Research Spotlight

Good economics makes for good politics

Does good economics make for good politics? Historically, In India, good economics has not been considered good politics. However, recent experiences in Gujarat, Bihar, Odhisha, Chattisgarh, Haryana and Madhya Pradesh have pushed forward the view that good economics can make for good politics as well. Nevertheless, anecdotes cannot substitute for systematic evidence.
Poonam Gupta and Arvind Panagariya study the 2009 Loksabha elections and find that candidates of incumbent parties in states that recorded high growth won more often than their counterparts in states that recorded low growth. The 2009 elections followed a period of high growth in the economy. Therefore, evidence from the 2009 elections may be an exception rather than the norm. To establish whether good economics makes for good politics in general in India, one needs to examine the evidence across several electoral cycles that span high and low growth periods. Moreover, with a voting population that contains a significant proportion of illiterate voters, a related question that arises is: does the positive effect of growth on reelection prospects hold only in states where the population of literate voters is high? Or does good economics make for good politics in states where the proportion of literate voters is low? Since literate voters are more likely to be aware of the state of the economy, the effect of good economics on good politics needs to be viewed through this prism. In a recent working paper, we examine these questions using data on all elections—state and national level—from 1980 until 2012.
Our analysis of over 20,000 election outcomes from every constituency over more than three decades reveals the following. First, high economic growth generated by the incumbent government significantly enhances the chances of ruling party candidates getting re-elected. Careful empirical analysis requires controlling for the effects of various other determinants of the incumbents’ success. Factors that are peculiar to a particular state, such as the leftist leanings of the electorate in states such as West Bengal and Kerala, can influence the incumbent’s ability to get reelected in that particular state. Moreover, a wave that sweeps the entire nation, such as the sympathy wave generated in the 1984 national elections after Indira Gandhi’s death, can affect all incumbents’ ability to get reelected. In arriving at our results, we control for such state level determinants as well as general time trends. Since a higher turnout ratio is considered bad news for the incumbent, we also control for the effect of voter turnout on the likelihood of ruling party candidates getting reelected. After controlling for all such determinants, we find that a 1% increase in growth achieved during the entire tenure increases the chances of reelection of every ruling party candidate by 0.6%.  Such an increase for every ruling party candidate translates into a significant difference particularly given three-pronged contests in most constituencies, which include candidates from the two main national parties and from the main regional party in a state.
Second, we find that creating a mirage of economic growth by pushing up growth in the election year cannot mask the incumbent's incompetence over the entire tenure. Specifically, a government that manufactures growth just in the year of election by inflating the economy is likely to be punished by the voters while a government that achieves high growth throughout its tenure is likely to be rewarded. The above finding goes contrary to political cycle theories where incumbents successfully manipulate economic variables just before the elections and are handsomely rewarded by myopic voters.  Nordhaus (1975) shows that politicians can exploit the Philips Curve, which posits a negative relation between inflation and unemployment in the short term, to create artificial prosperity just before the election. Since such growth is achieved by inflating the economy, it dies down quickly leading to post election recessions. Our analysis reveals that Indian Voters are not myopic in Nordhaus's sense. Indian politicians, take note.
Third, we examine the effect of the percentage of literate voters in a state on the ruling party candidates’ fortunes. We find that while incumbents are more likely to retain power in states where the population of literate voters is lower, higher economic growth translates into greater prospects of reelection particularly in states where the percentage of literate voters is higher. In other words, not only are illiterate voters more likely to vote back the incumbent to power, they also care relatively less about economic growth when voting back the incumbent power. In contrast, literate voters are less likely to vote back the incumbent to power and care a lot about the performance of the incumbent in generating economic growth when deciding whether to vote back the incumbent power or not.
A caveat: a statistical exercise such as ours highlights the average effects observed over more than 20000 election outcomes spanning more than three decades. However, a few exceptions that constitute statistical outliers do not necessarily undermine our findings.  For example, among the various states, Kerala and Tamil Nadu have the highest percentage of literate voters. Moreover, these states have also enjoyed high economic growth rates. Yet, they also exhibit high levels of bias against the incumbent. In sum, our analysis exhorts politicians to take note: good economics can indeed make for good politics!