Policy makers in developing countries have long emphasized improvements in agricultural productivity as a central strategy for promoting rural development. Ultimately, however, economic development hinges upon firm creation and shifting employment from the agricultural to the manufacturing and service sectors, a process often linked to urbanization and migration (Johnston and Mellor, 1961, Lewis, 1954, Gollin et al., 2002, Rostow, 1960, Kuznets, 1961, Studwell, 2013). It is, therefore, crucial to understand how gains in agricultural productivity impact non-agricultural development.

The paper ‘Irrigation and the spatial pattern of local economic development in India’ in the Journal of Development Economics studies the effects of long-term agricultural productivity shocks on local economic development in India. Access to irrigation is considered to be key to agricultural development, and the establishment of irrigation infrastructure has long been one of the leading forms of public investment in the sector in much of the developing world, and particularly so in India.

Since 1950, the Indian government has extended irrigation to close to 250,000 villages through the construction of large-scale dams and networks of canals that distribute river water to downstream villages. The analysis reveals that this widespread transformation had heterogeneous impacts on local economic development.

It studies the long-term impact of large-scale irrigation infrastructure on the composition of local economic activity in India.

The analysis uses high-resolution spatial data covering approximately 150,000 villages and towns and exploits spatial discontinuities in the coverage of irrigation projects. These results highlight the heterogeneous impacts that agricultural productivity gains can have on the patterns local economic development.

On the one hand, it provides evidence that these irrigation projects had a positive impact on agricultural productivity in villages by allowing them to expand crop production to seasons when it had previously been nonviable.

It furthermore shows that these areas experienced increases in population density and indicators of economic development (assets and nightlights), but there were no significant changes in the share of agricultural workers and only a modest increase in small-firm activity. 

The effects of irrigation in towns, however, were notably different. Towns differ from villages in being larger, more densely populated, and having economies oriented towards non-agricultural production and trade. The paper shows that towns located within project areas experienced declines in population density and indicators of economic development. Importantly, towns also experienced a substantial decline in the scale of manufacturing activity and the presence of large firms, as well as a shift in the labor force away from non-agricultural employment. 

The analysis is guided by a stylized, spatial economy model with endogenous manufacturing productivity growth that illustrates the effects of an agricultural productivity shock on the spatial and sectoral allocation of workers. The model indicates that these effects depend substantially on the type of region that is hit by the shock. A positive, permanent agricultural productivity shock slows down productivity growth in the manufacturing sector, which in the long-run can generate a reduction in population and real wages relative to a scenario with no shock, akin to Matsuyama (1992).

The extent to which this long-run mechanism dominates the short-run benefits of the agricultural productivity gain depends on the size of the manufacturing sector in a region, and the degree of labor mobility between regions of the country. As a result of the shock, regions that are primarily agricultural tend to have an increase in population in the long-run despite the slower manufacturing productivity growth, whereas more urbanized regions tend to experience the opposite effect.

It makes use of fine spatial data on more than 1500 major surface irrigation projects in India, which we merge with administrative village-level agricultural, demographic, and economic data, as well as remotely sensed land-use data. The boundaries of the areas served by these irrigation projects (called “command areas”) are primarily determined by engineering considerations related to topography. It exploits the discontinuity in program inclusion arising at the boundary of command areas, comparing locations proximate to one another but on opposite sides of the boundary, while controlling for geographic features and imposing sampling restrictions to ensure comparability.

The paper joins a growing literature on the causal impact of different forms of agricultural productivity gains on the broader economy, both across (McArthur and McCord, 2017, Gollin et al., 2021) and within (Hornbeck and Keskin, 2015, Bustos et al., 2016, Bustos et al., 2020) countries. In the context of India, Foster and Rosenzweig (2004) show that high rates of crop yield growth in India are correlated with lower industrial growth across a nationally representative sample of villages.

While existing studies conduct their analysis at relatively high levels of administrative aggregation, we contribute to the literature by using higher resolution data to show that impacts can dramatically vary in space and by baseline levels of urbanization.

It also contributes to the literature on the impacts of large irrigation infrastructure. Since the seminal work of Duflo and Pande (2007), a handful of papers have studied the impacts of surface irrigation projects on downstream areas, generally relying on exogenous variation in the geographical determinants of dam location for causal identification (Hansen et al., 2011, Strobl and Strobl, 2011, Blanc and Strobl, 2014, Olmstead and Sigman, 2015, Jones et al., 2022, Dillon and Fishman, 2019, Zaveri et al., 2020).

These studies have documented important effects of irrigation on agricultural output, income volatility, and poverty rates, but have not investigated impacts on non-agricultural economic activity, which is a primary focus of this paper. 

Two concurrent papers that also examine the impact of irrigation in India are worth highlighting. Boudot-Reddy and Butler (2021) examine the impact of groundwater (well) irrigation and find that it increases agricultural production and consumption, but does not re-allocate labor across sectors. Asher et al. (2021) study the impacts of canal irrigation using a spatial RDD, with elevation relative to the canal as the running variable, and similarly find positive impacts on agricultural productivity, population density, and measures of economic development in rural areas. 

The key distinction between the approach of these other papers and this one is in the treatment of towns. Boudot-Reddy and Butler (2021) focus exclusively on villages; while (Asher et al., 2021) do not distinguish between towns and villages in their RDD analysis, yielding results that are dominated by village effects, and provide a separate analysis estimating impacts on the India-wide distribution of town location and growth. In contrast, we show, both theoretically and empirically, that the effects of irrigation on the economic development of villages and towns are strikingly different.

In section 2, it presents a stylized spatial economy model to derive predictions for our empirical analysis. Data and summary statistics are discussed in Section 3, followed by the empirical strategy in Section 4. Finally, it presents the results in Section 5 and conclusion in Section 6.

The full article can be viewed here