Comments

Sebastián Galiani:

This paper by Lora and Olivera is very interesting. It investigates whether the Latin American governments that adopted market-oriented reforms during the late 1980s and early 1990s were rewarded with votes by the electorate. Certainly, this is an interesting and much-debated question. It is also a very hard one. There are many sources of complexity.

A virtue of the paper is that the authors keep the analysis at an explorative level. Identifying the impact of the reforms implemented during the 1990s on the voting behavior of the electorate is an intrinsically delicate matter, since there is no good way to assess how the same electorate would have voted had the reforms not been implemented (in precisely the way they were). Another serious difficulty is the lack of a more detailed theory to help disentangle the effects under consideration from the data.

The authors rely on cross-country panel data to attempt to control for factors that could potentially impact the effects they are interested in. Clearly, this is the best strategy for the data sets they use. They analyze both presidential and legislative elections, with separate empirical models.

A first question, which is very important for their analysis, is whether the electorate cares about policies per se, or only in relation to the effects they have on outcomes. This would determine whether the effect of economic reforms on voting behavior is a structural parameter or not. The authors assume that voters do care about policies per se, and that they vote on the basis of policies in the recent past.

Of course, the reforms of the 1990s were not implemented in a vacuum. They could be expected to affect variables such as growth and inflation and were also likely to hit unemployment, although transitorily. All of these variables are likely to influence voters, and thus should be included as controls in the regression models. The authors do so.¹ What is more, they also rightly compute the total effect (in addition to the direct effect) of the reforms on voting, using ancillary models that estimate the effects of the promarket reforms on these macroeconomic outcomes.

In presidential elections, the authors find that the incumbent's party is rewarded for reducing inflation, but once this is controlled for, the electorate seems to oppose market-oriented reforms.² Even when the total effect of these reforms is computed, it appears that reforming parties paid a price for adopting them. Given the relative importance of inflation among the explanatory variables, it would have been useful to check extensively the robustness of this result. Unfortunately, there are episodes of very high inflation that might be driving the results.

In any event, one question naturally arises: why did so many governments adopt promarket reforms? A common argument today is that structural reforms were supposed to deliver more growth than they did. However, even if this were the case, growth is found to have little effect on presidential elections. Thus, if one sticks to the estimated model, one needs to look elsewhere for the answer. Perhaps these reforms were also an essential component of the stabilization programs adopted by the countries that reduced inflation during this period, in a way not captured by the ancillary models reported in the paper? This seems to be the case for Argentina.³

Another important issue is how the reforms affect voting, once the main economic outcomes are controlled for. It is true that voting may well reflect the taste of the electorate for these policies. But it could also capture other things. It is likely to depend on the way these policies were implemented. It might perhaps capture distributive effects. For example, voters might not have fixed ideas about trade liberalization in general. Maybe, in deciding how to vote, an individual only considers his own economic situation, which could have been affected by this particular policy reform and is also affected by the macroeconomic performance of his country.

Consider privatization. Firms improved substantially after privatization. And consumers, in general, also benefited.⁴ But not everyone gained: displaced workers lost earnings and employment security...