Living standards in Indian states have been diverging for nearly three decades and it is the poorer regions that are bearing the brunt. Delhi, Goa and other states with high per capita income in 1980 have enjoyed high growth rates over the past 27 years, while Bihar, Assam and other low per capita income states are languishing. Such “unbalanced growth” is worrisome.
While there is little consensus on the fundamental causes of divergence, economists know more about the proximate causes. For instance, rich states spend more on infrastructure than poorer states. Not only this, recent empirical work in the Indian context shows that the richer states are rich because of higher infrastructure spending.
To get a sense of the raw correlation, the diagram plots an infrastructure index relative to real per capita state domestic product for several Indian states over 1980–2000. It is based on state-wise development spending — a proxy for social and physical infrastructure. This incorporates both economic and social expenditure categories ranging from education, water supply and sanitation, urban development, agriculture, and transport and communication. The regression line shows a positive (and significant) correlation between development expenditures and real per capita income across states in the 1980s and 1990s.
So, why do richer states spend more in infrastructure? And if more infrastructure spending leads to higher growth, what policy changes would incentivize politicians to spend more on it?
In my recent research (Journal of Economic Policy Reform, March 2008), I construct a simple model that offers a potential explanation for why infrastructure investments across Indian states vary and how this can lead to divergence in real per capita incomes. Divergence boils down to: how big the middle class is, the availability of formal sector finance, and corruption.
The basic idea is simple. People in the economy can invest in one of two sectors: a low return yielding subsistence sector (say, agriculture), and a modern formal sector of the economy (say, a small enterprise). The returns on investing in the modern sector depend on the stock of infrastructure, such as the quality of roads or availability of electricity. More infrastructure raises the returns from investing in the modern sector. However, to start a modern sector project requires people to incur a fixed cost, such as getting clearances for a project, or other shoe leather costs. If the fixed costs of investing in the modern sector are prohibitively high, people would opt to work in the subsistence sector even if the financial returns for investing in the modern sector are high. In contrast, the returns on investing in the subsistence sector do not depend on the stock of infrastructure. However, if people invest in the subsistence sector, they receive a consumption subsidy. Politicians finance infrastructure and the consumption subsidy in the economy by taxing modern sector profits. If a greater proportion of tax revenues is spent on the subsidy, less is left for infrastructure and, therefore, the more likely it is for people to invest in the subsistence sector since less infrastructure depresses the rate of returns on investing in the modern sector.
Since people may not have enough wealth to invest in the fixed costs of setting up a small enterprise, they could borrow from a local bank. Essentially, which sector people choose to invest in depends on their wealth levels: If they are poor, they invest in the subsistence sector. But if they are rich, they will take out a loan and invest in the modern sector.
Given that the sole objective of politicians and political parties is to win elections, what proportion of tax revenues will politicians spend on infrastructure and the subsidy?
The key insight from the analysis is that it will depend on the wealth of the median voter. If the median voter is rich enough to invest in the modern sector — where the quality of infrastructure determines the returns on investing in this sector — politicians spend tax revenues on infrastructure and nothing on the unproductive subsidy.
However, if the median voter is poor, or the fixed costs are high, and so he cannot invest in the modern sector, then he invests in the subsistence activity, and the analysis shows that the politicians’ election winning bundle is a high-subsidy-low-public good bundle.
What is important is that median voter wealth determines this outcome. This means that if the middle class of a state is sufficiently large, so that the median voter is more likely to lie in the middle of the distribution, this leads to more infrastructure provision and less expenditures on the unproductive subsidy. For states in which there is a large proportion of the population that is poor, the median voter is likely to have low wealth and the optimal policy mix involves a high subsidy and low provision of public goods such as infrastructure.
Casual observation seems to accord with this prediction. States with large middle class populations such as Delhi and Goa spend more on infrastructure.
But what happens over time? If the median voter’s wealth is low and the politicians’ election winning bundle is a high-subsidy-low-infrastructure bundle, low public investment makes it more likely that the median voter is in the subsistence sector in the next period as well. This is because high subsidies crowd out public investment, creating less opportunities for income generation, and lead to a low-wealth -level median voter in future periods. So, policies persist and the state economy has persistent levels of low public investment.
This basic insight can be extended to examine the role of corruption on divergence in per capita incomes as well. I interpret corruption as a leakage: i.e., what percentage of expenditures on infrastructure gets expropriated in the process of disbursement. It is well known in India that there is enormous leakage in the implementation of public investment programmes. Rajiv Gandhi famously quipped that only 15% of the intended disbursement reaches the intended recipients. Corruption means less is spent on infrastructure, making the median voter opt for the subsistence activity. Once again, the high-unproductive-subsidy-low-infrastructure policy persists because of corruption.
There are two main policy implications. The way to incentivize political parties to spend more tax revenues on infrastructure is to: 1) reduce the fixed cost of accessing the modern sector, and 2) reduce corruption. Both would allow people to invest in the modern sector of the economy which, in turn, induces politicians to spend more on public investment and less on subsidies as an election winning strategy.
For instance, the 2006 World Bank Development Policy Review noted “the vast majority of India’s poor still don’t have access to formal finance”. The analysis suggests that the relevant policy bundle is one where infrastructure provision is made in conjunction with a credit policy that reduces the fixed costs of investing in the formal sector. Hence, infrastructure provision and enhancing the reach of formal credit markets are strong policy complements. The analysis also suggests that eliminating rampant, institutionalized corruption would incentivize political parties to spend more on productive public investment and less on unproductive subsidies.
Chetan Ghate is an assistant professor, planning unit, Indian Statistical Institute, New Delhi. Comment at firstname.lastname@example.org