By Ali Cheema
THE emphasis on eliminating poverty through the pursuit of social justice in the prime minister’s hundred-day programme is a welcome change from President Musharraf’s policy, which favoured the trickle-down recipe of poverty alleviation.
The previous regime’s erstwhile economic gurus went about the business of calculating the percentage decline in mean poverty in relation to the increased rate of growth, which became its battle cry. The much-advertised official verdict was that poverty in Pakistan had declined by over 10 per cent in the last four years.
The merits and demerits of the official calculations notwithstanding (the veracity of the official calculations is subject to intense debate), by becoming obsessed with average reductions the previous regime’s policy approach to poverty failed to develop an appreciation of the deep structural constraints impacting poverty in Pakistan.
These structural constraints impact poverty in a number of ways. Their primary impact is the tremendous variation in household poverty that is caused at the district and sub-provincial level. These constraints have resulted in the creation of high-poverty districts that are stuck in ‘poverty traps’, where endemic poverty is persistent over the long run. The socio-economic channels through which growth ‘trickle down’ is said to happen remain extremely fragile in these districts. That is, growth alone has not and will not deliver in these districts.
Given these poverty characteristics, it is highly probable to have reductions in average poverty coexisting with endemic poverty in the high-poverty districts. This happens because growth-led poverty reduction occurs in the low-poverty districts, where channels of the trickle-down effect are strong. The upshot of this process is that the reduction in average poverty ends up widening the poverty gap between different types of households and between households in different districts, thereby compromising precepts of social justice and democracy.
This should not be read as an anti-growth argument. The argument is two-fold: (a) growth in itself is unlikely to dent endemic poverty in high-poverty districts; and (b) social justice requires that design of policy that uses the dividends of growth to mitigate and reduce endemic poverty in these districts as well. If the objective is to eliminate poverty, then policy should concern itself with reducing the large and persisting variation in poverty across districts in addition to reducing average poverty.
I illustrate this argument about sub-provincial ‘poverty traps’ with reference to Punjab using the Multiple Cluster Indicators Survey (2003-04) of the provincial government. The important point about this survey is that it is statistically representative at the district level and allows poverty inferences to be made at this level. The analysis that follows is based on work that I am currently doing with Ms Lyyla Khalid (Lahore University of Management Sciences) and Dr Naved Hamid (Lahore School of Economics).
In order to make the argument tractable I divide Punjab into four regions. The north consists of the four districts of Rawalpindi, Attock, Jhelum and Chakwal. The west consists of the districts of Mianwali, Bhakkar, Khushab, Layyah, Muzaffargarh, D.G. Khan and Rajanpur. The south consists of the districts of the old Multan division and the old Bahawalpur state. The centre contains the remaining districts of the province.
Using a poverty line based on the Economic Survey methodology we find that north Punjab has poverty head-count ratios (HCRs) of around 12 per cent and the centre has HCRs in the range of 20 per cent. In contrast, the west and the south of Punjab have HCRs of around 45 per cent, indicating the existence of poverty levels that are more than double that found in the north-centre districts of the province.
Interestingly, rural poverty calculations show a similar trend. In the north, rural poverty remains extremely low and more or less in line with overall poverty in the region. However, in the case of the centre, rural HCRs are 10 per cent higher than overall HCRs in the region at nearly 30 per cent compared to 20 per cent. In the south and the west rural HCRs are extremely high and in the range of 55 per cent.
The severity of poverty in the south-west, measured by the shortfall in household expenditure from the poverty line, is six times higher than the north and three times higher than the centre. This indicates that more than half the rural population of the south-west districts in one of Pakistan’s most developed provinces is living in conditions of endemic and abject poverty.
Nearly a decade ago, Haris Gazdar using a different dataset that was unrepresentative at the district level, found similar gaps in sub-provincial poverty in Punjab, with the south and west emerging as the endemic poverty belt of the province. In spite of the statistical issues with the dataset, it appears that he had found an essential insight into the structure of poverty in Pakistan, which has failed to find appreciation in policymaking.
Writing at the time of independence, Malcom Darling documented exactly the same sub-provincial gaps in poverty across Punjab. This suggests that the current regional gaps in poverty across Punjab appear to have persisted over the long run.
Furthermore, our analysis suggests that seven districts in Punjab have rural poverty HCRs of over 60 per cent. These include the districts of: Muzaffargarh; D.G. Khan; Rajanpur; Rahimyar Khan; Bahawalpur; Bahawalnagar and Lodhran. Together these seven districts constitute a crescent of endemic poverty at the bottom of the province. We also find that the severity of poverty tends to be the highest in this crescent.
Punjab’s endemic poverty crescent is in stark contrast to the districts of Sialkot; Jhelum; Rawalpindi; Chakwal; Gujrat; Lahore; and Attock, which have poverty HCRs of below 15 per cent and a low severity of poverty ratio. The extremes of poverty in Punjab in 2003-04 ranged from a poverty HCR of six per cent in Sialkot to HCRs of more than 65 per cent in Rajanpur.
These findings have important implications and raise important questions. The most important implication is that there are many Pakistans existing in the state of Pakistan. The people of these different Pakistans have different opportunities, aspirations, and access to different assets and endowments and are faced with different constraints. In some parts a majority of citizens are concerned with improving livelihoods, in others they are battling to survive the disease of abject poverty.
These differences in outcomes will clearly have important implications for the design of a programme of social justice in Pakistan. The findings suggest that the one-policy-fits-all approach adopted by the provincial and federal governments is unlikely to work. This will, in all likelihood, increase the poverty gap between different sub-provincial regions and different types of households.
It also suggests that on an index of abject poverty the placement of citizens from all districts is not equal and a policy emphasis on reducing averages may benefit residents of some regions more than others, as has been the case during the previous regime. Growing gaps at the sub-provincial levels are going to weaken the country politically.
The most obvious question that needs to be addressed is why does poverty tend to be endemic and persistent in the high-poverty districts and sub-provincial regions as opposed to the low-poverty districts. And, finally, what implications does this have for policy.
The writer is an associate professor of economics at the Lahore University of Management Sciences.
This piece was first published in the DAILY DAWN, PAKISTAN