Poverty Alleviation in Southern Punjab (Pakistan)

This paper investigates the factors affecting rural poverty using Logit regression modeling based on primary source of data in the project area of Asian Development Bank.According to the results, rural poverty can be alleviated by lowering the household size, persons per room and dependency ratio, improving education, more female labor force participation, higher household participation rate, improving assets and household’s access to market especially in remote areas. The government should pay special attention to basic infrastructure and market access facilities beside some other socio-economic and demographic variables to alleviate rural poverty in remote areas of Pakistan.

Introduction

Poverty alleviation has been one of the foremost objectives of development programs in many developing countries of the world for the last several decades. A considerable research has been carriedout on the issue of poverty alleviation and its long run social and economic effects in developed as wellas in developing countries. Since the issue of poverty alleviation is core on the agenda of allinternational organizations, the three international reports published by World Bank (2000), IFAD(2001) and ILO (2003) by the start of this century could be observed. The World Bank report(2000/2001) was primarily on the issue of poverty. This tradition became decennial since the reports of1980 and 1990 on the same subject. The International Fund for Agricultural Development (IFAD) Report on the rural poverty of the year 2001 was also published on the challenge of ending rural poverty. Indeed, it is in rural areas that poverty is the mostly pronounced with multidimensionalaspects (economic, social demographic and so forth). The International Labor organization (ILO)report presented in the 91 st session, 2003, of International Labor conference was also on the working out of poverty. This report also focused on the working poor and the poor are excluded from work.With this brief global importance of the issue of poverty, we now come to the experience ofPakistan. There is an ample evidence that poverty, which increased in 1960s declined rapidly in the1970s and 1980s has returned back in the 1990s and again it has some declining trend in 2000s [Amjadand Kemal (1997); Ali and Tahir (1999); Jafri (1999); Arif et al. (2001); (Government of Pakistan, 2006)]. Moreover, it is difficult to explain the declining trends in poverty as shown by World Bank (2000), through macro-level factors such as demographic dynamics that affect the labor force and dependency ratio, employment levels, real wage rates, workers’ remittances, assets ownership andaccess, and inflationary impact on food availability. Nevertheless the bulk of poverty exists in Pakistan but more than that in remote areas. The Asian Development Bank’s (ADB’s) current strategy is also based on Pakistan’s development priority of poverty reduction and the emphases in the PRSP.

The ADB’s Country Strategy and Program (CSP) 2002-2006 states this and translates shared objectives to assistance withinterventions in good governance (including devolution and sector and province-based reforms),sustainable pro-poor growth (rural development and employment creation) and inclusive socialdevelopment (education, health, water supply and sanitation, and social protection). However the majorfocus is to be on good governance. In addition, cross-cutting themes are to include sustainable environmental management and gender [ADB (2002)].

Since poverty alleviation is considered as an important issue of economic development in theliterature, attempts have been made to alleviate poverty by increasing the level of income of households. It has also been the declared goal of every government policy in Pakistan. Nevertheless,most of the efforts have been made to alleviate poverty in Pakistan at macroeconomic level in terms ofsustainable economic growth, facilities of job creation by improving the functioning of labor market,provision of high quality health and education services, provision of infrastructural facilities and safety nets in times of natural and financial crisis. However the least emphasis has been placed at micro orregional level poverty alleviation.

A large number of studies have also been conducted in Pakistan on poverty. Some of them haveeither concentrated more on counting numbers below the poverty line at national, provincial level, orhave explained poverty with reference to variables. The present study is the continuity of the previous studies with an emphasis on a different set of variables and study area at micro level. However somesignificant studies conducted on poverty alleviation are reviewed as follows. Kemal (1995) focused on evaluation of public strategies and policies for poverty alleviation. Heexplained four mechanisms to reduce poverty.

First was the introduction of new technologies andpromotion of large or small – scale firms. Second, the taxation policy should be reasonable. Third, the poormembers of society should get more opportunities so that income equality among households should beimproved.
Malik (1996) used micro survey data from a Punjab village and considered a large number ofrural and household specific variables besides land holding in an attempt to determine their role in raisinglevels of living of rural masses and subsequently to alleviate rural poverty.

Amjad and Kemal (1997) analyzed the correlates of poverty by using large number ofmacroeconomic variables. They also examined the impact of the structural adjustment policies on poverty in Pakistan. Malik et al. (2000) concluded that land redistribution as a direct policy measure is consideredto alleviate poverty in rural areas of Pakistan. An inference is drawn from the data analysis that the breaking of land concentration may help in achieving higher rates of agricultural growth and then consequently, poverty of rural areas of Punjab may be alleviated to some extent.

Azid et al. (2001)attempted to explain the role of female labor force participation in poverty alleviation through cottageindustry like cloth embroidery in rural Multan. They concluded that there is significant association betweenhours of female workers and household poverty.

Siddiqui (2001) stressed that the role of gender is very crucial for poverty reduction. She also concluded that improvement in human capital formation can be important in increasing women’seconomic involvement and a reduction in gender-based poverty.Chaudhry (2003) analyzed the issue of poverty in Bahawalpur district considering the extent,status, profile, correlates and micro determinants of poverty. To identify the factors affecting rural poverty, income regression and Logit regression models have been used for empirical data analysis. Itis concluded that efficient economic infrastructure, more agricultural output/production, promotion of non farm rural economy, better quality of life, health and educational facilities are considered thepolicy measures to alleviate rural poverty.

Chaudhry et al. (2005) conducted a study on the correlates and profile of poverty in the areas of Cholistan in Southern Punjab. They empirically analyzed and concluded that Cholistani land should bedistributed based on equality, improvement in livestock sector, improvement in socio-economic anddemographic variables are considered for reduction in poverty in remote areas of Pakistan like Cholistan. Chaudhry et al. (2006) also conducted a study on urban poverty alleviation through good governance in southern Punjab. They identified that good governance of economic infrastructural facilities among others will alleviate poverty in urban areas of south Punjab based on the results of a household survey data.

Arif (2006) analyzed the nationally implemented poverty reduction programs in Pakistan interms of disbursement of zakat (charitable donations), microfinance and provision of health services using data from Pakistan Socio-economic survey (PSES) for 2000/01. He concluded that public zakat currently covers far fewer households than the number actually eligible and problem is practical rather than theoretical. Most microfinance programs fail to target the poorest households, while healthservices have not reached the most disadvantaged areas of the country. He also provided broad guidelines and policy implications for improving the targeting efficiency of narrow-target programs ofpoverty reduction in general and zakat in particular.

Keeping in view some of significant literature review, we come to the scope and issues ofpresent study. This study is limited to the areas where Bahawalpur Rural Development Project initiatedby Asia Development Bank in Southern Punjab but is undertaken by emphasizing on poverty trends and largely on its determinants. Nevertheless major objectives of the present study are:
i. To present the brief description of significant projects for poverty alleviation and rural development particularly in Southern areas of Punjab.
ii. To show the profile and trends of rural poverty in Southern Punjab
iii. To document and analyze the factors affecting rural poverty alleviation in the areas ofBahawalpur Rural Development Project in Bahawalpur district.This paper is organized as follows.

In section II, we explain the brief description of povertyalleviation and rural development projects with an emphasis on the government polices. Section III states the issues of data and methodology used in this study. Section IV presents the trends of ruralpoverty in Southern Punjab. Results and discussions of the factors affecting rural poverty alleviationare elucidated in section V. The last section offers concluding remarks.

Poverty Alleviation and Rural Development Projects 1 :

A Brief Description with Government Policies Notwithstanding the improvements in poverty reduction in 1970s and 1980s, a substantial increase inpoverty had occurred between 1990 and 2000 had resulted in roughly a third of Pakistan’s populationbeing defined as poor 2 , the government produced an Interim Poverty Reduction Strategy Paper (I-PRSP) [GOP (2001)] which outlined the measures designed to reduce poverty and restore economicstability. The core elements were focused on engendering growth, improving human development andgovernance, income generating opportunities and attempting to reduce shocks affecting the poor.

During the implementation of I-PRSP, preparation for a Poverty Reduction Strategy Paper(PRSP) began which involved detailed analysis of data and consultation with a broad spectrum ofstakeholders. Analysis of data collected in 2001 from the combined Household Income-ExpenditureSurvey (HIES) and Pakistan Households Integrated Survey (PHIS) led to the following break down being estimated: non-poor, 15.3 percent, transitory non-poor 32.2 percent, transitory vulnerable 20.4percent, transitory poor 20.1 percent, chronically poor 10.4 percent and extremely poor 1.6 percent.The PRSP emphasized the importance of community involvement in planning, implementation and 1 In this section, data and material are drawn from the evaluation report of phase I of BRDP, May 2006. 2 See, for example, Amjad and Kemal (1997), Ali and Tahir (1999), Jafri (1999) and Arif et al. (2000).

monitoring and recommended group formation for the purposes. The importance of livestock andparticularly micro credit was also discussed and recognized.Using data from the 2001-2002 HIES, a recent ADB study 3 shows that the cotton/wheat agro-climatic zones in Punjab and Sindh have disproportionate shares of the poor in relation to population.With 17.5 percent of Pakistan’s rural population, the cotton/wheat zone of Punjab had 20.1 percent ofthe country’s poor. With the exception of Sindh, where the ratio was marginally higher, thecotton/wheat zone of Punjab (of which the three districts of Bahawalpur, Bahawalnagar and Rahim YarKhan form a major component) has proportionally more rural poor than anywhere else in Pakistan.

Nevertheless these observations shape the basis and justification of the study area in this paper.In view of the fact that southern areas of Punjab have received little attention not only at thelevel of policy but also at the level of empirical inquiry as compared to others. Keeping in view thehigh levels of poverty and other associated problems in southern Punjab, many projects have beenstarted in different regions. Punjab Resource Management Program (PRMP) aimed to improve short tomedium-term fiscal management in the province. Following on from this, the outputs of PRMP include:
(i) better strategic programming for poverty reduction, and
(ii) restructured and strengthenedgovernment, administration and human resource development.

As such, it was designed to impact onlocal government in Bahawalpur, Bahawalnagar and Rahim Yar Khan Districts. As part of its objectiveof reducing poverty in line with Government and Asian Development Bank (ADB) policies, theSustainable Livelihoods in Barani Areas Projects (SLBP) is intended to strengthen governance andbuild capacities within local government so that it can plan and implement rural development. The assumptions are that rural access roads, electricity and other sizeable social infrastructure are to beidentified at village and union levels, but are to be implemented by district governments. Howevercommunities or groups, assisted by rural support programs (RSPs), are themselves to plan and to implement smaller-scale activities.The Rural Access Roads Projects (RARP) was one of a series funded by the ADB and theWorld Bank aimed at rehabilitating rural roads across Pakistan.

The RARP focused on farm-to-marketroads, and its activities included preparing standard designs for them. The Bahawalpur RuralDevelopment Project (BRDP) has latterly copied the RARP design standards. The Punjab devolved Social Services Program (PDSS) aims, through two program loans and a grant, to strengthen devolvedsocial services in line with the Punjab Local Government Ordinance (PLGO) of 2001.

It focuses on theexpansion of health services, school enrolment and improved access to water supply and sanitation(WSS). It works through district governments.In line with the emphasis on credit in the PRDP, in 2000 the Microfinance Sector DevelopmentProgram (MSDP) was initiated. One of its aims was to support the formation of the Khushhali Bank,which specializes in microfinance (MF). The Bank is reported to be serving over 230,000 active clients in 2004. About 30 percent of its borrowers are women and the bank makes loans starting at PR10, 000(USD167) but which are available thereafter up to a limit of USD500.

It operates in 33 districts of thecountry, among which are Bahawalpur, Bahawalnagar and Rahim Yar Khan. The Pakistan PovertyAlleviation Fund (PPAF) is a national summit institution wholesaling financial services to eligible institutions including RSPs and other non-governmental organizations (NGOs). As of June 2005 it hadover 220,000 active sub-loans through 56 partner organizations in 96 districts of Pakistan, amongwhich are the three BRDP district [BRDP (2006)].
Data and Methodology

In this study, empirical analysis is based on the primary source of data collected from one of the areasof BRDP. Bahawalpur district forms one of the Southern parts of the Province of Punjab and is largelycovered by the BRDP. Keeping in view the difficulties of this remote area we selected one village 3 Sohail J. Malik, 2005, Agricultural Growth and Poverty: A Review of the Evidence, Islamabad, ADB Pakistan Resident Mission Working Paper No.2.

namely 145 DB consisting of 120 households by using simple random sampling technique from theuniverse of BRDP to conduct a household survey for primary data. This village is located on the linkroad from Tehsil Yazman via Kud Wala Banglaw to Lal Sohanra town of Bahawalpur Tehsil.As far as methodological issues are concerned, we considered a poverty line adopted by Malik(1992) instead of calculating a new poverty line to be used in this study and adjusted it using theconsumer price index (CPI) of annual changes in prices (Government of Pakistan, 2005-2006). Theresultant rural poverty line is Rs. 865.70 (2350 calories). Incidence of rural poverty among households is calculated by the well know method of Head Count Ratio in this study. More over the empiricalanalysis of the factors affecting rural poverty in the areas of Southern Punjab is carried out byemploying Logit Model. In a Logit Model, the endogenous variable is a dummy or categorical variablewith 1 representing household is poor and 0 if the household is non-poor. In the present study, weconsidered some significant quantitative variables besides some qualitative or dichotomous variables.The list of some selected variables for Logit model is given in table 1. Table 1: List of Variables Affecting Rural Poverty for Logit Model Analysis VariablesDescription of VariablesExplained Variable= 1 If household is poorPOVT= 0 If household is non-poor Explanatory VariablesHSIZSize of the householdDPNRDependency ratio EDUC Educational codes (Educational level of the household)FRTMFemale-male ratio (Members)FRTWFemale-male ratio (Workers)= 1, If household head is literate, andHHLT= 0, otherwisePARTParticipation rate (Employment variable)AGEHAge of the household head (Years)= 1, If household has landholding, andLDHL= 0, otherwiseLIVSPopulation of livestock per household= 1, If household has physical assets, andASET= 0 otherwise PROM Persons per room among the household= 1, If household produces for market and access to it, andHMRT= 0 otherwise It is hypothesized that household size (HSIZ), dependency ratio (DPNR), female-male ratio ofmembers (FRTM), age of household head in years (AGEH) persons per room among the households(PROM) have positive, while others have negative/inverse relation/impact on poverty alleviation basedon theoretical framework and literature. Profile and Trends of Rural Poverty in Southern Punjab Pakistan has 67.5 percent of its population residing in the rural areas.

Since poverty in Pakistan islargely a rural phenomenon, Punjab accounts for almost 56 percent of the country’s population. About36 percent of its rural population is poor which stands on the second highest in the provinces of Pakistan. However lower Punjab reveals the level of poverty (40 percent) not significantly lower thanrural NWFP and Balochistan [FBS (2002)]. It is also recognized in the report of International Fund forAgricultural Development (IFAD), 2001, on rural poverty that it is more sensitive and chronic in the rural areas of South Punjab in Pakistan. Multan, Bahawalpur and Dera Ghazi Khan Regions form thearea of South Punjab. Profile, patterns and trends of poverty have been well documented by Federal

The results of table 1 show that incidence of poverty, poverty gap and severity of povertyincrease during the 1990s and after that there is declining trend. The evidence also suggests that ruralpoverty falls by about 19 percent (73 percent to 54 percent) in the areas of BRDP. The results of thepresent study states that 55 percent households are poor. Our results are similar to that of BRDP(2006). The results of the table also suggest that poverty falls from 69.64 percent to 55 percent. Finally,it is concluded that poverty is alleviated in the areas of Bahawalpur region including Cholistan becauseof the government policies and the working of BRDP 4 .

Results and Discussions It is obvious from the discussion in last section that poverty is wide spread in the rural areas ofSouthern Punjab (Lower Punjab) as compared to Upper Punjab. Since the present study is significantlybased on the primary source of data, a Logit regression model is used to analyze the factors affectingrural poverty alleviation. The estimates of the Logit model are given in table 3. 4 See, for example, BRDP (2006)

The empirical result states that EDUC has the significant inverse association with a povertyprobability. It means that higher the level of education in a household, lower the probability of beingpoor household. Similar association is also observed for the literate household heads. The sign ofHHLT indicates that the literate head of household reduces the probability of being poor. Female-male ratio of members is considered as a sex ratio. Mostly female members in a householdin rural areas evade to work because of their customs and religious norms. Therefore their attitude toparticipation is rather discouraging. FRTM has positive and significant association with the probability ofhousehold’s poverty.
Similarly (FRTW) female-male ratio of workers is also calculated to address the issueof female participation ratio/gender issue. According to the results, there is a strong and significantaffirmation of the importance of female labour force participation in rural areas of Southern Punjab. Theincrease in this ratio leads to lower the poverty probability. The overall participation rate of a household is also imperative employment variable. In thepresent analysis, the participation rate is defined as the ratio of number of workers to number of adultsin a household. It has also the inverse relation with the probability of being rural poor at 10 percentlevel of significance. The numbers of livestock per household and physical assets (like tractor, tube
well, property other than land holding etc.) are significant variables and have inverse impact on theprobability of poverty. Age of household head (AGEH) and persons per room (PROM) are thesignificant variables at 95 percent of confidence level. The increase in both will aggravate theprobability of being poor.Finally, an important stochastic variable is considered to address the issue of secondarypoverty. Majority of the rural households is living on subsistence level; nothing is being produced forthe market in terms of agricultural production and livestock, and has no access to the market. If households produce for the market and have access to it then their per capita income will rise, manyemployment opportunities will be created and finally it will support to fill the gap between demand andsupply of these products in the country. According to the empirical results, household produces for market and access to it as a dummy variable (HMRT), is significant at 1 percent level. It has negativeimpact on the probability of poverty. This result seems to imply that households produce for themarkets have a lower probability of being poor than those who have not access to market.As most of the major descriptors of rural poverty are shown to be significant, the model doesnot do a particularly useful job of narrowing down the set of key poverty determinants.

To begin toimprove on this, the coefficients in the logistic regression can be represented as odds ratios. These oddsratios are also included in table 3 and explain the impact of a unit change in an independent variable onthe odds of being poor (the probability of being poor over the probability of being non-poor), holdingall other independent variables constant. A few clear pointers are thrown up by these ratios in the table. Concluding Remarks and Policy Implications Many efforts have been made to alleviate poverty at all levels in Pakistan since its inception evenhaving a reasonable economic growth rate in different years.

These considerable high growth rates andconsequences of the many projects on poverty alleviation have become the cause to fall in the levels ofpoverty but not to the extent as was expected by the governments and international institutions. It isalso observed that poverty alleviation efforts in terms of projects are on the right track at micro levelbut there is need to support them with macroeconomic stability in the country.Patterns of poverty differ by provinces and even within the province.
The data consistentlyshows that poverty is considerably higher in rural as compared to urban areas. Differences between therural areas of the same province are notable particularly in Punjab and NWFP [FBS (2002)]. Accordingto the estimates of the present study, 55 percent households are poor and this level of rural poverty issubstantially higher than other rural areas of Punjab Province.
It is also observed that poverty inSouthern Punjab is not a nature of primary, but it is secondary poverty as was differentiated byRowntree (1901) 5 .In the present study, we have also attempted to analyze the factors affecting poverty alleviationin Southern Punjab using Logit Model. The main findings of this empirical analysis are summarizedbelow:i. Household size, age of household head, persons per room in a household, dependency ratio and female male ratio of members in a household are strongly associated with rural poverty and thepresence of these variables increases the probability of being poor.ii. Education, female labor force participation, overall participation rate, Market access andproduction for the market, population of livestock and physical assets are the significantvariables and having negative impact on the probability of household’s poverty.Poverty alleviation became a major policy concern in Pakistan since its independence and particularly for the last seven years. Poverty reduction strategy was launched by the Government ofPakistan in 2001 in response to the rising trend in poverty during 1990s. As a result of this poverty 5 Rowntree (1901) distinguished between primary and secondary poverty. Primary poverty referred to those who did not have access to the resourcesto meet their subsistence needs and where as secondary poverty to those who seemingly did have the resources but were still unable to utilize these toraise themselves above the subsistence level.