The extent to which the electoral system distorted the results of last week’s Palestinian legislative elections becomes all the more clear when the analysis is conducted at the district level, with the full votes for losing as well as winning candidates. The complete data also make it clear how much stronger the cohesion of Hamas voters was than was the case for Fatah voters (as I suggested prior to the election that it would be), and how dependent Fatah was on the personal vote of some of its individual candidates to win as many seats as it did.

First of all, consider the distribution of party pluralities across the districts. Using the list vote, Hamas led in eight districts and Fatah in eight. The districts vary in their magnitudes (i.e. the number of seats they elect). The eight districts in which Hamas led elect 39 of the 60 nominal-tier seats, and the districts in which Fatah led combine for 21. So, Hamas did better in the more populated districts, as would be expected from its lead in the national tier in which the list-PR seats were allocated. But the extent to which Fatah was hurt by the specific electoral system used in the local districts is striking. In two districts where Fatah had the most list votes, it managed to elect no candidates, and in two others where it led in list votes it elected fewer candidates than Hamas.

Here are the vote and seat totals in the districts in which Hamas led. Both votes percentages and seats are given as Hamas/Fatah:

Jerusalem, 41.9 / 35.8, 4 / 0
Tulkarem, 43.5 / 39.3, 2 / 0
Nablus, 44.7 / 38.2, 5 / 1
Salfit, 35.4 / 34.4, 1 / 0
Ramallah, 42.2 / 38.8, 4 / 0
Hebron, 49.5 / 36.0, 9 / 0
N. Gaza, 46.9 / 41.4, 5 / 0
Gaza, 56.7 / 36.6, 5 / 0

Now the districts where Fatah led, again giving the Hamas figures first:

Jenin, 37.8 / 45.3, 2 / 2
*Tubas, 38.8 / 40.1, 1 / 0
Qalqilya, 32.4 / 53.5, 0 / 2
Jericho, 33.1 / 50.0, 0 / 1
*Bethlehem, 31.5 / 44.2, 2 / 0
*Deir al-Balah, 43.9 / 47.4, 2 /1
*Khan Younis, 44.0 / 48.1, 3 / 2
Rafah, 40.4 / 53.2, 0 / 3

(The starred districts are those in which the party that led the party list votes did not win the most seats.)

Notice that Hamas beat Fatah 35-1 in seats in the districts in which it led in the party vote, while Fatah managed only a 10-10 split of the seats in the districts where it was the leading party (other seats were won by independents).

The reason a party could run ahead of its competition in the list vote, yet fail to lead in the seat allocation lies, of course, in ticket splitting, whereby some voters give their party-list vote to one party, but cast some or all of their candidate votes for candidates of other parties (or independents).

We can analyze this phenomenon more closely by looking into the ratios of list to candidate votes for each of the leading parties. If a candidate has a personal vote, the ratio of his vote to his party’s will be greater than 1.00. If he is less popular than his party, the ratio will be less than 1.00.

In the Palestinian context, the nominal tier used an MNTV system (multi-seat plurality, with the voter allowed to cast as many votes as there are seats). So, we need to compute the ratio for the leading candidate and the last candidate (and we could do so for all a party’s candidates in a district, but let’s not get too carried away, fun though this is!).

Averaging across the 16 districts, the ratio of Hamas’s leading candidate’s votes to the party-list vote is 1.01. In other words, on average, the leading Hamas candidate obtained almost the same vote share as the party list. The lowest district ratio was .88 (Jericho) and the highest was 1.14 (Gaza). In twelve of the sixteen districts it was .95 or better.

For Fatah, on the other hand, the average ratio of the leading candidate’s vote to the party vote was .91, suggesting that Fatah candidates tended to be less popular than the party that nominated them. The lowest was .68 (Bethlehem) and the highest was 1.15 (Jericho, which elects a single member). In only four districts was it greater than 1, and in seven it was less than .90.

Taken together, these ratios reveal that Fatah’s leading candidates were less popular than the party as a whole, while those of Hamas were almost exactly as popular as the party. That fact alone tells us a lot about the extent to which this electoral system’s emphasis on individual candidates hurt Fatah. But that is not all. One of the challenges for a party under MNTV is to ensure that its voters use all their votes in the candidate races and cast them all for the party’s candidates. Voters are free to partially abstain or jump around and give some votes to candidates of one party and some to those of other parties or independents. Especially in districts that elect several members, MNTV thus poses a real challenge to a party’s candidate-recruitment and vote-mobilization efforts.

So, to determine the extent to which voters for each major party were loyal enough to give most or all of their votes to the party’s candidates, we can look at the ratio of the last candidate of the party to the party list votes in each district.

For Hamas, the ratio of the last candidate’s vote to the list vote averaged .89 across all thirteen multi-seat districts, while for Fatah it averaged .73. Two things stand out here. These figures show that there was a ballot “fall off” for both parties, either because voters got tired of marking ballots and stopped or because attractive candidates from another party pulled them away from some of their own party’s candidates. Yet note how much greater the Fatah fall off is for the last candidate. Not only did Fatah’s slate of candidates start off lower relative to its party list, but the numbers suggest a greater “tiredness” or “temptation” by other parties’ candidates for Fatah than for Hamas as voters cast (or didn’t cast) their multiple votes.

In some of the districts, the fall off for Fatah was really striking. In Jerusalem, the fourth and last Fatah candidate had just over half the votes of the Fatah list, and in Tulkarem the third and last had less than half.

This analysis underscores the extent to which Hamas had the more party-loyal electorate. They also show that Fatah had some individual candidates who were quite strong personally, or else they could not have broken through the more solid Hamas electorate. And they should remind us again that it was not the Palestinian people who gave Hamas such a sweeping victory (74-45 in seats, but only 45-41% in party votes). It was bad electoral-system design.