Why the voter146;s wise

Non-economic factors, while obviously and always important, can8217;t be automatically assumed to be the determining factor. Economics gets far less detailed attention than it deserves in Indian political analysis. If the voter is wise he will never ignore the economic context.

Madhya Pradesh is a good pick for two reasons. As in Delhi and Chhattisgarh, the incumbent won. And the verdict was strongly bipolar in terms of votes cast, not just seats won. The BJP and the Congress accounted for almost 75 per cent of votes. The BSP was more marginal in MP than in other states. If the two major parties attract most votes cast in constituencies, state-wise economic data has better analytical relevance. This is because state-wise party vote share is then a good analytical approximation of the verdict, which is of course determined by constituency-wise first past the post rule.

But MP beats Delhi and Chhattisgarh in being economically typical. The state is big and relatively poor. Agriculture dominates economic activity in MP slightly more than it does at the level of national averages. NSSO data classifying households by economic activity shows that in MP, of every 1,000 households, 479 report farming as chief activity, 94 are in non-farm category and 322 in farm labour category. The national averages for these categories are 384, 170 and 236, respectively.

So what was the economic context of the typical MP voter? To answer this, an important sub-question needs to be asked. What was the time horizon of the typical voter? That is, did he judge the incumbent on the basis of economic performance over the last few months leading up to the elections or its performance over the whole term?

We in the media have short-term memory and we assume so do voters. But a fascinating research paper by Stuti Khemani, published under the aegis of the World Bank8217;s Development Research Group, indicates that in state elections voters may be judging incumbents by their performance over the whole term. Voters react more reflexively to election-time contexts in national elections.

Khemani argues this is because voters judge state governments, which have large developmental roles, to be more directly responsible for their quality of life. Their votes in state elections are therefore more 8220;rational8221; and less 8220;myopic8221;. Khemani8217;s data set spans 1960 to 1992 for 14 major states. There8217;s no reason to assume these premises and conclusions on voting behaviour have changed since then. Indeed, post-reforms state governments have in general become more economics-conscious.

MP8217;s voters did face some not-so-happy economic contexts just before the election. CMIE data from the organisation8217;s Monthly Review of States in India shows that southwest monsoon started well in the state but faltered after June. This affected kharif sowing operations. Most crops registered a lower acreage than last year. Foodgrain distribution through subsidised schemes also registered a fall in the months before the election. In a relatively poor state with heavy dependence on agriculture, these are not pro-incumbency economic contexts. Unless we think, as Khemani did, that MP voters were thinking of the BJP8217;s whole five years. In that time frame, these data don8217;t show up the government badly.

The time frame issue is sharper for inflation data. We should look at consumer price indices CPI, which economists say are statistically better constructed than the wholesale price index and which capture consumer-relevant prices better. Inflation as measured by state-level CPI for agricultural labour went up from 1.2 per cent in 2003-04 to 8.7 per cent in April-July of this year the latest figures available. CPI for MP8217;s industrial workers went up in the same period from 2.0 per cent to 8.8 per cent. In the months before elections, therefore, MP8217;s aam voter did face high consumer prices. But the five-year averages of farm and non-farm CPIs for workers are around a moderate 6.5 per cent and the first three years of the BJP8217;s term saw below-average inflation. Over the whole term, MP8217;s voters didn8217;t suffer high inflation.

There8217;s some academic debate on whether voters blame state governments for inflation. But it8217;s hard to think of voters only blaming the Centre for high prices. State governments can8217;t remain unaffected. To the extent MP8217;s voters judged the incumbent on prices, the time horizon argument may indicate why inflation didn8217;t hurt the incumbent.

And when you look at roads 8212; good roads are paved with positive voting intentions because better connectivity radically changes aam voters8217; economic opportunities 8212; another clue to MP8217;s verdict may be found.

After a bad start, the state government rapidly expanded expenditure on roads 8212; from around Rs 140 crore in 2005 to almost Rs 600 crore in 2008. Spending on transport and communications, road projects come under this head, registered the second-highest compounded annual growth rate, 44.6 per cent, in the MP government8217;s major budget categories. If roads improved over the government8217;s term, voting decisions couldn8217;t have been indifferent to it.

Even from this very limited look at economic data it appears that questions like why incumbency can work and why inflation doesn8217;t necessarily hurt incumbents may appear more amenable to concrete answers than when seen from purely election-time political analysis. Rigorous analysis can give us far better clues to voter behaviour.

India8217;s many talented economists should venture into this underexploited research area: the much-celebrated but little-understood Indian voter.

saubhik.chakrabartiexpressindia.com