Most Americans in Israel are in favor of personal elections for the Knesset. It's what we are used to and, having experienced both proportional and direct elections, we know that directly elected politicians tend to be more responsive to their constituents. But a sober evaluation of the alternatives suggests that the issue is not as clear-cut as some think.

Consider, for example, the election reform bill recently proposed by four MKs from across a spectrum of political parties. Briefly, the idea is that only 60 of 120 MKs would be elected in the proportional method currently used; the other 60 will be elected one each from 60 regions in first-past-the-post (FPTP) elections. (FPTP means simply that each voter chooses one candidate and the one with most votes wins -- the method we are used to.) In other words, each voter has two ballots: one is for voting for a regional candidate affiliated with one of the parties and the other is for voting for a party.

Sounds pretty innocent. Sort of half-way between the current proportional method and the American method of voting for congressmen. What could be bad? But you might have second thoughts if you look at the results of a simple simulation done by my friend Avram D. Avram computed what the distribution of Knesset seats would have been in the 2003 and 2006 elections had the proposed new method been used instead of the old one. (He assumed that every voter would have simply used both ballots to vote for the same party; experience in other countries with similar methods suggests that this is a plausible assumption -- but see caveats below.)

Here is what would have happened (actual results are in parentheses):
In 2003, Likud (38) would have gotten 87 seats. Shinui (15), Shas (11), Ichud Leumi (7), Mafdal (6), Meretz (6) all would have been wiped off the map. In 2006, Kadima (29) would have gotten 58 seats and the Likud would have been wiped off the map.

How does the thought of Sharon with 87 seats grab your fancy? Or Olmert with 58?

The reason this happens is plain to see. A party that has a slight edge in each region gets all the regional seats in a disproportional manner. (In the case of the particular method proposed here, the problem is exacerbated by a stipulation that a party that doesn't win any regions is disqualified from getting any seats in the proportional part as well.)

It should be noted that what will really happen if the proposed method were adopted is a bit different than these numbers suggest. According to a straightforward principle known as Duverger's law, FPTP elections inevitably result in a two party system, or at the very least each region becomes a two-party race, because voters will simply stop voting (in the regionals) for any party that has no chance of winning. So it is likely that some of the votes that went to small parties in the regions where they lost badly would actually go to a larger party such as Likud.

While I understand perfectly well the huge advantage in terms of accountability conferred by direct elections, I am scared to death of the thought of any party having 87 seats. I think a plethora of parties is a blessing not a curse; it makes compromise both necessary and possible. (The more parties the more potential coalitions.) To be sure, I think datiim would be better served by having clout in the major parties than by being represented by ineffective politruks (same for Arabs, Russians, etc.), but even so I'm not ready to contemplate a two-party system here quite yet.

One very partial remedy would be for the regional elections to use some system other than FPTP. There are many such systems, each with advantages and disadvantages. In my next post, I'll give an overview.