Wednesday, May 16, 2007

Learning from the French

Sandy Levinson

George W. Bush has 612 remaining days in office. That's bad enough news, but consider the fact that the last 77 days of his office will, very likely be spent (and, of course, I believe much to be hoped), as a thoroughly repudiated President whose successor will be a Democrat committed to overturning many of his policies, beginning, of course, with Iraq (which is certainly not going to be resolved by November 4, 2008). This is, of course, a not uncommon reality of American politics. Consider in the 20th century alone the repudiations of Taft, Hoover, Carter, and George H. W. Bush (though by a candidate who did not get anything close to a majority of the vote). The most serious of these is probably Hoover's defeat by Roosevelt, for the United States did not have a truly functioning government, with regard to making economic policy, between November 1932 and March 4, 1933.

The 20th Amendment, of course, moved up inauguration day to January 20, but that's cold comfort for those of us who want George W. Bush out of office as soon as possible. Moreover, I continue to believe that no sane country would today adopt a presidential election process that includes such a hiatus between a repudiated President and his/her successor. (As with all of my criticisms of our defective Constitution, I'm not interested in Founder bashing. Nothing rides on whether the Founders had wonderful reasons in 1787 for doing every single thing they did, including compromising with slavery.)

The French offer us a model from which we might learn. Nicholas Sarkozy, elected by a majority of the electorate--more about this in a moment--nine days ago, was literally given the keys to the presidential palace this morning. The presumptive desire of the French electorate for whatever policy changes Sarkozy will bring will begin being realized immediately. Jacques Chirac will not have almost three months to torpedo them or to engage in foreign-policy adventurism--recall George H. W. Bush's sending of troops to Somalia during his interregnum--that will haunt his successor's presidency.

I had previously thought that such a sensible time between election and inauguration required a parliamentary system, where a prime minister would have a "shadow cabinet" in waiting, one of the realities that allows winning candidates for the PM in Great Britain to take office basically the day after the election. But Sarkozy had no such "shadow cabinet." Yet he was able to name his cabinet this morning--including the founder of Doctors Without Borders as the Foreign Minister. I take it that this proves that a rational political system--which ours is not--is fully able to generate and elect presidents who can, within roughly a week, make appointments. Indeed, isn't it absolutely ludicrous that we in the US are about to endure an 18-month-long campaign for the presidency without being told whom the candidates will be appointment to the Cabinet. Even if they don't want to reveal this publicly, can't we expect them to have the names in their pockets, for release within a week?

The French can also teach us something about majority rule. Many of our presidents got to the White House without winning support from a majority of the electorate. Since World War II, these include Truman, Kennedy, Nixon (1968), Clinton (both elections), and Bush (2000). That is not the case in France. The two-stage system of presidential elections assures that the winner will have a credible claim to majority support, something that we cannot claim of our own presidents. A good alternative to the two-stage election is the "instant runoff," by which voters would cast an Alternative Transferable Vote indicating their second (or subsequent) choices should their first choice not prevail.

I note that a number of states are passing the Fair Vote bill by which the largest states would torpedo the electoral college by giving their votes to the popular vote "winner." I'm all for torpedoing the electoral college, but it should be recognized that this scheme preserves one of its worst features, which is the possibility (indeed, near certainty over several election cycles) of electing candidates supported only by a minority of the population. It will also tend to reinforce the two-party duopoly that afflicts the US and prevent a third-party from demonstrating its appeal to parts of the electorate, who can cast their first-preference votes for such a party without being "spoilers" (as were, to their unending damnation, the Naderites in 2000). I support the Fair Vote bills inasmuch as they might generate a long-overdue conversation about the electoral college and break the logjam created by the inability of any amendment to get through the Senate. One might hope that some conversationalists would be willing to look outside our own narrow borders and look at how other respectable democracies elect their chief executives.


O.K., so we only need as much time as it took for Bush / Gore to be sorted out, right?

Thank you for your comment on the National Popular Vote plan in the last paragraph, which is the first commentary I have seen on that subject with which I basically agree.

But I wouldn't call instant runoff voting (the U.S. name) or the alternative vote (its name almost everywhere else) a "good alternative" to two round runoff. I would call it a much better alternative.

IRV not just quicker and cheaper than the French method, and it doesn't just overcome the problem that voter turnout is generally lower in one round than the other. Just as important, it also picks better finalists. For example, look at the last French presidential election (in 2002), in which the two finalists were candidates of the moderate right and far right, even though a majority of voters would have voted for the candidate of the moderate left in the second round -- if only there had been one.

I think that Bob Richard makes a very good point. Forced to choose, I think I'd go for the ATV "instant runoff" for the reasons given.

But Sandy, you forget that the Founders were Gods, of perfect wisdom, bringing down from God a system for the ages! Nothing they did could ever be improved upon by us mere mortals, pygmies to their beginning of time Giants! Didn't they live 900 years and weren't they 10 feet tall?

Bahh, conservatism is just rewashed medieval "thinking".

Charles, how about instituting systems that are simple, uniform and can count by hand the entire national vote in an evening? Like even in third world countries? But I guess our system is sacred, a holy electoral system bequeathed to us by Shammash, as it says on the stele.

I think you've clearly got a point about the extended period between the election, and the new President taking office. One which is much stronger than your preference for some kind of parlimentary system, and which in no way depends on the assumption that Bush is a uniquely bad President.

I also think IRV is decidedly superior to our current system; The electoral college really only made sense back when electors were, in fact, chosen by state legislatures. Just as the Senate only made sense under the same circumstances.

A change in the date the newly elected President takes office should be relatively easy to gather bipartisan support for. Much easier than proposals to make the President easily removable by a Congress of the opposing party. A change to IRV would be rather more difficult, since it clearly threatens the duopoly which currently has our politics in a stranglehold. Especially since it would be rather difficult to explain why it was good for the Presidency, but not for Congress.

Not a chance of getting it through Congress, I think. A Convention would be the obvious route, but, of course, that duopoly controls state legislatures, too. Further, I rather suspect that if the states every get their act together to call a constitutional convention, Congress will in some manner sabotoge it, by ignoring the call, or perhaps declaring their own membership to be the ideal delegates.

After all, they haven't seized the power to unilaterally "amend" the Constitution without state approval, only to let it go without a fight.

Professor Levinson:

What do you know, I think we have some common ground here.

I would love to see a run off if one Presidential candidate does not get a majority. Third party spoilers do undermine democracy. (See Perot and Nader).

I would also like to see the new President take office within, say, 30 days. That should be enough time to start assembling a cabinet and prepare the inaugural.

We should be able to do both with a statutory change.

Oh my God! Bart giving an opinion on something that isn't the pre-determined Party Line! Will wonders never cease?

The Maoists always warn about such independent thinking - you always run the risk of having committed an opinion that is "objectively incorrect". One thing leads to another, and before you know it, you may start to wonder whether the Party Line is actually "objectively correct".

Bart, I must warn you: bourgeois individualist morality lie this way. The Great Leader will not be pleased.

Really, Bart, reading Article II, section 1, I'm having a hard time figuring out how you get a runoff election for President without an amendment. The Constitution pretty unambiguously lays out the procedure, and doesn't leave any room *I* can see for a runoff election.

OTOH, while the date the President takes office is constitutionally fixed, Congress could certainly move election day closer to that fixed date. Though there'd be a one time interaction with the mandate that the President serves 4 years at a time; 50 months isn't 4 years on anybody's calender.

Aw, C'mon Brett. Bart doesn't have a pre-formatted, pre-digested opinion here. You can't expect him to have actually thought it out and posted a substantial opinion now, do you?

I appreciate Bart's kind words. I do believe, though, that we would have to face the amendment issue sooner or later, probably sooner. We would, after all, be getting rid of the electoral college system (and it is the EC that is the only plausible reason for the long hiatus. I'm not sure that the Constitution should state a specific date for inauguration at all, since, after all, and I say this with great sadness, one might envision a rerun of the Bush v. Gore kind of litigation. So one could simply say that "The newly elected president will take office 72 hours after the election is officially certified" (or something like that).

Incidentally, the gratifying postings by Bart and Brett underscore my own view that most of my suggestions about changing the Constitution are basically non-partisan. It's true that I despise George W. Bush, but my proposal to strip future presidents of the policy-based veto could, obviously, just as easily apply to Hillary Clinton as to Mitt Romney, etc. We should stop assuming that every single issue can easily be located along a "liberal-conservative" spectrum.

""The newly elected president will take office 72 hours after the election is officially certified"

God forbid someone could be made "President for life" by the simple expedient of failing to officially certify the election results. No, I kind of like the highly determinant nature of our elected terms.


Do you imagine that just arbitrarily holding up certification would work politically? Politics doesn't depend just on legal technicalities - someone who was obviously disregarding proper succession could not get any leverage to act. If he could, the inverse would work just as well - just stay in the White House and declare yourself dictator.

In practice, I've never heard of a situation were lack of certification in and of itself produced a coup - the coup was always military and such certification worked only as a fairly obvious fig leaf.

Just like Congress can decide not to seat the opposition, technically. Just doesn't happen, in practice.

Earlier I wrote, Thank you for your comment on the National Popular Vote plan in the last paragraph, which is the first commentary I have seen on that subject with which I basically agree. I'm afraid that statement might misrepresent my own view.

Prof. Levinson says this scheme would do nothing to solve the related problems of minority winners and two party duopoly. I think it would actually make the situation worse for small party and independent candidates. Historically, their campaigns have been either symbolic (Nader) or intended to throw the election into the House by getting electoral votes based on a strictly regional appeal (Thurmond, Wallace). The National Popular Vote plan eliminates the latter possibility altogether.

So, I'm in the odd position of being opposed to the interstate compact proposal itself, because of the effect on small parties and the absence majority rule, but hoping that it passes in enough states to get to the Supreme Court (where I think it will be ruled unconstitutional), because of the pressure that might create for action by Congress and the state legislatures. Such pressure might get help get runoffs, instant or otherwise, on the national agenda.

Also, I failed to mention that the most recent French presidential election is another example of how two round runoff can pick the wrong two finalists. See France's election flaw (registration required) by Steven Hill and Guillaume Serina for an argument that the centrist Francois Bayrou should have made the final -- and would have if the French used IRV.

Perhaps needless to say, I regard the elimination of the possibility of going to the House for resolution of the presidential race as a positive good, and not a defect. I can understand the political logic of a regional campaign that wishes to push the final decision into the House, but there's no reason to support it.

I agree with Mr. Richards on his other major point. We also agree that there are real differences between IRV and a two-stage system. I wonder if we both agree that either, whatever the inevitable flaws, would be better than the system being contemplated by Popular Vote, with its first past the post emphasis.

Professor Levinson, I gather from your post that you believe that a claim to majority is very important for a democracy, but what I do not understand is why. Fundamentally we are talking about changing vote counting methods, not underlying realities. Abandoning plurality voting won't change the fact that winners really only have a plurality of voter support. The only thing that will change is the claim of majoritarian mandate, but I am not so sure that is a good thing. Should every single president get to have a mandate? Is having a mandate such a good thing? I doubt there would be many who would argue that GWB was a better president in 2004 for having won a popular vote majority than in 2000 when he had not.

On IRV, my hang up is that I just can't get past the fact that it fails the monotonicity criterion. Also see Bob's comment about Francois Bayrou.

To Professor Levinson's comment about foreign intervention, I am not sure that the President's constitutional ability to execute repudiated foreign policies in January of 2009 represents a distinct flaw from the President's ability to execute repudiated foreign policies in November of 2006.

Prof. Levinson asks, I wonder if we both agree that either, whatever the inevitable flaws, would be better than the system being contemplated by Popular Vote, with its first past the post emphasis.

We do agree on this. My dilemma is that the National Popular Vote plan has a chance of success (ignoring, for the sake of argument, my view on what the Supreme Court is likely to think of it), while any plan that requires amending the Constitution has almost no chance.

There has been a little bit of discussion in electoral reform circles of a modified interstate compact which would require member states to conduct ranked ballot elections and pledge their electors to the IRV winner. The problem is what to do about non-member states, which would presumably continue to have plurality votes.

The idea has also been floated of using the interstate compact mechanism to do away with the unit rule in favor of proportional allocation of electors (as Maine and Nebraska have done unilaterally). This would overcome the barrier to proportional allocation that states doing it unilaterally are acting against their own self interest.

This has one advantage over NPV: it would not make the situation even worse than it is now for small party and independent candidates. But, as Prof. Levinson points out, it does so only at the expense of increasing the risk that elections would be decided in the House -- one state, one vote. He's right; I shouldn't be so cavalier about this.

I'm torn between the idea that NPV, if implemented, could break the logjam and lead to constitutional reform, and my desire to end the two-party duopoly. If NPV would contribute to the latter, it would only be in the very long run.

JimM47: Fundamentally we are talking about changing vote counting methods, not underlying realities.

Well, changing the vote counting method would change the winner of a few (not many) elections. I would call who wins part of the underlying reality of an election.

On IRV, my hang up is that I just can't get past the fact that it fails the monotonicity criterion.

Since you know what monotonicity is, you no doubt also know that all voting methods fail some plausible-sounding criteria (Arrow). And that all voting methods reward tactical (insincere) voting behavior in some situations (Gibbard-Satterthwaite).

Asserting that method X fails criterion Y is just idle chatter unless you are willing to also reveal what criteria are not met by the method you are proposing instead of X, as well as the circumstances in which your method compels tactical voting.

I would call who wins part of the underlying reality of an election.

Touche. I picked the wrong word. But my point remains, changing the vote counting method doesn't change the underlying realities that lead to candidates getting less than 50% of the vote in a first-past-the-post system.

all voting methods fail some plausible-sounding criteria and reward tactical (insincere) voting behavior in some situations

True, but...

1) The primary voter picking the centrist candidate more likely to prevail in the general election is engaged in tactical voting; but I see that form of tactical voting as the most socially desirable form of tactical voting.

2) I just can't summon the same gut feeling about a voting method's indeterminacy in the case of a condorcet paradox as I feel about IRV.

That doesn't mean the way we do it now is perfect, but if you are talking about implimenting a system that would allow IRV, there are other counting methods that seem more attractive.

Consider also that party primaries might serve desirable functions other than dealing with Independence of Irrelavent Alternatives, the main criteria that our current system fails.

JimM47: party primaries might serve desirable functions other than dealing with Independence of Irrelevant Alternatives, the main criteria that our current system fails.

We've wandered pretty far off the original topic. But I do have to say (1) I think the main criterion that the current system fails is majority rule; (2) while IRV allows you to eliminate party primaries, it doesn't require that you do so.

Post a Comment

Older Posts
Newer Posts