Printer Friendly

ARE THE DETERMINANTS OF FOOD INSECURITY FOR LANDLESS HOUSEHOLDS DIFFERENT FROM THAT OF OTHER RURAL HOUSEHOLDS?

Byline: Muhammad Khalid Bashir, Steven Schilizzi and Ram Pandit

This paper aims to investigate whether or not the determinants of food insecurity for landless households are different from that of other rural households in the Punjab province of Pakistan. Household level data were collected from 576 landless households representing 12 districts of the province. The data were analyzed in two stages: first, we measured households' food security status by calculating their food consumption; and in a second stage, a binary logistic regression was used to examine the determinants of their food security. The results suggest that about 27% of the sample households were measured to be food insecure. The analysis revealed that level of education of household head had the greatest impact on food security, followed by increases in monthly income. Conversely, family size had the greatest impact on increasing food insecurity, followed by the household head's age.

The analysis further reveals that these determinants of food insecurity are similar to those found for the rural households in the same region and other countries of the world, but their relative importance for food insecurity differs for landless households. These results suggest clear priorities for food security policy in the Punjab. Keywords: landless households, rural food security, determinants, Punjab, Pakistan

INTRODUCTION

Food insecurity is on the rise in developing countries where about 900 million of the world's 925 million undernourished people are living (FAO, 2010). Over 70 percent of such people are living in rural areas and are dependent on agriculture for their livelihoods. Rural areas have some unique characteristics including the limited number of markets, their accessibility and less diversity in terms of available food items which are affecting the food security of people living there (Morris et al., 1992). In many developing countries underinvestment in the agricultural sector, makes them more vulnerable to price instability. Since the late 1980s, a sharp decline was observed in the overall rate of growth in agricultural research and development investment in developing countries.

Investment in the agricultural sector has focused largely on exportable crops to generate foreign exchange, forcing countries to rely on continued low international food prices to meet national food demand which failed to fulfill the desired results (IAASTD, 2008).

Pakistan, on the contrary, achieved food self sufficiency in the 1980s (Gera 2004) and maintains its status of food self sufficient country in terms of total production (Bashir et al., 2007, 2012). The economy of Pakistan depends largely on its agricultural sector, which contributes about 22% to the national GDP and employs about 45% of the workforce.

Moreover, a significant proportion of total population (65%) still lives in rural areas and about 26% of the population is undernourished despite Pakistan being one of the largest producers of many agricultural commodities in the world and is self sufficient on food at national level(GOP, 2011; FAO, 2010; FAO, 2011).

Punjab is the most populated province in Pakistan, a home to more than 73 million people i.e. 55% of Pakistan's population (GOP, 1998). Agriculture is the key economic sector of the province. It has more than 3.8 million farms out of 6.62 million total farms in Pakistan and has the largest share (57%) to agricultural GDP of the country. However, the majority of the households in the province are landless (74%). Such households earn most of their income from non-agricultural sources. They are mostly engaged in informal activities that accommodate a large majority of unskilled, uneducated and less educated individuals. Landless households usually earn their livelihood from paid employment and self employment (Anwar et al. 2004). Such households are the most vulnerable ones to food insecurity (Yasin, 2000).

This study aims to answer the key question: are the determinants of food insecurity for landless households different from that of other rural households? To answer this critical question we must know the answer to the following research questions:

1. What levels of food security are experienced by the landless households of the province?

2. Which socio-economic factors correlate with and best explain the levels of food security of these households?

3. What is the relative importance of these socio-economic factors for food security of landless households?

4. Results of this study are expected to provide information that will help policy makers to formulate policies that will ultimately lead to food security of landless households in rural areas of Pakistan.

MATERIALS AND METHODS

Data collection: The Punjab province was dub-divided into three regions on the basis of geographical characteristics. Out of its 36 districts, the districts situated in the south that have desert and some mix of desert and plains (i.e. river plains) were kept together to formulate South Punjab region. The districts that are mostly plains were jointly termed as Central Punjab region and those districts that are situated 350-900 meters above the sea level formed the North Punjab region for this study. The regions were asymmetric in terms of district numbers i.e. 11, 17 and 8 in South, Central and North Punjab, respectively. The household level data were collected from one third of the total districts (i.e. 12).

A proportionate sampling was adopted to determine the number of districts for each region, which resulted in 3, 6 and 3 districts from south, central and north Punjab, respectively. These districts were selected on basis of homogeneity in population size, number of villages and availability of irrigation water.

One percent of the total villages were randomly selected from each district, which resulted in 72 villages (i.e. 6*12) as sample villages. From each village, 8 landless households were randomly selected that lead to a total sample size of 576 households (i.e. 72*8).

Household level information was collected using an interview schedule. Detailed information on various aspects relating to food security including household size, household type, household income, expenditures, ownership of livestock asset and food intake were obtained from the household heads during the interview.

Data analysis: Data were analysed in two stages. In stage 1, we calculate the household food security status; and in stage 2, we examined the data to find determinants of household food security of landless households.

Food security status of the landless households was measured by calculating their per capita calorie intakes using a 7 days recall method. The calculated calories were converted into per capita intakes after adjusting to adult equivalent units to cancel out the impacts of age and gender differences. The calculated per capita calorie intake was, then, compared to a threshold level defined by the Government of Pakistan for rural areas, i.e. 2450 Kcal/capita/day (GOP, 2003). The households whose per capita calorie intake were equal to or above this threshold level were considered as food secure households, otherwise not. Mathematically, it is defined as:

Where:

Yi is the food security status of ith landless household (1 for food secure and 0 for food insecure), and

The determinants of rural household food security for the selected household were identified using a binary logistic regression model. The binary form of the dependent variable i.e. '0' for food insecure and '1' for food secure, guided us to use this model (see for example Feleke et al., 2005; Babatunde et al., 2007 and Bashir et al., 2010). The probability of the occurrence of an event for more than one explanatory variable is directly estimated using this model (Hailu and Nigatu, 2007). Assuming a linear relationship between food security status and various explanatory

RESULTS:

Household food security: Table 1 shows the results for food security situation of the sample households in the Punjab province. According to the results, more than 27% of the sample households are measured to be food insecure.

Table 1. Food security status

###Frequency###Percent

Food insecure###156###27.1

Food secure###420###72.9

Total###576###100.0

Data source: Field survey 2010-11

Descriptive statistics: The results of descriptive statistics are presented in Table 2. A huge diversity has been observed in calorie intake, monthly income, household heads' age and household size of these households. On average, per capita calorie intake remained above the recommended intake (i.e. 2450 Kcal/person/day). Similarly, average monthly income was also above the minimum wage set by the government i.e. Rs 13,000 compared to Rs. 7,000.

Table 2. Descriptive Statistics

Variables###Min###Max###Mean###SD

Per capita calorie intake (Kcal/person/day)###590###4980###3006###879

Income (Rs.)###3000###48792###13210###6424

Age (Years)###23###75###45###10

Family size (Numbers)###2###18###6###2

Number of Earners (Numbers)###1###5###1###1

Large livestock animals (Numbers)###0###15###1###2

Small livestock animals (Numbers)###0###10###0.5###1

SD = Standard deviation; Data source: Field survey 2010-11

Determinants of household food security: The results of the binary logistic regression are presented in Table 3. Based on the results 5 variables are found to be the significant determinants of food security of landless households which are explained below using odds-ratios (ORs).

Household Mmonthly income (HHMIi): Household's monthly income is the total monthly income of the household from all sources. The coefficient of this variable is positive and significant implying a positive relationship between food security and monthly income of the household. The magnitude of coefficient is rather small which is converted to the value of the coefficient into OR for an increase in Rs 1000 as; exp0.0001*1000 = 1.105. An increase of Rs 1000 in monthly income of a household increases the chances of a household being food secure by about 1.105 times or by 10.5%.

Age of household head (AHHHi): The age of the household head has a negative sign showing an inverse relationship between the age of household head and household food security status. It indicates that one year increase in the age of household head decreases the chances of household being food secure by about 4.5%. The younger people are stronger than the elders and can perform tougher jobs in field. Moreover, households with older person as head of the household are the multigenerational households having more retired and/ or older persons to feed in the family. This may explain the negative effect of this variable on household food security.

Household size (HHSi): Household size also has a negative sign indicating an inverse relationship with food security. The coefficient of this variable suggests that one extra household member decreases the chances of a household becoming food secure by a factor 0.541. The odds-ratio (0.582) indicates that each one-member increase in household size decreases the odds of being a food secure household by 41.8%.

Education level of household head (EduMi and EduIi): Regression results indicate that the family with household heads having middle (8 years of schooling i.e. grade 8) and intermediate levels of education (10-12 years of schooling i.e. grade 10 or 12) has a positive impacts on household food security. These education levels increases the chances of a household being food secure by 99.9 and 177.1%, respectively. As pointed out in the introduction, the landless people generally lack higher education; they primarily depend on labour market for their income. For such households at least intermediate level of education may serve as a necessary condition to assure food security among landless rural households.

Model significance: In terms of predictive efficiency, the model predicted with about 80% accuracy (see Table 3 above). To check the goodness of fit of a logistic model's outcomes there are two alternatives: one is the two descriptive measures known as Cox and Snell R2 and Nagelkerke R2, and second is an inferential goodness of fir test known as Hosmer and Lemeshow (H-L) Test (Peng et al. 2002). The values of Cox and Snell and Nagelkerke R2 indicate that the model explains 27% and 39% of the variations in the data, respectively. These measures are also known as pseudo R2 and the results cannot be tested in an inferential framework (Menard 2000), hence are not a good measure of goodness of fit (Hosmer and Lemeshow 2000). On the other hand, the result of Hosmer and Lemeshow (H-L) test is non-significant at p greater than 0.05, suggesting the acceptance of the null hypothesis that the model fits to the data well.

Table 3. Results of Binary Regression

Variables###b###SE###OR

Household Monthly income (HHMIi)###0.0001###0.000###1.0001

Age of Household Head (AHHHi)###-0.046###0.012###0.955

Household Size (HHSi)###-0.541###0.070###0.582

Total Earning Household Members (TEHHi)###0.087###0.180###1.091

Household Type (HHTi)###-0415###0.308###0.660

Ownership of Livestock (large animals) (OLLi)###0.097###0.166###1.102

Ownership of Livestock (large animals) (OLSi)###0.006###0.211###1.006

Education Level of Household Head (primary) (EduP)###0.270###0.264###1.309

Education Level of Household Head (middle) (EduM)###0.693###0.419###1.999

Education Level of Household Head (up to intermediate) (EduI)###1.019###0.423###2.771

Education Level of Household Head (Graduation +) (EduG)###0.134###0.489###1.143

Constant###5.217###0.769###N/A

Model Prediction success###79.9%

Log-likelihood ratio test statistics###494.142

Cox and Snell R2###0.267

Nagelkerke R2###0.387

significant at less than 1 %; significant at less than 5 %; * significant at less than 10% | Data source: Field survey 2010-11

DISCUSSION

The incidence of food insecurity among the sample households is alarmingly high compared to an earlier study of Bashir et al. (2010) for Faisalabad district of the same province. They found that about 20% the sample households were measured to be food insecure. The incidence of food insecurity among landless households is higher than the average undernourishment in the country i.e. 26% (FAO, 2010).

The results for the determinants of rural household food insecurity of earlier studies from various countries are presented in Table 4 to compare the results of this study. Although we are focusing on a specific household category

Table 4. Results of Earlier Studies

Variables###Study###Economy###Methods###Coefficients###Interpretations

Household###Bashir et al.###Pakistan###Binary Logistic###0.00005###An increase of 1000 rupees in monthly income

Monthly###2012 A###(Pak Rupee) Regression###increases the chances of a household to become food

Income###secure by 5%

###Bashir et a!.###Pakistan###Multivariate###15.06###Households belonging to the income group of Rs.

###2010 A###(Pak Rupee) Logistic###500 ito Rs 10000 had 15 times more chances to

###Regression###become food secure compared to the households

###having zero income

###Sindhu et at###India###Binary Logistic###-0.00036###The chances of food insecurity are decreased by 30%

###2008 A###(md. Rupee) Regression###with an increase of Rs 1000 in the monthly income of

###households

###Onianwa and###USA###Binary Logistic###-0.06###The chances of food insecurity are decreased by 6%

###Wheelock 2006A###(US S)###Regression###with an increase of $1000 in the annual income of

###households with children

###Che and Chen###Canada###Multivariate###7.96###(low Households belonging to the lower income group had

###2002 A###(Can $)###logistic income)###8 times more chances to become food insecure as

###regression###compared to the households in upper middle income

###group

Age of###Bashir et al.###Pakistan###Binary Logistic###-0.032###An increase of one year in the age of household head

Household###2012 A###(years)###Regression###decreases the chances of a household to become food

Head###secure by 3%

###Bashir et al.###Pakistan###Multivariate###-1.808###Households headed by the heads belonging to 36 to 45

###2010 A###(Years)###Logistic###years of age group had 83% less chances of food

###Regression###security compared to the households headed by the

###heads belonging less than 35 years of age group

###Onianwa and###USA###Binary Logistic###-0.02###The chances of household food insecurity are reduced

###Wheelock 2006 C###(Years)###Regression###by 2% with an increase of one year in the age of

###household head

Household###Bashir et al.###Pakistan###Binary Logistic###-0.372###An increase of one household member in household

Size###2012 A###(Numbers)###Regression###size decreases the chances of a household to become

###food secure by 31%

###Bashir et al.###Pakistan###Multivariate###-4.056###Households belonging to having 7 to 9 household

###2010 A###(Numbers)###Logistic###members group had 97 percent less chances of

###Regression###becoming compared to those who belong to less

###household member group

###Sindhu et at###India###Binary Logistic###-0.6743###An increase of one household member increases the

###2008 A###(Numbers)###Regression###chances of food insecurity by 49%

###Amaza et at###Nigeria###Binary Logistic###-0.014###An increase of one household member reduces the

###2006 A###(Numbers)###Regression###probability of food security by 1.5%

Education###Bashir et al.###Pakistan###Binary Logistic###0.686###Household heads' Education level of up to

###2012 A###(years)###Regression###intermediate increases the chances of a household to

###become food secure by 98%

###Bashir et al.###Pakistan###Multivariate###1.857###Households whose heads were having an education

###2010 A###(Years)###Logistic (middle)###level of middle (8 years of schooling) had 6.4 times

###Regression###more chances of food security

###Ojogho, 2010 A###Nigeria###Multivariate###-1.503###The chances of food security increased by 78% with

###(Years)###Logistic (secondary)###an increase of educational level from primary to

###Regression###secondary

###Kaiser et al###USA###Binary Logistic###-0.34###The chances of a household to become food insecure

###2003 A###(Years)###Regression###were reduced by 29% with the mothers having higher

###education levels within households

but we would expect the general thrust of the qualitative results to be the same as in this study. These results have been grouped into two broad categories while comparing with the current results: first, the results that are corroborated by current study (marked as A in Table 4); and second, the results that are contradicted with current study (marked as C in Table 4). Almost all the results presented in Table 4 are corroborated except for the age of household head and household type. An earlier study by Onianwa and Wheelock (2006) in the USA found that increase in the age of household head improve the chances of household being food secure by about 2%. This contradiction in result may be due to the social and geographical differences between the countries (i.e. Pakistan and USA). Our results also contradict with an earlier study by Bashir et al. (2010) for Faisalabad district of Pakistan where joint family increases the odds of a household to become food secure by 5.287.

Usually the joint families have multiple income earners but in our sample, the majority of such households had only one income earner (60% households) that resulted in an extra burden on the limited income of the household resulting in negative impact.

Relative importance of the factors to food security: The governments in developing countries usually operate under limited budget and knowing which factors is relatively important will give insights on prioritizing the limited resources to targeted sector. Also, it will help to initiate the policy debates both at aggregate and disaggregate (i.e. macro and micro) levels. The relative importance of the factors identified above for landless household food security can be explained in terms of the comparison of the magnitudes of their coefficients (Bashir et al., 2012; Omotesho et al., 2007; Mengistu et al., 2009). This will compare these factors in terms of the effects they have on the food security of landless households. Based on this relative importance of factors, we find the rank of the factors in the following order:

1. Education levels of up to intermediate and middle increase the chances for a household to become food secure by 177% and 100%, respectively.

2. Increasing household size decreases the chances of a household to become food secure by 42%.

3. Increase of Rs 1000 in monthly income increases the chances of a household to become food secure by 10.5%.

4. Increasing age of household heads decreases the chances of a household to become food secure by 4.5%.

For rural households of the same region, Bashir et al. (2012) found that livestock assets are the second most important factor after education level. While the studies from other countries i.e. Nigeria and Ethiopia, the ranks were totally different than this study. Omotesho et al. (2007) found that household size was the most important factor to effect rural household food security in Nigeria. According to them, expenditures on food and access to health facilities were the 2nd and 3rd most important factors, respectively. Similarly, Mengistu et al. (2009) in Ethiopia found that livestock assets was the most important determinants of food security followed by marital status, inaccessibility to economic factors, household size and household income. The comparison of these studies based on relative importance of the factors of food security indicates that these factors vary between countries due to their varying socio-geographical conditions.

It is also expected that the ranks may vary in different regions within a country and for different groups of households.

Conclusions: Our findings indicate that food insecurity of landless households in rural areas of the Punjab province of Pakistan is on the rise. Monthly income and household heads' education levels improve food security, while it deteriorates with household heads' age and household size.

Table 5. Relative ranks

Ranks###Determinants

###Landless rural households###Rural households

###Current study###Bashir et al. (2010)###Mengistu et al. (2009)###Omotesho et al. (2007)

1###Education level (up to###Education level (up to###Livestock assets (bullocks)###Household size

###intermediate and middle)###intermediate)###

2###Household size###Livestock assets (small###Marital status###Expenditure on food

###animals)

3###Household monthly income###Household size###Inaccessibility to economic###Access to health

###factors###facilities

4###Age of household head###Household monthly income###Household size###Farm size

5###-###Age of household head###Household income###-

* polygamy or monogamy; average distance (in time) to markets (input, output, credit, etc.); -- no ranking

In addition, this is one of the first studies to rank the factors for their relative contribution to food security, providing policy makers an important 'to do list' for more effective policy design. Our results suggest reforms in the education system, along with improved family planning and income generating opportunities, should be most effective. Results also suggest that the ranking of determinants of food insecurity, as well as the contribution of income earned on other people's farms relative to income earned in towns, needs to be further explored in relation to food security.

REFERENCES

Amaza, P.S., J.C. Umeh, J. Helsen and A.O. Adejobi. 2006. Determinants and measurement of food insecurity in Nigeria: some empirical policy guide. Presented at international association of agricultural economists annual meeting, August 12-18, Queensland, Australia. (Available online at http://ageconsearch.umn.edu/ bitstream/25357/ 1/ pp060591.pdf).

Anwar, T., S.K. Qureshi and H. Ali. 2004. Landlessness and rural poverty in Pakistan. The Pakistan Development Review 43:855-874.

Babatunde R.O., O.A. Omotesho and O.S. Sholotan. 2007. Socio-economic characteristics and food security status of farming households in Kwara State, North-Central Nigeria. Pak. J. Nut. 6:49-58.

Bashir, M.K., M.K. Naeem and S.A.K. Niazi. 2010. Rural and peri-urban food security: a case of district Faisalabad of Pakistan. World App. Sci. J. 9: 403-411.

Bashir, M.K., S. Schilizzi and R. Pandit. 2012. The determinants of rural household food security in the Punjab, Pakistan: An Econometric Analysis, Working Paper 1203, School of Agricultural and Resource Economics, University of Western Australia, Crawley, Australia.

Bashir, M.K., Z.N. Bajwa, K. Bakhsh, A. Maqbool and M.W.A. Chattha. 2007. Food security perspectives in Pakistan. Proceedings of the International Conference on Productivity and Growth in Agriculture: Strategies and Interventions. University of Agriculture, Faisalabad, Pakistan. 150p.

Che, J. and J. Chen. 2002. Food insecurity in Canadian households. Health Reports 12:11-22.

FAO. 2010. The state of food insecurity in the world: addressing food insecurity in protracted crises. Food and Agriculture Organization, Rome, Italy.

FAO. 2011. Country rank in the world, by commodity. Food and Agriculture Organization, Statistics Division. (Available online at: http://faostat.fao.org/site/ 339/ default.aspx).

Feleke, S., R.L. Kilmer and C. Gladwin. 2005. Determinants of food security in Southern Ethiopia. Agric. Econ. 33:351-363.

Gera, N. 2004. Food security under structural adjustment in Pakistan. Asian Survey. 44:353-368.

GOP. 1998. Population census of Pakistan. Population Census Organization, Statistics Division, Government of Pakistan.

GOP. 2003. Economic survey of Pakistan, 2002-03. Ministry of Food and Agriculture. Finance Division, Economic Advisor's Wing, Islamabad, Pakistan.

GOP. 2011. Economic survey of Pakistan 2010-11. Finance Division, Ministry of Finance, Government of Pakistan. Hailu, A. and R. Nigatu. 2007. Correlates of household food security in densely populated areas of Southern Ethiopia: Does the household structure matter? Stud. Home Comm. Sci. 1:85-91.

Hosmer, D.W. and S. Lemeshow. 2000. Applied logistic regression. John Wiley and Sons, New York, USA.

IAASTD. 2008. Food security in a volatile world. International Assessment of Agricultural Knowledge, Science and Technology for Knowledge. (Available online at: http://www.agassessment.org/docs/ 10505_ FoodSecurity.pdf).

Kaiser, L.L., H.M. Quinonez, M. Townsend, Y. Nicholson, M.L. Fujii, A.C. Martin and C.L. Lamp. 2003. Food insecurity and food supplies in Latino households with young children. J. Nut. Educat. Beh. 35:148-153.

Menard, S. 2000. Coefficients of determination for multiple logistic regression analysis. The American Statistician 54:17-24.

Mengistu, E., N. Regassa and A. Yusufe. 2009. The levels, determinants and coping mechanisms of food insecure households in Southern Ethiopia: case study of Sidama, Wolaita and Guraghe Zones. DCG Report No. 55, the Drylands Co-ordination Group.

Morris, P.M., L. Neuhauser and C.C. Campbell. 1992. Food security in rural America: A study of the availability and costs of food. J. Nut. Educat. 24:52-58.

Ojogho, O. 2010. Determinants of food insecurity among arable farmers in Edo State, Nigeria. Agric. J. 5: 151-156.

Omotesho, O.A., M.O. Adewumi and K.S. Fadimula. 2007. Food security and poverty of the rural households in Kwarwa State, Nigeria. AAAE Conf. Proceed.: 571-575.

Onianwa, O.O. and G.C. Wheelock. 2006. An analysis of the determinants of food insecurity with severe hunger in selected Southern States. South. Rural Sociol. 21:80-96. Peng, C.Y., T.S.So, F.K. Stage and E.P. St. John. 2002. The use and interpretation of logistic regression in higher education journals: 1988-1999. Res. Higher Edu. 43:259-293.

Sindhu, R.S., I. Kaur and K. Vatta. 2008. Food and nutritional insecurity and its determinants in food surplus areas: the case study of Punjab State. Agric. Econ. Res. Rev. 21:91-98.

Yasin, M.A. 2000. An investigation into food security situation in rain-fed areas of district Rawalpindi. M.Sc. (Hons.) Thesis, Dept. Agric. Econ., Univ. Agric., Faisalabad, Pakistan.

2Institute of Agricultural and Resource Economics, Faculty of Social Sciences, University of Agriculture, Faisalabad, Pakistan; 2School of Agricultural and Resource Economics, Faculty of Natural and Agricultural Sciences, University of Western Australia. Australia., *Corresponding author's e-mail: khalid450@uaf.edu.pk
COPYRIGHT 2012 Asianet-Pakistan
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2012 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Publication:Pakistan Journal of Agriculture Sciences
Geographic Code:9PAKI
Date:Sep 30, 2012
Words:4492
Previous Article:ACUTE TOXICITY OF METALS MIXTURES FOR FISH, Catla catla, Labeo rohita AND Cirrhina mrigala.
Next Article:THE CONTRIBUTION OF VARIOUS NON-FARM INCOME SOURCES IN REDUCING INCOME INEQUALITY ACROSS DIFFERENT FARM SIZES: A COMPARISON OF BARANI PUNJAB AND...
Topics:

Terms of use | Privacy policy | Copyright © 2021 Farlex, Inc. | Feedback | For webmasters |