The dangers of proportional coalitions

OKONOMOS

What sort of a political and electoral system should a country have? One like ours that is based on the majority rule, or one that based on proportional representation as in, say, Ireland and Germany? Germany's electoral system is called additional member system "" only half the seats are filled by proportional representation and the other half by first-past-the-post system.

Fully proportional represented systems are Ireland, Israel and several continental countries. Many Indians, fed up with the instability of the coalitions during 1996-1999, had advocated the latter.

Economists generally do not like to conduct research such a question but a recent paper* by Torsten Persson, Gerard Roland and Guido Tabellini does precisely that. And it presents an opposite view, namely, that proportional systems are no good.

They have examined economic and political data from 50 parliamentary democracies, which they say bears them out that coalitions spend more than non-coalition governments. As befits economists, this is their main case against the proportional system.

They have worked out a theoretical model as well, where "party structures, government coalitions and fiscal policies are endogenously determined", which seems reasonable enough. Their model predicts that contrary to popular belief "" which holds that proportional representation is good for the fiscal "" it is the clear-majority thing that is so.

Thus, "majoritarian elections are associated with smaller government spending, smaller budget deficits and smaller welfare states, compared to proportional elections.

The reason they give is also verifiable empirically, namely, that majoritarian systems "reduce party fragmentation and, therefore, the incidence of coalition governments." India is certainly moving in this direction. They go on to say that this does not mean you do not get fragmentation in the non-proportional systems, which too is true of India.

If fragmentation occurs, it is only because "supporters are unevenly distributed across electoral districts". This also fits in with the Indian situation where regional and other parties are mostly descendants of the Congress. (It also suggests that a similar fate might await the BJP, perhaps sooner rather than later, as personal ambitions of the heavy weight politicians overcome party loyalties. Kalyan Singh is a case in point).

Be that as it may, the authors' central contention is that proportional representation will induce an increase in the number of parties because everyone has a guaranteed opportunity to participate in the government.

This seems right. In contrast, no such guaranteed incentive exists in the majoritarian system, which is therefore less likely to lead to a proliferation of parties in the long run. Indeed, after a splintering there is likely to be a consolidation, if only to reap scale economies in electioneering.

There is also the related issue of movements towards a two-party system. Political scientists have for long agreed that a system of proportional representation acts to prevent the emergence of a two-party system while a majoritarian system will eventually lead to it. This seems largely true of India at the state level.

Indeed, even at the central level there is movement in this direction with the formation of the National Democratic Alliance (NDA) on the one hand, and the so-called Secular Alliance on the other.

The authors say "a coalition government allows voters to discriminate at the polls between the coalition parties, and this creates electoral conflicts within the government coalition and that intra-government conflict induces higher spending under both electoral rules." We have seen this as well in India.

The authors do not, however, address an important issue. This is one of the time it will take a country to attain the bliss associated with a two-party system. Britain and the US are not very good examples because of historical reasons. And, although Britain has had more than two parties, the US hasn't. The European countries are more representative.

The authors also do not examine a three-party system of the sort that prevailed in Britain just before the First World War and the sort that prevails now in states like Tamil Nadu. Here, basically, if two parties ally in a majoritarian system, the third is dead even if it polls a very substantial number of votes.

Another assumption, which is understandable, is that political parties represent only economic and perhaps moral or ideological interests and not social ones like the BSP or the RJD do. Hopefully, some clever Indian economist will read this paper and figure out things that fit India even more closely than this one does.

How Do Electoral Rules Shape Party Structures, Government Coalitions, and Economic Policies? Working Paper 251, November 2003, IGIER - Università Bocconi, Via Salasco 5, 20136 Milan, Italy, http://www.igier.uni-bocconi.it