|Watch those states carefully|
|Monday, 04 April 2011 00:00|
Economists have argued, for a long time, that it is economic growth that determines the outcomes of elections. The higher the growth, the higher the chances of a government getting re-elected. While that could explain the Congress party coming back to power in 2009 with 42% more seats than it got in 2004 (from 145 in 2004 to 206 in 2009), how does it explain the BJP losing so badly in 2004 (its seats tally fell from 182 in 1999 to 138 in 2004) despite all the hype over higher growth and India Shining? Former chief economic advisor, Arvind Virmani, one of the big proponents of the economy-leads-to-votes theory, explains this by arguing that GDP growth during the BJP years was actually lower than the average of the previous period—as compared to 6.7% between 1994-95 and 1998-99, growth in the Vajpayee years was a lower 5.6%; others refine the argument by bringing in inflation, jobs … A new study,
by Poonam Gupta at ICRIER and Arvind Panagariya at Columbia University, brings in an altogether new
dimension, one that could be more than mildly worrying for the ruling Congress party in the context of how it has done in various state elections in the recent past.
Gupta and Panagariya, like the other economists, begin by proving that higher GDP growth results in higher chances of winning elections. So, using data for the 2009 general elections, they say incumbent parties in high-growth states won 85% of the seats they contested. In contrast, they say, incumbents won just 52% and 40% of the seats they contested in medium- and low-growth states. From the old days where Prannoy Roy and Vinod Dua made ‘anti-incumbency factor’ a household term, Gupta-Panagariya are tying to do the same with ‘pro-incumbency factor’!
But, and here’s the twist, the duo bring in not just incumbency at the national level, but argue that state-level incumbency has an effect on the national or all-India results. So, while the older economic models (Arvind Virmani, Surjit Bhalla, among others) would say the Congress would win more seats if GDP growth picked up, Gupta-Panagariya argue that if growth in Bihar is faster than that for all-India, the Lok Sabha seats in Bihar will go to Nitish Kumar’s JDU and not to the Congress. This is why, they point out, that even while India’s GDP growth picked up in 2009 when the Congress was in power at the Centre, the party managed to win just nine of out 72 seats in the states of Bihar, Orissa and Chhattisgarh—these states, whose growth was rising, were all ruled by non-Congress parties.
Gupta-Panagariya argue that the wealth of candidates is important, as is their criminal record (the richer you are, and the more criminal cases you have, the greater are the chances of your winning!), but there’s a caveat to this. Much of the individual characteristics really apply to low-growth states; in the higher-growth states, what matters more is whether the candidates are affiliated to a major national party, and whether that party is the incumbent one or not. All other factors being constant, they say, a candidate belonging to the state-incumbent party has a 35-40% greater chance of being elected than the candidates of other parties. This relationship is greater in faster-growth states and non-existent in the slowest-growth ones.
If the growth of GDP in a state is 1% higher than the national average for the last few years, their regression analysis tells them, the candidates of state-incumbent parties have a 6 percentage point higher probability of winning than do candidates of non-incumbents. If the growth difference is 2 percentage points, this probability rises by 12 percentage points.
So, if the Congress is not in power in states like Uttar Pradesh, Bihar, Madhya Pradesh, Orissa, Gujarat and Karnataka, and these states manage to register higher growth than the national average, the Gupta-Panagariya model would suggest the seats here will go to opposition parties.
Of course, there is some hope here as well. The authors have tried to map the Congress party’s extra 61 seats in the 2009 elections in terms of an incumbency/non-incumbency and high-growth/medium-growth/low-growth matrix. They find the Congress won 1 extra seat in a high-growth state that it was in power in; it won just 2 extra seats in a non-Congress high-growth state; it won 40 seats in low-growth states where it was not the incumbent party (Prannoy’s anti-incumbency factor is alive and kicking in low-growth states!)—it won 8 seats more in the medium-growth states that it was an incumbent for and 10 in the medium-growth states it was not the incumbent party. In which case, the Congress’s big hope lies in states like UP not being able to up their growth game.
None of this is to suggest the model is infallible—JDU MP NK Singh, who was in the audience where the model was being presented, made a valid point when he said that the model would be really robust if it could explain why, despite all the years of non-performance, Lalu Prasad managed to get re-elected. Certainly, it would be interesting to see how the model takes care of coalitions, and it would be interesting to see whether voters are reacting to GDP growth or whether they respond more to public works programmes (Virmani’s model takes care of this); the model seems to assume inflation is irrelevant and that the distribution effects of growth are not as important as many would think, and it seems to assume caste is not as important as our politicians make it out to be. This may well be true, but could just as well be wishful thinking. The model is young, and needs a lot more testing. But there can be little doubt politicians who ignore growth are being short-sighted. That’s the lesson politicians from both the Congress and the BJP would do well to keep in mind.