In rural Bangladesh, mobility of farmers between farm and non-farm activities has increased. This mobility is from farms to non-farms and not the other way round. As far as primary occupation is concerned, the proportion of cultivators has declined substantially and that of agricultural wage labourers has become almost half. Among rural non-farm activities (RNFA), the occupations that attracted most are rickshaw/van pulling and self-employed mechanical activities. The findings from a panel data show that expansion of rural roads, advent of shallow tube wells and power tillers etc. have been increasingly creating employment opportunities in rural areas. It is also shown that about one-tenths of the earners depend on these two sources for earnings as against a insignificant per cent decades back. Besides, services also significantly attracted one-fifths of the earners as against 15 per cent in earlier periods. By and large, unemployment in rural areas is now estimated to be 3.0-4.0 per cent, compared to about 7.0-8.0 per cent in earlier periods.
The occupational mobility from farm to non-farm activities has partly been facilitated by improvement in quality of human capital. The various surveys conducted so far found notable increase in school enrolment for children, and increase in average years of schooling for  working age population. But not surprisingly, those who have no formal schooling continue to be employed in farming (if the household owns land) or work as agricultural wage labourers (from the landless and marginal landowning households). However, those, who have attended or passed secondary schools, appear to have been pulled out of cultivation to be engaged in relatively more remunerative services or business ventures.
An important issue in the context of poverty alleviation is dependence of the resource-poor households on RNFA. In the base year in 1980s, 41 per cent of workers from the functionally landless households were engaged as agricultural wage labourers. At that time, only about 13 per cent of households reported farming as principal occupation. Over time, however, the land-owning households began to leave lands and, depending on attained education and accessing credit from institutional and non-institutional sources, went out in search of business and other sources of income. Meantime, a part of their cultivated lands went to the hands of the functionally landless households as rented-out lands which, in course of time, embraced these lands as one of the important sources of livelihoods. But some of them also moved out from agriculture and took up RNFA for livelihoods. For example, the proportion of workers engaged in rural non-farm activities has increased from 42 per cent in 1988 to roughly 60 per cent in recent years. But while the destinations of the workers from the land-owing households had been from cultivation to services and trading, those of the landless households had been from agricultural labour to low-paid non-agricultural activities such as transport operation, construction work and wage labour in agro-processing enterprises. In this case, lack of education and access to finance appear to be the major constraints for land-poor households. The lack of access to education and credit, possibly, confined them to low productive points. That is, they have fallen into the vicious circle of poverty with low wage, low income and low productivity cycle.
As mentioned earlier, with the passage of time, poor households in rural Bangladesh have increasingly been linked to RNFA for livelihoods. This is not surprising given their perceived fluctuations in earnings following from swings in output levels. Thus, more often than not, poor households embrace multiple occupations as a shield against fluctuations in household income. To drive home the point, a comparison of our observations pertaining to RNFA with those in other areas needs a mention. It could be observed that manufacturing and services such as weaving, pottery, food preparation and processing, domestic and personal services and unskilled non-farm wage labour etc. typically require low investment but account for a greater share of income for the rural poor households than their wealthier counterparts. The reverse, however, holds true for transport, commerce and upgraded manufacturing etc. where higher investments result in higher income for the wealthier households. In another finding on RNFA, some economists emphasised non-farm income for the poor as a means to help household income firmed up and stabilise in drought years. In a similar vein, some others, in reference to several villages in the semi-arid tropics of India, found that non-agricultural self-employment and labour market earnings became increasingly important source of income during the 1980s by increasing mean income and dampening household income variability.
The estimates of the total number of days of employment in non-farm activities reveal an increase in the duration of employment for most of non-agricultural activities. The duration of employment for agricultural labour had been declining over time. Assuming 6 hours as full-time employment, it could be estimated that the degree of under-employment marginally declined from 30 per cent to 28 per cent.
In this context, we invoke the 'push' and the 'pull' debate. Questions are raised whether the surplus labour and low productivity in agriculture is pushing labour towards relatively low productive RNFA or not. In this case, the rural non-farm sector may be identified as the 'employer of last resort'. But as evidences reveal, it is not only the landless lacking employment in agriculture who have stepped into rural non-farm activities but also the households with better resource positions.
The estimates of labour productivity obtained from one survey shows that productivity of RNFA is 40 to 50 per cent higher than the agricultural wage rate even in non-farm activities that needs very little physical and human capital, such as construction work, rickshaw pulling and shop keeping. In services and business enterprises, average labour productivity is observed to be 2 to 2.5 times higher than the agricultural wage rate. The labour productivity in business and service sector activities is, however, substantially lower for workers belonging to the functionally landless households than for those belonging to the medium and large landowning households. However, as it appears from our data, the resource-poor households are engaged in trade and services at the lower end of the productivity scale, presumably due to lack of access to capital and education.
Since productivity of non-farm labour is higher than agricultural wage rate, the mobility of the poor workers from agriculture to the non-farm sector is contributing to an increase in productivity and earnings. In this case, possibly, we can hypothesise that there is evidently a 'pull' situation where the non-farm sector is luring the labour to relatively higher paid jobs. In other words, observations from surveys lead to the impression that as opposed to earlier years when labour was 'pushed out' of agriculture to embrace low-productive non-farm jobs, non-farm opportunities are in fact 'pulling' labour from agriculture in recent times.
The writer, a former Professor of Economics at Jahangirnagar University, is Chair, Department of Economics and Social Science (ESS), BRAC University. abdul.bayes@brac.net/
abdulbayes@yahoo.com
Â