Accounting for Poverty in Africa: Illustration with Survey Data from Nigeria

No 149 – May 2012

John C. Anyanwu

Editorial Committee Rights and Permissions

All rights reserved. Steve Kayizzi-Mugerwa (Chair) Anyanwu, John C. Verdier-Chouchane, Audrey The text and data in this publication may be Ngaruko, Floribert reproduced as long as the source is cited. Faye, Issa Reproduction for commercial purposes is Shimeles, Abebe Salami, Adeleke forbidden. The Working Paper Series (WPS) is produced by the Development Research Department of the African Development Bank. The WPS Coordinator disseminates the findings of work in progress, Salami, Adeleke preliminary research results, and development experience and lessons, to encourage the exchange of ideas and innovative thinking among researchers, development practitioners, policy makers, and donors. The findings, interpretations, and conclusions expressed in the Bank’s WPS are entirely Copyright © 2012 those of the author(s) and do not necessarily African Development Bank represent the view of the African Development Angle de l’avenue du Ghana et des rues Bank, its Board of Directors, or the countries Pierre de Coubertin et Hédi Nouira they represent. BP 323 -1002 TUNIS Belvédère (Tunisia) Tel: +216 71 333 511 Fax: +216 71 351 933 Working Papers are available online at E-mail: [email protected] http:/www.afdb.org/

Correct citation: Anyanwu, John C, Accounting for Poverty in Africa: Illustration with Survey Data from Nigeria, Working Paper Series N° 149, African Development Bank, Tunis, Tunisia.

AFRICAN DEVELOPMENT BANK GROUP

Accounting for Poverty in Africa: Illustration with Survey Data from Nigeria

John C. Anyanwu1

Working Paper No. 149 May 2012

Office of the Chief Economist

1 John C. Anyanwu is a Lead Research Economist at the Development Research Department, AFDB ([email protected])

Abstract

Apart from presenting the poverty Moslem. The variables that are profile, this paper examines the negatively and significantly correlated correlates of poverty with multivariate with the probability of being poor are: models that predict the probability of age of the household head, quadratic of being poor using data from the Nigerian household size, residence in an urban National Consumer Survey (NCS) of area, post-secondary (tertiary) 2003/2004. The probability of a education attainment, being a Christian, household being poor was examined for and residence in the south south, the nation as a whole, as well as male- southeast, south west, and north east headed and female-headed households zones of the country. Based on the and for urban/rural geographical areas. results, we recommend a number of In particular, the variables that are policy interventions (including a broad positively and significantly correlated poverty reduction framework) necessary with the probability of being poor to reduce poverty in Nigeria and similar nationally are: household size, lack of African countries. education, residence in the North Central zone, being single, and being a

Keywords: Poverty, Africa, Nigeria. JEL Classification: I32, I38

I. Introduction

One of the targets for reducing extreme poverty in Africa involves halving the proportion of people living in absolute poverty from 48 percent in 1990 to 24 percent by 2015. Available data so far indicate that it is only the North African countries of Algeria, Egypt, Libya, Morocco and Tunisia as well as Mauritius that have already met this target. Available data indicate, for example, that in Sub-Saharan Africa, the $1.25 a day poverty rate has shown no sustained decline over the whole period since 1981, starting and ending at roughly 50 percent at 2008 purchasing power parity (PPP) – the highest in the world (Figure 1). Indeed, in absolute terms, the number of poor people nearly doubled from 205 million in 1981 to 386 million in 2008 (Figure 2). If current trend continues, the proportion of people living in extreme poverty in Africa as a whole would be about 39 percent by 2015 – far greater than the targeted 24 percent.

Source: Author, using data from The World Bank (2012): http://siteresources.worldbank.org/INTPOVCALNET/Resources/Global_Poverty_Update_2012_02-29-12.pdf

Nigeria is one of those African countries that will not be able to reach the target poverty MDG target by 2015. Successive governments in Nigeria have initiated measures aimed at poverty (rural and urban) reduction since 1980. These include: the Green Revolution (1980); programs to alleviate the pains of Structural Adjustment Program (SAP) through the Directorate of Food, Road and Rural Infrastructure (DFRRI) and the National Directorate of Employment (NDE) (1986); the People’s Bank of Nigeria (1990) (see Anyanwu and Uwatt, 1993); community banks; the Better Life Program (BLP); Family Support Program (FSP) and Family Economic Advancement Program (FEAP); establishment of National Agricultural Land Development Authority (NALDA) (1993) as well as the Agricultural Development Programs (ADP) and the Strategic Gains Reserves Programs (SGRP). Another key measure was the establishment of the Poverty Alleviation Program (PAP) (2000) which later metamorphosed into the Poverty Eradication Program (PEP) and culminated in the National Poverty Eradication Program (NAPEP) (2001). NAPEP has been organized around four schemes, namely, the Youth Empowerment Scheme (YES), Rural Infrastructure Development (RIDS), Social Welfare Schemes (SOWESS) and the National Resource Development and Conservation Scheme (NRDCS).2 We have also had periodic reviews of salaries/wages and tax rates and allowances as well as pensions for increase the purchasing power of civil and public servants. In addition, there is the Interim Poverty Reduction Strategy Paper (IPRSP) with the aim of building on the gains of PAP and PEP. One of the recent measures that attracted a lot of attention was the National Economic Empowerment and Development Strategy (NEEDS), which was built on the interim PRSP. This medium term strategy (2003-2007) derived from the long-term goals of poverty reduction, wealth creation, employment generation and value re-orientation, being a national coordinated framework of action in close collaboration with the state and local governments and other stakeholders. The main strategies were anchored on: empowering people (Social Charter or Human Development Agenda); promoting private enterprise, and changing the way the government does its work (Reform Government and Institutions). The equivalent of NEEDS at State and Local Government levels were the State Economic Empowerment and Development Strategy (SEEDS) and Local Government Economic Empowerment and Development Strategy (LEEDS). Though some of measures and reforms made some positive impacts, they proved unsustainable while at the same time failed to result in sustainable poverty reduction. A number of factors have been identified to contributing to the failure of these measures to achieve sustainable poverty reduction, including poor targeting mechanisms, lack of focus on the poor, program inconsistency, apparent disconnect between the government and the poor, poor implementation, and corruption (Adogamhe, 2010; Ugoh, and Ukpere, 2009; Ogwumike, 2002). Indeed, these earlier efforts to address poverty failed largely because they were badly implemented and had no particular focus on the poor in terms of design and implementation. Thus, an understanding of the various dimensions and determinants of poverty in Nigeria is a precondition for effective pro-poor development strategies in the country.

This paper, therefore, examines the correlates of poverty with multivariate models that predict the probability of being poor using data from the Nigerian National Consumer Survey (NCS) of 2003/2004. The data covers 36 states and Abuja FCT. It comprises a large sample size of 19,158 usable households. The comparison of this data set to previous ones is summarized in Table 1.

2 See Ugoh and Ukpere (2009) and Obadan (2002) for full detailed discussion. 6

The analysis is useful, first, to verify the relative role of the various factors in determining poverty status, and second, to recommend policy changes to reduce poverty incidence in the country. The probability of a household being poor was examined for the nation as a whole, as well as male-headed and female-headed households and for urban/rural geographical areas.

Table 1: Sample Sizes for NCS Data Sets, 1980 -2004

Year Sample Design Urban Rural Total

Three Stages-towns, EAs, No (%) No (%) 1980 Households 5,582 54.3 4,698 45.7 10,280 1985 Two Stages- EAs, HHs 5,273 56.6 4,044 43.4 9,317 1992 Two Stages- EAs, HHs 3,978 41.0 5,719 59.0 9,697 1996 Two Stages- EAs, HHs 3,037 21.1 11,358 78.9 14,395 2003/ Two Stages- EAs, HHs 4,646 24.2 14,512 75.8 19,158 2004 Source: Federal Office Statistics (1999), National Bureau of Statistics (2005) and NBS Data Files.

Thus the further contents of the paper can therefore be adumbrated as follows. Section II discusses the incidence of poverty in Nigeria using the nationwide survey results (2003/2004) while Section III presents the empirical estimates of the determinants of poverty in Nigeria. In section IV, first, a regression for the entire country was estimated, second, geographic area regressions were estimated (urban and rural), third, the probability of being poor is presented for male-headed and female-headed (gender) households. Section V concludes the paper with policy implications.

II. Nigeria’s Poverty Profile: Trend and Dimension

2.1. Trend in Poverty Incidence

Table 2 shows the national levels and population in poverty from 1980 to 2010 - see also Figure 3. Starting from 28.1 per cent in 1980, national poverty reached 66.9 per cent in 1996 before falling to 54.4 per cent in 2003/2004 – and then reaching a peak in 2010 to 69 per cent. However the population in poverty continues to rise – from 18.3 million in 1980 to 68.7 million in 2003/2004 and 112.5 million in 2010. Poverty incidence is even worse when measured using international poverty line – population below $1.00 in PPP terms in 2010 was 61.2 per cent while those below $1.25 a day by 2003/2004 was 64.41 per cent and 68 per cent in 2010. The population below $2 a day in 2010 was 84 per cent (World Bank, 2011).

Table 2: Trends in Poverty Levels, 1980-2010

Estimated Poverty Year Total Population in Poverty Incidence population

1980 28.1 65 18.26 1985 46.3 75 34.73 1992 42.7 91.5 39.07 1996 66.9 102.3 67.11 2004 54.4 126.3 68.7 2010 69 163 112.5

Source: Author’s estimations from Data Obtained from the Federal Office of Statistics (FOS)/National Bureau of Statistics (NBS) and NBS (2010)

Source: Author’s estimations from Data Obtained from the Federal Office of Statistics (FOS)/National Bureau of Statistics (NBS) and NBS (2010).

Table 3: Relative Poverty Incidence by Sector, 1980-2010

Year Urban Rural 1980 17.2 28.3 1985 37.8 51.4 1992 37.5 46 1996 59.3 71.7 2004 43.2 63.3 2010 61.8 73.2

Source: Author’s estimations from Data Obtained from the Federal Office of Statistics (FOS)/National Bureau of Statistics (NBS) and NBS (2010)

Table 3 presents the relative poverty incidence in urban and rural Nigeria from 1980 to 2010 - see Figure 4 also. Urban poverty was only 17.2 per cent in 1980 but reached a high of 59.3 per cent in 1996 before falling to 43.2 per cent in 2003/2004 – still more than double its 1980 level. It reached a peak of 61.8 per cent in 2010. On the other hand, rural poverty stood at 28.3 per cent in 1980, reaching a high of 71.7 per cent in 1996 before decreasing slightly to 63.3 per cent in 2003/2004, also more than double its 1980 level. As with the urban poverty, rural poverty reached a peak of 73.2 per cent in 2010. The Table also shows that in all the years, rural poverty incidence had dominated urban poverty. Thus, Nigerian poverty is largely a rural phenomenon. Nigerian poverty depth and severity are not only high but rising, until recently (Table 4). More importantly, rural poverty was more widespread, deeper, and more severe than urban poverty throughout the period, 1980-2004.

Table 4: Depth and Severity of Poverty by Sector (%)

1980 1985 1992 1996 2004 Variable Depth Severity Dept Severity Dept Severit Depth Severity Depth Severity h h y

9.0 4.3 16.3 7.8 16.4 8.6 30.4 17.4 21.8 11.9 National 5.2 2.3 12.1 5.4 13.5 6.7 26.3 15.0 16.7 9.2 Urban 9.5 4.6 18.9 9.3 18.3 9.8 33.0 18.9 25.8 14.1 Rural

Sources: Source: Author’s estimations from Federal Office of Statistics (FOS)/National Bureau of Statistics (NBS) Data.

2.2. Some Dimensions of National Poverty in Nigeria

The index of poverty used in this section is headcount index (incidence). Table 5 shows the distribution of headcount poverty by gender of household head, household size, education, age group, occupation groups, zone, state, and religious affiliation and marital status.

2.2.1 Poverty and Gender of Household Head

Though poverty was more pronounced in the female-headed household in 1980 (29.1 per cent as against 26.9 per cent for the male-headed households), this picture changed since 1985 and male headed households have demonstrated higher incidence of poverty till date. The widest gap – 20.0 per cent - was in 2003/2004 when national poverty of male-headed households was 56.5 per cent against 36.5 per cent for female-headed households (Table 5).

2.2.2 Poverty and Household size Nigerian poverty is high for large households. Tables 5 demonstrate that there is correlation between the levels of poverty and the size of the household. While households with one person showed the least incidence of poverty, households with more persons especially those with 20 persons and above showed the highest incidence of poverty. For instance, in 2003/2004, the incidence of national poverty with the least size (i.e. one person) was 12.6 per cent. This is against households with more than 20 persons whose incidence of poverty was estimated at 87.1 per cent in that year.

2.2.3 Poverty and Education of Household Head

Nigerian poverty is high for those with little or no education. Table 5 also shows that the level of education is an important determinant of poverty. For instance, household heads with no education have a higher proportion of poverty than those with at least primary education. For instance, among household heads with no education, their proportion in terms of poverty was 74.1 per cent in 1996 and 68.7 per cent in 2003/2004. For those with post-secondary (tertiary) education, their proportion was 47.8 per cent and 26.3 per cent, respectively during those two years.

2.2.4 Poverty and Age Group of Household Head

For poverty and age group of household head, the figures generally show that levels of poverty increase as we move up the age ladder. However, between 1992 and 2004, after the age group 45 to 54 years, poverty tended to decline marginally (Table 5).

2.2.5 Poverty and Occupation Characteristics of Household Head

Table 5 shows that heads of households’ occupation characteristics varied during the survey years. While those in administrative occupations had the highest incidence of poverty in 1980 (44.7 per cent), in 1985, it was those in production/transportation sector (46.5 per cent). In 1992 agriculture/forestry sector household heads had a modestly high poverty incidence (47.8 per cent). In 1996, the production/transportation workers had the highest poverty incidence of 72.8 per cent while in 2004 it was the agriculture/forestry sector workers with 67.0 per cent.

Table 5: Headcount of Poverty by Sector and Household Head Characteristics (%)

1980 1985 1992 1996 2004

National 27.1 46.3 42.7 66.9 54.4 Location Urban 17.2 37.8 37.5 59.3 43.2 Rural 28.3 51.4 46.0 71.7 63.3 Gender

Male-headed HH 26.9 47.3 43.1 67.7 6.5 Female-headed HH 29.1 38.6 39.9 59.9 36.5

Household Size 1 person 0.2 0.6 2.8 12.0 12.6 2-4 persons 8.8 19.3 19.4 51.5 39.3 5-9 persons 30.0 50.4 45.3 75.0 57.9 10-20 pers. 51.0 71.3 66.1 88.5 73.5 20+ persons 80.9 74.8 93.3 95.0 87.1 Education Level

None 29.6 51.3 46.4 74.1 68.7 Primary 24.8 49.7 43.3 60.5 48.7 Secondary 18.5 40.6 30.3 53.5 44.3 Post Sec. 21.4 26.3 25.7 47.8 26.3 Age group

15- 24 16.2 25.3 28.6 38.9 36.4 25- 34 17.7 33.4 28.5 53.3 49.9 35- 44 26.6 45.9 42.1 65.5 55.6 45- 54 27.2 49.6 45.7 72.1 56.2 55- 64 39.6 55.7 48.2 71.4 55.8 65+ 28.7 49.1 49.4 70.4 51.4 Occupation

Prof. / Tech. 17.3 35.6 35.6 52.9 34.2 Administration 44.7 25.2 22.2 22.6 45.3 Clerical 9.9 29.0 34.3 61.9 39.2 Sales worker 14.9 36.5 33.4 58.5 44.2 Service Ind. 21.1 37.9 38.2 72.8 42.8 Agric./Fores. 31.4 53.4 47.8 72.5 67.0 Prodn. /Trans. 23.1 46.5 40.7 65.6 42.5 Manuf./Proces 12.3 31.6 33.2 50.6 44.2 Students/Appr 15.5 36.8 41.7 62.7 41.9 Others 1.4 40.4 42.7 53.2 49.1 Zone

North- East 35.6 54.9 54.0 66.8 72.2 North- West 37.7 52.1 36.5 68.1 71.2 Central 32.2 50.8 46.0 66.2 67.0 South- East 12.9 30.4 41.0 67.7 26.7 South- West 13.4 38.6 43.1 66.9 43.0 South- South 13.2 45.7 40.8 66.6 35.1

Source: Author’s Computation from National Consumer Survey 1980, 1985, 1992, 1996 and 2004.

2.2.6 Zonal Levels of National Poverty

Table 5 also shows the headcount poverty by zones. It shows that the Northwest had the highest level of poverty for the periods, 1980 and 1996, while the Northeast had the highest in 185, 1992, and 2004. On the other hand, the Southeast had the lowest level of poverty during the periods, 1980, 1985, and 2003-2004 while interestingly, the Northwest had the lowest poverty level in 1996. Though poverty distribution among the zones was roughly equal in percentage terms in 1996, the lowest was in North Central zone. On the basis of the overall general trend over the years, many analysts have thus argued that Nigerian poverty is largely a Northern phenomenon.

2.2.7 Poverty in the States and Federal Capital Territory (FCT)

While poverty incidence increased in nine states between 1996 and 2004, the highest increase was in Jigawa whose headcount index rose from 71 per cent to 95 per cent during the period. Also, as Figure 5 shows, Jigawa State had the highest incidence of poverty in 2003-2004 while Bayelsa at 19.98 per cent had the lowest – close to Anambra State’s 20.11 per cent. A key characteristic of Nigerian poverty by State is that poverty incidence is larger in Northern States than in the Southern States. Indeed, out of 20 States with poverty incidence above 60 per cent, 19 were in the North with only Lagos Sate in the South.

Figure 5: Headcount of Poverty by State and FCT (%), 2003-2004

Source: Author’s Computation from National Consumer Survey 2004.

2.2.8 Poverty and Religious Affiliation and Marital Status of Household Head

Poverty in Nigeria differs by both religious affiliation and marital status. Poverty was highest among Moslem heads of households among the major religions while polygamous household heads experienced the highest level of poverty with respect to civil status. In 2004, poverty incidence of household heads for the different religions was as follows: Christian (40.41 per cent), Moslem (68.81 per cent), Traditional (56.61 per cent), and others (48.17 per cent). For marital status, poverty incidence of household heads was: divorced (43.40 per cent), married (monogamous) (52.22 per cent), married (polygamous) (73.21 per cent), separated (32.10 per cent), widowed (36.98 per cent), never married (21.26 per cent), and others (informally married) (46.30 per cent).

III. Factors Driving Poverty In Nigeria

3.1 Empirical Specification

In this section, we investigate the key factors that have had a positive impact and negative impact on the Nigerian poor with a view to addressing the related question of whether and how poverty can be sustainably reduced as well as distilling the lessons learned for tackling the problem of poverty in the country and perhaps elsewhere in Africa.

The discussions in section 3 above have relied largely on tabulated data, exploring relationships between variables without holding other factors constant. Although many of the relationships in the data seem clear, correlations among key variables potentially could obscure the relationship between poverty and a single factor of interest. Consequently, it is useful to analyze the impact of the relevant variables on poverty holding all other factors constant. This implies the need to separate the effects of correlates.

We approach this problem through the application of multivariate analysis, using a logistic regression in accordance with the basic principles of discrete choice models on the 2004 data set. In order to explore the correlates of poverty with the variables thought to be important in explaining poverty a logistic regression model was estimated, with dependent variable being the dichotomous variable of whether the Nigerian household is poor (1) or not poor (0). The explanatory variables considered important in the analysis of poverty were: personal characteristic (age and its square; gender - household is male- or female-headed), demographic characteristic (household size and its square), educational attainment (no education, primary, secondary, and post-secondary), occupation (professional, administrative, clerical, sales, services, agriculture/farming, production, manufacturing, “others”, and being student/apprentice), geographical residence (zones – northeast, northwest, central, southeast, southwest, and Southsouth), marital status (single and married), and religious affiliation (Christian, Moslem, Traditional, and “Other” religions).

It is argued that poverty increases at old age as the productivity of the individual decreases and the individual has few savings to compensate for this loss of productivity and income. This position is consistent with those of Gang, Sen and Yun (2004), Datt and Jolliffe (1999), and Rodriguez (2002). It is generally argued that women are more prone to poverty due principally to low education and lack of opportunity to own assets such as land. The feminization of poverty – a phenomenon, which is said to exist if poverty is more prevalent among female-headed households than among male-headed households – has been the focus of many studies in recent times. Some of the reasons advanced for this existence of feminized

13 poverty include: the presence of discrimination against women in the labor market, or that women tend to have lower education than men and hence they are paid lower salaries (see Anyanwu, 2010).

The literature is also full of evidence that large households are associated with poverty (World Bank, 1991a, b; 1996; Lanjouw and Ravallion, 1994; Cortes, 1997; Szekely, 1998; Anyanwu, 1997, 1998a, 2005, 2010; and Gang, Sen and Yun, 2004). The absence of well- developed social security systems and low savings in developing countries (especially those in Africa) tends to increase fertility rates, particularly among the poor, in order for the parents to have some economic support from children when parents reach old age. This is one of the rationales for parents to increase the number of children so that they will have high probability of getting support when they are old. Also, as Schultz (1981) had indicated, high infant mortality rates among the poor tends to provoke excess replacement births or births to insure against high infant and child mortality, which will increase household size.

In addition, the literature shows that education increases the stock of human capital, which in turn increases labor productivity and wages. Since labor is by far the most important asset of the poor, increasing the education of the poor will tend to reduce poverty. In fact, there appears to be a vicious cycle of poverty in that low education leads to poverty and poverty leads to low education (see also Bastos et al, 2009). The poor are unable to afford their education, even if it is provided publicly, because of the high opportunity cost that they face. Many times they cannot attend school because they have to work to survive. Indeed, Plamer-Jones and Sen (2003) and Anyanwu (2005, 2010) have found, rural households in India whose main earning member does not have formal education or has attended only up to primary school are more likely to be poor than households whose earning members have attended secondary school and beyond. However, Sadeghi et al (2001) have noted higher levels of education were not seriously needed in rural areas where only a few well-educated people live.

It is hypothesized that occupation has a high correlation with poverty because occupations which require low amounts of capital, either human or physical, will be associated with low earnings and therefore with higher poverty rates. Location of residence also matters. In particular, due to more job opportunities in urban areas, poverty tends to be lower in urban than rural areas.

It has been posited that marriage brings an array of benefits (Waite and Gallagher, 2000): in economic terms, since marriage generally adds a potential earner to the household, it seems obvious that marriage should increase the economic well-being of members of the family, including the children. Married women living in male-headed households have the prospect of enjoying larger family income because these families have a larger number of earning members and especially a larger number of earning male members. A long-term marital relationship may also mean higher permanent income and a larger build-up of consumer durables, factors that could limit the extent of economic hardship experienced in downturns in the economy. In addition, married couples may be more easily able to draw on relatives for help in difficult situations (Lerman, 2002).

Religion affects poverty since it embodies a great deal about a person's general approach and outlook to the world. The first broad reason why religion and poverty are related is because religion affects wealth indirectly through its very strong effect on important processes such as educational attainment, marriage, decisions to have kids, how many kids people have and women's decisions to work or stay home with their kids ((Darnell and Sherkat 1997; Lehrer 1999; Keister, 2007). Religion affects these behaviors and processes, and they, in turn, affect

14 household income, expenses and the amount of money left over to save. The second broad reason is that religion can affect poverty and wealth directly by influencing intergenerational processes (the transfer of both religious ideas and wealth from parents to children) , social relations (contacts made through religious group who can provide information, capital and other resources) and orientations toward work and money (Keister, 2011).

Thus, in the model, the response variable is binary, taking only two values, 1 if the Nigerian household is poor, 0 if not. The probability of being poor depends on a set of variables listed above and denoted as x so that:

Prob(Y  1)  F(' x)

Prob(Y  0)  1 F(' x)...... (4.1)

Using the logistic distribution we have: e 'x Prob(Y  1)  1 e 'x  ( ' x)...... (4.2) where  represents the logistic cumulative distribution function. Then, the probability model is the regression:

E[y / x]  0[1 F(' x)] 1[F(' x)]

 F(' x)...... (4.3)

The results are meant to strengthen and clarify the descriptive analysis as well as point to factors that can lead to the sustainability of poverty reduction in Nigeria. The dependent variable is defined as 1 if average per capita household expenditure is below the poverty line and 0 if it is above the poverty line (see also Mason, 1996; Anyanwu, 1997, 1998a, 2005, 2010; 2011; Anyanwu and Erhijakpor, 2009, 2010; Rodriguez, 2002; Ghazouani and Goaied, 2001; and Gang, Sen and Yun, 2004).

3.2 Empirical Results

Our empirical results for the 2004 national data are summarized in Table 6. Since the logistic model is not linear, the marginal effects of each independent variable on the dependent variable are not constant but are dependent on the values of the independent variables (see Greene, 2003). Thus, to analyze the effects of the independent variables upon the probability of being poor, we looked at the change of odds ratio as the dependent variables change. The odds ratio is defined as the ratio of the probability of being poor divided by the probability of not being poor. This is computed as the exponent of the logit coefficients (e  ). In addition to the odds ratios, the predicted probabilities are also presented. All odd ratios greater than one means that the associated variables are positively correlated with the probability of being poor while odd ratios lower than one means that the associated variables are negatively correlated with the probability of being poor.

3.2.1 Empirical Results for the Overall National Data

This section presents regression results for the probability of being poor for Nigeria as a whole. The model was estimated to determine which variables are relevant to the poverty

15 classification and which variables are not. Table 6 presents the empirical results for the overall national data. The results provide strong support for the descriptive analysis above.

National Poverty and Age It is argued that poverty increases at old age as the productivity of the individual decreases and the individual has few savings to compensate for this loss of productivity and income. However, the relationship between age and poverty may not be linear, as would be expected that incomes/expenditures would be low at relatively young age, increase at middle age and then decrease again. Thus, according to the life-cycle hypothesis, we would expect that poverty is relatively high at young ages, decreases during middle age and then increases again at old age (Datt and Jolliffe, 1999; Rodriguez, 2002; Gang, Sen and Yun, 2004). However, we found no evidence for the life-cycle hypothesis since the quadratic term was insignificant throughout and therefore dropped. But increases in household head’s age significantly reduce poverty in Nigeria. This result is not surprising given the composition of the data where about 85 percent of the sample falls within the ages of 20 to 64 years – the productive years in Nigeria. The cultural practice of extended family system whereby children and other younger family members help their parents and the aged in the family give added fillip to this finding. In addition, these results agree with those of Rhoe et al. (2008) for Kazakhstan in Central Asia.

National Poverty and Household Size The literature is full of evidence that large households are associated with poverty (World Bank, 1991a, b; Lanjouw and Ravallion, 1994; Cortes, 1997; Szekely, 1998; Anyanwu, 1997, 1998a; Gang, Sen and Yun, 2004). The absence of well-developed social security systems and low savings in developing countries (especially those in Africa) tends to increase fertility rates, particularly among the poor, in order for the parents to have some economic support from children when parents reach old age. This is one of the rationales for parents to increase the number of children so that they will have high probability of getting support when they are old. Also, as Schultz (1981) had indicated, high infant mortality rates among the poor tend to provoke excess replacement births or births to insure against high infant and child mortality, which will increase household size. In this study, we find household size is positively and significantly related to poverty. It is also the most important factor determining poverty in Nigeria. In addition, while household size increases national poverty, this is at a decreasing rate, with the non-linear (quadratic) relationship. From our results on Table 6, an increase of one in the size of the household increases the odds of being poor by 6.82 times, given other variables. Also, households that increase by ten persons have a probability of 0.87of being poor, and a probability of 0.13 of not being poor. Our results are consistent with those of Gupta and Dubey (2003), Schoummaker (2004), Aassve et al (2005), Kates and Dasgupta (2007), and Rhoe et al. (2008).

National Poverty and Rural-Urban Location A number of studies, including the World Bank (1990, 2001) and the African Development Bank (2002) have indicated that poverty in Africa (and other developing countries) is higher in rural areas than in urban areas. Some of the reasons advanced for this include that historically government policy has been biased against rural areas; rural areas are heavily dependent on agricultural production, which in Africa is characterized by low labor productivity and hence low incomes; and natural disasters such as flooding and drought tend to affect rural areas more heavily than they affect urban areas.

Table 6: Determinants of Poverty in Nigeria, 2004 Variables Coefficient z-value Odd Ratio Predicted Probability Age Age -0.18 -2.78** 0.84** 0.46** Household size HHSize 1.92 17.59*** 6.82*** 0.87*** HHSize squared -0.18 -4.55*** 0.84*** 0.46*** Gender Male Female 0.06 0.91 1.06 0.59 Location Rural Urban -0.60 -11.67*** 0.55*** 0.35*** Education None 0.45 6.58*** 1.56*** 0.61*** Primary 0.12 1.20 1.13 0.53 Secondary 0.04 0.59 1.04 0.51 Post-Secondary -0.77 -7.51*** 0.46*** 0.32*** Occupation Professional -0.35 -0.93 0.71 0.42 Admin Clerical -0.21 -0.55 0.81 0.45 Sales -0.22 -0.58 0.80 0.44 Services -0.06 -0.16 0.94 0.48 Agriculture 0.14 0.37 1.15 0.54 Production -0.17 -0.44 0.84 0.46 Manufacturing -0.12 -0.31 0.89 0.47 Others 0.05 0.14 1.05 0.51 Student/Apprentice 0.42 1.012 1.53 0.60 Zones Northeast Northwest -0.12 -1.99** 0.89** 0.47** Central Southeast 0.15 2.42** 1.16** 0.54** Southwest -1.66 -21.13*** 0.19*** 0.16*** Southsouth -0.64 -9.24*** 0.53*** 0.35*** Marriage -1.19 -15.82*** 0.30*** 0.23*** Single Married 0.23 3.35*** 1.25*** 0.56*** Religion Christian Moslem -0.19 -1.87* 0.83* 0.45* Traditional 0.25 2.39** 1.29** 0.56** Other -0.17 -0.50 0.84 0.46 Constant -1.65 -4.02*** Pseudo R2 = 0.2441 LR chi2(22) = 6483.64 Prob > chi2 = 0.0000 Log likelihood = -10037.423 N = 19158

*** Significant at 1% level; ** Significant at 5% level; * Significant at 10% level. Source: Author's Estimations from NCS Data of 2004.

Our estimates confirm that Nigeria’s poverty is largely a rural phenomenon – just as the descriptive statistics above show. This is because our results show that there is a statistically significant negative effect of urban dwelling on the probability of being poor. This means that that the probability of being poor increases if the household is located in a rural area. From the odds ratio results in Table 7, residing in urban areas decreases the odds of being poor by 0.55 times more with the probability of not being poor estimated at 0.35. These results agree with those of Rhoe et al. (2008) for Kazakhstan.

National Poverty and Gender The feminization of poverty – a phenomenon, which is said to exist if poverty is more prevalent among female-headed households than among male-headed households – has been the focus of many studies in recent times (see Bastos et al, 2009). Some of the reasons advanced for this existence of feminized poverty include: the presence of discrimination against women in the labor market, or that women tend to have lower education than men and hence they are paid lower salaries. Our results show that gender does not matter in determining national poverty in Nigeria.

National Poverty and Education The literature shows that education increases the stock of human capital, which in turn increases labor productivity and wages. Since labor is by far the most important asset of the poor, increasing the education of the poor will tend to reduce poverty. In fact, there appears to be a vicious cycle of poverty in that low education leads to poverty and poverty leads to low education (see Bastos et al, 2009). The poor are unable to afford their education, even if it is provided publicly, because of the high opportunity cost that they face. Many times they cannot attend school because they have to work to survive. Our results indicate that having no education significantly increased the level of poverty in Nigeria. On the other hand and more pleasantly, holding a post-secondary education in Nigeria significantly reduces poverty. From the odds ratio results in Table 6, not having any formal education increases the odds of being poor by 1.56 times more with the probability of being poor estimated at 0.61. But possessing a post-secondary education decreases the odds of being poor by 0.46 times more with the predicted probability of not being poor estimated at 0.32.

National Poverty and Occupation It is hypothesized that occupation has a high correlation with poverty because occupations which require low amounts of capital, either human or physical, will be associated with low earnings and therefore with higher poverty rates. We find, curiously, that no occupation surveyed significantly affects poverty in Nigeria. As also confirmed by the odds ratios and predicted probabilities, no occupation appears important in reducing poverty.

National Poverty and Zonal Location Our results indicate that zonal location matters in explaining rural poverty in Nigeria. Location in the Northeast, Southeast, Southwest and south south zones of Nigeria has a statistically significant negative effect on the probability of being poor. Contrariwise, the results show that location in the North central zone increases the probability of being poor. From the odds ratio results in Table 6, in 1996, being a resident of the Northeast, Southeast, Southwest and south south reduces the odds of being poor by 0.89, 0.19, 0.53 and 0.30 times more with the probability of being non-poor estimated at 0.47, 0.16, 0.35 and 0.23, respectively. But living in north central (Niger, Kwara, Kogi, Benue, Plateau, and Nassarawa) increases the odds of being poor by 1.16 times more with the predicted probability of being poor estimated at 0.54.

National Poverty and Marital Status Our results indicate that households headed by unmarried (single) ones have a statistically significant positive effect on the probability of being poor. In fact, being single increases the odds of being poor by 1.25 times more with the probability of being poor estimated at 0.56.

National Poverty and Religious Affiliation From the results on Table 6, we can see that households headed by Christians have statistically significant negative effect on the probability of being poor. Being a Christian household head reduces the odds of being poor by 0.83 times more with the probability of being non-poor estimated at 0.45. On the other hand, households headed by Moslems have statistically significant positive effect on the probability of being poor. Being a Moslem household head increases the odds of being poor by 1.29 times more with the probability of being poor estimated at 0.56. Traditional and other religious affiliations did not yield statistical significant relationships.

3.2.2 Empirical Results for Specific Geographical (Urban/Rural) Areas

The determinants (or correlates) of poverty can vary from one segment of the population to the other. For instance, marital status can have a different impact for an urban household compared to a rural household. The logistic regressions for rural and urban areas (separately) included the same independent variables as the national regression (excluding the geography variables). The results are presented in Tables 7 (rural) and 8 (urban).

Most of the significant parameters in the individual geographic areas show the same tendency as each other and as the results at the national level. Since most of the tendencies are largely similar, only notable differences are discussed here.

While age is statistically significant in the urban areas, it is not in the rural areas. Household size increases poverty at a decreasing rate in the rural areas but not so in the urban areas. In addition, the effect of household size and no education status are stronger in rural than urban areas. In rural areas, residence in the north central significantly reduces rural poverty while in the urban areas it significantly increases urban poverty. Also, while households headed by unmarried person significantly increases poverty in the rural areas, marital status is not significantly correlated with poverty status in urban areas.

Table 7: Determinants of Poverty in Nigeria: Rural Households, 2004 Variables Coefficient z-value Odd Ratio Predicted Probability Age Age -0.04 -0.58 0.96 0.49 Household size HHSize 2.09 15.68*** 8.04*** 0.89*** HHSize squared -0.21 -4.40*** 0.81*** 0.45*** Gender Male Female -0.04 -0.52 0.96 0.49 Education None 0.51 6.41*** 1.67*** 0.63*** Primary 0.18 1.58 1.20 0.55 Secondary 0.13 1.47 1.14 0.53 Post-Secondary -0.51 -3.60*** 0.60*** 0.38*** Occupation Professional -0.04 -0.06 0.96 0.49 Admin Clerical 0.12 0.20 1.13 0.53 Sales 0.34 0.54 1.40 0.58 Services 0.29 0.46 1.33 0.57 Agriculture 0.47 0.75 1.59 0.61 Production 0.58 0.90 1.78 0.64 Manufacturing 0.11 0.17 1.12 0.53 Others 0.20 0.31 1.22 0.55 Student/Apprentice 0.92 1.46 2.50 0.71 Zones Northeast Northwest -0.22 -3.25*** 0.80*** 0.44*** Central Southeast -0.19 -2.60** 0.82** 0.45** Southwest -1.90 -20.83*** 0.15*** 0.13*** Southsouth -1.28 -13.51*** 0.29*** 0.22*** Marriage -1.41 -15.95*** 0.24*** 0.19*** Single Married 0.24 3.02** 1.27** 0.56** Religion Christian Moslem 0.0001 0.00 1.00 0.50 Traditional 0.46 1.30 1.59 0.61 Others Constant 0.24 0.66 1.28 0.56 -2.36 -3.20*** Pseudo R2 = 0.2566 LR chi2(22) = 5125.88 Prob > chi2 = 0.0000 Log likelihood = -7425.4412 N = 14512

*** Significant at 1% level; ** Significant at 5% level; * Significant at 10% level. Source: Author's Estimations from NCS Data of 2004

Table 8: Determinants of Poverty in Nigeria: Urban Households, 2004 Variables Coefficient z-value Odd Ratio Predicted Probability Age Age -0.41 -3.09*** 0.66*** 0.40*** Household size HHSize 1.26 6.51*** 3.52*** 0.78*** HHSize squared 0.02 0.27 1.02 0.50 Gender Male Female 0.07 0.55 1.07 0.52 Education None 0.31 2.21** 1.36** 0.58** Primary 0.06 0.29 1.07 0.52 Secondary -0.11 -0.88 0.89 0.47 Post-Secondary -1.08 -6.79*** 0.34*** 0.25*** Occupation Professional -0.44 -0.94 0.65 0.39 Admin Clerical -0.31 -0.66 0.73 0.42 Sales -0.37 -0.81 0.69 0.41 Services -0.15 -0.32 0.86 0.46 Agriculture 0.13 0.28 1.14 0.53 Production -0.43 -0.90 0.65 0.39 Manufacturing -0.23 -0.47 0.80 0.44 Others 0.01 0.02 1.01 0.50 Student/Apprentic 0.03 0.06 1.03 0.51 e Zones Northeast Northwest -0.09 -0.64 0.92 0.48 Central 1.11 8.09*** 3.04*** 0.75*** Southeast -1.31 -6.17*** 0.27*** 0.21*** Southwest 0.16 1.22 1.17 0.54 Southsouth -0.83 -4.64*** 0.44*** 0.31*** Marriage Single Married -0.15 -1.07 0.86 0.46 Religion Christian 0.03 0.03 1.03 0.51 Moslem 0.29 0.24 1.33 0.57 Traditional Others -0.67 -0.53 0.51 0.34 Constant -1.61 -1.24 Pseudo R2 = 0.1768 LR chi2(22) = 1065.17 Prob > chi2 = 0.0000 Log likelihood = -2480.3128 N = 4646

*** Significant at 1% level; ** Significant at 5% level; * Significant at 10% level. Source: Author's Estimations from NCS Data of 2004.

3.2.3 Empirical Results for Gender

The determinants (or correlates) of poverty can also vary by gender. For instance, level of education attainment can have a different impact for a male-headed household compared to a female-headed household. The logistic regressions for male-headed and female-headed 21 households (separately) included the same independent variables as the national regression (excluding the gender variables). The results are presented in Tables 9 (male-headed households) and 10 (female-headed households). Again, most of the significant parameters in the individual gendered households show the same tendency as each other and as the results at the national level. Also, since most of the tendencies are similar in a number of respects, only notable differences are discussed here.

Table 9: Determinants of Poverty in Nigeria: Male-Headed Households, 2004 Variables Coefficient z-value Odd Ratio Predicted Probability Age Age -0.17 -2.50** 0.84** 0.46** Household size HHSize 2.14 15.65*** 8.47*** 0.89*** HHSize squared -0.24 -5.12*** 0.79*** 0.44*** Location Rural Urban -0.57 -10.15*** 0.57*** 0.36**** Education None 0.45 6.50*** 1.57*** 0.61*** Primary 0.10 0.96 1.11 0.53 Secondary 0.05 0.63 1.05 0.51 Post-Secondary -0.72 -6.80*** 0.49*** 0.33*** Occupation Professional -0.64 -1.64* 0.53* 0.35* Admin Clerical -0.46 -1.16 0.63 0.39 Sales -0.61 -1.54 0.55 0.35 Services -0.30 -0.74 0.74 0.43 Agriculture -0.03 -0.09 0.97 0.49 Production -0.47 -1.15 0.63 0.39 Manufacturing -0.37 -0.91 0.69 0.41 Others -0.22 -0.54 0.81 0.45 Student/Apprentice 0.02 0.05 1.02 0.50 Zones Northeast 1.49 18.10*** 4.42*** 0.82*** Northwest 1.61 18.85*** 4.99*** 0.83*** Central 1.64 21.73*** 5.16*** 0.84*** Southeast Southwest 1.02 12.60*** 2.77*** 0.73*** Southsouth 0.48 6.35*** 1.61*** 0.62*** Marriage Single Married -0.46 -5.49*** 0.63*** 0.39*** Religion Christian -0.01 -0.03 0.99 0.50 Moslem 0.44 1.34 1.56 0.61 Traditional Other 0.11 0.33 1.12 0.53 Constant -2.98 -5.58*** 2 Pseudo R = 0.2351 LR chi2(22) = 5316.50 Prob > chi2 = 0.0000 Log likelihood = -8649.4005 N = 16370 *** Significant at 1% level; ** Significant at 5% level; * Significant at 10% level. Source: Author's Estimations from NCS Data of 2004.

While age is statistically significant in male-headed households, it is not in female-headed ones. Urban dwelling significantly reduces poverty in male-headed households while rural residence is strongly and positively correlated with poverty in female-headed households. Not going to school at all significantly increases poverty in male-headed households while it is 22 not significantly correlated to poverty in female-headed households. While residence in the various zones is highly correlated to increased poverty in male-headed households, it is only in north central that this is true in female-headed households since residence in the other zones strongly and significantly reduces poverty in female-headed household. Being married is related to lower levels of poverty in male-headed households, while marital status is not significantly correlated with poverty status in female-headed households. Also, while being inn a Christian-headed household is related to lower levels of poverty in female-headed households, religious affiliation is not significantly correlated with poverty status in male- headed households.

Table 10: Determinants of Poverty in Nigeria: Female-Headed Households, 2004 Variables Coefficient z-value Odd Ratio Predicted Probability Age Age -0.29 -1.51 0.75 0.43 Household size HHSize 1.97 8.06*** 7.19*** 0.88*** HHSize squared -0.19 -1.65* 0.83* 0.45* Location Rural 0.59 4.41*** 1.80*** 0.64*** Urban Education None 0.46 1.41 1.58 0.61 Primary 0.24 0.67 1.28 0.56 Secondary 0.01 0.03 1.01 0.50 Post-Secondary -1.04 -2.48** 0.35** 0.26** Occupation Professional -1.25 -0.80 0.29 0.22 Admin Clerical -1.68 -1.06 0.19 0.16 Sales -1.63 -1.05 0.20 0.17 Services -1.77 -1.13 0.17 0.15 Agriculture -1.79 -1.16 0.17 0.15 Production -1.83 -1.17 0.16 0.14 Manufacturing Others -0.71 -0.43 0.49 0.33 Student/Apprentice -0.83 -0.54 0.44 0.31 Zones Northeast Northwest -0.25 -0.75 0.78 0.44 Central 1.21 5.17*** 3.35*** 0.77*** Southeast -1.49 -6.43*** 0.23*** 0.19*** Southwest -0.69 -3.02** 0.50** 0.33** Southsouth -1.13 -4.90*** 0.32*** 0.24*** Marriage Single -0.06 -0.50 0.94 0.48 Married Religion Christian -0.44 -1.84* 0.64* 0.39* Moslem 0.01 0.03 1.01 0.50 Traditional Constant -0.11 -0.07 2 Pseudo R = 0.2357 LR chi2(22) = 809.12 Prob > chi2 = 0.0000 Log likelihood = -1311.7151 N = 2784

*** Significant at 1% level; ** Significant at 5% level; * Significant at 10% level. Source: Author's Estimations from NCS Data of 2004.

IV. The Policy Proposals For The Reduction Of Poverty In Nigeria

This paper has examined the correlates of poverty with multivariate models that predict the probability of being poor using data from the Nigerian National Consumer Survey (NCS) of 2003/2004. The probability of a household being poor was examined for the nation as a whole, as well as male-headed and female-headed households and for urban/rural geographical areas. Our results and analyses above suggest that policy interventions are necessary to reduce poverty in Nigeria and other African countries on a sustainable basis.

First, we found that increases in household head’s age significantly reduce poverty in Nigeria. This result gives support to the continued and intensification of the “solidarity” (a form of “social security”) with the Nigerian family system whereby children and other younger family members help their parents and the aged in the family financially and materially till death comes.

Second, given that poverty increases with the number of household members (or family size), there is urgent need to intensify family planning services efforts and activities in Nigeria so as to improve knowledge, acceptance and practice (KAP) of family planning. This will involve not only increased financial outlay but also research on fertility determinants as well as decentralized planning, delivery and supervision of family planning services (Anyanwu et al, 1998b, c).

Third, in the educational sector, there is very urgent need to re-orientate the thinking and value system of both parents and their children through mass educational campaign regarding the importance of education and the need for parents to insist on their children (male and female) going to school (at least up to first degree) before seeking employment or going into business. In addition, apart from quantitative expansion (including through private participation and public- private partnership), there is urgent need for a fundamental reform of content (e.g. curriculum reforms, availability of school books equipment/facilities, and other teaching materials) towards more emphasis on skill acquisition and problems faced by the poor. It will also be necessary to devise means to assist poor households with school fees, textbooks and other school materials for their children. Non-formal education programs should also be expanded to help the poor gain literacy and most importantly, to acquire skills. These will have to be complemented with increased employment opportunities through public works and infrastructural development so as to encourage children to go to school and hence have greater assurance of finding jobs on graduation.

Evidence has shown that conditional cash transfers and expenditures (for education, for example given our results that education is important in reducing poverty) are effective safety nets (Figures 6 and 7) and levers of poverty reduction and redistribution (see Levy, 2006; Kanbur, 2008; Anyanwu and Erhijakpor, 2010). Using community-based approaches, some important development successes have been achieved under conditional cash transfers, including those that dealt with nutrition in Tamil Nadu, total sanitation in parts of Bangladesh and Indonesia, oral re-hydration in Bangladesh and Egypt, and the reduction of the burden of several neglected tropical diseases in sub-Saharan Africa. Successes occur when conditional cash transfers achieve the best outcome, at the lowest cost and in a sustainable manner (Skolnik, 2011). Indeed, recently, Rosenberg (2011a, b) had extensively discussed success stories in in using cash transfers to reduce poverty in Brazil and Mexico. Improving access to education will reduce poverty both by increasing individual productivity and by facilitating

24 the movement of poor people from low-paying jobs in agriculture to higher-paying jobs in industry and services. More importantly, public spending on education, when targeted toward the poor, can produce a double dividend, reducing poverty in the short run and increasing the chances for poor children to access formal jobs and thus break free from the intergenerational poverty trap. Increasing educational levels (and its quality) should be accompanied by a strong investment climate to ensure that productive jobs are created for the newly educated. Another recent successful example has come from Africa: Miller (2011) has shown that cash transfers in Malawi benefited both the recipients, non-recipients and local businesses given that the transfers strengthened local markets by providing a steady source of customers and cash. Figure 6: Social Safety Nets and Poverty Reduction

Source: Ouerghi (2008)

Fourth, given that occupations overall are not reducing poverty, there is therefore urgent need to encourage productivity and access in both farm and non-farm occupations. This can be done through direct input supply, strengthening and expanding of agricultural research and extension services, adapting agricultural technology and extension services to poor farmers, and by improving physical infrastructure such as rural roads and irrigation. At the same time income sources diversification and labor market training should be encouraged.

Fifth, government should design socio-economic policies to promote long-term employment. Government can assist rural dwellers through increased and broadened National Agricultural and Rural Development Bank’s, Community Banks' and Employment Creation Fund's financial assistance for micro- and small-scale enterprises, complemented by school curricula orientation towards skill acquisition, among other measures. 25

Sixth, since poverty in Nigeria does have important spatial implications, geographic targeting (especially in the Northwest and rural areas) can play an important role in government anti- poverty efforts. Moreover, geographically targeted programs are attractive partly because they are more cost-effective than untargeted programs. Thus, making financial capital, physical infrastructure (especially roads and electricity) and technological innovation available in poor zonal and rural areas will lead to important contribution to government's efforts to reduce poverty in Nigeria.

Figure 7: Conditional Cash Transfer – A Good Example of Appropriate Social Safety Net for Nigeria

Conditional Transfers

Targeted transfers conditional on school attendance or preventative health care Intended Beneficiaries Targeting Methods

• Poor and vulnerable families • Means or proxy means and/or with low level of human capital • Categorical Geographic and/or • Community (together with one of above)

Key Design Features

• Same as cash • Efficient way to verify compliance

Advantages Disadvantages

• Supports income of the poor • Effectiveness influenced by • Can improve school existing education/health attendance and/or health care infrastructure use • Administratively demanding – needs sophisticated targeting by monitoring of compliance

Appropriate Context

• Health and education services are available • Poor are not making use of them • Administrative constraint not too big

Source: Adapted from Ouerghi (2008)

Seventh, the solution to poverty in Nigeria is not less government but more. This requires not only the political will to execute its own policies but also to empower the poor themselves to initiate, design, execute and manage their own priorities. This multi-dimensional empowerment involves political empowerment (through public administration institutions, village and neighborhood councils, participation in democratic processes, and hence with a voice and right to vote), economic empowerment (through easy access to economic resources and institutions: provision of basic assets-equity-enhancing land reform measures, micro-credit, physical

26 infrastructure, extension services, etc), and social empowerment (e.g. provision of secondary basic needs, especially education and health; and involvement of the poor in non-governmental organizations (NGOs), private voluntary organizations (PVOs), and other community-based and grassroots institutions).

In conclusion, sustainable reduction of poverty in Nigeria requires an effective broad policy framework that will increase opportunities, enhance capabilities, promote security, and engender empowerment and participation (Figure 8) (see also Bamberger, Blackden, Fort, and Manoukian, 2002). Increasing opportunities will involve helping the poor earn their way out of poverty by improving their economic conditions. Specific measures to achieve this include promoting micro-enterprise development; generating income-earning opportunities; improving access to geographically inaccessible areas, especially the Niger Delta areas; proper and timely maintenance of damaged roads, including rural ones; addressing challenges in vocational training and skills acquisition; increased investment in rural infrastructure; improving access to labor markets, including through on-the-job training; improving access to productive assets, including credit and land; and engaging in land reforms so as to improve land tenure and land access for the landless poor.

Enhancing capabilities involves improving human conditions and quality of life so as to meet the Millennium Development Goals (MDGs), the key of which is extreme poverty reduction by 2015. Specific actions required to achieve this will include improving access to education, particularly in rural areas and educationally-disadvantaged parts of the country (especially in the Northern parts of the country) coupled with early childhood development; improving access to health and nutrition; improving access to water and sanitation; improving energy generation and its access; and taking measures to effectively implement the national population policy so as to check rapid population growth. Other related measures are creating social investment funds; further tax reforms for greater equity and progressiveness; and increasing redistributive social spending using the recent huge oil revenue earnings following the latest boom in oil prices and resources that will accrue to the recently approved Sovereign Wealth Fund (SWF).

The promotion of economic and physical security involves the minimization of vulnerability as well as social, economic and environmental risks. Measures to attain this include instituting safeguards against socio-economic shocks and natural disasters; putting in place basic social safety nets; setting up social protection programs; constructing and distributing at affordable costs low-income housing; establishing effective HIV/AIDS prevention and treatment measures; reforming, retraining and adequately arming the Nigeria Police Force to fight rising insecurity especially from kidnapping and incessant bombings from the Boko Haram Sect in the North; and checking environmental degradation, especially desertification in the North and soil erosion in most parts of the South-East.

Finally, engendering empowerment and participation involves inclusive decision-making, gender equality and ensuring unfettered political participation. To achieve this, effective measures will involve strengthening the independence of the justice system and increased accountability; making anti-corruption bodies more independent to promote fairness, justice and equity; making the legal systems more accessible through mobile courts, legal aid, and working with customary institutions; the promotion of the rule of law at all levels of government; building social capital and strengthening the organizational capacity of NGOs and community development organizations; and initiating and implementing “gender budget initiatives” as practiced in South Africa and Tanzania. Other measures include enhancing

27 political empowerment through strengthening of public administration institutions, village and neighborhood councils, promotion of participation in democratic processes and thus the voice and right to vote; and introducing institutional reforms to make government institutions more accountable and responsible, strengthening civil society organizations, and encouraging communities (especially rural women) to participate in public policy discourse and dialogue.

Figure 8: Broad Policies for Reducing Poverty in Nigeria

Policies Required Interventions Outcomes

 Promoting economic growth and sound macroeconomic policies Increase  Equal access to labor markets Improved economic Opportunities  Access to productive resources (land, credit, etc…) conditions  Reducing women’s travel and time burdens  Increase private and public investment, especially in infrastructure

 Access to education Enhanced  Access to health and nutrition Enhance human capital  Access to water, sanitation and energy for women and men Capabilities and quality of  Skill enhancement and vocational training life  Conducive business environment

 Helping poor women and men manage risk Greater  Managing economic crises and natural disasters economic and Promote Security  Protection from civil and domestic violence (e.g. from Boko physical Haram Islamic Sect) security  Setting up of social investment funds and safety nets

Increased Engender  Making state institutions more responsive to poor women and men political Empowerment  Removing barriers to political participation for women and men participation and Participation  Sound governance and accountability and gender equality

References

Adogamhe, Paul G. (2010) "Economic Policy Reform & Poverty Alleviation: A Critique of Nigeria’s Strategic Plan for Poverty Reduction," Poverty & Public Policy: Vol. 2: Iss. 4, Article 4.

Aassve, A. et al (2005), Poverty and Fertility in Less Developing Countries: A comparative Analysis, ISER Working Paper 2005-13.

African Development Bank (2002), African Development Report 2002: Rural Development for Poverty Reduction in Africa, Oxford University Press, Oxford.

Anyanwu, J. C. and Uwatt, U. B. (1993), "Banking for the Poor: The Case of the People's Bank of Nigeria", African Review of Money, Finance and Banking, No.1, 87-103.

Anyanwu, J. C. (1997), "Poverty In Nigeria: Concepts, Measurement and Determinants", in Nigerian Economic Society (NES), Poverty Alleviation In Nigeria, Proceedings of the 38th Annual Conference, NES, Ibadan, 93 – 120.

Anyanwu, J. C. (1998a), "Poverty of Nigerian Rural Women: Incidence, Determinants and Policy Implications", Journal of Rural Development, Vol.17, No.4, 651 - 667.

Anyanwu, J. C. et al (1998b), The Role of Men in Family Planning in Nigeria: A Case Study of Edo State, NISER, Ibadan.

Anyanwu, J. C. et al (1998c), The Role of Men in Family Planning in Nigeria: A Case Study of Cross River State, NISER, Ibadan.

Anyanwu, J. C. (2005), “Rural Poverty in Nigeria: Profile, Determinants and Exit Paths”, African Development Review, Vol. 17, Issue 3, December, 435-460.

Anyanwu, J. C. (2010), “Poverty in Nigeria: A Gendered Analysis”, African Statistical Journal, Vol. 11, November, 38-61.

Anyanwu, J. C (2011), “Towards reducing poverty in Nigeria: The case of Igboland”, Journal of Economics and International Finance, Vol. 3, No.9, September, 513–528.

Anyanwu, J. C. and Erhijakpor, A. E. O. (2009), “The Impact of Road Infrastructure on Poverty Reduction in Africa”, in Thomas W. Beasley (Ed), Poverty in Africa, Nova Science Publishers, Inc., New York, 1-40.

Anyanwu, J. C. and Erhijakpor, A. E. O. (2010), “Do International Remittances Affect Poverty in Africa?, African Development Review, Vol. 22, No. 1, March, 51-91.

Bamberger, M., Blackden, M., Fort, L. and Manoukian, V. (2002), “Gender” in Klugman, Jeni (ed.) (2002). A Sourcebook for Poverty Reduction Strategies, Washington DC: The World Bank.

Bastos, A. et al (2009), “Women and Poverty: A Gender-Sensitive Approach”, Journal of Socio-Economics, Vol. 38, Issue 5, October, 764-778.

Chen, S. & Ravallion, M. (2008), The Developing World Is Poorer Than We Thought, But No Less Successful in the Fight against Poverty, World Bank Policy Research Working Paper 4703, Washington, D. C., August.

Cortes, F. (1997), “Determinants of Poverty in Hogares, Mexico, 1992”, Revista Mexicana de Sociologia, Vol. 59, no. 2, April-June, 131-160.

Darnell, A. and Sherkat, D. E. (1997), "The Impact of Protestant Fundamentalism on Educational Attainment", American Sociological Review, 62:306-315.

Datt, G. and Jolliffe, D. (1999), Determinants of Poverty in Egypt: 1997, FCND Discussion Paper No. 75, October.

Federal Office of Statistics (1999), Poverty Profile for Nigeria, 1980-1996, Lagos, February, 1999.

Fields, G.S. (1997), Poverty, Inequality, and Economic Well-Being: African Economic Growth in Comparative Perspective. Paper presented at the AERC Poverty Training Workshop, Kampala.

Foster, J., J. Greer, and Thorbecke, E. (1984), A Class of Decomposable Poverty Measures, Econometrica, 52(3), pp. 761-776.

Gang, I. N., Sen, K., and Yun, M-S (2004), Caste, Ethnicity and Poverty in Rural India. (See: www.wm.edu/economics/seminar/papers/gang.pdf)

Ghazouani, S. and Goaied, M (2001), The Determinants of Urban and Rural Poverty in Tunisia, June.(See: www.erf.org.eg/html/goaied_Ghazouani.pdf)

Greene, W. H (2003), Econometric Analysis, 5th Edition, Prentice Hall, New York.

Gupta, N. D. and Dubey, A. (2003). Poverty and Fertility - An Instrumental Variables Analysis on Indian Micro Data, Scandinavian Working Papers in Economics. No 03-11: November, 2003.

Kanbur, R. (2008), ‘Poverty and Distribution: Twenty Years Ago and Now’, paper presented at the 3rd African Economic Conference, African Development Bank, Tunis, November.

Kates, R. W. and Dasgupta, P (2007), African Poverty: A Grand Challenge for Sustainability Science, PNAS, 104(43):16747-16750.

Keister, L. A. (2007), Conservative Protestants and Wealth: How Religion Perpetuates Asset Poverty, Duke University, Department of Sociology, September 2007. http://www.soc.duke.edu/~lkeister/conservativechristians.pdf

Keister, L. A. (2011), Faith and Money: How Religion Contributes to Wealth and Poverty, Cambridge University Press, Cambridge.

Lanjouw, P. and Ravallion, M. (1994), Poverty and Household Size, Policy Research Working Paper 1332, World Bank, Washington, D. C.

Lehrer, E. L. (1999), "Religion as a Determinant of Educational Attainment: An Economic Perspective", Social Science Research, 28:358-379.

Lerman, R. I. (2002), Impacts of Marital Status and Parental Presence on the Material Hardship of Families with Children, Paper prepared for the U. S. Department of Health and Human Services' Office of the Assistant Secretary for Planning and Evaluation under HHS Grant Number 00ASPE359A, July. http://www.urban.org/UploadedPDF/410538_MaterialHardship.pdf

Levy, S. (2006), Progress against Poverty: Sustaining Mexico’s Progresa-Opportunidades Program, Brookings Institution Press, Washington DC.

Mason, A. D (1996), "Targeting the Poor in Rural Java", IDS Bulletin, Vol. 27, No. 1, January, 67-82.

Miller, Candace M. (2011) "Cash Transfers and Economic Growth: A Mixed Methods Analysis of Transfer Recipients and Business Owners in Malawi," Poverty & Public Policy: Vol. 3: Iss. 3, Article 3.

National Bureau of Statistics (NBS) (2005), “Poverty Profile for Nigeria,” National Bureau of Statistics, Lagos. Available on line: http://www.nigerianstat.gov.ng/Connections/poverty/POVPreliminary.pdf

National Bureau of Statistics (NBS) (2010), The Nigeria Poverty Profile 2010 Report, National Bureau of Statistics, Abuja.

Obadan, M. I. (2002), “Poverty Reduction in Nigeria: The Way Forward”, CBN Economic & Financial Review, Vol.39, No. 4.

Ogwumike, F. O. (2002), “An Appraisal of Poverty Reduction Strategies in Nigeria”, CBN Economic & Financial Review, Vol.39, No. 4.

Okojie, C.E.E., Anyanwu, J. C., Ogwumike, F.O. and Alayande, B.A. (2001), Poverty in Nigeria: Analysis of Gender Issues, Access To Social Services and The Labour Market, Submitted to AERC (Nairobi, Kenya), Collaborative Research Project on Poverty, Income Distribution and Labour Market Issues In Sub-Saharan Africa.

Ouerghi, A. (2008), The Design and Implementation of Effective Social Safety Nets For Protection and Promotion, Seminar sponsored by Social Protection and Poverty Reduction Division African Development Bank, Tunis, December 2.

Palmer-Jones, R. and Sen, K. (2003), What Luck Has Got to do With It: A Regional Analysis of Poverty and Agricultural Growth in Rural India, Journal of Development Studies, Vol. 40 No. 1.

Rhoe, V., Babu, S. and Reidhead, W. (2008), “An Analysis of Food Security and Poverty in Central Asia – Case Study from Kazakhstan”, Journal of International Development, 20, 452–465.

Rodriguez, J. G. (2002), The Determinants of Poverty in Mexico, (see www.gdnet.org/pdf/2002AwardsWinners/GrowthInequalityPoverty/Jorge_garza_rodriguez_p aper.pdf).

Rosenberg, T (2011a), “To beat back poverty, pay the poor”, The New York Times (NYT), January 3.

Rosenberg, T (2011b), “Helping the World’s Poorest for a change”, The New York Times (NYT), January 7.

Sadeghi, J. M. et al (2001), Determinants of Poverty in Rural Areas: Case of Savejbolagh Farmers in Iran, World Bank Working Papers 0112.

Schoummaker, B. (2004), Poverty and fertility in sub-Saharan Africa: Evidence from 25 countries. http://www.brunoschoumaker.be/PAA2004schoumaker.pdf: accessed on 16 June 2011.

Schultz, T. P. (1981), Economics of Population, Addison-Wesley, Reading, MA.

Skolnik, R (2011), Conditional Cash Transfers – Learning as We Go, 21 March 2011: http://endtheneglect.org/2011/03/conditional-cash-transfers-%E2%80%93-learning-as-we-go/ Szekely, M. (1998), The Economics of Poverty, Inequality and Wealth Accumulation in Mexico, St. Anthony’s Series, New York.

Waite, L. J. and Gallagher, M. (2000), The Case for Marriage, New York: Doubleday.

World Bank (1990), The World Bank Annual Report 1990, The World Bank, Washington, D. C.

World Bank (1991a), Assistance Strategies to Reduce Poverty: A Policy Paper, The World Bank, Washington, D.C.

World Bank (1991b), Indonesia: Strategy for a Sustained Reduction in Poverty, The World Bank, Washington, D.C.

World Bank (1996), Nigeria: Poverty in the Midst of Plenty, The Challenge of Growth With Inclusion: A World Bank Poverty Assessment, May 31, The World Bank, Washington, D.C.

World Bank (1998), Poverty Profile for Nigeria: 1985-1996, December. http://www4.worldbank.org/afr/poverty/pdf/docnav/03007.pdf#search='poverty%20in%20nig eri'

World Bank (2001), World Development Report 2000/2001: Attacking Poverty, The World Bank, Washington DC.

World Bank (2011), World Development Indicators 2011, The World Bank, Washington DC.

World Bank (2012), “An update to the World Bank’s estimates of consumption poverty in the developing world” http://siteresources.worldbank.org/INTPOVCALNET/Resources/Global_Poverty_Update_20 12_02-29-12.pdf (Accessed 9 March 2012).

Ugoh, S. C. and Ukpere, W. I. (2009), “Appraising the trend of policy on poverty alleviation programmes in Nigeria with emphasis on a National Poverty Eradication Programme (NAPEP)”, African Journal of Business Management, Vol. 3, No. 12, December, 847-854.

Recent Publications in the Series

nº Year Author(s) Title

John C. Anyanwu, Yaovi Role of Fiscal Policy in Tackling the HIV/AIDS Gassesse Siliadin and 148 2012 Epidemic in Southern Africa Eijkeme Okonkwo

Ousman Gajigo, Emelly Gold Mining in Africa: Maximization Economic 147 2012 Mutambastere and Guirane returns for countries Nadiaye Pietro Calice, Victor M. Bank Financing to Small and Medium Enterprises 146 2012 Chando and Sofiane In East Africa: Findings of A Survey In Kenya, Sekioua Tanzania, Uganda And Zambia Jeremy D. Foltz and 145 2012 Assessing the Returns to Education in the Gambia Ousman Gajigo An Analysis of the Impact of Financial Integration on Economic Activity and Macroeconomic 144 2012 Gabriel Mougani Volatility in Africa within the Financial Globalization Context Does Good Governance Create Value for Thouraya Triki and Olfa 143 2011 International Acquirers in Africa: evidence from US Maalaoui Chun acquisitions Thouraya Triki and Issa Africa’s Quest for Development: Can Sovereign 142 2011 Faye Wealth Funds help? Guy Blaise Nkamleu, Always Late: Measures and Determinants of 141 2011 Ignacio Tourino and Disbursement Delays at the African Development James Edwin Bank Adeleke Salami, Marco Stampini, Abdul Kamara, Development Aid and Access to Water and 140 2011 Caroline Sullivan and Sanitation in sub-Saharan Africa Regassa Namara Giovanni Caggiano and The Macroeconomic Impact Of Higher Capital 139 2011 Pietro Calice Ratios On African Economies Politique Economique Et Facteurs Institutionnels Cédric Achille Mbeng 138 2011 Dans Le Développement Des Marchés Obligataires Mezui Domestiques De La Zone CFA Does Aid Unpredictability Weaken Governance? 137 2011 Thierry Kangoye New Evidence from Developing Countries