Latin American Politics & Society 46.2 (2004) 69-72
[Access article in PDF]
Making Sense of a Model:
A Response to J. Samuel Valenzuela
Josep M. Colomer
I am very pleased to start a dialogue with Samuel Valenzuela, although, because he reveals that he was one of the anonymous reviewers of my article, I have to point out that he was the only one out of five who disagreed with the editor's decision to publish it. I hope that the dialogue might extend to other interested readers who can find my approach more helpful and challenging from the beginning.
Valenzuela declares himself to be for what I call "an explicitly comparative, theory-driven analysis, which is more characteristic of the social sciences literature," but I am not sure that he understands some basic features of this approach. Rather typically for outsiders, he finds that a formal model is "simplistic" and has "an abstract flatness," and then, instead of trying to exploit the insights from the model for further discussion and applications, he refers to a number of unrelated empirical details and specifications. There is an unavoidable trade-off: the more general the model is—and it must be general in order to permit comparative analysis—the "simpler" and "flatter" it has to be. But generality implies a focus on the core variables of the problem, which, once well identified, should help the observer make sense of all the details and their relative importance or irrelevance. I do not think that any of the empirical details specified by Valenzuela help him deny the validity of my model at all.
First, about the model. It is valid not only for social "classes," as Valenzuela interprets; this can be one of the applications, but I am referring to "groups of voters" with different political preferences, which, of course, may include ethnic, territorial, religious, or other defined groups. Actually, I explicitly say in the article that "distances" in the model may reflect social structures, values, or ideological opinions. I also refer to "coalitions of voters" in a way rather usual in political science and sociology literature, while, of course, more formal political coalitions are formed by parties and leaders. Regarding the threatening character of a group of voters, it will certainly depend on the status quo. Thus, for example, the introduction of women's voting rights, which has traditionally not been considered strongly threatening for a rather moderate or conservative status quo, as I comment in my article, can be more destabilizing if the status quo is a government held by anticlerical forces, as Valenzuela argues. But again, these are different possible [End Page 69] applications of the same simplified model, which show nothing but the "abstract" model strength.
Now about electoral rules. I do not "put together" proportional representation and absolute majority rule but make clear that they are electoral rules for different institutions, respectively Congress and the presidency. As explicitly remarked in the article, while proportional representation can be highly inclusive because it permits representation of multiple parties and groups of voters with different preferences, which become "multiple partial winners" (in my usual wording), absolute majority rule (typically with a second round) for the presidency is only "less bad" than simple plurality rule. This is true because although absolute majority rule can induce the formation of broad multiparty coalitions, it finally produces a "single absolute winner," as irremediably happens with any presidential electoral rule. Valenzuela wonders what the difference is between "simple majority" and "plurality." It is very clear: "simple" majority rule requires more than half the votes (as distinct from "qualified" majority rules requiring two-thirds, three-fourths of other thresholds higher than half), while (simple) plurality does not require any specific number or proportion of votes but only more votes than any other alternative (also, some requirements, such as 40 or 45 percent of votes, can be called qualified plurality rules).
Finally, in spite of these latter two points and many others in the text, my commentator complains about my not having developed a proper discussion of electoral rules...