An Asian Development Bank (ADB) expert Dr Sohail Jehangir Malik said the agricultural growth without specific interventions, targeting small farmers and rural non-farm households, might not alleviate poverty for much of the poor in rural Pakistan. In his second working paper on "Agricultural growth and rural poverty", published recently by the ADB' s Pakistan Resident Mission (PRM), he suggested an explicit strategy for the development of rural non-agriculture sector, which was currently lacking in Pakistan's Poverty Reduction Strategy Papers and development plans.
He said the skewed land distribution was one of the major obstacles, hindering the rapid reduction of rural poverty. Consequent to external shocks or disasters, the probability of remaining poor was much higher amongst poor households than it was for non-poor households, he added.
The incidence of poverty was low in zones where the percentage of incomes from wages and salaries and transfer incomes was high, he mentioned.
Jehangir Malik said the agriculture sector in Pakistan was a major determinant of overall economic growth because of its significant contribution to Gross Domestic Products (GDP), export earnings, employment, and food availability.
Agricultural growth helps reduce poverty directly through an increase in farmers' income, he said, adding the indirect effect of agricultural growth on poverty reduction occurred through the spending of farmers' income on locally produced non-agricultural goods.
He mentioned the relevant literature argued that the increased demand for those goods led to a much larger increase in employment, which was the main vehicle for poverty reduction in rural areas.
However, in the case of Pakistan, the reasonable agricultural growth in the 1990s could not be translated into poverty reduction. Rather, the consensus is that the percentage of rural persons living below the poverty line had increased over time and especially since the late 1990s.
It appeared that the interplay of several factors caused that paradox, he added.
Jehangir Malik said the inability of small farmers to diversify crop production, particularly in the cotton/wheat belt of Sindh and southern Punjab where poverty was relatively high, also contributed to an increase in rural poverty.
Moreover, in the presence of a highly skewed land distribution, the benefits arising from agricultural growth in the 1990s did not reach small farmers and sharecroppers, instead accruing primarily to large farmers.
He pointed out a substantial decline in using casual labour and real wages in the 1990s were other important determinants of the persistent high rural poverty in question.
" Employment opportunities in the rural non-agriculture sector were inadequate. Income from this sector is derived from many diverse activities. The poor tend to be found in low-productivity activities that demand low levels of skill and little education; whereas within wage work, the poor are generally involved in unskilled labour in the construction sector," he added.
In view of the relatively high incidence of poverty in the cotton/wheat zones of Punjab and Sindh, Jehangir Malik suggested a poverty reduction strategy, focusing on education, skill development, job creation, and health care, needed to be designed for those areas.
Given that land distribution was highly skewed in those zones, he said it was necessary to concentrate more on income sources that were independent of land to reduce rural poverty. That suggested providing land less and small farmers with easy access to credit, technology, and information, he added.
He said "Transitory poverty" could be reduced if policy interventions aimed at levelling out income fluctuations, adding a reduction in chronic poverty was possible through large and sustained growth in household incomes.
For the former, the availability of micro-credit would be an effective tool, while for the latter, targeted public works programmes could help reduce chronic poverty.
He pinpointed the Poverty reduction was linked strongly to employment, adding the exploitation of labour in situations of poor governance and thin, inadequate labour markets was a major cause of increasing poverty.
The fact that rural real wages actually declined in Pakistan despite significant external and internal out-migration during the 1980s bears testimony to the fact that there is no minimum wage protection in the rural sector.
In addition, the legislative framework for the protection of workers does not apply, as these laws do not apply to the agriculture sector. Agricultural labour is thus deprived of benefits such as social security and old-age pension benefits. Minimum wage laws needed to be set for the agriculture sector and all legislation, including workers' protection and non-wage benefits, made applicable to the agriculture sector, he proposed.
The ADB expert said the government had followed an interventionist pricing policy for the agriculture sector, adding although those pricing policies had been designed to favour small farmers, studies argue that they achieved the opposite. Such policies should be revised for the benefit of small farmers, he added.
A comparison of Indian and Pakistani Punjab shows that higher productivity has been achieved in the former, mainly due to greater efforts to tackle resource degradation.
Greater efforts need to be directed towards the conservation of natural resources. One measure would be to educate and encourage farmers through incentives to move to more sustainable practices such as diversified crop rotation and the cultivation of legumes, he suggested.
Finally, Jehangir Malik said the research on the linkages between agricultural growth, rural development, and poverty reduction required more desegregated data at various levels.
It was necessary for the government to mobilise additional resources for data collection and to co-ordinate the existing sources of data collection to maximise synergies, he said, adding those data gaps were also reflected in the gaps in existing and ongoing research.
HE MENTIONED FOLLOWING KEY AREAS REQUIRING FURTHER RESEARCH INCLUDE:
a) A more comprehensive and desegregated analysis of trends in rural poverty using region-specific poverty lines estimated on a scientific basis;
b) b) An analysis of the dynamic aspects of rural poverty and its determinants;
c) c) An analysis of the private and public incomes of the poor and the variations in these;
d) d) An evaluation of the efficiency of publicly provided goods and services in reaching the rural poor, especially the recent measures addressing the vulnerability of the poor under the Poverty Reduction Strategy Papers.
In Pakistan, Jehangir Malik said, achieving high agricultural growth is considered one of the key factors in poverty reduction. The reported growth rate of 4.6 percent in the agriculture sector in the 1990s was reasonable. However, estimates of poverty based on a headcount measure show an increase in rural poverty during the 1990s.
The present study explores the reasons why the benefits of agricultural growth in the 1990s could not be translated into poverty reduction.
A review of the literature shows that despite high growth rates in the agriculture sector in the 1960s, poverty increased in rural areas because the initial beneficiaries of agricultural subsidies and new technology emerging during this period were generally large farmers. Hence, the increased agricultural growth was not reflected in reduced levels of poverty.
The 1970s witnessed a decrease in the incidence of rural poverty, largely due to private investment in agriculture and the heavy emigration of rural-based workers to the Middle East. The resulting increased foreign remittances are cited as one of the major reasons behind the falling poverty trends witnessed during this period. These trends continued through the 1980s largely due to the sustained inflow of remittances and relatively better performance of the agriculture sector.
Several recent studies concur on a trend of increasing rural poverty since the late 1980s. Certain other studies, which found rural poverty fluctuated during the 1990s, also indicate that by the year 2001, the incidence of poverty was considerably higher than in the early 1990s. The consensus is that the percentage of rural persons living below the poverty line has increased over time and especially between FY1997 and FY2001.
This study also examines regional variations in rural poverty. Despite methodological differences, the results of recent studies consistently indicate the lowest levels of poverty for barani (rain-fed) Punjab in various years, while the high poverty levels have generally been observed in the cotton/wheat zones of Sindh and southern Punjab.
Together, these two zones account for over 33 percent of the poor in Pakistan, but only about 29 percent of the country's total population (17.5 percent in cotton/wheat Punjab and 11.2 percent in cotton/wheat Sindh). Barani Punjab has only a little more than half its share of poor as compared to its share of the population. Relatively low poverty levels in the barani areas of northern Punjab are attributed first to certain socio-economic characteristics of barani areas, including the lowest dependency ratio, the highest levels of literacy (particularly female literacy), and the lowest number of unpaid family workers.
Second, the rural areas in this region are well integrated with prosperous urban centers that have strong linkages to the services sector.
Third, a significant proportion of the region's labour force is employed in both the armed forces and government sector.
Finally, due to the high incidence of domestic and overseas migration, remittances contribute a significantly higher proportion to total household income in the barani areas of Punjab.
One of the major contributions of this study is an analysis of income sources: it identifies five major sources in rural Pakistan, including wages/salaries, transfer income, crop income, rental income, and livestock income. Crop income accounts for 67 percent of the total income in cotton/wheat Sindh and 64 percent of the total income in cotton/wheat Punjab. These ratios are highest across the two zones, showing that the highest incidence of poverty is in zones that rely most on crop incomes.
The incidence of poverty is low in zones where the percentage of incomes from wages and salaries and transfer incomes is high. It appears that poverty is greater in zones where the possibility of diversifying incomes to manage risk is limited. Sources of income vary with poverty status.
Crop income is an important source for non-poor households, particularly in irrigated areas of Punjab, while poor households generally rely on wages and salaries.
Within the farm sector of the cotton/wheat zone, both non-poor and poor households depend on the cotton crop, although the former diversify crop production by growing sugarcane, particularly, in Sindh. The share generated by rice seems to contribute significantly to the income of poor and non-poor households in rice-growing areas of Sindh and Punjab while maize is an important source of income for households located in the north-west Frontier Province (NWFP) and in barani Punjab.
A majority of poor, rural, land less households derives their non-farm income from the construction sector where nearly half the employed persons were found to be under-employed. The services sector, on the other hand, appears to be the most important source of non-farm income for better-off households. Moreover, the wholesale and retail trade, and transport and communications sectors also contribute significantly to the non-farm incomes of non-poor households.
The findings of this study, based on data from the Household Integrated Economic Survey (HIES) for FY2002, show that income from livestock occupies a small proportion of overall household income (less than two percent), and that it increases only marginally across income quintiles. The highest livestock income was reported in the cotton/other zone of Sindh and Balochistan and the lowest in mixed Punjab.
These findings, however, are not consistent with the results of earlier studies, which show a much higher contribution of livestock income to total household income. More importantly, the national accounts show that 40 percent of the value-addition to the agriculture sector is from livestock income, and that the contribution is equal to that of crop income.

Although it is likely that the HIES does not provide an accurate valuation of livestock produce consumed at home or gifted out, much more in-depth research using all available data sets is needed on this issue.
The study also examines the dynamics of rural poverty, ie the movement into and out of poverty. While more rural households have fallen into poverty since the 1990s, many may also have escaped it.
The results of the International Food Policy Research Institute (IFPRI) five-year panel show that although 21 to 29 percent of households fall below the poverty line in any given survey year, the probability of entering poverty ranges between 0.15 and 0.24, while that of overcoming poverty varies between 0.43 and 0.51. The sample entry and exit probabilities for the five pairs of sequential years used in the study indicate an increase in the probability of moving into poverty and a decline in that of escaping it over the survey years.
Consequent to external shocks or disasters, the probability of remaining poor is much higher amongst poor households than it is for non-poor households. Chronically poor households remain poor for longer periods than those, who are transitorily poor.
With respect to the agricultural growth and poverty nexus, this study shows that the estimates of agricultural growth reported by the Pakistan Economic Surveys during the 1990s were highly overstated. The incorrect accounting of livestock, fishing and forestry value-added in some years has contributed to the high reported rate of 4.6 percent per annum during the 1990s. However, the high growth rates reported for these sectors in this period were not backed by data from the Census of Agriculture or by the desegregated data in the Pakistan Economic Surveys themselves.
After adjusting for these overstated sub-sectors growth rates, the growth rate of the agriculture sector over the 1990s declines to only 3.1 percent per annum.
Assuming a population growth rate of 2.5 percent over this period, this implies a paltry 0.6 per cent per annum per capita growth rate. With such low growth per capita, the increase in rural poverty in the 1990s was inevitable.
A major portion of agricultural growth is dependent on the production of cotton and wheat, Pakistan's two major crops. Cotton production is concentrated in a few districts of the country where land distribution is highly unequal: expanding the production of this crop would theoretically benefit large farmers. Moreover, given the large variability in the production of cotton and wheat, the growth in value-added from these two crops over the 1990s is only 1.3 percent per annum, implying a negative per capita growth in value-added.
For the small farmer in particular, the reliance on cotton with few avenues for diversification means increased poverty. Persistent drought and the lack of irrigation in these regions have also resulted in a substantial decline in cotton production, otherwise the mainstay of many households in this region. The rise in poverty in the 1990s seems to be a result of increase in the number of poor in the cotton/wheat zones of southern Punjab and Sindh.
Agricultural production in Pakistan is highly erratic; one good year may be preceded or followed by a bad year(s). The available research indicates that, for small farmers, the possibility of overcoming poverty after a bad year declines even if it is followed by a very good year.
A considerably higher incidence of transitory poverty indicates the vulnerability of households close to the poverty line. A minor shock may push them into poverty while a little assistance (in terms of employment or income) can help them escape it.
The skewed land distribution is one of the major obstacles hindering the rapid reduction of rural poverty. More than half the total farms are smaller than five acres in size. Excessive land fragmentation and the sub-division of landholdings from generation to generation are causing a persistent decline in farm size, and, therefore, in agricultural productivity. Smaller farms tend to be less diversified and so more vulnerable to poverty. In particular, the dependence of small farmers on a highly erratic cotton crop across most of the Cotton Belt seems to be a major cause of rural poverty.
It is necessary to provide small farmers with technology and information to help smooth out the variability in cotton production that they are subject to, and to give them access to technology and markets to enable them to diversify their incomes.
The highly skewed land distribution in rural Pakistan results in sharecropping, which is detrimental to poverty reduction. The incidence of poverty among sharecroppers has been found to be considerably higher than those who cultivate their own land (even small farmers) or cultivate the land of others at a fixed rent. The prevailing tenancy arrangements in different parts of the country lead to the exploitation of sharecroppers. While various tenancy laws exist, there is no proper mechanism to monitor the implementation of these laws.
Finally, mechanisation in agriculture has increased considerably; reducing the use of on-farm hired labour. Agricultural workers' real wages show a decline and this, too, has contributed significantly to the rise in poverty.