Why is it so hard to predict Israeli elections?

Why is it so hard to predict Israeli elections?

After weeks of polling showing him with a comfortable lead, it appears that Benjamin Netanyahu will sneak back into office with his Likud/Yisrael Beiteinu alliance taking just 31 seats. (The two parties had more than 40 between them before the vote.) Overall, exit polls suggest right-wing parties will take 61-62 seats with 58-59 for the center-left.

The big surprise of the day will probably be the rise of the centrist Yesh Atid party, which came in second with 18-19 seats. Yesh Atid had gotten relatively little covereage in the Israeli andinternational media (certainly compared to Naftali Bennet’s annexationist "Jewish Home" party) until it began to rise in the polls just days before the vote. Yesh Atid is led by journalist-turned-politican Yair Lapid, son of the legendary secularist crusader Tommy Lapid.  Yair may not go as far as his dad in expressing stridently anti-religious views, but his apparent success should lead to some questioning of the popular "death of Israeli secularism" narrative.

There’s likely to be quite a bit of discussion in the coming days about why no one saw Lapid coming. Though of course, as Netanyahu well knows, even the exit polls in Israel are not always that reliable. On May 30, 1996, most Israelis went to sleep believing that the heavily favored Shimon Peres, who had served as caretaker Prime Minister since the assassination of Yitzhak Rabin, six months earlier, would remain in power. They awoke to find that Likud’s then young and inexperienced leader, Netanyahu, had pulled off a historic upset. 

As the Times of Israel‘s Raphael Ahren notes, it wasn’t the last time Israeli pollsters would get it wrong: 

In 2006, most pre-election polls shortly before election day predicted two seats for the Pensioners’ party; four polls said the party wouldn’t get any seats at all. It won seven – a substantial showing given that the Knesset has only 120 members.

The final polls ahead of the 2009 elections – published four days before the voting day opened – were all wrong to a greater or lesser degree. Four out of six polls predicted 23 seats for Likud, which ended up winning 27. Four surveys forecast Yisrael Beytenu winning 19 mandates, four more than it actually received.

According to Haaretz, the final polls before of the 2006 and 2009 elections erred by an average of 18 and 19 Knesset mandates, respectively. Nate Silvers, they were not.

So what makes predicting Israeli elections so tough? Ahren’s very thorough article blames small sample sizes — Israeli pollsters often survey as few as 400-500 people — as well as poor methodology: In a year when voter apathy was high, most pollsters surveyed "eligible" voters rather than "likely" voters, for instance. 

The blogger "Carl in Jerusalem" says Israelis are poor poll-takers by character:

This is a country in which polling is not always accurate. Respondents often deceive the pollsters. Some respondents – like me – slam down the phone (otherwise you get called nearly every day because Israel is very politically active and there’s a relatively small population).

But I suspect it’s not that Israeli polls are particularly bad or Israelis are particularly uncooperative than the fact that Israel is simply a very small country with very complicated politics. Before this election, there were 17 parties in a Knesset with only 120 seats. A party going from 3 to 12 seats, as Jewish Home appears to have, counts as an earth-shattering development. In short, pollsters have a lot less margin for error in a country with fewer people than New Jersey.

There’s also the fact that political alliances in Israel have a habit of shifting quickly and — particularly this year — it’s much harder to define any given party’s natural base. As Noah Efron recently wrote:

Unsurprisingly, this growing diversity within parties has come at a time when old voting blocs have begun to disintegrate. Although the million or so immigrants from the former Soviet Union still tend to oppose territorial compromise with the Palestinians and reject welfare-state economic policies that recall the brutal socialism of their birthplace, their political affiliations are increasingly spread across the political spectrum. Israeli Palestinians, though they are largely ignored by Jewish media and politicians during elections, will vote in larger numbers than ever for majority-Jewish parties, chiefly Labor. Settlers’ votes are also spread among more political parties than in the past (though almost exclusively on the right). The same is true of Mizrahim, Haredim, and, no less, secular cosmopolitans.

The decline of the old voting blocs has come with a decline of old ideologies as well, on the right and on the left. Among the most notable losers in the Likud primaries was Benny Begin, son of Menachem Begin, the legendary founder of Likud and its first prime minister. The younger Begin represented perhaps the last of the old Likud ideologues, whose commitment to retaining the West Bank was matched, perhaps incongruously, by a commitment to liberal democracy blind to religious and ethnic background. Most politicians today, on the right and left, insist that they maintain consistent political opinions; few, however, will cop to having an ideology. In the past, ideology was de rigueur; now it is vaguely déclassé.

Taken together, these trends suggest that Israeli politics have recently lost definition and grown shaggier. They have changed from a French garden, sharp of line and in fine trim, into an English garden in which the shrubs and the trees have expanded into one another, and a skein of ivy stretches from this plant to that.

Efron’s primary argument is that there are dramatic developments happening just below the surface in Israeli politics that are a lot more complicated than they look from they outside. Today’s results would seem to bear that out. 

(Plus, to be fair to Israeli pollsters, even Nate Silver hasn’t fared so well when trying to apply his model outside the two-party, data-rich environment of American politics.)