Arend Lijphart's work

I have been told that Arend Lijphart is the world expert on electoral systems. What was his major contribution?

1 Like

It seems he has a somewhat substantial Wikipedia page and quite a few works with enticing titles:

I think these would be interesting works to read and discuss here. His most notable work is “Patterns of Democracy,” which I have just now purchased.

I read the wikipedia page and skimmed a few papers but I could not find out what actual contributions he made. If somebody is going to claim he is “the best researcher in the world on this topic” I would think that he should have made a breakthrough. Had they claimed Douglas R. Woodall to be the best I could point to his work on Free Riding.

What he has done is studied the economics of countries with various different systems. It is outlined here and from this account he did not really have any great finding.

After looking into it more it seems clear that he does support PR. I don’t necessarily disagree with him, but I would have to do more research. It seems a somewhat contentious issue, although according to Lijphart it is mostly agreed upon by political scientists in general.

Political scientists like PR so of course they all agree that it does things that there is no hard evidence for.

Why might political scientists like PR if there are no evidence-based reasons to support it?

I did not say that there is “no evidence-based reasons to support” PR. I said that there is no evidence-based reasons to believe PR does things like lower poverty, curb inflation, lower crime, lower unemployment, ect.

My point was that advocates of PR would be biased towards believing that PR does these miraculous things because they like PR. Why political scientists like PR is another topic.

1 Like

I am currently reading Lijphart’s book Patterns of Democracy. I think there is a lot more discussed there than PR. On one hand I have heard Lijphart address a broad wholistic approach to Democracy, while on the other I have heard him seem to place PR on a pedestal—again, not something I necessarily disagree with. Maybe he will explain why. I’ll let you know what I personally think after reading.

But in any case, I think his expertise is more in general comparative systems of government than in electoral systems, and I would wager whoever told you that he is the world expert on electoral systems is mistaken in terms of what we here mean when we say “electoral systems.” He does illustrate some of the key political issues that tamp down on electoral reform, but as far as I’ve read he has yet to suggest any ways to address them.

One thing to note however is that PR systems are more stable in the sense that once they are established, it’s virtually unheard of for any non-fringe organizations to form demanding a change away from PR. That obviously isn’t true of more majoritarian systems with disproportionate representation. As to whether systems with PR in place perform better along whatever metrics you might consider, that is at least one tenth of the subject of Lijphart’s book, and I’ll have to analyze his arguments on that topic for myself and research what other experts or interested parties have to say about it.


I just read Anderson’s critique, and the page is actually misleading in three ways:

  1. While it mentions that Lijphart found a correlation between PR and equality, it ultimately claims (without support) that PR worsening both inflation and unemployment (by any amount?) would be a “fairly clear conclusion that PR is worse”. Evidently, the author is not “one of those who says “economic inequality is bad,”” (aka humans) or at least thinks stable inflation is worse.
  2. It fails to put Lijphart’s initial work on the relationship between PR and efficiency in context. The contemporary view was that PR reduced efficiency, and Lijphart’s initial findings of no relationship (other than a possible increase in employment), were significant in dispelling that myth.
  3. It confuses the results of Anderson’s two analyses. In the first analysis, he found no significant unemployment effect of PR. Unsatisfied with that result, he changed methodology midstream, declared Switzerland an outlier, excluded it, and got the result he was determined to. But another result of excluding Switzerland was the appearance of a large (6 times PR’s) and significant unemployment effect of central bank independence. uses central bank independence’s effect (or lack thereof) in the first analysis and PR’s effect in the second, which is inconsistent and makes PR appear worse than it did in either analysis.

To Anderson’s credit, he evidently understood what did not, which is that his empirical findings alone aren’t enough to take the advantage away from Lijphart. On their own, they’re akin to pointing out that guns are nonlethal when you control for bullet wounds. Luckily, Lijphart himself had already admitted that PR and corporatism share a cause, rather than PR causing corporatism, so that just leaves central bank independence. But, strangely, Anderson’s lengthy discussion of its relationship to PR was purely philosophical, even semantic. At no point did he offer a reason to suppose PR affecting central bank independence is any less plausible than any other explanation for the correlation.

What we certainly want to avoid is going into this thinking Lijphart has to prove PR’s benefits beyond a shadow of a doubt, as if he’s prosecuting a capital crime. The question is, quite simply, is PR a good bet? The correlations Lijphart found make it a good bet until proven otherwise. I’m also curious, since you conceded PR may have some positive effects, what you imagine those positive effects may be? The reason I ask is, I always thought the world was interconnected. I didn’t think it was separated into airtight compartments–a poverty compartment, an inflation compartment, a crime compartment, an unemployment compartment…What effects could PR possibly have that would have absolutely no ripple effects on these indicators?

Thanks for this.

I clearly think that PR is good for some things like representation but I have doubts about the effect on government running the economy better. Would you disagree with the summary on

After all the sound and fury from Lijphart, Crepaz, and Anderson, my net conclusion is nobody won and I see no clear evidence that the PR/not-PR choice has any effect on inflation, unemployment, or economic growth.

No evidence does not mean no effect but that the effect is smaller than our power to see it. As you said before though, no effect is a win for people who would argue PR is not worse.

1 Like

I would agree that there’s no clear evidence of a growth, inflation or unemployment effect. There is clear evidence (statistically significant at the .01 level) of an equality effect and a peace effect, but Lijphart didn’t show them to have the right chronology (PR then the effect). also mischaracterized the equality effect as “less wealth disparity”, when Lijphart actually measured “income inequality”. Wealth disparity is overwhelmingly driven by inheritance in all major countries, but most income is labor income; consequently, there’s a significant meritocratic factor of income disparity in some countries (especially those with low wealth disparity), so I’m not going to fault a country just for that.

I’m also beginning to appreciate Anderson’s objection that an independent central bank (Lijphart’s 10th characteristic) isn’t a veto (it’s dictator of its domain) and thus should not have been associated with consensus. I would even go further. A supreme court’s monopoly on constitutional interpretation (Lijphart’s 9th characteristic) isn’t a veto either. Nothing but the power of faith (and, occasionally, the threat of court-packing) prevents the supreme court from unilaterally making law by “reading it into the constitution”, of which its blasphemous for a mortal to accuse a justice but routine for justices to accuse each other. Luckily, though, Lijphart separated his first five characteristics (including PR) from his other five, so the dubious ones don’t distort the results.

I think the biggest problem with Lijphart’s analysis is that he doesn’t take into account the disagreement point, which becomes more important the closer we get to consensus rule. For even if we assume an agreement will be reached, the disagreement point influences parties’ relative bargaining power and thus what agreement will be reached.

For example, for most legislation in the US, the disagreement point is the status quo. But for the annual budget, the disagreement point is government shutdown. In one case, consensus characteristics advantage those who benefit most from the status quo; in the other, they advantage those who benefit most from (or are harmed least by) government shutdown. The two resulting power distributions are arguably more different from each other than either is to a majoritarian power distribution. Indeed, they’re arguably opposites: in one case, consensus characteristics are barriers to repealing the law (because it has no expiration date); in the other, they’re barriers to renewing the law.

I agree with your augments but my major issue with the studies comes from a different angle. It is a false dicotomy to say there are PR countries and non-PR countries. The difference in the forms of PR are as large as the difference from FPTP to one of them. Most countries which have a “PR system” are party list. I would argue that the effect on parliamentary formation is about as large between STV and closed party list as FPTP and either of those. This is because the effect on partisanship and strategy are largely different when you vote for a party vs a person. Also, this is assuming a system where the Prime Minister is the leader of the government. This means the USA and Canada should not be in the same non-PR group. Nancy Pelosi leads the house (parliament) not the government so even if the house was elected from a “PR system” it would not really dictate as much government policy. There are likely examples of PR presidential systems that exist like that. In any case, my point is that the system of government formation is likely to have an effect on the same scale as if the system is PR or not and same goes for the types of PR. I do not think Lijphart adequately controlled for this effect.

Lijphart doesn’t dichotomize countries into PR and non-PR. He trichotomizes voting systems into PR, semi-PR and majoritarian, but it doesn’t affect the countries’ scores. The Gallagher Index is the disproportionality component of the scores.

Nor does he assume the Prime Minister is head of government. For presidential systems (including the 3 that use party list), he uses the geometric mean of legislative disproportionality and presidential disproportionality. Perhaps that’s still too forgiving of presidential systems, in that it doesn’t take into account that the President is usually more powerful than the legislature. On the other hand, parliamentary executives are never as proportional as parliamentary legislatures, and the distribution of executive posts probably better reflects even the legislative component of the distribution of parliamentary power than the distribution of seats in parliament does.