Vivek Dehejia: The true 'cost' of elections

The money spent on vote-seeking should not be reckoned a cost, but merely a costless transfer from electoral candidates to the public

With the cost of elections soaring in India - estimates suggest that total spending by all candidates this time around could be as high as $5 billion - many worry that money, both legally and illegally spent, is corroding the electoral process.

Are these concerns justified?

Let us leave aside the legality of candidates offering voters money to cast a vote in their favour - it is illegal - as well as the morality, which clever philosophers could no doubt argue both sides of, and focus instead on the economics.

Here, the untutored idea that a large amount of money spent by candidates "bribing" voters must, necessarily, be bad for the electoral outcome, as well as represent a large economic cost to society, must be interrogated. As we shall, both propositions will be found wanting.

Since the ballot is secret in India, leaving aside those cases in which voters may be intimidated into voting for a particular candidate - which could be a real concern in some situations - a rational voter will accept money or gifts from as many candidates as offer them to him and then will simply vote his conscience once in front of the electronic voting machine.

If that is accepted, then bribes offered by candidates will, to a first order of approximation, have no impact on the electoral outcome. This leads directly to the fallacy of the second proposition.

If voter behaviour is not affected by bribes, then the large sums spent by candidates to woo voters represent, in terms of economic analysis, merely a transfer, rather than an economic cost. The confusion arises because "cost" is used, in everyday language, to represent an accounting cost.

So while the money spent by an electoral candidate will show up on his books (at any rate, the unofficial one not shared with election officials!) as a cost, as far as society as a whole is concerned, it should not be reckoned a cost, but merely a costless transfer from candidates to the public - a form of redistribution, in fact.

This begs the question: if bribing voters doesn't work, why do candidates try to do it anyway?

Here, we need to modify our simple model slightly. Suppose most voters will vote their conscience, but a small minority will vote out of loyalty to whichever candidate gives them the most money. The trick is that candidates don't know who exactly these "irrational" voters are.

In that event, it will be rational for all candidates to give voters as much money as they can afford, even though candidates know that most of this money will be "wasted" on voters who will vote their conscience anyway, in the event that the small number of "irrational" voters may swing the election result their way.

Again, though, the large proportion of money spent represents a transfer, with a small proportion, arguably, reckoned as socially wasteful if it sways the small number of irrational voters away from voting their conscience.

As such, if they could agree, all candidates in a particular constituency would be better off if they all refrained from spending money, assuming that such an agreement were enforceable. But there's the rub: every candidate would have an incentive to renege and try to swing the few irrational voters his way, so the agreement would break down and every candidate would resume spending (this is a version of what is known as the "prisoners' dilemma" in game theory.)

This result - that money spent on vote-seeking largely represents a transfer, rather than a loss, to society - is reminiscent of a famous debate on the costs of "rent-seeking" behaviour.

In a now-classic 1974 paper, economist Anne Krueger argued that "rents" from quantitative trade restrictions - equivalent to revenues from a tariff - should be reckoned as a cost to society, rather than being considered revenue for the government, as conventionally assumed by economists.

This assumption, of course, magnified the estimates of the costs of rent-seeking by a large order of magnitude since the "rectangle" of revenue is much larger than the "triangle" of welfare losses conventionally associated with a tariff or quota.

Krueger reasoned that if the "market" for rent-seeking is competitive and risk-neutral, putative rent-seekers will dissipate the entire amount of rent available in an attempt to sway policymakers, with the end result that this amount of money is wasted rather than simply transferred from rent-seekers to the government.

Subsequent analysis, by Jagdish Bhagwati and others, suggested that Krueger's result of 100 per cent loss was, at most, an "upper bound" on the true losses from rent-seeking. To the extent that lobbying the government is done "efficiently" (for instance, via an envelope stuffed with currency, rather a 10-hour, 10-course dinner), and transactions costs are absent, the "losses" from rent-seeking are, once again, more appropriately understood as a transfer from rent-seekers to the government, rather than a loss to society.

By the same token, $5 billion spent on electioneering is, at most, an upper bound on the true costs to society. Like it or not, most of it is a form of redistribution.

The writer is an economics professor at Carleton University in Ottawa, Canada, and is co-author of Indianomix: Making Sense of Modern India (Random House India, 2012)
Marginalia makes research from the academic world accessible to all our readers