Optimal Spatial Policy: Policies for low-wage areas in India

A paper in last year's edition of Quarterly Journal of Economics (QJE) by Fajgelbaum and Gaubert (2020) caught my eye (and brought me my first blog post!). The paper asks if the existing distribution of productive activities and workers within countries is optimal. By optimal, economists are usually referring to Pareto efficiency, a state where no individual can be made better off without making someone else worse off. Spatial concentration of economic activity and regional divergence has been the norm in the 21^st century. Agglomeration literature within economics emphasize the benefits of firms producing similar or complementary goods locating close to each other and this is the standard justification for the spatial concentration seen within large economies. Some within this literature have expressed the view that agglomeration gains in some areas have come at the expense of agglomeration losses in other areas (see Kline and Moretti, 2014). India itself seems to have witnessed such a pattern right from the 19th century, when agrarian decline occured simultaneously with the emergence of prosperous commercial centres like Kolkata, Mumbai and Varanasi. Today, the problem of concentration of economic activity in a few states in India is without question a major area of policy correction. Countless redistributive policies have provided short term relief but fallen prey to capacity constraints in the areas of governance. This is directly connected to lack of high-skill workers in low-wage areas whether in government or private firms. This paper suggests sorting of workers so that high skill workers raise productivity in low-skill areas. We will first try to understand the model of spatial sorting, its predictions and policy recomendations. Following that, I will reflect on the Indian scenario.

Did the writers of Swades make an economically sound argument concerning spatial disparities?

Central to the model of optimal spatial sorting in the paper, is the idea that the consumer-worker of of a certain skill level generates efficiency and utility spillovers onto other consumer-workers within the city. This could occur through gains in productivity and the simultaneous increase in public amenities, congestion, pollution and the cost of living. Efficiency gains occur because high-skill workers generate positive spillovers onto low-skill workers and also amongst each other. Gains in efficiency and public amenities increase utility, whereas an increase in congestion reduces utility. Under optimal conditions, the net change in utility due to these spillovers is balanced out by the opportunity cost of attracting workers to a location. For eg. the opportunity cost of attracting a high-skill worker in a low wage area to a high wage area is greater because of the positive spillover the worker can have on low-skill workers in the low wage area. 

These productivity and efficiency spillovers are not internalised (i.e. taken into their decision functions) by firms and workers and so the real-life sorting of heterogeneous workers can be sub-optimal. Productivity gains of new workers then come at the cost of negative spillovers from increase in cost of living and pollution/congestion along with simultaneous decline in areas from where high-skill workers migrate out.  This can occur even while allowing for income-redistributive policies run by the government and remittances sent home by workers. The authors, therefore, recommend a policy of labour wage subsidies and income taxes wherein, workers in high-wage areas pay higher income taxes which, are then used for making transfers to workers as per their skill-type and location, as a form of wage subsidy. For eg. the model predicts a higher optimal subsidy for workers of a certain type that are scarce within a region. Such a system will lead to welfare gains and an equilibrium that features lesser wage inequality across regions and a greater proportion of high-skill workers in smaller cities. It is important to note here, that the predictions of the model don't necessarily imply that a NASA astrophysicist could be incentivised to live in a rust-belt region (recall Shah Rukh Khan in Swades?). The authors incorporate the industrial mix of different regions and the corresponding region-wise labour-efficiency funtions in predicting optimal allocation.

Reflecting on the Indian scenario - India has just a few large metropolises and they are some of the most crowded cities in the world. These cities have a much higher concentration of firms and skilled workers than the rest of the country. Congestion costs in these big cities are likely to be going up, whereas GDP growth has been slower, in recent years. It may be possible that the current spatial allocation of economic activity in the country is sub-optimal. If that is so, it is worthwhile to see whether the policy recommendation by Fajgelbaum and Gaubert could have an impact on welfare. To enable high-skill workers to work in low-wage areas, the policy will have to be able to cater to such workers' aspirations which, may imply a very high wage subsidy and infrastructure spend. Higher rates of income tax in large cities may reduce profitability of key firms, as firms will have to pay more to retain talent. At the moment, only public-sector wages can compensate high-skill workers in low-wage areas. This is not ideal for sustainability of such a policy. The industry mix in non-metropolitan regions may not be ideal for implementing such a policy in India. However, the paper does provide an improved framework to understand the dynamics of India's economic geography and its optimality. A very exciting and important contribution to spatial economic literature!