Review of Economics and Development Studies, Vol. 6 (2) 2020, 425-438 425 Socio-Economic and Demographic Factors of Poverty Alleviation in Pakistan: A Case Study of Southern Punjab a Rashid Ahmad, b Muhammad Zahir Faridi a Assistant Professor/ PHD Scholar, School of Economics, Bahauddin Zakriya University Multan, Pakistan Email: rashidahmad@bzu.edu.pk b Professor, School of Economics, Bahauddin Zakariya University Multan, Pakistan Email: zahirfaridi@bzu.edu.pk ARTICLE DETAILS ABSTRACT History: Accepted 29 May 2020 Available Online 15 June 2020 This study aims to explore the socio-economic and demographic determinants of poverty in Southern Punjab by using the cross sectional data consisting of 785 household heads. Binary logistic regression and ordinary least square method are used for estimation. The findings exhibit that the variables like family system, household size, presence of disease and status of employment of household head are positively and significantly related to poverty whereas household head age, rural-to- urban migration, years of schooling, number of earners, women status of work, remittances, the physical assets value and ownership of house significantly and negatively influence the likelihood of poverty and positively influence the per capita income of the households in Southern Punjab. The study also provides the comparison of regional and division level. It is concluded that DG Khan division is the poorest among all the divisions of the southern Punjab. In DG Khan Division, the households have less education, high dependency ratio. In rural areas of southern Punjab, there is more poverty as compare to urban areas. The rural poverty is due to many factors like high dependency rate, lower level of education, adoption of profession, lower per capita income, dissaving. It is suggested that education should be promoted, employment opportunity should be provided so that dependency rate may be reduced, rural areas should be restructured by provision of basic necessities of life. © 2020 The authors. Published by SPCRD Global Publishing. This is an open access article under the Creative Commons Attribution- NonCommercial 4.0 Keywords: Poverty, Households, Logistic Regression, Pakistan JEL Classification: A13, C13, I32 DOI: 10.47067/reads.v6i2.220 Corresponding author’s email address: zahirfaridi@bzu.edu.pk 1. Introduction Poverty has many dimensions which have been discussed by many economists and social scientists. These dimensions may be no shelter, malnourishment, not having capabilities to visit a Review of Economics and Development Studies, Vol. 6 (2) 2020, 425-438 426 doctor for medical checkup, being ill, unemployment, not able to pay school fee, lack of representation, powerlessness, illness due to the dirty water, having little assets such as land and others, high expenditure shares on staple food and many interconnecting obstacles which lead to develop human beings. Poverty is the "incapability to maintain a minimum living standard anticipated concerning basic consumption needs or some amount of income required for satisfying them [World Bank (2006)]. Sen (1976) is among the pioneers to conceptualize the concept of Multidimensional Poverty as it is an alternative and direct way to use the concept of capability and functioning to define poverty. Multidimensional poverty is a relatively stable measure as it does not fluctuate due to inflation because there are difficulties in adjusting data for prices and inflation. As definition and response of poverty have constantly being changed as a result of developments in knowledge and results of events over the period, so it is not a new phenomenon for the developing counties. United Nations (UN) member countries signed the series of eight targets as a part of The Millennium Development Goals (MDGs). One of them was to alleviate poverty by 2015. Not only is this, but the subsequent Sustainable Development Goals (SDGs) (UN, 2012) also targeting on ending poverty. On the other side, there has been a great debate on the extent of poverty either it is psychological, social, physical, or political as a largely economic interpretation of poverty is embraced in the agenda of SDGs and the MDGs. For the development of a nation, poverty reduction should be the state foremost obligation as its part of the objective of Millennium Development Goals (MDGs). There is a persistent decrease in household poverty since the last decade in Pakistan. It is estimated that 24%population of Pakistan's lives below the national poverty line; out of which 31% live in rural areas and 13% belong to urban areas. When poverty is measured on the basis is of the multidimensional poverty index, 38.8% of the population live under the poverty line, is poor, out of which 9.4% live in urban areas and 54.6% reside in rural areas. Having aim at poverty reduction, poverty can drive the choice of policies (Chambers 1995; Wratten 1995; Kanbur and Squire 1999). At different times and different places, poverty can be compared which gives insight into poverty in the community (Ravallion 1992). Measuring poverty can help policymakers to focus efforts to reduce poverty or to direct resources or as if it is not measured properly then it does not reflect reality and will lead to misinformation to the policymakers. This research empirically investigates the determinant of household poverty of households belongs to Southern Punjab (Pakistan). This study focuses on the following objectives.  To investigate the socio-economic and demographic determinants of household poverty through the poverty line approach of Southern Punjab (Pakistan)  To investigate the socio-economic and demographic determinants of household poverty through per capita income approach of Southern Punjab (Pakistan)  To recommend policies which are based on the micro-level findings of the study 2. Literature Review This section highlights the studies available in literature that explore the factors of poverty at national and international levels. Sandefur and Sakamoto (1988) highlighted the structure of household and income as well for America and India. The estimation was done through primary data by using a questionnaire for the year 1980. The authors concluded that during this estimation period, Indians preferred to live in a family; household having coupled with children, as compared to the black or white in the United States. The greater frequency of the family household, specifically couples having children among Indians and Americans would be considered to design social policies. Ahmed (1995) analyzed income inequality in different occupations and worked on Pakistan's PCI (per capita income) based on Review of Economics and Development Studies, Vol. 6 (2) 2020, 425-438 427 1992-1993. The results showed that in the various occupations or professions, the lowest level of inequality was observed in Baluchistan, and the highest inequality in NWFP province. Lanjouw and Ravallion (1995) elaborated on the association between poverty and the size of the household by taking into consideration the case study of Pakistan. Primary data were used for the year 1991 through a survey. The results showed the existence of a strong negative relationship between the consumption and the size of the household. Azhar and Ahmad (2001) studied the basic shortcomings of different poverty reduction methods and focused on the role of the political economy and the shift from the wealthy to the poor. The results showed that poverty alleviation policies and programs were a real concern for the political economy. The basic premise of poverty alleviation policy was to focus on how information and incentives were integrated into the program design. The technical poverty alleviation method ignored the background of policy design. These two different approaches could be bridged by best utilizing two interdependent elements. Datt and Jolliffe (2005) analyzed the influence of poverty in Egypt. Primary data were collected through a questionnaire for the year 1997 and were used to estimate the results. Findings revealed that the household size had a negative impact on the dependent variable living; the standard of the people. While other independent variables including education, intergenerational effects, and owned land had a positive relationship with the dependent variable. Furthermore, education was considered a crucial factor for the living standard. Khalid et al (2005) examined the determinants of food poverty in Pakistan. The study decomposed the estimates on the urban or rural levels. The microdata used in this study from 1998 to 1999 was taken from HIES (Household Integrated Economic Survey and PIHS (Pakistan Integrated Household Survey). Multivariate analysis was used to determine food poverty in Pakistan. The results indicated that on average 40 percent of households were poor at the national level. In rural areas, poverty was comparatively higher with 46 percent of the households falling below the poverty line, while in urban areas 41 percent of households were poor. Cheema et al. (2008) estimated changes in the intensity, severity, and incidence of poverty in the Punjab region below the provincial level. Foster, Greer, and Thorbecke were used to measure poverty. It was estimated that there was a highly impoverished enclave in the southern and western parts of Punjab. With very few exceptions, the incidence and severity of poverty in most areas of the enclave were extremely high and there was average poverty in every two families. The enclave also performed poorly at the district level development indicators. Choudhry et al. (2009) analyzed the impact of the social and economic factors of Tehsil Jatoi (Muzaffargarh District). Poverty profile and econometric methods were used for empirical analysis. It was suggested that dependency on family, household size, land ownership, participation, and livestock quantity had a significant impact on the incidence of poverty. Jamal (2011) estimated the poverty through non-income deprivation indicators for Pakistan. The estimation period was from 2008 to 2009. Findings showed that fifty-seven percent of people in the year 2008-09 living in Pakistan were in the state of the multiple deprivations. The rural incidence was high, however, twenty-six percent of the urban population was facing poverty in terms of the indicators that had been used in the multidimensional construction. Nisar et al. (2013) investigated Pakistani issues such as income inequality and poverty and their impact on the family. Using multiple logit models, the 2008-09 dataset was used for estimation. The Review of Economics and Development Studies, Vol. 6 (2) 2020, 425-438 428 results showed that land and livestock as an asset, male as a head of household, secondary education, employment, and women's empowerment greatly reduced the chances of short-term poverty. During this period, income distribution deteriorated due to the gap between low-income and high-income families. Jayamohan and Kitesa (2014) used cross-sectional data from 1999 to 2000 and 2004 to 2005 to investigate the relationship between poverty and gender in Ethiopia. The results showed that the number of poor households, in both the male-headed households and the female-headed households, declined during the period 1999-2000 to 2004-2005, with a lower the reduction of the female-headed household rate. The feminization of poverty was, therefore, a weak theory of the Ethiopian city. Haq et al. (2015) measured the causes of family poverty affecting rural areas in Vehari (Pakistan). The data were collected from 350 selected households through multi-level random sampling and for analysis binary logistic regression models were used. It was analyzed that nearly 34.8% of the respondents in the region were poor. Socio-economic empowerment, dependency ratio, family size, household economic activity, access to finance, marital status, and head of household age had a positive impact on the incidence of poverty in Vehari. 3. Data and Methodology This section presents the data sources, sampling technique, poverty line recruited in this study, model specification, and description of variables. 3.1 Data Sources and Sampling Technique About 785 household heads were interviewed from southern Punjab. From the formula, the sample size of 785 was calculated which is as given below.             Ne ppz e ppz SS 2 2 2 2 )1(* 1 )1(* The total household of southern Punjab is almost 4930998 (Pakistan Bureau of Statistics, 2019). The sample size of 785 was calculated at a 95 % confidence interval and a 3.5 % margin of error. To calculate the sample size from each district out of 785, propionate sampling is used. There are three divisions of southern Punjab such as Multan, Bahawalpur, and DG khan division and each division consists of four districts. The division with a higher number of households has a higher proportion to be included in the sample. In the second stage, convenience sampling is used to collect information from the household head. 3.2 Poverty Line In the current study, the poverty line is drawn on the basis is of the World Bank definition. The World Bank defines poverty in absolute terms. The household is living in extreme poverty if it falls below the poverty line of US$1.90 per person per day. While calculating the poverty line based on the current dollar rate, it is estimated that the household having income 8664 Rupees per month or more than 8664 Rupees are classified as non-poor and the household having income less than 8664 Rupees are classified as poor. 3.3 Model Specification and variable1 description: Review of Economics and Development Studies, Vol. 6 (2) 2020, 425-438 429 Determinants of households poverty of overall Southern Punjab and divisions of southern Punjab like Bahawalpur, Multan and DG Khan based on the poverty line are explored by using Binary Logistic regression. The operational model is specified as follows:        OWNHLNSATSRIMIWSOS NOERNSOEMPYOSHHPODMIGAGEHHHHSIZEFSYS fPov ,,, ,,,,,,,, In order to explore socio-economic and demographic determinants household poverty based on per capita income in the three divisions and Southern Punjab by using the same variables. we have used ordinary Least Square method in the log linear form. The functional form of operational model is as follows:        OWNHLNSATSRIMIWSOS NOERNSOEMPYOSHHPODMIGAGEHHHHSIZEFSYS fPCI ,,, ,,,,,,,, Table 1: Description of Variables Variables Variables Description Dependent Variable LNPCI Natural Log of Per Capita Income It is calculated by dividing the total household income on total household size POV Poverty =1 if the household head per capita income is lesser than 8664 rupees per month. =0 if household head per capita income is more than 8664 rupees per month Explanatory Variables AGEHH household head age household head age expressed in years HHSIZE Household size Total number of peoples residing in a household AREA Area of residence It describes the region of residence i.e rural or urban if the household lives in urban areas = 1 otherwise = 0 MIG Migration Rural to urban migration =1 if household migrated from the rural area to urban areas = 0 if not YOSHH Years of schooling of the household head It describes the educational level or schooling of household head in years FSYS Family system The family system is estimated as nuclear or joint =1 if joint = 0 if nuclear RIMI Remittances = 1 if receives remittances =0 if not NOERN Number of earners A total number of earner sin households. It including own, spouse and other members OWNH Ownership of house Either house is owned or rented = 1 if own house Review of Economics and Development Studies, Vol. 6 (2) 2020, 425-438 430 =0 if not WSOS Working status of spouse =1 if the Spouse is participating in the labor force =0 if Otherwise LNASTS Natural log of the value of assets It includes all the assets of the household like land, car, house, property, television, tractor, etc. SOEMP Sector of employment It discusses the sector of employment of household either it is formal or informal = 1 if informal sector employee = 0 if formal sector employee POD Presence of diseases Either the person has a disease or not. = 1 if presence of diseases = 0 if not 3.4 Data Analysis Data is analyzed through descriptive and econometric technique. Descriptive statistics shows maximum, minimum, mean, median and standard deviation. Correlation matrix shows degree of association among variables. Econometric technique measures the relationship between the dependent and independent variables by log-linear regression and lohgisti9c regression technique. 3.4.1 Descriptive Analysis: Descriptive analysis may be done through descriptive statistic, poverty trends and correlation analysis which is as follows. 3.4.1.1 Descriptive Statistics The descriptive statistics of variables that are used in this analysis are presented in Table 2 that consists of a minimum, maximum, mean, and standard deviation of variables. Table 2: Descriptive Statistics Variables Minimum Maximum Mean Std. Deviation AGEHH 20 90 48.12 11.546 YOSHH 0 20 9.85 4.723 HHSIZE 2 19 6.19 2.162 NOERN 1 7 2.02 1.152 WSOS 0 1 0.22 0.417 MIG 0 1 0.13 0.339 RIMI 0 1 0.12 0.322 SOEMP 0 1 0.45 0.497 POD 0 1 0.22 0.412 OWNH 0 1 0.89 0.316 FSYS 0 1 0.47 0.500 PI 0 1 0.39 0.487 LNASTS 20000 77550000 4382464.46 7909851.730 INOME 6000 506000 71745.94 59431.120 Review of Economics and Development Studies, Vol. 6 (2) 2020, 425-438 431 Source: Author’s Calculations Based on Household Survey Data, 2019 3.4.1.2 Poverty Trends Table 3 explains the poverty trends based on the international poverty line which is estimated in Pakistani rupees at 8664 in 2019. The household is poor if the income of the household has less than 8664 rupees per month and is non-poor if the income of the household having more than 8664 rupees per month. It is found that 38.59 percent of residents of the Southern Punjab living below the poverty line the incidence of poverty is higher in rural areas (46.13) than in urban areas (29.05). Among the divisions, the incidence of poverty is higher in the DG khan division it is estimated at 43.14 percent. Table 3: Poverty Incidence (Poverty Line 8664) Division Area Poor Non-Poor Total Households Poverty Incidence Southern Punjab Total 303 482 785 38.59 Rural 201 233 434 46.13 Urba n 102 249 351 29.05 Multan Total 89 166 255 34.90 Rural 59 72 131 45.03 Urba n 30 94 124 24.19 DG Khan Total 107 141 248 43.14 Rural 72 101 143 50.34 Urba n 35 70 105 33.34 Bahawalpu r Total 107 175 282 37.94 Rural 70 90 160 43.75 Urba n 37 85 122 30.32 Source: Author’s Calculation from the Survey data, 2019 Note: National Poverty line as estimated at Rs. 8664.0 per person per month 3.4.1.3 Correlation Analysis Correlation analysis is used to check the multicollinearity problem and degree of association among each set of variables. The results of the correlation analysis of south Punjab are presented in Table 4. To evaluate the multicollinearity issue, it is suggested that if the value of the coefficient of correlation between two independent variables in a model is greater than 0.80 there exists a problem of multicollinearity otherwise there is no multicollinearity in between two independent variables in a model. Finding show that there is no issue of multicollinearity among the variables. Review of Economics and Development Studies, Vol. 6 (2) 2020, 425-438 432 Table 4: Correlation Matrix PI AGEHH YOSHH HHSIZE RIMI MIG SOEMP POD FSYS OWNH NOERN LNASTS WSOS PI 1.000 AGEHH -0.06 1.00 YOSHH -0.49 -0.04 1.00 HHSIZE 0.22 0.28 -0.14 1.000 RIMI -0.23 0.10 0.08 0.001 1.000 MIG -0.22 -0.012 0.21 -0.04 0.10 1.000 SOEMP 0.41 0.004** -0.47 0.05 -0.09 -0.12 1.000 POD 0.25 0.052 -0.24 0.12 -0.03 -0.06 0.15 1.000 FSYS 0.22 0.13 -0.08 0.44 -0.03 -0.05 0.05 0.12 1.00 OWNH -0.22 0.12 0.14 -0.06 0.05 -0.02 -0.13 -0.12 -0.003 1.00 NOERN -0.18 0.24 0.07 0.45 0.11 0.07 -0.10 -0.01 0.20 0.06 1.00 LNASTS -0.30 0.08 0.16 0.000** 0.12 0.12 -0.14 -0.05 -0.01 0.11 0.09 1.00 WSOS -0.26 0.08 0.16 0.15 0.01 0.12 -0.13 -0.038 0.06 0.07 0.47 0.13 1.000 **. Correlation is significant at the 0.01 level (2-tailed). *. Correlation is significant at the 0.05 level (2-tailed). Review of Economics and Development Studies, Vol. 6 (2) 2020, 425-438 433 3.4.2 Econometric Analysis This section provides the estimation of both specifies models. Binary Logistic regression model that uses a dummy variable to determine the factors of occurrence or non-occurrence of poverty in a household. A household is poor if the per capita income of the household falls below the poverty line otherwise non-poor. Ordinary Least Square (OLS) method is employed for determining the poverty level of household by using household per capita income. 3.4.2.1 Determinants of Household Poverty in Southern Punjab Binary Logistic Table 5 demonstrates the binary logistic estimates of household poverty. The household poverty is used as a dependent variable, measured by a dummy variable is equal to “one” when household falls below the poverty line and equal to “zero” when household falls above the poverty line, while explanatory variables consist of the family system (FSYS), size of the household (HHSIZE), household head age (AGEHH), rural-to-urban migration (MIG), presence of disease (POD), household head schooling years (YOSHH), head of the household’s working sector (SOEMP), total earners in a household (NOERN), women status of work (WSOS), remittances (RIMI), value of physical assets value (ASTS), and ownership of the house (OWNH). The McFadden R-squared indicates the goodness of fit of the model and it means that the variation in dependent variable due to the explanatory variables is 0.46 percent while the remaining 0.54 percent is random variation. The LR-statistic (482.72) shows the overall significance of the model. Likewise, the McFadden R-square and LR-statistic of other divisions of Southern Punjab can be observed from Table 5. We have observed that the variables like family system, household size, presence of disease, and sector of employment of household head positively and significantly affect the likelihood of poverty whereas the variables household head age, rural-to-urban migration, the household head years of schooling, a household having several earners, women status of work, remittances, the physical assets value and ownership of house significantly and negatively influences the possibility of poverty of household in Southern Punjab. In Bahawalpur division, the demographic variables household size and status of employment of the household head originated to be positive and significant determinant while household head schooling years, a household having several earners, women status of work, remittances, and physical assets value are the factors that statistically significant and negatively influences the poverty in Bahawalpur division. Considering the Multan division it is found that the variables household size, presence of disease, and sector of employment of household head are significant and encouraging factors of the likelihood of household poverty while the variables household head age, rural-to-urban migration, household head years of schooling, household having number of earners, spouse participation and physical assets value are the discouraging and significant factors of the likelihood of household poverty in Multan division. In the DG khan division, it is originated that the variables, family system, household size, and sector of employment of household head are the factors that positively and significantly impacted the likelihood of household poverty while the variables rural-to-urban migration, household head years of schooling, a household having many earners, remittances, physical assets value and ownership of house negatively and significantly contributing aspects of the likelihood of household poverty in DG khan division. Review of Economics and Development Studies, Vol. 6 (2) 2020, 425-438 434 Table 5: Binary Logistic Estimates of of Household Poverty in Southern Punjab Dependent variable: Poverty Status of Household = 1 if household lies under a poverty threshold, = 0 otherwise Explanatory Variables Southern Punjab Bahawalpur Division Multan Division DG Khan Division Constant 7.444** (6.249) 11.246** (4.783) 9.360** (3.458) 4.097* (2.277) Demographic Variables FSYS 0.850** [0.201] (3.628) 0.344 [0.081] (0.897) 0.707 [0.161] (1.320) 1.313** [0.322] (2.958) HHSIZE 0.403** [0.095] (5.546) 0.477** [0.112] (4.057) 0.441** [0.100] (2.608) 0.493** [0.121] (3.571) AGEHH -0.026** [-0.006] (-2.547) -0.009 [-0.002] (-0.549) -0.076** [-0.017] (-3.360) -0.002 [0.000] (-0.145) MIG -1.289** [-0.305] (-2.999) -1.019 [-0.240] (-1.300) -1.347*** [-0.306] (-1.771) -1.356*** [-0.333] (-1.682) POD 0.844** [0.200] (3.130) 0.654 [0.154] (1.436) 1.073*** [0.244] (1.783) 0.750 [0.184] (1.629) Socio-Economic Variables YOSHH -0.188** [-0.045] (-6.923) -0.144** [-0.034] (-3.005) -0.237** [-0.054] (-3.351) -0.201** [-0.049] (-4.559) SOEMP 1.243** [0.294] (5.519) 0.925* [0.218] (2.381) 1.506** [0.342] (2.725) 1.885** [0.462] (4.550) NOERN -0.531** [-0.126] (-4.100) -0.534* [-0.126] (-2.250) -0.905** [-0.206] (-2.895) -0.478* [-0.117] (-2.342) WSOS -0.961** [-0.228] (-2.986) -1.113*** [-0.262] (-1.650) -1.613*** [-0.366] (-1.881) -0.789 [-0.193] (-1.604) RIMI -1.954** [-0.463] (-3.651) -1.483*** [-0.349] (-1.870) -2.389 [-0.543] (-1.617) -1.844*** [-0.452] (-1.849) PASTS -0.428** [-0.101] (-5.880) -0.811** [-0.191] (-5.213) -0.348** [-0.079] (-2.151) -0.278** [-0.068] (-2.722) OWNH -1.121** [-0.265] (-3.101) 0.807 [0.190] (1.406) -1.129 [-0.257] (-1.265) -1.981** [-0.486] (-2.747) McFadden R 2 0.461 0.471 0.478 0.440 LR Statistic 482.72 176.33 190.71 149.36 P-value (LR-Stat) 0.000 0.000 0.000 0.000 Sample Size (N) 785 282 255 248 Source: Author’s Calculations by Using E-Views (Statistical Software). Review of Economics and Development Studies, Vol. 6 (2) 2020, 425-438 435 Note: The values in the brackets and parenthesis are marginal effects and z-statistic values respectively. *5 percent level of Significance, **1 percent level of Significance, ***10 percent level of Significance 3.4.2.2 OLS Estimates of Determinants of Household Poverty Based on Per Capita Income Table 6 portrays estimation of the factors that affect the per capita income and in turn household poverty status in Southern Punjab. The dependent variable used in a model is the log of household’s per capita income while the explanatory variables comprised of the family system (FSYS), size of the household (HHSIZE), household head age (AGEHH), rural-to-urban migration (MIG), presence of disease (POD), household head schooling years (YOSHH), household head employment status (SOEMP), number of earners in a household (NOERN), women status of work (WSOS), remittances (RIMI), physical assets value (PASTS), and ownership of the house (OWNH). The R-squared is 0.455 it reveals that explained variation in the dependent variable is 45.5 percent while the remaining 54.5 percent is due to the other variables that are not added in a model. The value of F-statistic (53.722) and it shows that the model is highly significant statistically overall. Likewise, the value of R-squared and F-statistic of various divisions of Southern Punjab are 0.53,0.46 and 0.47 respectively. The results show that family system, size of the household and status of employment of household head are the factors that reduces the household income per capita and enhance poverty. The variables household head age, rural-to-urban migration, the household head schooling years, a household having several earners, women status of work, remittances, physical assets value and ownership of house are the aspects that positively and significantly influences the household welfare. These factors reduces the poverty level in Southern Punjab. In the Bahawalpur division, the household size and employment status of the household is negatively linked to the income per capita and directly related to the household poverty while the age of the household head, household head schooling, total employed members, women status of work, remittances, and physical assets value are the factors that positively and significantly influenced the household per capita income and inversely linked to the poverty probability in the Bahawalpur division. In the Multan division, it is found that the variables like household size and employment status of household head negatively and significantly impact the household per capita income and positively influences poverty whereas the household head age, household head schooling, total employed members, spouse participation and remittances positively and significantly affects the per capita income. As a result, poverty in Multan division has declining trend. In the DG khan division, it is originated that size of the household and employment status of the household are the factors that negatively and significantly influence the household income per capita and directly impact the household poverty whereas the age of the household head, years of schooling of the head of the household, household having many earners, remittances, physical assets value and ownership of house increase the per capita income of the household. When per capita income increases, poverty level reduces in Dera Ghazi Khan division. Review of Economics and Development Studies, Vol. 6 (2) 2020, 425-438 436 Table 6: OLS Estimates of Household Poverty Based on Per Capita Income Dependent Variable: Household Per Capita Income Explanatory Variables Southern Punjab Bahawalpur Division Multan Division DG Khan Division Constant 7.861** (36.069) 7.473** (22.422) 7.840** (16.671) 7.932** (22.386) Demographic Variables FSYS -0.083*** (-1.853) -0.060 (-1.035) -0.135 (-1.290) -0.102 (-1.201) HHSIZE -0.110** (-9.332) -0.102** (-6.668) -0.085** (-3.151) -0.141** (-6.539) AGEHH 0.008** (4.695) 0.004*** (1.935) 0.009* (2.326) 0.010** (3.039) MIG 0.156** (2.562) 0.140* (1.641) 0.168 (1.460) 0.192 (1.560) POD -0.002 (-0.056) -0.028 (0.425) -0.065 (-0.584) -0.058 (-0.664) Socio-Economic Variables YOSHH 0.036** (7.267) 0.034** (4.991) 0.038** (3.048) 0.037** (4.441) SOEMP -0.338** (-7.368) -0.227** (-3.743) -0.293** (-2.613) -0.465** (-5.886) NOERN 0.138** (6.104) 0.088** (2.849) 0.159** (3.380) 0.172** (4.307) WSOS 0.169** (3.001) 0.190* (2.301) 0.271* (2.360) 0.007 (0.079) RIMI 0.244** (3.838) 0.168*** (1.950) 0.415** (3.481) 0.051 (0.389) LNASTS 0.062** (4.698) 0.114** (5.300) 0.040 (1.429) 0.055** (2.619) OWNH 0.136* (2.076) 0.002 (0.030) 0.168 (1.158) 0.342** (2.731) R 2 0.455 0.526 0.461 0.474 Adjusted R 2 0.446 0.505 0.434 0.447 F-Statistic 53.722 24.932 17.256 17.689 P-value (F-Stat) 0.000 0.000 0.000 0.000 Sample Size (N) 785 282 255 248 Source: Author’s Calculations by Using E-Views (Statistical Software). Note: The values in the parenthesis are t-statistic values. *5 percent level of Significance, **1 percent level of Significance, ***10 percent level of Significance Review of Economics and Development Studies, Vol. 6 (2) 2020, 425-438 437 4. Conclusions The main purpose of study is to investigate the factors that alleviate poverty in southern Punjab. The study is based on survey. The study concludes that family system, household size, presence of disease and employment status of household have become the cause of poverty whereas the Household head age, rural-to-urban migration, the household head years of schooling, household having number of earners , women status of work, worker remittances, the physical assets value and ownership of house significantly reduce the poverty in Southern Punjab. In Bahawalpur division household size and occupation of the household head escalates poverty significally. The variables like household head, schooling years, a household having number of earners, women status of work, remittances, and physical assets value are the factors that significantly decreases poverty level. Similarly in Multan and D.G Khan division, the same situation is found. The poverty is caused by household size an occupation of the household. The rest of the variables have decreasing impact of the poverty. But the more intensity is observed in D.G Khan division. It is recommended that household size should be reduced by taking action on birth control program. Moreover health and education facilities should be provided at the nearest distance of household residence. In rural areas of southern Punjab, the government should provide basic facilities so that the economic status of the household may be improved and poverty intensity may reduce. There should be provision of employment opportunities in formal as well as informal sector especially in D.G Khan Division because it is poorer as compared to other divisions of Southern Punjab. References: Ahmad, M. A. I. (2001). Poverty alleviation and the third world. Pakistan Economic and Social Review, 39(1), 49-56. Chambers, R. (1995). Poverty and livelihoods: whose reality counts?. Environment and urbanization, 7(1), 173-204. Chaudhry, I. S., & Malik, S. (2009). The Impact of Socioeconomic and Demographic Variables on Poverty: A Village Study. Lahore Journal of Economics, 14(1), 39-68. Cheema, A., Khalid, L., & Patnam, M. (2008). The geography of poverty: Evidence from the Punjab. The Lahore Journal of Economics, 13(2008), 163-188. Datt, G., & Jolliffe, D. (2005). Poverty in Egypt: Modeling and policy simulations. Economic Development and Cultural Change, 53(2), 327-346. Jamal, H. (2011). Assessing poverty with non-income deprivation indicators: Pakistan, 2008-09. The Pakistan Development Review, 50(4), 913-926. Jayamohan, M. K., & Kitesa, A. T. (2014). Gender and poverty–an analysis of urban poverty in Ethiopia. Development Studies Research. An Open Access Journal, 1(1), 233-243. Kanbur, R., & Squire, L. (2001). The evolution of thinking about poverty: exploring the interactions. Frontiers of development economics-The future perspective., 183-226. Khalid, U., Shahnaz, L., & Bibi, H. (2005). Determinants of poverty in Pakistan: A multinomial logit approach. The Lahore Journal of Economics, 10(1), 65-81. Lanjouw, P. and M. Ravallion. 1995. "Poverty and Household Size." The Economic Journal, 105, November: 1415-1434.45. Nisar, R., Anwar, S., Hussain, Z., & Akram, W. (2013). An investigation of poverty, income inequality and their shifters at household level in Pakistan. Journal of Finance and Economics, 1(4), 90-94. Rahman, R. I., & Islam, R. (2013). Female labour force participation in Bangladesh: trends, drivers and barriers. International Labour Organization, DWT for South Asia and Country Office for India. Sandefur, G. D., & Sakamoto, A. (1988). American Indian household structure and Review of Economics and Development Studies, Vol. 6 (2) 2020, 425-438 438 income. Demography, 25(1), 71-80. SEN, A. K. (1976): "Poverty: An Ordinal Approach to Measurement," Econometrica, 44, 219-231. Ul Haq, R., Jahangeer, A., & Ahmad, A. (2015). Out-migration in rural Pakistan: does household poverty status matter?. The Pakistan Development Review, 54 (4), 315-329. Wratten, E. (1995). Conceptualizing urban poverty. Environment and urbanization, 7(1), 11-38.