poverty reduction by period, percentages) 20 11.5 11.2 10 2.5 1.7 0.0 0 -10 -5.8 -1.6 -1.7 -20 -10.4 -11.9 -30 -20.2 -23.1 2002-2008 2008-2014 2002-2014 Growth Redistribution Price Total Source: World Bank using IOF-2002/03, IOF-2008/09 and IOF-2014/15 The increasing role of services in the and the contribution to jobs in the economy economy and favorable macroeconomic jumped from 15 to 24 percent (Table 2.1). conditions contributed to faster consumption growth after the late 2000s. 19. The transition of workers from agriculture into services has contributed to faster increase 18. The increasing role of the services sector, an in the standards of living after the late 2000s. outcome of the ongoing process of structural Decomposition analysis shows that growth in transition, has created a path to jobs outside labor productivity has been the main engine of agriculture. In recent years, the sources of growth economic growth in the last two decades (Table have gradually shifted away from agriculture. 2.2).11 In contrast, changes in employment levels Between the late 1990s and the middle of the and labor force participation have had a negligible 2000s, output growth was pulled by investments contribution. Labor productivity growth, in turn, in capital intensive industrial activities (largely has been driven in recent years by the shift of jobs “megaprojects” in extractive, export-oriented away from agriculture and into sectors with higher industries) with relatively higher value-added but productivity. Back in 1996, shortly after the end low job creation. The other emerging economic of war, 86.6 percent of workers were primarily activity, the services sector, also began to play a engaged in agriculture. That share fell to 71 percent larger role in the economy, both in terms of output by 2014 and most of that shift was absorbed by the and employment. Between 2008 and 2014, its service sector, where productivity is over six times share of GDP increased from 49.8 to 55.7 percent larger despite high levels of informality. Table 2.1. The services sector is gradually playing a greater role in the economy (GDP and jobs composition across economic sectors) Sector shares of GDP 1996 2003 2009 2014 Agriculture 38.1% 31.4% 30.5% 25.5% Industry 10.2% 21.1% 19.7% 18.8% Services 51.8% 47.7% 49.8% 55.7% Total 100% 100% 100% 100% Sectors shares of jobs 1996 2003 2009 2014 Agriculture 86.6% 80.5% 80.4% 71.0% Industry 4.4% 3.4% 4.7% 4.9% Services 9.0% 16.1% 15.0% 24.0% Total 100% 100% 100% 100% Source: World Bank Jobs Diagnostics (2017) 11 A “growth accounting exercise” can be used to decompose GDP per capita growth into four components: productivity, the employment rate, the labor participation rate and the ratio of the working age population to the total population. 13 mozambique poverty assessment | 2. the inclusiveness of economic progress Table 2.2. Labor productivity growth is the single greatest contributor to growth in GDP per capita (sources of GDP per capita growth) 1996-2014 1996-2003 2003-2008 2008-2014 Annual Growth of GDP per capita 4.85 5.41 5.30 3.83 % Yearly Contribution to Growth of: Productivity (Y/E) 5.36 5.01 6.30 4.89 Employment Rate (E/LFP) -0.07 0.27 -0.27 -0.30 Participation Rate (LFP/WAP) -0.34 0.28 -0.49 -0.87 Demographic Change (WAP/P) -0.09 -0.15 -0.24 0.11 Source: World Bank Jobs Diagnostics (2017) 20. Macroeconomic conditions were favorable Data on internal migration is scarce but for private and public consumption growth. numbers from the populaiton census of 2007 Macroeconomic expansionary policies indicate that 8 percent of the Mozambicans provided the right conditions for faster private live in a district different from the one in which consumption growth in the period 2008/09- they were born and half of them are located 2014/15. Public expenditures, measured as in a province outside the place of birth.12 a proportion of the GDP, increased steadily Decomposition analysis suggests that that most between 2008 and 2014, raising from 24 percent of the poor that left their rural homes stayed to 39 percent. Mozambique also experienced poor after they settled in urban areas.13 The several years of expansionary monetary policy gains in consumption growth that lifted people over the past decade. Annual credit growth to out of poverty were concentrated on individuals the private sector averaged 23 percent between that already lived in rural or urban areas and did 2009 and 2015. All this happened along a sharp not migrate. Scarce employment opportunities, increase in external inflows of resources. Foreign skills mismatch, low productivity and high costs direct investments into Mozambique increased of living are factors that undermine the chances continuously, reaching almost 40 percent of for rural migrants to improve their livelihoods GDP in 2013, up from 5 percent in 2008. after settling in urban centers.14 Internal migration had a small effect on Other factors such as location, poverty reduction. demographic structure and limited ownership of and lower returns on assets 21. While Mozambique is slowly becoming continue to keep people in poverty. more urbanized, rural-urban migration appears to have contributed little to poverty reduction. 22. Geographic location, demographic An increasing share of the population now lives structure, education, type of work and isolation in urban areas owing in part to migration flows matter for poverty. As noted before, poverty form rural areas, which are pulled mostly by is overwhelming rural in Mozambique. As of the prospect of better economic opportunities. 2014/15, the poverty rate is 24 percentage 12 The lack of information about the place of origin of migrants in the Census 1997 and the Census 2007 does not allow establishing the share of migrants that moved from rural to urban areas. 13 This analysis follows the methodology proposed by Ravallion and Huppi (1991). It decomposes changes in poverty over time into “intra-regional effects” (poverty changes within urban and rural areas assuming no migration between the two of them), “inter-regional effects” (allowing for changes in the distribution of the population between rural and urban areas keeping poverty rates constant) and an “interaction” term that can be interpreted as a measure of the correlation between the population shifts and the intra-regional changes in poverty. 14 After controlling for human capital and occupation, earnings are not significantly higher in urban areas than in rural areas. On average, nominal earnings are 26 percent higher in urban areas than in rural areas, not enough to offset the differences in the costs of living. The undermines the possibility of a potential urban wage premium among the unskilled (World Bank, 2017c). 14 2. the inclusiveness of economic progress | mozambique poverty assessment points higher in rural areas than in urban areas, transportation and markets are systematically 56 percent and 32 percent, respectively. Poor lower among poor households (Table 2.3). households are larger, having on average nearly 1.1 more members. In addition to family size, 23. Poor households are also characterized the age structure of poor households implies by having limited ownership of basic assets higher levels of dependency since they have and earning lower returns on them. Despite relatively more children in ages 0 to 14. The level having improved over time, endowments such of schooling of the household head is associated as physical, financial and human capital remain with the poverty status of the family. Household lower among poor households relative to the heads that are poor have on average 0.8 fewer non-poor. In addition to this, decompositions years of education than those that are not poor. of consumption growth show that returns on The sector of work also matters. Regression these endowments also help explain the welfare results show that employment in agriculture gap, particularly in recent years.¹⁵ The better off remains a strong predictor of poverty. Coverage are found to benefit relatively more from higher and accessibility to critical infrastructure and returns on their endowments relative to the services such as electricity, water and sanitation, returns earned by the poor. 15 This study performed regression analysis to disentangle the changes in consumption into changes in two components: 1) household endowments such as demographic characteristics, education, experience, assets, access to basic services, location, proximity to markets and occupation, among others; and 2) returns to those characteristics such as returns to education, experience, land productivity, etc. More details of the methodology employed are available in the long version of this poverty assessment. 15 mozambique poverty assessment | 2. the inclusiveness of economic progress Table 2.3. The livelihoods of the poor differ from those of the non-poor in many key aspects Variable Poor Non-Poor Significance Significance (t-test) (Model) Household socio-demographics Age of household head 43.5 43.9 ** *** Female-headed (%) 23.3 24.2 ** *** Household size 6.8 5.7 *** *** Proportion of children aged 0 to 14 (%) 55.0 42.8 *** *** Lives in rural area (%) 79.0 58.2 *** *** Years of education of household head 3.5 4.3 *** *** Household head is illiterate 19.7 12.8 *** *** Sector of work Head works in agriculture (%) 74.6 49.7 *** *** Head works in manufacturing (%) 3.2 4.5 *** Head works in services (%) 6.6 20.2 *** Head employed in private sector (%) 1.9 8.9 *** *** Head employed in public sector (%) 6.3 13.5 *** Access to services Improved water 48.0 69.2 *** *** Improved sanitation 14.0 42.4 *** *** Electricity 11.5 42.0 *** *** Within 30 minutes of walking distance to … Road (%) 46.3 66.6 *** *** Market (%) 48.3 65.8 *** *** School (%) 66.5 77.8 *** *** Health facility (%) 64.4 71.3 *** Asset ownership Household has a car 0.4 8.2 *** *** Household has a bicycle 40.1 35.4 *** *** Household has a T.V. 11.0 43.1 *** *** Household has a fridge 4.4 30.7 *** *** Household has a phone 46.1 75.3 *** *** Note: Column t-test shows significance values from a standard unconditional t-test of differences between the means. Column Model shows significance values from a binary dependent variable (poor =1, = 0 otherwise) model (probit) controlling for all variables shown and province fixed effects. *, **, and *** indicate significance level at 10%, 5%, and 1%, correcting for the clustered nature of the errors in the probit regressions. Source: World Bank using IOF-2002/03, IOF-2008/09 and IOF-2014/15 16 3. evolution of living conditions and economic mobility | mozambique poverty assessment 3 Evolution of Living Conditions and Economic Mobility The average household in Mozambique the increase in school enrollment has gone has better standards of living today than in hand with an increase in educational at the turn of the century, but there are attainment. Figure 3.1 shows two snapshots of still major gaps. the school attainment across education levels for the population 20-65 years old, one for 24. Several education indicators show 2002/03 and the other one for 2014/15. The improvements in human capital accumulation increase is evident, with fewer people without since the early 2000s. Mozambican children education and instead a higher share of the are now more likely to participate in school population that either accumulated some than before. Back in 2002/03, 43 percent years of primary and secondary education or of children ages 5 to 14 were not enrolled completed both levels. Overall, average school in school, a value that fell to 24.2 percent attainment increased to 5.1 years of schooling, in 2014/15. There has also been a decline in up from 2.4 in 2002/03. late and overage enrollment. Furthermore, Figure 3.1. Higher school participation is slowly increasing educational attainment (school attainment by educational levels) (2002/03) (2014/15) Higher Education 0.3 0.3 Higher Education 3.5 2.1 Some technical 1.4 0.6 Some technical 2.2 1.1 Completed secondary 2.8 1.8 Completed secondary 10.0 7.3 Some secondary 5.1 3.0 Some secondary 16.2 11.7 Completed primary 12.4 13.4 Completed primary 20.1 14.3 Some primary 25.1 26.2 Some primary 38.9 41.8 None 53.0 54.2 None 9.2 21.8 Male Female Male Female Source: World Bank using IOF-2002/03 and IOF-2014/15 17 mozambique poverty assessment | 3. evolution of living conditions and economic mobility 25. Health outcomes such as life expectancy, 54.3 percent and 47.7 percent, respectively. infant and maternal mortality, and morbidity are also moving in the right direction. Since 2001, 26. Access to basic services such as water and longevity has increased by almost 9 years from 48.8 sanitation improved but large disparities in to 57.6. The infant mortality rate, expressed as the coverage remain across different groups of the number deaths per thousand live births, fell from population. Almost 70 percent of the population 99.1 in 2003 to 68.1 in 2011. Child mortality rates has access to safe water, a 28-percentage point have followed a comparable downward trend. increase from the level in 2002/03. Regarding Over the same period, maternal mortality from any access to improved sanitation, nearly 4 in 10 cause related to or aggravated by pregnancy have households were covered in 2014/15, twice the fallen too, from 804 to 596 deaths per 100,000 coverage level in the early 2000’s. Improvements in live births (Figure 3.2). The fraction of workers that access to electricity are positive (increased from 12.2 reported not working due to sickness dropped percent to 40.9 percent) but overall electrification from 16.8 percent in 2002/03 to 13.9 percent in rates are low (Figure 3.3). There are, however, large 2008/09. The improvement in health outcomes is gaps between income groups and rural and urban associated with a modest increase in access to and areas. Household location and income levels are utilization of health services. In 2011, 90.8 percent strong determinants of access to basic services. of pregnant women underwent an ante-natal For instance, access to safe water and sanitation check, a higher fraction than in 2003 (84.5 percent). among urban households is 89.4 percent and Over the same period, the proportion of children 69.4, respectively. The corresponding rates for under five with full immunization coverage raised rural households are 46.6 percent and 7.5 percent. from 43 percent to 46 percent. More births were Similarly, nearly 60 percent of urban households are delivered in a health center or with the assistance connected to the distribution network compared to from a health professional in 2011 than in 2003, 15.1 percent amongst rural households (Figure 3.4). Figure 3.2. Infant and maternal mortality rates have fallen (infant mortality: infant deaths per 1,000 live births) (maternal mortality per 100,000 live births) 120 99.1 1,000 804 100 800 80 68.1 596 600 60 40 400 20 200 0 0 2003 2011 2003 2011 Source: World Bank using DHS-2003 and DHS-2011 Figure 3.3. Access to basic services is Figure 3.4. Location is a strong determinant of improving but are yet far from access to basic public services being universal 80% 69.7% 100% 89.4% 52.7% 80% 69.4% 60% 60.0% 42.6% 39.3% 40.9% 60% 46.6% 40% 25.5% 22.1% 19.7% 40% 20% 12.2% 15.1% 20% 7.5% 0% 0% Improved Improved Electricity Improved Improved Electricity Water Sanitation Water Sanitation 2002/03 2008/09 2014/15 Rural Urban Source: World Bank using IOF-2002/03, IOF-2008/09 and IOF-2014/15 Source: World Bank using IOF-2014/15 Source: World Bank using IOF-2014/15 18 3. evolution of living conditions and economic mobility | mozambique poverty assessment 27. Households have also experienced especially in rural areas. There have also improvements in their housing conditions and been improvements in ownership of bicycles, the ownership of basic assets. All indicators motorcycles, TVs and, above all, cellphones, measuring the quality of housing such as which record the most marked increase. The improved floor, improved roof and improved share of households having a mobile phone has walls show positive development between multiplied by 15 from 4 to 61 percent, including 2002/03 and 2014/15, providing evidence for increased ownership among poor households, rising living standards. Along the same lines, which reached 46 percent in 2014/15 (Figure Mozambicans now own more assets than 3.5). While in general these positive trends are what they used to own in the past. Ownership observed in urban and rural areas, they are of traditional household items such as beds, more marked in the former, and especially for irons and fridges, among others, has increased, households from the top of the distribution. Figure 3.5. Ownership of traditional and modern assets has increased (traditional assets) (modern assets) 0.6 0.54 0.7 0.61 0.5 0.46 0.43 0.6 0.38 0.5 Percentage Percentage 0.4 0.34 0.29 0.28 0.4 0.3 0.3 0.28 0.2 0.18 0.2 0.10 0.04 0.05 0.06 0.05 0.03 0.1 0.1 0.04 0.00 0.00 0.02 0.0 0.0 Iron Bed Bicycle Radio Cellphone TV Fridge Computer Motorcycle Car 2002/03 2014/15 2002/03 2014/15 Source: World Bank using IOF-2002/03 and IOF-2014/15 Multidimensional poverty has been if it is deprived in a given number of indicators falling, but more so in the last decade – for this analysis the threshold is set at 3 out and in urban areas. of the 8 indicators.17 The findings show a drop in multidimensional poverty, from 92.8 percent 28. While it has fallen, multidimensional in 2002/203 to 71 percent in 2014/15, but most poverty remains high, with 7 in 10 of the gains were achieved in the period of Mozambicans still deprived in several faster poverty reduction (2008/09- 2014/15) key aspects of human welfare. Looking (Figure 3.6). The gains are concentrated in at deprivations in multiple dimensions of urban areas, where the share of the population wellbeing (human capital, access to services, experiencing multidimensional deprivation fell housing conditions, asset ownership and steadily from 78.6 percent to 32.0 percent. In monetary poverty) all at once sheds light on contrast, progress has been noticeably slower how they overlap and possibly reinforce each in rural areas, where 9 in 10 rural households other.16 A household is multidimensionally poor are poor in multidimensional sense. 16 The eight indicators used for this part of the analysis are the following: education (no household member completed primary schooling, at least one school-age child in the household is out of school), access to services (no access to electricity, improved water and improved sanitation); housing conditions (poor quality dwelling ), asset ownership (no ownership of at least two of the following assets: fridge, TV, phone, bicycle, car or motorcycle) and the prevalence of monetary poverty (household’s consumption per capita is below the poverty line). 17 Results of the analysis are qualitatively similar for higher values of this threshold. 19 mozambique poverty assessment | 3. evolution of living conditions and economic mobility Figure 3.6. The prevalence of multiple deprivations has declined but mostly in urban areas (households experiencing three or more monetary and/or non-monetary deprivations) Total Urban Rural 92.8% 7.2% 78.6% 21.4% 99.6% 0.4% 2002/03 85.5% 14.5% 57.2% 42.8% 97.9% 2.1% 2008/09 71.0% 29.0% 32.0% 68.0% 89.4% 10.6% 2014/15 Poor Non-poor Source: World Bank using IOF-2002/03, IOF 2008/09 and IOF 2014/15 The incidence of both monetary and analyzed. For example, 49.4 percent of the non-monetary poverty is strongly monetary poor live in dwellings that lack correlated. access to safe water whereas 27.1 percent of the monetary non-poor are deprived in 29. Compared to households above the this indicator (Figure 3.7). Data from previous poverty line, monetary poor households are surveys reveal that the strong association also remarkably more likely to be deprived between monetary and non-monetary poverty in each of the non-monetary indicators has changed little over time. Figure 3.7. Non-monetary deprivations continue to be larger among the monetary poor (2002/03) (2014/15) Water Water deprived deprived Schooling 2 100.0% Electricity Schooling 2 100.0% Electricity deprived deprived deprived 75.0% deprived 75.0% 50.0% 50.0% 25.0% 25.0% Schooling 1 Sanitation Schooling 1 Sanitation deprived 0.0% deprived 0.0% deprived deprived Quality Quality Asset dwelling Asset dwelling deprived deprived deprived deprived Poor Non-poor Poor Non-poor Note: Schooling deprived 1 = no member in the household completed at least 5 years of education, Schooling deprived 2 = at least one primary-school age children out of school. Source: World Bank using IOF-2002/03 and IOF-2014/15 20 3. evolution of living conditions and economic mobility | mozambique poverty assessment The chronic poor continues to be the 3.8). In other words, as of 2014/15, more than 4 largest welfare group. in 10 individuals are both unable to afford basic food and non-food baskets and are deprived 30. The share of people in chronic poverty has in at least three core, non-monetary measures fallen but continues to be the largest group, of human welfare (education, access to basic signaling a poverty trap problem. In the absence services, housing conditions and ownership of panel data, the overlay of monetary and non- of basic assets). A continued persistence of monetary poverty can be used to categorize these deprivations is expected to keep trapping the population into welfare groups with high these households into a condition of monetary and low risks of staying poor.¹⁸ Results of this poverty. Most of the households likely to be analysis show that the share of Mozambicans in chronically poor are in rural areas (84.9 percent), chronic poverty fell from 53.6 percent to 46.3 particularly in the provinces of Zambezia, Niassa percent between 2002/03 and 2014/15 (Figure and Nampula. Figure 3.8. The chronic poor remains the largest welfare group in the population (welfare groups,2002/03) (welfare groups, 2014/15) 0.6% 5.5% 22.8% Transient poor Better off 6.3% Better off Transient poor 40.3% 53.6% 24.7% 46.3% Not poor but Chronic poor Not poor but Chronic poor deprived deprived Source: World Bank using IOF-2002/03 and IOF 2014/15 18 More specifically, the population can be divided into four groups: 1) the chronic poor, namely those that are poor in multidimensional and monetary sense and thus are less likely to depart the condition of poverty; 2) the not poor but deprived is comprised of households whose consumption is above the poverty line but are multidimensionally poor; 3) the transient poor corresponds to households that are not deprived in any of the non-monetary dimensions despite being consumption poor; and 4) the better off represents households that are not poor by either approach. 21 mozambique poverty assessment | 4. inequality of opportunities 4 Inequality of Opportunities There is an increase in the availability of shows that Mozambique has registered an increase human and economic opportunities in the in the coverage of basic opportunities, but some Mozambican society, but their allocation groups of people have remarkably lower chances across the population is still largely unequal. of capitalizing on these opportunities.20 As noted before, the coverage of indicators capturing human 31. Economic progress has brought more capital, access to basic services and quality housing economic opportunities, but not everyone in opportunities have increased. Yet, a large share of the population can seize them. A context where this coverage would need to be reallocated from poverty reduction has gone in hand with higher the more advantaged to the less advantaged inequality makes it important to measure the groups to achieve equality of opportunity: 3.8 availability of basic opportunities and how equitable percent in education, 15.9 percent in water, 34.6 these opportunities are distributed across the percent in sanitation, 53.3 percent in electricity and population. The Human Opportunity Index (HOI)19 5.0 percent in quality housing (Table 4.1).21 Table 4.1. The distribution of opportunities is highly unequal but is slowly improving (fraction of opportunities unequally distributed across the population) Education Water Sanitation Electricity Quality Housing 2002/2003 8.7% 16.9% 55.3% 67.4% 12.3% 2008/2009 5.1% 16.6% 46.3% 65.2% 12.5% 2014/2015 3.8% 15.9% 34.6% 53.3% 5.0% Source: World Bank using IOF-2002/03, IOF-2008/09 and IOF 2014/15 19 The HOI can shed light on the influence of personal “circumstances” –exogenous variables such as gender, race or place of birth for which individuals have no control or responsibility– on the access that people get to the basic services that are necessary for achieving a fully productive life. The index has two components. The first one measure the average coverage rate of basic services. The second component –the equity of opportunity distribution– measures the gap in access rates for a certain service in a group defined by personal “circumstances” relative to the average access rate for that service for the whole population (Barros et al. 2009). The second component discounts the average coverage rate by the fraction of the opportunities that needs to be reassigned from the better-off groups to the worse-off groups to attain equal opportunity in the population under study. The higher the inequality in the allocation of opportunities, the higher the rate of discount. 20 The five opportunities considered in the index for children ages 5 to 11 are: 1) the child is enrolled in primary education, 2) the household uses either piped water, public tap or mineral/bottle water for human consumption; 3) the dwelling is connected to a sewer system or has access to a sceptic tank and/or improved latrine; 4) the energy for lighting is electricity; and 5) the housing material is adobe, cement and/or brick. The seven children circumstances defined are: 1) location, 2) gender of the child, 3) child’s area of residence (urban or rural), 4) per capita household consumption, 5) years of schooling of the family head, 6), number of siblings and 7) if the child lives in either a single-parent or two-parent household. 21 Inequality of opportunity in sanitation, for instance, implies that 34.6 percent of the total available coverage would have to be reallocated among the six circumstance groups (as shown below, mostly from urban to rural households) to equalize the probability of access across all children. 22 4. inequality of opportunities | mozambique poverty assessment 32. There is large variation in the availability Maputo Province have access to the grid of opportunities across provinces and compared to 13.6 percent in Zambezia across personal circumstances within (Figure 4.1). Furthermore, the opportunities provinces. The increasing coverage of basic available are also more unevenly allocated opportunities masks large regional variation. in the poorest provinces. 67.2 percent of the Provinces where monetary poverty has been coverage in electricity in Zambezia needs to be historically high (such as Zambezia, Nampula reallocated away from the better-off groups to and Niassa) show lower availability of human achieve equal coverage. On the other side of and economic opportunities. The largest gap the spectrum is Maputo Province, where 13.2 of the dimensions analyzed is in electricity. percent of the coverage is unequally assigned As of 2014/15, nearly 8 in 10 households in across personal circumstances (Figure 4.1). Figure 4.1. Human opportunities are more unequally allocated in the poorest provinces (unequal distribution of opportunities across personal circumstances and poverty rates by province, 2014/15) (Electricity) (Water) 80% 80% 70% 70% 60% 60% Poverty rate Poverty rate 50% 50% 40% 40% 30% 30% 20% 20% 10% 10% 0% 0% 0% 20% 40% 60% 80% 0% 5% 10% 15% 20% 25% Unequal allocation of opportunity Unequal allocation of opportunity The chances of Mozambican children (98.4 percent) of the unequal access to water later in life are largely influenced by their is driven by location (59.3 percent), household location and family background. per capita consumption (31.9 percent) and head’s human capital (7.2 percent). These three 33. Location, household consumption and characteristics are also important drivers of parental education drive the inequality of inequitable distribution of opportunities with opportunity. Three personal circumstances relatively high HOIs, such as education (94.8 beyond the control of children are the most percent) or housing (89.8 percent) (Figure salient in explaining the severity of opportunity 4.2). This pattern has remained constant over deprivation: whether the household is located time. Decomposition analysis for the HOI in in an urban center, household’s per capita 2002/03 shows that back then this sub-set of consumption and the school attainment of circumstances was also the largest contributor the household head. For instance, almost all to the overall inequality of opportunity. 23 mozambique poverty assessment | 4. inequality of opportunities Figure 4.2. Location, consumption and parental education drive the inequality of opportunity (contribution of each circumstance to inequality of opportunity, 2014/15) 100 Water Electricity Education 76.3 80 70.5 Sanitation Housing 62.7 59.3 Percentage 60 36.2 31.9 40 31.0 28.1 27.4 17.9 20 13.0 7.2 6.3 5.5 2.0 0 Urban Household Head's school Child gender Number of Both parents consumption attainment siblings Source: World Bank using IOF-2002/03, IOF-2008/09 and IOF 2014/15 24 5. productivity, market development and vulnerability in agriculture | mozambique poverty assessment 5 Productivity, Market Development and Vulnerability in Agriculture Productivity in agriculture is low, partly 94 percent of the poor are primarily engaged driven by low input intensity and weak in agriculture. But productivity in this sector is market orientation. low by global and regional standards (Figure 5.1). The gap between average cereal yields 34. Agriculture, the mainstay of Mozambique’s in Mozambique and global averages is large economy, especially among the poor, is and has been growing by more than 2 percent characterized by low levels of productivity. annually between 2000 and 2009 (World Bank The country is rich in natural endowments 2016c). There are also large productivity gaps well suited for agriculture such as extensive with respect to other sectors in the economy. fertile land, abundant water and favorable Data for 2014 shows that the productivity climate. The agricultural sector accounts for level in agriculture was about one-third of the around 25 percent of the GDP and employs average productivity for the whole economy nearly 75 percent of the labor force. Close to (World Bank, 2017a). Figure 5.1. Average maize yields are lower in Mozambique than in other neighboring countries (maize yields in kilograms per hectare) 6,000 5,000 4,000 3,000 2,000 1,000 0 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 Southern Africa Eastern Africa Mozambique Source: FAOStat 25 mozambique poverty assessment | 5. productivity, market development and vulnerability in agriculture 35. Most of the rural poor are smallholders consumption and only a few (13 percent) engaged in subsistence agriculture with low commercialize their production or grow cash utilization of modern inputs. The median crops. Rates of adoption of productivity- farm size in the Integrated Agriculture Survey enhancing technologies (such as fertilizers or (AIS-2015),22 was 1.27 hectares. Most of the improved seeds) appear to be remarkably low, crops that smallholders grow is for their own as shown in Figure 5.2. Figure 5.2. There is low adoption of modern agricultural inputs among farmers in Mozambique 100% 80% 60% 40% 20% 6.6% 5.7% 4.3% 2.0% 1.0% 0% Pesticides Inorganic Herbicides Irrigation Improved Seeds Fertilizer Source: World Bank using AIS 2015 36. Poverty tends to be higher in provinces than those in the rest of the country (Figure 5.3). where agricultural productivity is lower. Maize In Nampula and Zambezia, for instance, where is the most commonly grown crop and is thus a headcount poverty rates are particularly high, the useful proxy indicator to measure productivity. A average maize yield was 593 kg per hectare. In comparison of maize yields per hectare across the rest of the country, the average yield equals regions in Mozambique shows that farmers in the 951 kg per hectare, almost twice the productivity poorest provinces are less productive, on average, levels of the poorest provinces. Figure 5.3. Poverty rates are higher in provinces with lower maize yields per hectare (all provinces) (excluding Maputo City and Maputo Province) 80% 70% Provincial Poverty Rate Provincial Poverty Rate 70% 60% 60% 50% 50% 40% 30% 40% 20% 30% 10% 0% 20% 400 600 800 1,000 1,200 1,400 1,600 1,800 2,000 2,200 2,400 500 550 600 650 700 750 800 850 900 950 1,000 Average Maize Yields (Kg/Ha) Average Maize Yields (Kg/Ha) Source: World Bank using AIS 2015 and IOF 2014/2015 22 The survey is administered every other year by the Ministry of Agriculture and Food Security with technical support from the University of Michigan. It collects data on socio-demographics, farm size, agricultural production outcomes and market access from 7,485 agricultural households. Like the household budget surveys (IOFs), data from the Agricultural Integrated Survey is representative at the national and provincial levels. 26 5. productivity, market development and vulnerability in agriculture | mozambique poverty assessment 37. Farmers with higher use of modern staple crops. Data from the AIS survey shows technologies and stronger market orientation that the median maize yields per hectare for the are more productive. Adoption of technologies first group is 33 percent higher than the median such as irrigation, fertilizer, and pesticides maize yield for farmers who did not produce among Mozambican farmers is correlated cash crops (Figure 5.4). In addition, smallholder with higher levels of productivity. For instance, farmers also have low access to extension and maize yields per hectare are 27 percent higher credit services. Data from the AIS-2015 shows among farmers that apply fertilizers compared that 6 percent received information from an to those who do not (Figure 5.4). Similarly, agricultural extension program and less than 1 farmers that cultivate cash crops are generally percent obtained agricultural credit. more productive than those that only cultivate Figure 5.4. Modern inputs and market orientation are correlated with higher agricultural productivity (maize yields per hectare) 1 1 Cumulative Frequency Cumulative Frequency .8 .8 .6 .6 .4 .4 .2 .2 0 0 2 4 6 8 10 2 4 6 8 10 Log Maize Yields Log Maize Yields Fertilizer No Fertilizer Cash Crop No Cash Crop Note: Cumulative distribution functions trimmed at 1st and 99th percentiles. Vertical lines show mean values. Source: World Bank using AIS 2015 Several aspects constraint the goods further compound the need for farmers commercialization of agricultural output. to access market smoothing mechanisms. Farmers who sell their surplus harvest tend to 38. Even if farmers increased their agricultural sell immediately after harvest, often creating a productivity, several factors hinder their ability to market glut and pushing down prices. Conversely, participate in market. Low storage and processing during the lean season, few farmers are selling capacity are critical constraints to reduce post- their production, and this leads to higher food harvest loss, strengthen market development and prices. Isolation and transport costs are another increase food security. Data from 2015 shows major barrier to access input and output markets. that 56 percent of the population do not own Nampula and Zambezia, the two provinces with any type of silo and hardly any farmers posses the highest poverty rates, are the provinces where knowledge related to processing of agricultural rural households face longer travel times to reach goods. Seasonal price fluctuations of agricultural markets and other basic services (Figure 5.5). 27 mozambique poverty assessment | 5. productivity, market development and vulnerability in agriculture Agriculture is particularly risky in Figure 5.5. The more isolated a province Mozambique, hindering output and rural is from the nearest market the livelihoods. higher is its poverty rate 39. Climatic shocks exert both direct and 80% indirect effects on agricultural output and 70% rural livelihoods. Agriculture by nature is a Provincial poverty rate 60% risky activity but more so in Mozambique 50% where the incidence of weather shocks is 40% high by regional –and even global– standards. 30% The country is often subject to erratic rainfall, 20% droughts, floods, cyclones, pests, and 10% diseases. In 2015, almost 8 in 10 farmers lost 0% part of their crops, animals or productive 5 15 25 35 45 55 65 75 85 assets due to climatic shocks. Drought is the largest risk, affecting a large share of farmers, Average travel time to nearest market (minutes on foot) with devastating effects on crops. Floods and Source: World Bank using AIS 2015 cyclones are also common, both of which inflict high damages on farm infrastructure and crops. In fact, the relationship between the occurrence of shocks and maize yields Figure 5.6. Maize yields per hectare are lower suggest that droughts and floods are negatively for farmers that experienced associated with crop productivity. Yields droughts and/or floods among farmers that experience droughts Drought Flood and floods are on average 8 and 18 percent 0% lower, respectively, compared to unaffected farmers –even after controlling for differences -4% in observable characteristics. (Figure 5.6) -8% -7.7% -12% -16% -20% -18.1% Notes: productivity measured by maize yields in Kg/Ha. Source: World Bank using AIS 2015 28 6. human capital, labor force and jobs | mozambique poverty assessment 6 Human Capital, Labor Force and Jobs Human capital is increasing but progress is As of 2003, average educational attainment in uneven across areas and population groups. Mozambique was 2.6 years of schooling among the adult population (21 years and older). Starting 40. The availability of basic education services from that low base, Mozambique more than has expanded in hand with increases in school doubled the average educational attainment for attainment. Investments in public education as all adults, to 5.7 years in 2015 (Figure 6.1). a share of GDP increased from 2.7 to 3.1 percent between 2009 and 2014, reaching levels above 41. However, progress in raising human capital the average for several countries in sub-Saharan varies across regions and income groups. Most Africa. School fees for primary education were schools in rural areas cover only the first level of abolished in 2004. The increase in the number primary education (grades 1 to 5) whereas most of of schools and teachers has been met by an the schools offering secondary education (grades increase in the demand for education, partly 8 to 12) are in urban areas. Enrollment rates in driven by demographic trends. Increased supply secondary school have more than doubled and demand in education has resulted in higher relative to the early 2000s, reaching 38 percent, levels of enrollment, attendance in primary benefitting proportionally more the upper half of education, literacy and school attainment. As of the distribution. Overall, children from households 2014/15, over 90 percent of primary-aged children below the poverty line are on average around 30 reported attending primary school. More schools percent less likely to go to school compared to and higher attendance rates have resulted in non-poor households. Rural children have 47.8 increased educational attainment as measured by percent lower probability of attending school than the average years of schooling of the population. urban children. Figure 6.1. Educational attainment in Mozambique is increasing across the board (average years of schooling for adults 21 years and older) 6.9 8 6.3 6.1 Average years of 7 5.7 5.0 5.0 6 4.8 4.8 4.4 schooling 4.1 5 3.8 3.6 3.0 3.1 4 2.6 2.2 2.4 2.2 1.9 3 1.6 1.3 2 1 0 Total Male Female Urban Rural Above Below Poverty Line Poverty Line 2003 2009 2015 Note: Average years for 2003 were calculated based on the highest education level achieved by the respondent (e.g. completed lower primary = 5 years of schooling). Source: IOF 2002/2003, IOF 2008/2009, IOF 2014/2015 29 mozambique poverty assessment | 6. human capital, labor force and jobs Low rates of school completion and quality accumulation. Figure 6.2 shows the fraction of continue to constraint skill development, children 6 to 21 years old in 2014/15 that attend particularly among the poor. that primary or secondary school by grade and whether they belong to a household that is 42. Progress in increasing access to education below or above the poverty line. Dropout rates is overshadowed by underperforming start to pick up at age 14 and are more marked efficiency in the sector and quality constraints. for poor children relative to the noon-poor. Less Late enrollment is widespread. Per the school than half of the Mozambican children that start system, the expected age of a first-grade primary education manage to complete it, over student is six years. However, almost half of 20 percentage points below the rate for sub- the children had never been to school at this Saharan Africa (World Development Indicators). age. Even at age nine, 15 percent of children Mozambique also faces major challenges in were still not enrolled in primary school terms of school quality and student learning. (Demographic and Health Survey, 2011). High As of 2014, student learning outcomes were dropout rates, especially among the poor, dismally low—with only 6.3 percent of third- also reduce the efficiency of the educational grade students mastering the required reading system and create inequalities in human capital abilities (World Bank 2014). Figure 6.2. The risk of dropping out of school is higher for children from poor households (share of children not attending school regularly by age) 100% Total Poor Not-Poor Percentage of children not studying 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 Age of individual Note: Sample includes individuals who have studied previously and are no longer studying, thus measuring actual dropouts and not those who never went to school in the first place. The sample also excludes individuals who have already completed secondary school. Source: World Bank using IOF 2014/2015 Investing in education offers households return masks differences across areas and a path out of poverty. educational levels. The returns to schooling are highest in urban areas and at the secondary and 43. The welfare of households is positively tertiary levels of education, which provides an associated with the school attainment of indication of excess demand for skilled labor. the household head, above all for those An additional year of education increases that reside in urban areas and transition into consumption by more than twice in urban secondary and higher education. Estimates areas relative to the average return estimated of the returns to schooling indicate that on for rural households, 10.6 percent compared average each additional year of education to 5.1 percent (Figure 6.3). Similarly, the increases household consumption per capita highest payoffs are concentrated in secondary by nearly 9 percent. However, the average and tertiary education (11.1 percent and 12.1 30 6. human capital, labor force and jobs | mozambique poverty assessment percent, respectively), around four times larger differentials in returns signal a growing demand than the returns experienced by workers with for skilled labor in Mozambique, particularly in some or completed primary education. These urban areas. Figure 6.3. Skilled workers in urban areas experience the highest returns to schooling 14% 12.1% 12% 10.6% 11.1% 10% 8.8% 8% 6% 5.1% 4% 2.9% 2% 0% Total Urban Rural Primary Secondary Tertiary Regions Education levels Notes: The rates of returns shown were obtained from estimating a model of consumption per capita on years of schooling (as a continuous variable and grouped by education levels) and a set of covariates that includes age and experience (linear and quadratic terms), dummies for area of residence and province fixed effects. Source: World Bank using IOF 2014/2015 A slow process of structural change is 45. Livelihood patterns are largely influenced gradually shifting the job structure away by the sector of work and the ability of the from agriculture and into urban areas. ability of workers to enter the higher paying non-agricultural jobs. In the absence of 44. The modest effects of the structural comprehensive income data, information transformation on employment are highly on household consumption provides an concentrated in the capital city, Maputo, approximation of the earnings of household and its surrounding areas, possibly driving heads across different types of jobs. Figure 6.4 its faster poverty reduction. The share of shows that the median consumption per capita jobs in agriculture has steadily declined from of households whose head is employed primarily 86.6 percent in 1995/96 to 71.5 percent in in non-farm wage jobs in the private and public 2014/15. Much of the transition away from sectors is 68 percent and 143 percent higher than agriculture has taken place within non-farm those working primarily in agriculture. Non-farm self-employment (highly represented in retail self-employment is also correlated with higher trading through household enterprises), living standards compared to employment in whereas private wage-based activities are agriculture. But the opportunities to get higher largely concentrated in the services sector. paying non-agricultural jobs are skewed towards Regional employment patterns show that the urban, male and more educated workers. Not shift into higher quality jobs is concentrated in surprisingly, workers that are poor –even the the capital city. Maputo Province concentrates urban ones– are also highly underrepresented almost 40 percent of private wage jobs in in wage and skilled employment. The clustering the country even though it accounts for 12 of formal enterprises and jobs in Maputo City has percent of the total population. Commercial led to higher labor productivity and opportunities and financial activity is largely clustered in for sustained poverty reduction, but they are Maputo city. largely confined to the capital city. 31 mozambique poverty assessment | 6. human capital, labor force and jobs Figure 6.4. Per capita expenditures are higher in households with jobs outside agriculture (Median household expenditure per capita by sector and type of job of household head, 2014/15) 58.0 60 50 42.0 40.0 40 30 23.8 20 10 0 Agriculture Non-farm Self- Non-farm Wage Non-farm Wage employment Employment Employment (private sector) (public sector) Source: World Bank using IOF 2014/2015 32 references | mozambique poverty assessment References Government of Mozambique. 2016. “Poverty and Well-Being in Mozambique - Fourth National Poverty Assessment (IOF 2014/15)”, Ministry of Economy and Finance. Mozambique Ministry of Agriculture (2011). Plano Estratégico para o Desenvolvimento do Sector Agrário (PEDSA) para 2011-2020. Mozambique Ministry of Economy and Finance (2016). “Pobreza E Bem-Estar Em Moçambique: Quarta Avaliação Nacional. (IOF 2014/15). Mozambique National Institute of Statistics (2015). Statistical Yearbook. World Bank. 2014. Mozambique - Public expenditure review : addressing the challenges of today, seizing the opportunities of tomorrow. Public Expenditure Review (PER). Washington, DC. World Bank. (2016a). Searching for the demographic dividend in Mozambique: an urgent agenda. Washington, D.C.: World Bank Group. https://hubs.worldbank.org/docs/imagebank/ pages/ docprofile.aspx?nodeid=26519678 World Bank. (2016b). “Mozambique – Systematic Country Diagnostic”. Washington, D.C.: World Bank Group. World Bank. (2016c). “Accelerating Poverty Reduction in Mozambique: Opportunities and Challenges”. Policy Report. Washington, D.C. World Bank (2017a). “Let’s Work Mozambique Country Pilot – Jobs Diagnostic Report”. Policy Report. Washington, D.C. World Bank (2017b). “Agriculture productivity, market access and vulnerability in Mozambique: Insights from spatial analysis”. Policy Report. Washington, D.C. World Bank (2017c). “Accelerating Urbanization to Support Structural Transformation in Mozambique – Mozambique Urbanization Review”. Policy Report. Washington, D.C. 33