Information about Sub-Saharan Africa África subsahariana
Chapter

2 Halving the Poverty Rate by 2015

Author(s):
Shanaka Peiris, and Jean Clément
Published Date:
May 2008
Share
  • ShareShare
Information about Sub-Saharan Africa África subsahariana
Show Summary Details
Author(s)

Mozambique has been remarkably successful in reducing poverty over the past decade. Sustained, broad-based growth resulted in a 25 percent decline, equivalent to 4 percentage points a year, in the poverty head-count between 1996 and 2002. Mozambique’s impressive poverty reduction performance was driven not only by the high rate of economic growth but also by the character of this growth. At the macro level, achieving growth without an increase in inequality was key, since this allowed a modest private consumption growth to efficiently reduce poverty, leaving room for a strong increase in investment to generate future growth. At the micro level, broad-based, labor-intensive growth in the nonagricultural sectors diversified household incomes, which underpinned the pro-poor growth performance. This chapter draws lessons on the linkages between growth and poverty reduction and analyzes the major factors associated with poverty reduction in Mozambique during 1996–2002. It uses both macro-level and nationally representative expenditure and income household survey data to illustrate those relationships in recent years and project future performance. The case of Mozambique also offers important lessons for other low-income sub-Saharan African countries, and we summarize these in our conclusions.

The chapter is structured as follows. First, it looks at the poverty and growth record over 1996–2002, paying special attention to relative trends across space and sectors of activity in the economy. Second, it analyzes the micro-level determinants of consumption growth and poverty reduction. Third, given the importance of rural poverty reduction over the period, particular attention is given to the sources of rural income growth, including the continued role of crop income and the increased importance of income diversification. Fourth, using assumptions with respect to sectoral growth and employment structure, we project the incidence of poverty for the years 2007 and 2015.

In spite of the progress, more than 50 percent of Mozambique’s population remains poor. We conclude that although agricultural growth was crucial for poverty reduction, particularly in rural areas, diversification of livelihoods away from agriculture was also important in reducing vulnerability and improving welfare. If economic growth supporting increases in the productivity of labor and other factors is sustained over the next decade, and is coupled with continued expansion of social services to the poorest, Mozambique is likely to make good progress toward achieving the United Nations’ Millennium Development Goals (MDGs), in particular, the goal of halving absolute poverty by 2015. Efforts will need to be made to ensure that growth and poverty reduction continue to be achieved without increased inequality. This will primarily involve achieving continued increases in labor productivity and income for smallholder farmers.

Poverty and Growth Record

Between 1996 and 2002, Mozambique’s growth performance contributed to reducing the number of people in abject poverty by nearly 30 percent, from 69 percent to 54 percent, according to its own national measure of the minimum living standard.1 The reduction was relatively more accentuated in rural areas (Figure 2.1). As expected, growth was efficient at poverty reduction because it raised incomes in the areas and sectors where the poor were concentrated. The 15 percentage point decline in poverty can be disaggregated into component parts, indicating the importance of particular regions or sectors of activity (Table 2.1). Incomes and consumption per capita increased rapidly, leading to an 11.1 percentage point decline in the poverty headcount for households engaged primarily in agriculture—the main activity of the poor—and this accounted for 75 percent of the decrease in poverty. Regionally, poverty fell sharply in the center of the country, which is not the poorest part but is relatively populous.

Figure 2.1.Poverty Rates in 1996 and 2002

(Pecent of total population)

Source: National Expenditure Surveys, 1996 and 2002.

Table 2.1.Decomposition of Changes in Poverty by Location and Sector, 1996–2002
Levels and

changes

(In percent)
Population shares
19962002
Mozambique
Poverty in 199669.4
Poverty in 200254.1
Total change in poverty from

1996 to 2002

(percentage points)
–15.3
Regional decompositionBy region1
Change in poverty in north–3.7North32.332.3
Change in poverty in center–11.7Center41.941.9
Change in poverty in south0.12South25.725.7
Total intraregional component (percentage points)–15.3
Population shift (regional migration)0.0
Interaction component (residual)0.0
Urban-rural (consistent 2002 definition)By area of residence
Change in urban poverty–3.6Urban29.032.0
Change in rural poverty–11.6Rural71.068.0
Total intrasectoral component (percentage points)–15.2
Population shift (urban-rural migration)–0.22
Interaction component (residual)0.11
Aggregate sectors (by head of household)2By sector of employment of head of household
Change in agriculture poverty–11.1Agriculture78.771.3
Change in industry poverty–0.8Industry7.65.3
Change in Service 1 poverty–0.9Service 19.417.5
Change in Service 2 poverty–0.9Service 24.15.8
Total intrasectoral component (percentage points)–13.9
Population shift (sector shift)–1.5
Interaction component (residual)0.1
Source: National Expenditure Surveys, 1996 and 2002.Note: Table entries are changes in the incidence of poverty, in percentage points, that are attributable to the group. A negative number indicates a reduction in poverty.

North includes Niassa, Cabo Delgado, and Nampula; center includes Sofala, Tete, Manica, and Zambezia; and south includes Gaza, Inhambane, Maputo Province, and Maputo City.

Individuals are assigned to the sector where the household head is employed. If the head is not employed, they are assigned to the sector of employment of the other oldest adult. If nobody is employed (fewer than 5 percent of all cases), they are assigned to agriculture. Service 1 includes trade, transport, and services; Service 2 includes health, education, and public administration.

Source: National Expenditure Surveys, 1996 and 2002.Note: Table entries are changes in the incidence of poverty, in percentage points, that are attributable to the group. A negative number indicates a reduction in poverty.

North includes Niassa, Cabo Delgado, and Nampula; center includes Sofala, Tete, Manica, and Zambezia; and south includes Gaza, Inhambane, Maputo Province, and Maputo City.

Individuals are assigned to the sector where the household head is employed. If the head is not employed, they are assigned to the sector of employment of the other oldest adult. If nobody is employed (fewer than 5 percent of all cases), they are assigned to agriculture. Service 1 includes trade, transport, and services; Service 2 includes health, education, and public administration.

Urban poverty fell by less than rural poverty, but since the share of the population in urban areas is low, the total poverty impact of this development was small. Key reasons for the slower pace of urban poverty reduction were inflation and a depreciating currency, which contributed to the increased costs of imported goods and of urban services such as transportation (owing, among other things, to higher fuel costs). Since these items are important in the consumption basket of the urban poor (rice is a staple food in urban areas), real consumption growth was slower in those areas.2 Migration was not a factor in poverty reduction, but the shifts in livelihoods were: the shift of some households out of agriculture as their main source of income and into other sectors reduced poverty by 10 percent.

Households used the increased income to diversify their consumption toward nonfood items (Table 2.2). Since households had more income, meeting basic needs took up proportionally less of it, and the share of food in total expenditures fell in both rural and urban areas in favor of nonfood items. The additional income was increasingly spent to buy consumer durables and to fix up dwellings. By 2002, the majority of households in both rural and urban areas had radios; rural households also acquired bicycles, while in urban areas households bought TVs, clocks, and a few motorbikes. The vast majority of households in 2002 were able to afford housing with durable walls, and in urban areas the majority of residential buildings had durable roofs.

Table 2.2.Nonmonetary Measures of Welfare
AllUrbanRural
199620021996200219962002
Food share686163507166
Durable goods
Radio296543652465
TV59122321
Clock244336602034
Motorbike122311
Bicycle134011231550
Housing
Durable wall318550872285
Durable roof16254056711
Source: National Expenditure Surveys, 1996 and 2002.
Source: National Expenditure Surveys, 1996 and 2002.

During the same period, national inequality (measured by consumption per capita) remained about the same (Table 2.3). In other words, the growth process helped the poor as much as the rich and, as a result, growth was efficient in reducing poverty.3 At about 0.40 on a per adult equivalent basis, and 0.42 on a per capita basis, Mozambique’s Gini coefficient is moderate.4 It is above those of South Asian countries China and Indonesia, and of African countries Tanzania and Malawi, but is one of the lowest in Africa and in the developing world as a whole. In sub-Saharan Africa, the Gini coefficients of Kenya, Uganda, and Zambia are all higher, as are those of most countries in Central America, Central Asia, and the Middle East.5 There was a modest increase in inequality in Mozambique’s urban areas, which helped account for the lower poverty reduction observed there. By contrast, rural inequality was not a factor in poverty reduction, nor was the rural-urban gap—which, in any case, is also quite low in Mozambique compared with other countries in the region—important.

Table 2.3.Measures of Inequality, 1996/7 and 2002/3
Gini coefficientTheil index
Area of residence1996200219962002
Urban0.4340.4630.3730.462
Rural0.3500.3630.2310.256
All0.3810.4030.2860.343
Decomposition of the Theil index in within- and between-group inequality

(groups defined by urban/rural)
19962002
Within-group inequality0.2800.335
Between-group inequality0.0060.008
Percent of within-group inequality97.997.7
Source: National Expenditure Surveys, 1996 and 2002.
Source: National Expenditure Surveys, 1996 and 2002.

Why was Mozambican economic growth pro-poor? One reason was the growth of labor productivity in agriculture. The rate of annual growth of production in agriculture was only about half the growth rate of GDP overall, but the sector shed labor into other sectors so that average labor productivity rose, fueling income growth for households whose main source of income was in this sector (Table 2.4) as well as for households that depended on this sector for subsistence and food security.

Table 2.4.Summary Statistics on GDP, Poverty, and Consumption by Sector, 1996–2002(In percent)
199620021996–2002
Share of GDPAnnual average growth rate
GDP100.0100.09.2
Agriculture134.426.94.8
Industry16.026.118.6
Services (private)40.938.28.0
Services (public)8.78.89.3
Share of labor forceAnnual average growth rate
Labor force100.0100.00.8
Agriculture189.781.7–0.8
Industry3.43.1–0.6
Services (private)5.012.317.1
Services (public)1.82.98.7
Average labor productivityAnnual average growth rate
Total4.06.58.4
Agriculture11.52.15.6
Industry18.254.119.9
Services (private)31.620.2–7.2
Services (public)19.319.80.4
Poverty headcountAnnual change in poverty rate
Household69.254.1–4.0
Agriculture172.658.2–3.6
Industry65.454.0–3.1
Services (private)54.644.4–3.4
Services (public)56.032.9–8.5
Sources: International Monetary Fund; and National Expenditure Surveys, 1996 and 2002.

Agriculture includes domestics.

Sources: International Monetary Fund; and National Expenditure Surveys, 1996 and 2002.

Agriculture includes domestics.

Labor in both rural and urban areas moved into the services sectors, and growth in these sectors was pro-poor as well. Primarily private sector services, such as retail trade, transport, food preparation, financial services, and telecommunications, saw their share of the labor force more than double. Primarily public services also increased as a share of the labor force, but this sector remained small. Incomes grew very rapidly in the public sector, however, so the growth of income brought the poverty rate for households headed by a public sector employee down to only 33 percent. Even though it accounted for only a very small percentage of the labor force, this sector contributed to poverty reduction as well. Average earnings were higher in the service sectors than in agriculture, and earnings growth was strong despite falling productivity.6

The industrial sector suffered net job destruction owing to the closing or downsizing of a number of parastatal manufacturing enterprises. The layoffs, combined with increased private investment, led to a strong growth in value added and major improvement in productivity per worker. The construction sector boomed during the period, contributing to the high rate of economic growth overall in the industrial sector. Construction doubled its share of total employment to 2.0 percent, and average earnings in this sector rose. A number of new businesses started up as well.

A key development behind rising labor productivity was the withdrawal of youth labor from the labor force. This involved primarily youth between the ages of 10 and 18, who mostly either did not enter the labor force or were able to leave the labor force to go to school. The share of those aged 10 years and above who reported being in school as their primary activity increased from 17 percent to 22 percent, and many in this group were girls, who were no longer required exclusively to stay home and help with domestic tasks. The ability of households to support these dependents while they are in school is another indication of the strength of the improvements in household welfare. As noted later, this has led to major improvements in human capital formation in Mozambique and should contribute to stronger economic growth in the future. It has also led to a strong increase in employment in the education sector.

In sum, the overall productivity increase and the structural changes in Mozambique’s economy strongly supported poverty reduction. Most sectors with strong growth in output (for example, construction, trade, and education) had an even stronger growth in employment. The exceptions were in the industrial sector (Figure 2.2). In agriculture, where the majority of the population works, the decline in employment was pro-poor. By the end of the period 1996–2002, agriculture accounted for only about one-fourth of Mozambique’s GDP.

Figure 2.2.Average Annual Growth Rates of Output and Labor Force by Sector, 1996–2002

(In percent)

Source: National Expenditure Surveys, 1996 and 2002.

Looking forward, Mozambique needs to sustain the pattern of labor-intensive growth in the nonagricultural economy, as more shifts in the structure of the labor force out of agriculture will be required in the future to continue the improved poverty-reduction performance. This is because in 2002, 80 percent of the labor force was still working in the agricultural sector despite its low and falling share in GDP. To get a clearer picture of this structural shift and its implications for households, we now turn to the micro foundations of this poverty and growth performance and focus on how households, and their livelihoods, have changed over the decade of poverty reduction.

Behind the Numbers: Micro-Level Determinants of Poverty Reduction

Household Livelihoods and Employment Patterns

Aggregate growth and structural change affect households through changes in a complex set of factors known as livelihoods. Livelihoods, as conventionally defined, comprise “people, their capabilities and their means of living, including, food, income, and assets,” where assets refer to both tangible (such as productive resources and household goods) and intangible assets (such as rights, claims, and access to such resources).7 The livelihood strategies that individuals and households adopt tend to reflect the opportunities and assets (natural, physical, financial, human, and social) available to them; their remuneration (monetary or otherwise) from these activities determines the economic changes in household welfare. Poorer households, with smaller asset bases, tend to have fewer livelihood options. Changes in livelihood strategies represent the response of households to the macro events analyzed previously; livelihood changes in households feed back into sectoral and aggregate economic performance.

Relative monetary poverty or wealth is a condition of the household at a given moment in time and results from the demographic structure of the household, its assets, and its income (including income in kind). In Mozambique, the dominant factors behind poverty reduction were the changes in the structures of both urban and rural livelihoods, which, especially in rural areas, were associated with improvements in welfare. By contrast, during 1996–2002, the demographic structures of households changed very little. The number of dependents did increase by 20 percent on average, however, owing to the withdrawal of household members from the labor force. The increase in dependents was greatest in the poorest households, where the number of dependents per worker (economic dependency ratio)8 rose from about 1.1 to 1.5 in 2002. It is an indication of the strength of the Mozambican recovery that such a large increase in dependents in the poorest households did not result in increasing inequality.

The dominant labor force trends during the period between the two surveys were the growth of private wage employment and of nonagricultural self-employment (Table 2.5). Overall, wage employment grew modestly, as public sector wage employment fell by 3 percent per year, on average. Private wage employment more than made up for this loss, however, growing at 18 percent a year for six years. This trend was most evident in urban areas, which averaged 19 percent growth; but even in rural areas, private wage jobs grew by 10 percent (from a very small base). The contraction in government jobs hit rural areas hardest, since the cashew-processing plants were shut down and the growth in private wage jobs was slower. In urban areas, the share of private wage jobs increased from one-third to two-thirds of total wage employment. Most of the private wage jobs are in the industrial sector (including mining and construction), but a growing share are in the services sector. The growth in wage jobs appears to have contributed to poverty reduction, since, for the most part, workers employed in private wage jobs were overrepresented in upper-income households.9

Table 2.5.Growth Rate of Labor Force (Ages 10 and Higher), by Type of Employment (Rural/Urban), 1996–2002(In percent)
Type of employmentShare of all

workers
Growth

rate

1996/97

2002/03
RuralGrowth

rate

1996/97–

2002/03
UtbanGrowth

rate

1996/97–

2002/03
1996/972002/031996/972002/031996/972002/03
Agriculture (all)188.881.5–0.694.893.1–0.371.353.5–1.9
Self-employed,

nonagriculture
3.88.114.32.03.610.69.019.016.5
Wage employment7.410.46.73.23.30.719.627.58.9
Public4.83.8–3.22.11.3–6.912.89.7–1.7
Private2.66.617.71.12.09.86.817.820.9
All economically active100.0100.00.8100.0100.00.0100.0100.02.9
Source: National Expenditure Surveys, 1996 and 2002.

Agriculture excludes those employed in the public sector.

Source: National Expenditure Surveys, 1996 and 2002.

Agriculture excludes those employed in the public sector.

The growth of nonagricultural self-employment was widespread, since it occurred in both rural and urban areas. Nearly two-thirds of this employment was in urban areas, however. Although the source of demand for the products and services of rural nonfarm self-employment was most likely the growth in agricultural sector incomes, the shift of the labor force into this sector—which grew overall by 16 percent a year—was a major factor in rural poverty reduction, since average incomes in this sector are 50 percent above those in agricultural self-employment. In urban areas, 95 percent of self-employment was in the services sector in traditional informal sector activities such as wholesale and retail trade, transport (taxis, small buses), catering and restaurants, repairs, lodging, and gardening. In rural areas, the sector is more diverse, including manufacturing based on natural resources (for example, charcoal, small wood furniture) or agricultural products (for example, milling, brewing), as well as services similar to those provided in the urban sector.

Women continued to participate actively in the economy. As a result of the trend of increased school enrollment for girls, the educational level of women in the labor force rose (Table 2.6). But still, in 2002, 86 percent of economically active women reported having not attended school or completed EP1 (“Primary Education 1”—the first 5 years of formal education). The distribution of women in the labor force by type of job did not change much between 1996 and 2002. Women’s access to paid jobs was about the same between the two years—the growth rate was higher in 2002, but it had increased from a tiny base. Women have expanded into self-employment, especially in urban areas. Nonetheless, the dominant sector of employment for women remains agriculture, either as self-employed persons or as family workers (Table 2.7). By remaining in agriculture, women facilitate the diversification of sources of income for the household while ensuring food security. This pattern of household income generation may have advantages, but it may also be reducing women’s access to cash income relative to men’s. This may support the maintenance of gender inequities and may have implications for future efforts to shed labor and raise labor productivity in agriculture.

Table 2.6.Educational Distribution of Labor Force, by Gender(In percent)
Highest level of

education
Share of labor forceAverage annual growth

1996–2002
19962002
AllMaleFemaleAllMaleFemaleAllMaleFemale
Never attended or completed EP180.570.888.478.168.085.60.3–0.20.2
EP113.119.08.213.418.49.31.1–0.12.7
EP24.77.32.65.37.63.42.71.15.3
SG11.01.70.41.83.00.811.010.412.2
SG20.40.60.10.91.60.418.917.821.8
Technical0.30.50.10.50.80.16.98.9–1.5
Post-secondary0.00.10.00.10.20.018.716.626.6
Total100.0100.0100.0100.0100.0100.00.80.40.7
Source: National Expenditure Surveys, 1996 and 2002.Notes: EP1 denotes primary education, Grades 1 through 5; EP2 denotes primary education, Grades 6 and 7; SG1 denotes secondary education, First Cycle/Grades 8 and 9; SG2 denotes secondary education, Second Cycle/Grades 10 through 12.
Source: National Expenditure Surveys, 1996 and 2002.Notes: EP1 denotes primary education, Grades 1 through 5; EP2 denotes primary education, Grades 6 and 7; SG1 denotes secondary education, First Cycle/Grades 8 and 9; SG2 denotes secondary education, Second Cycle/Grades 10 through 12.
Table 2.7.Type of Employment, by Gender and Sector, 1997–2003
Type of employmentShare of sector,

by gender

(In percent)
Number of

employed

(Thoukandkk
Percentage

change
Total share of

labor force

(In percent)
19972003199720031997–200319972003
Agriculture, female58.761.24,3444,283–1.452.749.9
Agriculture, male41.338.83,0562,717–11.137.131.7
Self-employed nonagriculture, female31.941.599287189.41.23.3
Self-employed nonagriculture, male68.158.521240591.42.64.7
Public sector wage, female15.420.7506634.00.60.8
Public sector wage, male84.679.3272254–6.73.33.0
Private sector wage, female11.816.82595276.80.31.1
Private sector wage, male88.283.2189472149.72.35.5
All economically active,

female
54.855.14,4924,7375.5
All economically active,

male
45.244.93,7083,8634.2
Total100.0100.0
Source: National Expenditure Surveys, 1996 and 2002.Notes: Agriculture includes domestic and unpaid family workers in agriculture. Self-employed workers in other sectors include unpaid family workers if the industry they work in is different from agriculture.
Source: National Expenditure Surveys, 1996 and 2002.Notes: Agriculture includes domestic and unpaid family workers in agriculture. Self-employed workers in other sectors include unpaid family workers if the industry they work in is different from agriculture.

A Closer Look at Rural Income Structure and Growth

The increase in rural incomes was a major force behind Mozambique’s poverty-reduction performance. Yet rural poverty is still severe and wide-spread. Sustaining rural poverty reduction may prove to be a serious challenge in years to come. Therefore, it is important to develop a better understanding of how rural households generate their incomes and what changes have occurred in the structure of that income over time, which justifies taking a closer look at the changes in production patterns and sources of income growth for households. We use income data from nationally representative rural household income surveys for the same years, 1996 and 2002. Although mirroring the trends described previously, these data provide details on households’ income-earning strategies, showing the structure of rural household incomes and how it has changed over time.

The main source of agricultural growth in Mozambique has been production of basic food crops, but export-oriented cash crop production is also expanding. Between 1997 and 2002, basic food crop production grew at an average rate of about 3–4 percent annually. Maize and millet production showed the greatest increases, followed by sorghum, beans, rice, and cassava. The number of households engaged in production of export crops and the total cash crop production for export have also increased; particularly important are sugarcane (for larger enterprises), cashews, copra, cotton, and tobacco. Within the crop sector, production was considerably more diversified in 2002 than in 1996, as the mean number of crops grown per household increased from fewer than five crops in 1996 to almost eight crops in 2002.

The shift into a more diverse crop mix and into higher-value crops has resulted in improvements in crop incomes for all income groups. Rural household income—defined as the value of crop production and sales, and earnings from nonfarm self-employment and wage income—showed a pro-poor trend similar to that of rural consumption growth (Figure 2.3). Rural households are overwhelmingly smallholder farms. Crop income is the most important income source, especially for poorer households, and growth in income from this source explains a great deal of the increases (nearly 80 percent) in mean income for the poorest groups (Table 2.8).

Figure 2.3.Average Annual Growth Rates of Rural Household Income per Adult Equivalent, by Income per Adult Equivalent Quintiles, 1996–2002

(In percent)

Source: Mozambique, National Rural Income Surveys, 1996 and 2002.

Table 2.8.Structure of Rural Household Income, by Source and Income Quintile, 1996 and 2002(Percentage of income by source, quintile, and year)
Quintiles of net

household

income/adult

equivalents
Survey

years
Crop

income
Livestock

sales
Wage

labor
Nonfarm

enterprise
Total
Quintile 1199692.51.62.83.1100.0
200285.34.42.18.3100.0
Quintile 2199687.52.01.78.8100.0
200284.43.61.810.2100.0
Quintile 3199681.91.41.814.9100.0
200275.83.86.314.1100.0
Quintile 4199678.21.12.418.4100.0
200261.03.914.620.5100.0
Quintile 5199674.81.02.122.0100.0
200238.72.925.932.5100.0
All rural households199683.01.42.113.4100.0
200270.93.89.116.2100.0
Source: Mozambique, National Rural Income Survey (TIA), 2002.
Source: Mozambique, National Rural Income Survey (TIA), 2002.

Although it is not obvious from national consumption data, analysis of rural income data shows that differences between richer and poorer households in rural areas widened between 1996 and 2002. Rural households in all income groups increasingly relied on nonagricultural sources such as livestock, wage labor, and nonfarm enterprise income (Tables 2.9 and 2.10). The probability of a household engaging in wage labor and nonfarm enterprises increased over the period and was directly correlated with income levels in each year. Increased crop income from export crop sources is also correlated with income. Households in areas with better infrastructure were more likely to have noncrop income sources. Regression analysis shows that a shorter distance to a paved road increases the size of a household’s noncrop income as well.

Table 2.9.Sources of Rural Household Income (All and Major), by Quintile, 1996 and 2002(In percent)
Quintiles

of net

household

income/

adult

equivalents
Households earning

income from source
Households earning

highest proportion of

income from source
Survey

year
Crop

income
Livestock

sales
Wage

labor
Nonfarm

MSE
Crop

income
Livestock

sales
Wage

labor
Nonfarm

MSE
Total
Quintile 11996

2002
98.8

100.0
9.1

20.2
15.8

6.5
9.1

27.9
95.0

89.5
0.9

2.8
1.2

1.2
2.9

6.5
100.0

100.0
Quintile 21996

2002
99.6

100.0
14.8

29.5
19.0

7.0
21.8

34.4
93.0

89.4
0.2

1.6
0.7

1.4
6.1

7.6
100.0

100.0
Quintile 31996

2002
99.8

100.0
14.3

32.9
18.6

14.3
34.3

40.6
82.1

79.2
0.8

2.1
2.1

6.4
15.0

12.2
100.0

100.0
Quintile 41996

2002
99.7

100.0
16.1

36.2
27.2

24.4
49.8

51.6
81.3

59.5
0.3

2.6
1.0

16.6
17.4

21.3
100.0

100.0
Quintile 51996

2002
100.0

100.0
15.2

36.1
24.2

36.7
61.8

61.5
76.5

34.3
0.7

1.6
0.6

28.6
22.2

35.5
100.0

100.0
All rural
households1996

2002
99.6

100.0
13.9

30.6
21.0

16.5
35.4

42.0
85.6

72.7
0.6

2.2
1.1

9.7
12.7

15.4
100.0

100.0
Source: Mozambique, National Rural Income Survey (TIA), 2002.Note: MSE denotes medium-size and small enterprises.
Source: Mozambique, National Rural Income Survey (TIA), 2002.Note: MSE denotes medium-size and small enterprises.
Table 2.10.Sources of Growth in Rural Household Income, by Quintile, 1996–2002
Annual

quintiles of net

household

income/adult

equivalents
Annual

growth rate

in income,

1996–2002
Change in mean total income attributable

to each source, 199u-20021
Crop

income
Livestock

sales
Wage

income
Nonfarm

enterprise
Total
Quintile 16.577.25.0–3.020.0100.0
Quintile 22.977.06.05.012.0100.0
Quintile 32.879.27.913.9–1.0100.0
Quintile 44.139.49.138.413.1100.0
Quintile 512.4–8.04.055.049.0100.0
All households8.79.95.046.538.6100.0
Sources: Mozambique, National Rural Income Survey (TIA), 2002; Boughton and others (2006a).

Percent of the total change attributable to the source. Calculated as a share of each source change in the total change.

Sources: Mozambique, National Rural Income Survey (TIA), 2002; Boughton and others (2006a).

Percent of the total change attributable to the source. Calculated as a share of each source change in the total change.

Most of the increase in crop income has been achieved through extensive agricultural practices, such as area expansion. Yet the area cultivated (averaging 1.4 hectares) is still too small for income maximizing. One reason farms remain small is that current agriculture practices—most Mozambican farmers make very little use of technologies such as improved seeds, or of chemical inputs such as fertilizers, pesticides, and herbicides—do not support larger farm sizes as labor shortages emerge. Failure to modernize means that basic food crops widely grown by smallholder farmers, predominantly for subsistence, have exhibited relatively stagnant yields (output per hectare) (Walker and others, 2004; Benfica, 2006; Benfica, Tschirley, and Boughton, 2006; and Boughton and others, 2006a). The conundrum for Mozambican agriculture is that although land area expansion is still important to achieving the needed returns through larger-scale production, the ability of households to expand that area may be limited under the currently available technology. There will have to be gains in productivity at the farm level, including the use of labor-saving technologies, if agricultural growth is to be sustained and to continue contributing to rural poverty reduction.

Smallholders can increase land productivity and crop income through diversification into profitable cash crops, production of many of which is tied to contract farming schemes. The use of productivity-enhancing technologies (particularly fertilizers) in Mozambique (where rural credit and input markets are poorly developed) is generally associated with such crops and schemes.10 If expansion of production takes place with improved technologies, particularly labor-saving ones involving animal traction and irrigation and chemical inputs capable of increasing yields, the effects can be significant for increasing both households’ food security and their ability to market their crops, and invest, off the farm. Furthermore, the contribution of agricultural growth to rural poverty reduction can be maximized if roads and access to markets are improved, resulting in better market integration across subregions. Evidence suggests that access to markets (both domestic and regional cross-border) improved over the period, and prices have tended to converge in domestic markets, indicating increased market integration, but further progress is needed.

Finally, off-farm income, particularly from wage labor and nonfarm self-employment, has played an important role in income growth and poverty reduction. Lack of opportunities in this sector may jeopardize sustainable poverty reduction over the longer run. Therefore, efforts to increase households’ access to those opportunities, in both the agricultural and the nonfarm sectors, are important. Increased wage-earning opportunities would come about through expanding the activity of larger farming households, such as commercial smallholders (predominantly in export cash crops), capable of generating employment. Increasing the availability of nonfarm wage-employment opportunities through public investment (for example, on public works) could generate direct and indirect poverty-reduction effects through increased effective demand in the local economy. Participation in nonfarm businesses is important for all income groups. It will, therefore, be important to promote the development of a more diversified nonfarm business sector in rural areas to achieve sustained rural income growth and poverty reduction. Focusing on business with backward and forward production linkages capable of generating multiplier effects can have a much more substantial and broad-based impact on rural poverty. Past experience shows that continued public investment in rural infrastructure is key.

Overall Determinants of Consumption Levels in Mozambique

After the structural changes described above, what are the key factors that determined the levels of consumption in Mozambican households in 2002/03? To answer this question, we ran a regression, using as the dependent variable the log of household consumption per adult equivalent.11 We used a broad set of independent variables in the analysis, and our results can be interpreted as the household production function for consumption.12 We included some variables that may be partly endogenous, such as household composition, because we still wanted to control for the independent part so that it did not pollute the other coefficients. We also included the gender of the head of household; the presence of any disabled adults or children; and the marital status, education, and sector of activity of the head. We estimated separate regressions for urban and rural areas, since we found the structures were quite different. To control for regional effects, we used dummy variables for districts (which show up as district fixed effects).13

With the exception of those for older household members, most variables on household composition are significantly negative—the more people in the household, the lower the consumption per capita. It is noteworthy that the effect of different demographic groups on household consumption is roughly the same in rural and urban areas (Table 2.11). The only exception is for the number of men between 15 and 59 years old. This suggests that men in rural areas may not bring in as much in terms of household consumption as they take out, but that men in urban areas do. Possibly there are better opportunities for men in urban areas to add to household consumption, corrected for all other variables that may affect opportunities. Having disabled children in the household does not seem to affect household consumption, but the presence of disabled adults has a negative effect on household consumption in rural areas.14 This is as expected, since it adds to the dependency burden of the household. In rural areas, the age of the household head has a negative effect on household consumption. A widowed head (regardless of gender) significantly reduces consumption in urban areas. In rural areas, we find that living in a household with a married female head results in higher consumption. We suspect that these households have a husband who could be a migrant worker sending remittances to his family or that they are actually polygamous households.

Table 2.11.Consumption Regressions with District Fixed Effects, 2002
Dependent variable:

log consumption
UrbanRuralSignificance of

difference
CoefficientSignificanceCoefficientSignificance
Household demographics
Number of children aged 0–5–0.061***–0.045***
Number of children aged 6–9–0.093***–0.076***
Number of children aged 10–14–0.106***–0.108***
Number of men aged 15–59–0.003–0.064******
Number of women aged 15–59–0.021**–0.028***
Number of adults>600.025–0.028
Any disabled adults–0.052–0.100***
Any disabled children0.020–0.052
Age head of household0.006–0.007****
Age head square–0.0000.000****
Head female0.092–0.186
Head’s marital status1,2

Married
–0.027–0.141
Polygamous0.001–0.024
Divorced0.030–0.058
Widowed–0.277**–0.127
Added effect of female head

on marital status
Female and married0.1020.385**
Female and polygamous0.0360.199
Female and divorced–0.2090.057
Female and widowed0.2370.171
Head’s education3
Some education0.129***0.062****
Completed EP10.234***0.131*****
Completed EP20.451***0.298*****
Completed SG10.715***0.695***
Completed SG21.142***0.542******
Head’s employment sector4
Mines0.231***0.174
Manufacturing0.0140.275*****
Construction0.0360.038
Transport0.293***0.660******
Trades0.304***0.296***
Services0.113***0.158***
Education–0.0720.283******
Health0.267***0.341***
Public administration0.156***0.132
Constant9.049***10.174***
District fixed effects5YesYes
Observations4,0014,695
Adjusted R-squared0.3640.374
Source: National Expenditure Survey, 2002.Notes: Three asterisks (***) denote significance at 1 percent; two asterisks (**) denote significance at 5 percent; and one asterisk (*) denotes significance at 10 percent. EP1 denotes primary education, Grades 1 through 5; EP2 denotes primary education, Grades 6 and 7; SG1 denotes secondary education, First Cycle/Grades 8 and 9; SG2 denotes secondary education, Second Cycle/Grades 10 through 12.

Head’s marital status. We included interaction terms with the gender of the household head. The first set of coefficients on marital status represents the total sample effect. The interacted terms represent the marginal effect for female-headed households. If the interaction terms (head female*X) are significantly different from zero, the total effect for female heads is the effect obtained from the first set of coefficients plus the interaction effect.

Base category = single male.

Base category = no education.

Base category = head I agriculture.

In 1996, 128 districts were covered; in 2002, 144 districts.

Source: National Expenditure Survey, 2002.Notes: Three asterisks (***) denote significance at 1 percent; two asterisks (**) denote significance at 5 percent; and one asterisk (*) denotes significance at 10 percent. EP1 denotes primary education, Grades 1 through 5; EP2 denotes primary education, Grades 6 and 7; SG1 denotes secondary education, First Cycle/Grades 8 and 9; SG2 denotes secondary education, Second Cycle/Grades 10 through 12.

Head’s marital status. We included interaction terms with the gender of the household head. The first set of coefficients on marital status represents the total sample effect. The interacted terms represent the marginal effect for female-headed households. If the interaction terms (head female*X) are significantly different from zero, the total effect for female heads is the effect obtained from the first set of coefficients plus the interaction effect.

Base category = single male.

Base category = no education.

Base category = head I agriculture.

In 1996, 128 districts were covered; in 2002, 144 districts.

Education of head of household has the expected positive signs, with rising returns, reflecting the relative scarcity of secondary and post-secondary education (in rural areas the number of post-secondary observations is very low, as is demand for this level of education, which may account for the declining returns). Returns are higher in urban areas for all levels of education. For upper secondary, the return shoots up in urban areas—the growing industrial and service sectors need more skilled labor than is being produced. Controlling for education, having the head work in mining (urban), manufacturing (rural), transport, trade, services, education (rural), health, and public administration (urban) increases household consumption compared with work in agriculture. The premiums are significantly lower in urban areas for manufacturing, transport, and education. Working in education seems to be equal to working in agriculture (controlling for education) in urban areas, which seems to suggest an emerging teacher pay issue in urban areas.

In sum, the conditions appear promising for continued poverty reduction in Mozambique. Structural change in the economy is translating into higher standards of living at the household level. Education, a key determinant of household living standards, is expanding rapidly. Even with disappointing results on completion in poor households, the levels of education of both males and females in the labor force will rise. In rural areas, though, the low completion rates will slow down the progress in poverty reduction, since the returns to acquiring just a few years of lower primary education are very low. Economic activity and employment are expanding in areas that are labor intensive and provide a strong boost to consumption (for example, trade, transport, publicly financed services), since there is a significant increase in effective demand. In addition, some of the expanding activities, particularly manufacturing (food processing), have important forward linkages with agriculture that may help boost growth in that sector. Although this regression does not allow for many inferences to be drawn about prospects in the agricultural sector, analysis by Walker and others (2004) has indicated that increased access to markets and land; productivity-enhancing inputs and technologies; and improved education, which can help sustain and increase labor productivity, will be important for future growth in the agricultural sector.

Prospects for Future Poverty Reduction

Mozambique has made good progress to date in reducing poverty and is likely to achieve the Millennium Development Goal of halving the poverty headcount by 2015. As stated in previous sections, sustained economic growth in various sectors and the diversification of the household economy through labor-force migration away from agriculture have been important in this process. In this section, we undertake simulations under alternative scenarios regarding GDP growth and sector migration to assess the prospects for future poverty reduction. We divide the households into two groups according to sector of employment of the head: agriculture and nonagriculture. We assume that household consumption per capita grows at the same rate as GDP in the sector of employment of the head of household15 and that growth is distribution neutral—that is, there is a constant elasticity of poverty with respect to growth (Fox, Bardasi, and Van den Broeck, 2005).

As presented in Table 2.12, we make assumptions on sector GDP growth and sector migration to describe three alternative scenarios and the respective projections with respect to poverty reduction. Under all scenarios, nonagricultural GDP grows faster than agricultural GDP, but different assumptions are made regarding the actual pace and rate of mobility of households across sectors, particularly from agriculture to nonagriculture. The following scenarios are assumed:

Table 2.12.Assumptions and Poverty Projections, by Sector of Employment
ScenariosAssumption on Sector GDP and employment
Annual sector

GDP frowth fates
Annual sector population

growth rates1
Poverty Projections
Poverty headcount
2002–062007–152002–072008–15200220072015
Scenario 1Optimistic GDP growth with sector diversification
Agriculture7.33.20.80.8584029
Nonagriculture7.55.75.15.1444037
544031
Scenario 2Optimistic GDP growth without sector diversification
Agriculture7.33.22.42.2584540
Nonagriculture7.55.72.42.2443320
544234
Scenario 3Pessimistic GDP growth without sector diversification
Agriculture6.32.22.42.2584848
Nonagriculture6.54.72.42.2443626
544542
Sources: World Bank GDP estimates; National Expenditure Surveys, 1996 and 2002.

Based on the 2002 population; the growth rates assumed here result in the National Institute of Statistics of Mozambique’s projected total population for 2007 and 2015.

Sources: World Bank GDP estimates; National Expenditure Surveys, 1996 and 2002.

Based on the 2002 population; the growth rates assumed here result in the National Institute of Statistics of Mozambique’s projected total population for 2007 and 2015.

  • (1) Optimistic sector GDP growth with sector migration. Under this scenario, we assume that agricultural GDP grows at an average annual rate of 7.3 percent in 2002–06,16 and 3.2 percent in 2007–15, while nonagricultural GDP grows at 7.5 percent and 5.7 percent, respectively, during the same periods. It is assumed that employment in the agricultural sector grows at 0.8, and employment in the nonagricultural sector at roughly 5.1 percent a year, implying that labor migrates into nonfarm employment over the projected period.
  • (2) Optimistic sector GDP growth without sector migration. Under this scenario, we keep optimistic assumptions regarding sector GDP growth. Regarding sector employment over the period, we assume the unlikely outcome that it grows at the same pace in both the agricultural and the nonagricultural sectors (by an annual average of 2.4 percent for 2002–07 and 2.2 percent for 2008–15)—that is, sector migration does not occur. Since diversification was a great deal of the past poverty-reduction story, we want to test the consequences of no diversification on poverty-reduction prospects.
  • (3) Pessimistic GDP growth without sector migration. This is the worst-case scenario overall. It assumes that GDP growth in both sectors is slower (1 percent less than in the other scenarios), with nonagricultural GDP still growing faster than agricultural GDP, and that sector migration does not occur—that is, employment in both sectors grows at the same pace.

The poverty projections for 2007 and 2015 under each of these scenarios are presented in the last columns of Table 2.12 (by sector and aggregate) and in Figure 2.4 (aggregates only).

Figure 2.4.Poverty Headcount Projections Under Alternative Scenarios of Sectoral Structure and GDP Growth, 2007 and 2015

(In percent)

Source: Mozambique, National Expenditure Surveys, 1996 and 2002; and GDP projections.

Overall, the results of the simulations emphasize the importance of sustaining economic growth and creating an enabling environment for the continued transformation of the economy. Sustained economic growth rates in Scenarios 1 and 2 result in significantly lower overall poverty rates—31 and 34 percent, respectively—in 2015.

As described in Scenario 1, if people move out of agriculture, as they have in the past decade, and growth continues, poverty overall, and particularly among those staying in agriculture, will decrease faster than if diversification does not occur. As stated previously, that has, in part, to do with the fact that labor productivity will increase rapidly in the sector. It should be noted, however, that increased productivity of labor (output per worker) has been made possible owing to a simultaneous expansion in the land area, and not to technological innovation, which has resulted in stagnant yields (output per hectare). Continued progress will depend, therefore, on the degree of intensification in the sector in years to come. Likewise, continued expansion in output in the nonfarm sector, which would have important effects on poverty reduction, is more likely to occur and be sustained if labor is more productive.

The conclusion here is that if growth can be sustained and inequality remains at more or less the same level, it should be possible for Mozambique to reach the Millennium Development Goal of halving the poverty rate by the year 2015.

Conclusions

Poverty in Mozambique has declined by both monetary and nonmonetary measures. The national headcount ratio was reduced from 69 percent to 54 percent over 1996–2002. The drop was more substantial in rural areas than in urban areas. Overall, inequality increased only slightly over the period. What elements of Mozambique’s growth and development process were most important? We have identified the following factors.

First, growth was faster in nonagricultural labor-intensive sectors. This growth structure allowed households to become less dependent on low-productivity agriculture, gaining a larger share of income from private wage jobs and nonfarm enterprises. Although most households with at least two earners have agriculture as a primary or secondary activity, agriculture is usually engaged in by women (with 90 percent of women working in it) and less often by men. The absolute size of the agricultural labor force fell, and labor productivity (output per worker) increased, contributing to Mozambique’s strong poverty-reduction performance.

Second, increased employment in higher-productivity activities increased all incomes, not just those at the top. In rural areas, subsistence agriculture still provides more than half of total income, but the rest comes primarily from sales of agricultural products and employment income. In urban areas, the fastest-growing sectors of employment were trade and private services, which were fast-growing sectors of GDP as well. Third, the quality of the labor force increased. Younger labor force participants left the labor force to go to school, and potential labor force participants stayed out of the labor force, and in school, longer. Education is still important for increased consumption, regardless of household size or location.

What are the prospects for continued pro-poor growth and poverty reduction? Rural income growth was crucial to the relatively bigger reduction of poverty in rural areas, and sustainable growth in crop income through productivity increases and crop diversification remains important for pro-poor growth. Smallholders can increase land productivity and crop income through diversification into profitable cash crops; policy and programs should be directed toward this goal. Continued access to land and improved access to markets, both domestic and foreign, and technology will be important in transforming Mozambican agriculture from low-productivity/subsistence to high-productivity/commercial and taking households out of poverty. Improvements in rural infrastructure are a crucial piece of the puzzle.

Second, promotion of self-employment opportunities and the creation of private sector wage jobs are important in both rural and urban areas. Off-farm income, particularly wage labor, played an important role in income and consumption growth, particularly for the highest-income households. Although wage labor jobs may not be available to the majority of households in the near future (the growth of formal sector jobs is still expanding from a low base), income diversification is crucial for rural poverty reduction. Investments in human and physical capital that increase the productivity of the self-employed sector will be critical to future success. Future growth may depend as well on the development of supporting institutions (for example, micro credit, savings cooperatives, and associations).

Our projections suggest that if Mozambique is able to sustain strong economic growth over the next 10 years, the poverty-reduction MDG can be met. Labor-intensive growth in the nonagricultural sector, which allows labor to continue to shift out of agriculture, is important in reducing poverty. Poverty reduction strategies should focus on the following: (1) expanding access to education (along with improving the quality of education); (2) achieving a growth performance balanced between agriculture and nonagricultural sectors; (3) expanding income-earning opportunities for poor households in the nonagricultural sector, allowing them to continue to diversify their sources of income outside agriculture; and (4) improving agricultural productivity through sustained intensification and increased access to markets.

Mozambique remains a very poor country—still one of the poorest in Africa. It offers important lessons on pro-poor growth, primarily for the low-income sub-Saharan African countries. The main lesson is that there is no substitute for a balanced growth strategy. Growth in the nonagricultural sector allows households to diversify income sources, while improvements in average labor productivity in crop agriculture support smallholder incomes. Within this context, four other lessons stand out.

  • First, it is important to create a good investment climate for private investment in smallholder agriculture. This means improving market infrastructure, increasing access to land, and encouraging contract-grower schemes that bring new technology into the sector. Unless smallholders are allowed to diversify and modernize, there is a risk they will be left out and inequality will increase.
  • Second, by encouraging the parallel development of a larger commercial farm sector, wage-labor opportunities arise that provide rural households with a source of cash income that they can invest or use for household needs.
  • Third, supporting the growth of the microbusiness sector (very small family businesses) is critical to pro-poor growth. In Mozambique, this happened by default—the government did not try to interfere in this sector.
  • Fourth, education pays off and is a critical building block for any poverty reduction strategy. But households need to have enough income security to send their children to school. Once Mozambican households had this security, young people began to stay in school longer.
Bibliography

    BenficaRui2006“An Analysis of Income Poverty Effects in Cash Cropping Economies in Rural Mozambique: Blending Econometric and Economy-Wide Models” (unpublished; Ph.D. dissertation; East LansingMichigan: Michigan State University, Department of Agricultural Economics).

    • Search Google Scholar
    • Export Citation

    BenficaRuiDavidTschirley and DuncanBoughton2006“Interlinked Transactions in Cash Cropping Economies: The Determinants of Farmer Participation and Performance in the Zambezi River Valley of Mozambique”paper presented at the 26th International Association of Agricultural Economists Conference, Gold CoastAustraliaAugust12–18.

    • Search Google Scholar
    • Export Citation

    BoughtonDuncanDavidMatherDavidTschirleyThomasWalkerBeneditoCunguara and EllenPayongayong2006a“Changes in Rural Household Income Patterns in Mozambique 1996–2002 and Implications for Agriculture’s Contribution for Poverty Reduction”MINAG/DE/DAP Research Report No. 61E (Maputo: Ministry of AgricultureDirectorate of Economics/Department of Policy Analysis). Available via the Internet: http://www.aec.msu.edu/fs2/mozambique/wps61e.pdf

    • Search Google Scholar
    • Export Citation

    BoughtonDuncanDavidMatherChristopherB. BarrettRuiBenficaDaniloAbdulaDavidTschirley and BeneditoCunguara2006b“Market Participation by Rural Households in a Low-Income Country: An Asset-Based Approach Applied to Mozambique” forthcoming in Faith and Development.Available via the Internet: http://www.aem.cornell.edu/faculty_sites/cbb2/Market%20Participation%20Mozambique%20Dec2006.pdf

    • Search Google Scholar
    • Export Citation

    ChambersR. and G.Conway1991“Sustainable rural livelihoods: practical concepts for the 21st century,”IDS Discussion Paper No. 296 (Brighton: IDS).

    • Search Google Scholar
    • Export Citation

    FoxL.E. Bardasi and K. Van denBroeck2005“Poverty in Mozambique: Unraveling Changes and Determinants,”Africa Region Working Paper Series No. 87 (Washington: World Bank).

    • Search Google Scholar
    • Export Citation

    JamesRobert C. ChanningArndt and KennethR. Simler2005“Has Economic Growth in Mozambique Been Pro-Poor?” (Maputo: Mozambique Ministry of Planning and Finance, Direcção Nacional do Plano e Orçamento; Purdue University; and International Food Policy Research Institute).

    • Search Google Scholar
    • Export Citation

    MaximianoNelsonChanningArndt and KennethSimler2005“Qual foi a Dinâmica dos Determinantes da Pobreza em Moçambique” (Maputo: Moçambique Ministry of Planning and Finance, Direcção Nacional do Plano e Orçamento).

    • Search Google Scholar
    • Export Citation

    Mozambique Ministry of Planning and Finance (MPF) International Food Policy Research Institute (IFPRI) and Purdue University2004“Poverty and Well-Being in Mozambique: The Second National Assessment” (Maputo).

    • Search Google Scholar
    • Export Citation

    WalkerThomasDavidTschirleyJanLowMariaTanqueDuncanBoughtonEllenPayongayong and MichaelWeber2004“Determinants of Rural Income in Mozambique in 2001–2002”MINAG/DE/DAP Research Report No. 57E (Maputo: Mozambique Ministry of Agriculture, Directorate of Economics/Department of Policy Analysis). Available via the Internet: http://www.aec.msu.edu/fs2/mozambique/wps57e.pdf

    • Search Google Scholar
    • Export Citation
1For a description of the methodology, see MPF, IFPRI, and Purdue University (2004).
2See MPF, IFPRI, and Purdue University (2004) for a further discussion of the observed increase in urban poverty.
3For further details, see Fox, Bardasi, and Van den Broeck (2005); and James, Arndt, and Simler (2005).
4The Gini coefficient is defined as a ratio of the areas on the Lorenz Curve.
5Comparison number from the World Bank’s POVCAL database.
6Falling productivity, often viewed negatively, is a mathematical certainty for a sector that experiences a rapid growth in the labor force without a similar growth in physical capital or technological change. In the nonagricultural sectors of a country with incomes as low as Mozambique’s, such an outcome is normal and positive from a poverty-reduction standpoint.
8The economic dependency ratio is the ratio of the number of people not working to the number of people working.
9It is possible that those employed in wage jobs in 2002 were living in high-income households in 1997, so, in the absence of panel data, we cannot be sure about the total effect this had on poverty. If wage jobs had not been created, however, the increase in labor income surely would have been lower.
10Households engaged in the cultivation of those crops tend to either apply them to food crops or follow rotation recommendations that contribute to increased yields in those crops. In addition to that, nongrowers of those crops in growing areas also tend to apply them to their food crops (Benfica, 2006; Benfica, Tschirley, and Boughton, 2006).
11A log specification (1) reduces the effect of outliers on the variables, thus producing a normally distributed variable; and (2) allows the coefficient to be interpreted as the marginal percentage effect of the independent variable on household consumption.
12The analysis of the determinants of poverty in Maximiano, Arndt, and Simler (2005), using the same data but slightly different variables or variable specifications, leads to broadly the same results. Especially with respect to the importance of education, the results strongly confirm each other.
13The survey covered 128 districts in 1996 and 144 districts in 2002.
14In 1996/97, disabled adults had a negative effect on consumption only in urban areas.
15We are grateful to Maria Teresa Benito-Spinetto for the GDP growth rate projections.
16GDP growth rates for 2004, 2005, and 2006 for the nonagricultural sector are estimated/projected to be 8.5 percent, 6.6 percent, and 8.6 percent, respectively; GDP growth rates for the agricultural sector (including fisheries) are 8.0 percent, 6.6 percent, and 6.7 percent, respectively.

    Other Resources Citing This Publication