Such a trend is confirmed by Chart 2, which indicates the difference between the richest (Goa) and poorest (Bihar) states. Throughout the 1980s, the per capita NSDP in Goa was around 4.5 times that of Bihar. In the 1990s, however, this difference increased sharply and continuously, reaching a peak of nearly 11 times in 1998-99. In the subsequent and most recent decade, the ratio fell slightly but stabilised around a high level, at an average of nearly ten times.

Clearly, the economic reforms that began in 1991 were associated with processes that generated rising horizontal inequality, which peaked around the close of that decade. In the 2000s, state per capita income divergences remained high, but did not keep increasing.

This was then reflected also in broader regional differences as well. Chart 3 groups the states into five major regions as follows:

Some of this can be related to the nature of the aggregate growth process in the country from the 1990s, which was heavily biased in favour of corporate expansion. The regions with a greater spread of large capital in organised activities – such as the western and southern regions, which include the states of Maharashtra and Tamil Nadu respectively – therefore showed more rapid growth in per capita incomes.

This process can be further unbundled into an examination of the growth rates of per capita income by decade for the poorest and richest states, so as to provide more insights into what exactly was happening in these three periods. Chart 4 provides data on annual compound growth rates of per capita NSDP, calculated by taking three year averages of the beginning and end of each decade. The six poorest large states are those that have been referred to as ''BIMAROU'', while the richest states include the western states of Goa and Maharashtra and the northern states of Punjab and Haryana.

Several interesting points emerge from this chart. First, growth rates of per capita income in the 1980s were broadly similar across the richest and poorest states, at between 2 and 3 per cent per annum. Although Assam experienced slightly lower growth and Goa and Maharashtra slightly higher growth in this decade, the differences were not large. In the 1990s, Goa and Maharashtra in particular grew much faster than the poorer states, accentuating the gap. Indeed, in this period Bihar and Assam showed stagnation/decline in per capita incomes, while Uttar Pradesh also had slow growth. However, growth also slowed down in Punjab and Haryana.

The most recent decade indicates an acceleration in expansion of per capita incomes across all of these states. While the fastest growth was experienced in Maharashtra, surprisingly some of the poorer states – particularly Orissa followed by Bihar – also had relatively rapid growth. The base effect meant that this did not translate into reduction in state-wise inequalities, but therefore they also did not increase further.

Obviously, increasing per capita income need not translate into better performance in terms of poverty reduction if the growth within the state has been unequally distributed. However, in fact Orissa does also show significant reductions in poverty as well, according to the latest National Sample Survey of 2009-10.

One interesting feature of the past decade in particular has been that this rapid aggregate growth has generated a change in the relative ranking of states. (Of course the caveat that these NSDP figures are not strictly comparable across states should be borne in mind here.) In particular, some of the previously middle-income states have grown rapidly enough in the last decade to overtake Punjab, which earlier had the second highest per capita NSDP in the country after Goa.

Chart 5 shows that five states have overtaken Punjab in terms of per capita income by 2009-10. Both Maharashtra and Haryana now have significantly higher per capita NSDP, though that is not really so surprising since they were both relatively high income states even in the 1980s and they both overtook Punjab at the turn of the century.

But there are other surprises, particularly Tamil Nadu and Kerala. In 1980-81, per capita income in Punjab was 56 per cent higher than in Tamil Nadu and 27 per cent higher than in Kerala – quite substantial differences. The advantage of Punjab remained through most of the period, really until the middle of the 2000s. However, in the latter half of the 2000s, faster growth in both of these states meant that they now have higher per capita incomes, with the difference between 6-8 per cent.

This is not really due to income stagnation in Punjab, since Chart 5 indicates that even in Punjab there was an acceleration of growth from around 2005-06. Rather, it was because growth in these other states was even faster from the early part of the last decade.

What explains this movement, and these differing trends across the decades? Obviously this is a complex issue and many factors would have contributed, both at an all-India level as well as state-specific factors. Much more research is required to delve into the causes of these varying trends. However, some broad hypotheses can be formulated.

The initial period of economic reform was one in which the state – at both Central and State levels – significantly reduced its own spending on both consumption and investment as shares of GDP. This is confirmed by Chart 6, which provides aggregate public spending data from the national accounts. The various liberalisation measures introduced in the early 1990s generated a much greater role for private investment, which did actually rise to fill the gap, but did so in a way that reinforced or aggravated existing regional inequalities. This can be expected, since market incentives tend to follow the hysteresis created by earlier patterns of investment and thereby lead to enhanced regional (or state-wise) concentration of economic activity.

The latest decade has been rather different, however, because it has been marked particularly by some revival of public investment, as Chart 6 illustrates. Public capital investment (by both Centre and States) fell continuously as a share of GDP from the peak of nearly 12 per cent in 1986-87 to as low as 6.1 per cent by 2002-03 (just above half of the previous peak rate). Since then there has been some recovery and increase, such that by 2009-10, the rate (at just above 9 per cent) was similar to that achieved in the early 1990s before the economic reform programme began in earnest.

If this argument is developed, it can then be surmised that an important means of reducing regional and spatial income differences is through increasing public investment. This also leads to a somewhat different understanding of the nature of the recent aggregate economic success of the country as a whole: from the generally acclaimed but somewhat simplistic role ascribed to private investment, to a more balanced and nuanced appreciation of the important role of public spending.

*This article was originally published in the Business Line, 14 May 2012, and is available at
http://www.thehindubusinessline.com/opinion/columns/c-p-chandrasekhar/arti
cle3418631.ece