Presidents vs. Parliaments: The Virtues of Parliamentarism

The critical comments that Professor Horowitz and Professor Lipset have offered on my essay provide stimulating contributions to the debate over the respective merits of various forms of democratic politics. This debate is most timely, as controversy seems to be subsiding about the merits of democracy versus other types of government. My essay, itself an abbreviated version of a much longer paper still in progress, was meant as a spur to further study of the problem.¹ By raising more questions than can be answered given the current state of our knowledge about how democracy works, Horowitz and Lipset confirm the need for more research and reflection.

To avoid any misunderstanding, I must stress that I did not argue that any parliamentary system is ipso facto more likely to ensure democratic stability than any presidential system. Nor was I suggesting that any parliamentary regime will make better policy decisions than any presidential government, which would be an even harder case to make. There are undoubtedly bad forms of both these types of government. My essay did not discuss possible new forms of presidentialism, confining itself instead to the existing democratic presidential systems and excluding detailed consideration of the United States, which I consider quite exceptiona.² I do not think that I have constructed a "straw-man" version of presidentialism; my analysis is based on careful study of many prominent presidential systems, though I did not include the Nigerian and Sri Lankan versions of presidentialism that Professor Horowitz so skillfully discusses. Yet my article (like Horowitz's comments) also omits consideration of the many possible varieties of parliamentarism, and of the complex issues surrounding semipresidential or semiparliamentary systems with dual executives. These deserve separate analysis.

I agree with Professor Horowitz that the study of democratic regimes cannot be separated from the study of electoral systems, and acknowledge that my analysis does not cover all possible methods of presidential election. The Nigerian system represents a unique method of presidential election that might be applicable in federal states, particularly multiethnic ones, but I doubt very much that one could justify it in more homogeneous societies, even in the federal states of Latin America. My analysis concentrates on the two most common methods of election: the simple majority or plurality system, and the two-candidate runoff. The case where an electoral college may make a decision irrespective of the popular vote is left out, as is the very special case of Bolivia. The Bolivian Congress chooses among presidential candidates without regard to their popular vote totals, a practice that has certainly not contributed to either political stability or accountability in that country. I also refrained from mentioning the practice of directly electing a plural executive or a president and vice-president to represent two different constituencies (of Greek and Turkish Cypriots, for example). My argument concerns the likelihood of certain patterns of politics in the most common types of presidential systems, and does not attempt an exhaustive analysis of all types of directly elected executives. The patterns in question are likely to contribute to instability or difficulties in the performance of presidential executives. I use the word "likelihood" to stress that those consequences need not be present in each and every presidential system, or lead to the breakdown of democracy itself. On the contrary, recent experience shows that even rather inept democratic regimes stand a good chance of surviving simply because all relevant actors find the nondemocratic alternatives to be even less satisfactory.

Horowitz stresses that the majoritarian implications of presidentialism-the "winner-take-all" features that I have emphasized-may also be present in parliamentary systems with plurality elections in single-member districts, especially under the two-party systems that so often go together with Westminster-style parliamentary government. In societies that are polarized, or fragmented by multiple cleavages, a multiparty system with proportional representation may allow the formation of alternative coalitions (as in Belgium, for example), and thus forestall dangerous zero-sum outcomes.

As for parliamentary systems with plurality elections, Mrs. Thatcher is certainly a first above unequals, like a president, and probably has more power than...