The main benefits of infrastructure come from marketing of inputs and outputs. Market orientation has increased over time more in villages with developed (good roads) rather than underdeveloped roads. For example — as shown by a survey of households spreading over 62 villages, the share of marketed paddy in 1988 was 44 per cent in developed villages. But by 2008, the share shot up to 56 per cent. On the other hand, distress sales in developed villages have drastically gone down. The reason may be that factors that fuel distress sales have weakened in developed villages possibly due to infrastructural development.
The impacts of infrastructure on some key economic indicators can be captured by regression analysis. While there could some correlation between infrastructure and economic variables, the establishment of a causal link warrants that the dependent variables be regressed upon the explanatory factors. To this effect, Dr Mahabub Hossain and Abdul Bayes regressed the factors on access to paved roads and electricity. The authors have taken non-farm incomes as the dependent variables and regressed upon roads and electricity (independent variables).
The regression appears as a good fit. The adjusted R2 at 0.57 implies that, 57 per cent of the variations in non-agricultural income is influenced by independent factors. For example, non-agricultural capital and workers, the presence of member working abroad and electricity emerge as highly significant factors affecting rural non-farm income. Access to quality roads does not appear as a powerful factor in explaining the variations in income levels. However, taking the dominant variables into account, the researchers ran a second round step-wise regression. The regression function emerges as better fit showing an adjusted R2 value at 0.69 and F=415. This time, again, non-agricultural worker and non-agricultural capital continue to impinge significant impact on household income. Likewise, agricultural capital and member working abroad also showed up as significant contributors to non-farm income.
An interaction of electricity and non-farm worker (to avoid multicollinearity) shows that the variable is highly significant in explaining variations in non-farm income. It means that access to electricity enhances the productivity of rural non-farm workers and, thus, contributes to increased income. The estimated elasticity on this count is observed to be 0.17, implying that there occurs a rise of income by 17 per cent following the availability of electricity. Similarly, interaction between roads and non-farm worker also positively affects income and as per the elasticity coefficient, the income rises by 6.0 per cent due to access to roads. Overall, access to electricity and roads tend to increase household income by 23 per cent.
It is thus observed that access to quality roads and electricity (or to any of them) is likely to positively affect non-farm income of households. The policy implication of this observation is obvious: government should invest in rural infrastructure – especially in the construction of roads and providing electricity – with a view to raising the income levels of rural households. Fortunately, these findings are in consort with those for China and India.
The impact of infrastructure on poverty is also well researched. Various studies eloquently exposed the differential impacts of infrastructure on poverty level in India. According to the researchers, additional expenditure on roads is found to have the largest impact on poverty reduction as well as a significant effect on productivity growth in rural India. The access leads to larger benefits to the rural poor and emerges as dominant “win-win” strategy. The researchers suggest that the government of India should pick up roads for investment, in case of any choice has to be made in allocating scarce funds among alternative investment purposes. More specifically, they focused on agricultural research and extension of roads as the panacea to poverty reduction.
Bangladesh’s experience also stands close to the Indian experience. For example, in a survey of households spreading over 62 villages, all sample households experienced decline in extreme poverty over time, but the reduction was higher in developed villages – with good roads – compared to others with bad roads. For example, between 1988 and 2008, extreme poverty declined by 2 percentage point per annum in developed villages as opposed to roughly 1 percentage point in other villages. In fact, the poverty reduction in developed villages was higher than the estimated national average. The same trend holds true in the case of other measures of poverty.
As far as inequality of income is concerned, the results are not seemingly so encouraging. First, between 1988 and 2008, inequality in income – as reflected by the higher gini ratio of household income – has increased in all sample villages. Noticeably, however, the distribution of income over time appears to be strikingly skewed in developed villages. Maybe, the opportunities created by better roads and electricity has largely been reaped by the top income deciles in developed villages.
If we look at the sources of inequality, we observe that inequality is higher where trade and business are major sources of income. For example, in developed villages, trade and business contributed 40 per cent of the inequality (relative contribution to pseudo gini) in 2008 compared to 13 per cent in 1988. In semi-developed villages also, trade and business continued to impart a pervasive influence in generating inequality. Since, expansion of trade and business is mainly a function of financial and human capital, it is not unlikely that upper income groups in rural areas tend to overtake others to contribute to greater inequality.
On the other hand, cultivation as a source of inequality declined in all villages possibly due to the fact that small and marginal farmers are engaged in it and their share is growing over time. Again, the inequality in income from non-rice crops increased during the same period of time for all villages. But income from wage labour has been more equaliser than other sources of income. And finally, income from remittances continues to remain as the most dominant source of inequality (pseudo gini ratio 0.60-0.70) in rural areas. However, there seems to be a marginal decline in the index in developed villages compared to under-developed ones.
By and large, the fall in head count of poverty over time in Bangladesh owes partly to the growth of rural roads and networks plus increased access to electricity. While exports of RMG and remittances played a vital role in alleviating poverty in Bangladesh, the contribution of roads and electricity should not be relegated to the back seat.
Abdul Bayes is a Professor of Economics at Jahangirnagar University.