Explaining the U-curve in low-income countries using unique data
In low-income countries like India, Indonesia, China, Tanzania, and Colombia, agriculture is dominated by farms that are, on average, much smaller than those in more developed countries, where farming is typically on a large scale. The relationship between farm productivity and farm size also shows a global pattern: the smallest farms in low-income countries are more productive than farms that are slightly bigger (decreasing returns to scale, in technical terms), while in the developed world, productivity increases with scale.
While the inverse relationship between farm size and productivity in low-income countries is well-established, a new study by Mark Rosenzweig, the Frank Altschul Professor of International Economics and an EGC affiliate, and coauthor Andrew D. Foster of Brown University, shows that the U-shape relationship between scale and productivity that is observed across richer and poorer countries occurs within a single low-income country – that is, productivity falls as farm size increases from its smallest unit, and then rises as farm size increases, after a threshold. What has remained a puzzle until now is what drives this pattern: why is the relationship between farm scale and productivity U-shaped? Why are small farms more productive than those which are slightly bigger but far less productive than the larger farms observed in high-income countries?
The authors take advantage of unique data from an Indian panel survey, which uniquely contained a substantial fraction of larger farms, to show that this U-curve relationship is driven by two factors: the transaction costs of labor (i.e. the cost of hiring additional farm workers) and the scale-economies of machine capacities.
Explained in simple terms: on a very small farm, family members work the land, participate in wage work on other farms, and operate their farm efficiently. As farm size increases, the family members increase their level of on-farm work until they are unable to spend any more time working (this point of complete self-sufficiency is known as autarky). Because hiring additional labor comes with additional transaction costs (finding workers and travel costs) and thus lowers net income, the family continues to work the land as farm size increases (decreasing labor to land ratio, or productivity) until such point that the benefit of hiring additional workers outweighs the cost, and productivity starts to increase – which is where we observe the bottom of the curve. After this point, productivity rises with farm size, as larger farms can take advantage of machines that have higher capacity at larger scales and lower labor use – mirroring the economies of scale that are well-observed in developed countries.
Increasing farm size could expand agricultural output and reduce poverty
The discovery of this U-shaped pattern within a single context means that what researchers had previously understood about farming in low-income countries was wrong: that is, the smallest farms are not necessarily the most productive. In short, there are too many small farms – especially in low-income countries. The research implies that increasing farm size through consolidation would ultimately increase agricultural output, while substantially reducing poverty in agriculture. The authors illustrate this point using their model to show that the minimum farm size in India that would maximize return on land using only locally- available machinery is 24.5 acres – over 7 times the mean farm size in India today. If all farms were this size, the number of farms in India would decrease by 82%, total output would increase by 42% and income per farm worker would increase by 68%. However, there would only be a 16% decline in the total agricultural workforce, owing to the under-utilization of labor in most small-scale farms.
A graph from Foster and Rosenzweig (2021) showing real profits per acre by owned area in India, with the different lines taking into account differences in plot quality and farmer characteristics. Data come from the International Crops Research Institute for the Semi-Arid Tropics (ICRISAT) survey of India, 2009-14.
The authors call attention to the fact that their findings are consistent with a study the government of India made over 70 years ago. In 1947, the United Provinces Zamindari Commission concluded that there were too many small farms in India and that this had a negative effect on the country’s output and poverty levels. Today, however, India’s government is in support of the large number of small land-holdings, and as recently as last year, the Minister of Agriculture stated that while Indian farms are smaller than they have ever been, output is increasing so that there is no evidence that small scale farming is not a cause of the low level of agricultural productivity. Until now.