Does America Need a New 'Science of Politics'?

Professor Buckley argues in “American Exceptionalism” that presidents cause countries with the office to realize less freedom on average than countries with prime ministers. Below I explain why neither Buckley’s theoretical claims nor the empirical evidence he provides persuades me that his conclusion is warranted.

Before digging into his argument, however, I do want to appreciate Buckley’s approach to the topic: Answers to questions of constitutional and institutional design are more contingent than many commentators, let alone the public, often allow. I am second to none in my admiration for what the U.S. framers created. They nonetheless advanced contestable theoretical and empirical claims. Just as Publius wrote in The Federalist #9 that the “science of politics” had “received great improvement” in his day, so I would not think that he would begrudge the possibility that that improvement might continue. At the least, it behooves each generation to learn that the framers were correct rather than merely to believe it on their word. I think they would appreciate a continuing conversation among those whose aim it is to retain “the excellences of republican government” while seeking to lessen or avoid “its imperfections.” And that’s the spirit in which I take Buckley’s argument, and that’s the spirit in which I intend these comments.

Buckley advances three reasons to hypothesize why presidential systems might “threaten political freedom.”

  • As heads of state, presidents enjoy a prestige and status denied prime ministers in a parliamentary system, and might exploit this to assume greater powers.
  • Prime ministers are more accountable for their misbehavior than presidents, who hold office for a fixed period of time and are freed from daily scrutiny before a House of Commons.
  • There is a greater possibility of deadlock in a presidential system, with its checks and balances. These invite a president to step in and assume greater powers.

None of these is claims is theoretically persuasive, however. Of first importance is that we recall that Buckley is making a comparative argument. So the question is not simply whether presidentialism has this or that effect, the question is whether there is reason to think that this or that effect is greater (or less) than what it would be in a parliamentary system. Buckley compares separation-of-power (SOP) systems in which the chief executive is elected separately from legislators to a parliamentary systems in which the chief executive is elected by legislators. In the SOP systems, the chief executive may face one or more legislative chambers dominated by political opponents. In parliamentary systems, the chief executive rules as a result of having the support of a majority of legislators on major policy positions.

Or to put it another way: Because of the nature of the relative systems, prime ministers rule more consistently with majority support in their legislatures than presidents rule with majority support in legislatures in their separation-of-power systems.

Given this, then the expectation must be that prime ministers receive legislative support for their proposals more often from the legislative majority that elected them to the office than presidents receive from legislatures that may or may not have supported their election. If I were an executive intent on maximizing my power, I would much prefer a dependable legislative majority than having something as ethereal as “prestige” that may, or may not, induce an opposition legislature to adopt my policy proposal.

That said, I am open to evidence that the prestige of a president can in fact overawe legislative opposition more often than a prime minister’s legislative majority supports his platform. But if this is the case, there is nonetheless a simple remedy that would come not from rejecting a separation-of-power system, but rather extending it: Separate out the position as “head of state” from the president, and give it to another office. The U.S. could add another office – let’s call it the Chancellor – who is head of state, and who receives all of the trappings and prestige of the office, thereby removing it from the office of the president. Voila. No more “head of state” prestige for presidents to manipulate.

Secondly, Buckley argues that prime ministers are “more accountable for their misbehavior” than presidents are because the prime ministers lose power once they lose support from their legislative majority. Buckley apparently appeals to the British example in which the prime minister receives “daily scrutiny” from the opposition. Beyond the fact that even in Britain for decades now the Prime Minister stands for questions only once or twice a week (rather than daily), I am unsure how many other parliaments share this practice. I know of no reason that it be inherent in parliamentary practice.

More importantly, however, is that these are questions from the minority in a parliamentary body. In contrast, in severing the link between majority support in one or more legislative chambers, as separation-of-power systems can do, oppositional majorities can bring the full resources of a legislative chamber to bear in investigating the executive: hearings, independent investigations, and opposition to legislation until their questions are answered. Further, in the U.S. system, Congress has endowed its standing committees with the power to compel witnesses to testify and provide documents. It is unclear to me that the legal power that an oppositional legislative chamber has to investigate executive corruption in the U.S. is, ex ante, less apt to uncover corruption in the executive branch than are a few questions put to the PM once a week, questions that can be easily dodged by a rhetorically astute politician. Compare, for example, the compelled testimony of a witness at a congressional hearing in the U.S. with the information gained at presidential press conferences.

Finally, Buckley claims that the checks and balances in a separation-of-power system can create gridlock that invites an expansion of presidential power. It is again critical to keep the comparative claim in focus: Chief executives in parliamentary systems have presumptive majorities in their legislatures. The chief executive in a separation-of-power system does not. The expectation would again be that it is easier for the chief executive in a parliamentary system to extend his power with the inherent support of the legislature than that a chief executive in a presidential system could extend his power in the teeth of legislative inaction or opposition.

At worst, theoretical expectations of executive power between parliamentary and SOP systems would be a toss up; at best, expectations would be that parliamentary systems have the effect at which they aim: to align executive power closely with legislative majorities and, therefore, ceteris paribus, executive power would be expected to be greater in parliamentary systems than in SOP systems.

We now turn to Buckley’s empirical evidence. This deserves a longer treatment than can be provided here. I focus on only a few items.

While Buckley includes several control variables, from what I can tell from his description of the statistical model in another paper, there are several important control variables not included in his model.

First, I do not see the inclusion of other institutional variables pertinent to separation-of-power systems. Two obvious ones include judicial review and a legislative branch sharing aspects of executive power as the Senate does in the U.S. The whole effect of checking institutions in SOP systems may be greater than the sum of the individual effects. Relatedly, I would also like to see the results without semi-presidential systems being grouped with presidential systems in the data. They are different institutions, after all.

Secondly, If one looks at a map of types of governments (see e.g., http://en.wikipedia.org/wiki/File:Forms_of_government.svg) it appears that government type is not randomly distributed across the globe. This raises a host of potentially confounding factors for Buckley’s empirical analysis. For example, given the groupings of countries with presidential systems, perhaps the policy preferences of the voters (or of the legislators) are running the dependent variable. Buckley does not include a variable that controls for the policy preferences of the people in the different countries from which he draws data. Policy preferences cannot be assumed to be homogeneous across diverse nations. As a result, his executive-type variable may be picking up policy preferences resulting from correlated needs in countries that are geographically clustered.

Finally, and relatedly, I wonder about the direction of the causal arrow. Conceding Buckley’s statistical association, I wonder whether the direction of the causal arrow actually goes from less freedom to presidential institutions rather than in the direction he posits, from presidential institutions to less freedom.

What do I mean? After World War II, numerous countries that were decidedly not democratic nonetheless placed “democratic” in their names, presumably as a form of “big lie” telling. Even today, is there likely any place on earth less democratic than North Korea, a.k.a., the “Democratic People’s Republic of Korea.” And it is not the more democratic for holding rigged elections.

Similarly, I would suggest that it is precisely because of the long-serving U.S. example that presidential systems have a certain cachet. Just like the least democratic countries are apt to names themselves “democratic” in an attempt to mystify their true reality, less-free countries may be apt to adopt presidential institutions with the same purpose in mind. “See, we must be free,” they hope their people will believe, “because we have a president just like the free U.S.A.” So Buckley’s statistical association could be conceded, but the observed association might occur for almost the opposite reason than he forwards.

And just to be clear: While my own suspicions are that the formal type of the chief executive has little marginal influence on the freedom of a country – free societies, as well as their opposites, can have parliamentary or separation-of-power systems – I remain open to evidence of the possibility that Buckley asserts. But neither the theoretical arguments nor the empirical evidence in this study are sufficient to move me from my agnosticism.