As you may have seen, Israeli Prime Minister Benjamin Netanyahu scored a clear victory in last night’s Israeli election that looks all but certain to keep him as Prime Minister in the next government. It is important to note that Netanyahu scored not one but what amounts to two major upsets last night.
The last polls published on Friday before the pre-election ban on polls showed the Zionist Camp with roughly a four seat margin over Likud. When the first exit polls were released on Tuesday afternoon at 4 PM eastern, they told a dramatically different story: both ZC/Labor and Likud locked in a tie at 27 or 28 seats.
That was a clear victory against expectations for Netanyahu. And it put him in a better position to form a coalition. But it was nearly as clear a victory as initial press reports suggested.
Only that was not the end of the story. Far from it. The exit polls turned out to be dramatically off as well. The final votes gave Likud 30 seats to ZC/Labor’s 24, a 6 seat margin. (Final tallies since I wrote this piece now suggest the final Likud number may be 29 mandates, not 30.) The Zionist Camp’s final number is almost exactly where the consensus of polls put them on Friday. The Haaretz poll of polls gave them 25 seats; they ended up with 24. But Likud got – depending on which polls you’re looking at – as many as 50% more seats than the polls predicted.
There is no getting around the fact that this is a big, big upset. The polls were totally wrong.
So what happened? There are three potential explanations – and probably three of them play some role.
1. The polls were simply wrong all along. This is almost never the case in a closely polled US election. In fact, polls are almost never this wrong anywhere in a closely polled contest. But the dynamics of an Israeli election are different. It’s not a binary question. It’s not if you don’t vote Democrat, you must vote Republican. Voters have a range of potential options and there are various levels of strategic voting. If you think Zionist Camp has it locked up, you may be more likely to vote for Yesh Atid, assuming that it will change the complexion of center-left coalition, not defeat it. On the right, you may prefer the Jewish Home party. But if it means blocking a center-left government you may happily choose to vote for Likud. (This is precisely what happened with many voters. And it seems to be the major story of the night.) Add to this that in Israel there are big chunks of the population that are difficult for pollsters to get in touch with. Put it all together. And polls in Israel simply are not close to as accurate as they are in the US. So maybe they were just off all along.
2. Maybe the polls were right on Friday. But since there weren’t any polls after Thursday night there were four days where the trend may have shifted dramatically. We just didn’t know it. In this scenario, the polls were right: they just couldn’t tell us what was happening during those four crucial days.
3. Netanyahu made a late and intense push for the votes of the far right. There were at least three major components of this: a) a series of alarms that he was being driven from office by a nefarious cabal of foreign governments, tycoons, the media and NGOs, b) a categorical refusal to ever allow the existence of a Palestinian state and then c) election day race-baiting scare tactics warning that Arab Israelis were deluging the polls and threatening to remove him from office.
As you can see, these explanations overlap. They may all be true. It’s irrelevant for the final result. The votes are the votes. But it would be interesting to know whether the polls were simply off or whether Netanyahu’s late go-for-broke charge turned the tide in his favor. We may learn more in the coming days as campaign or other groups release tracking or persuasion call data. But there are reasons to think that Netanyahu really did dramatically turn the numbers in his favor.
Remember, for weeks and especially in the last few days, Netanyahu was not trying to grow the center-right bloc. He was trying to cannibalize the other right wing parties to add to Likud’s number. He succeeded.
Almost all of Likud’s gains came from other parties on the right. But not totally.
Let’s take the Haaretz poll of polls as our final poll baseline. If you add up all the right wing and religious parties you get 55 seats. The final election results come to 57 seats. The polls were pretty accurate for the size of the right/religious bloc. And they were pretty on the mark for the all the parties beside Likud and the further right Jewish Home party. In fact, if you compare the final poll of polls to the final result I think I am right that beside those two parties, every party’s final results were no more than one seat (either higher or lower than its final poll result). The one exception is the Yahad party, which missed the electoral threshold and didn’t get into the Knesset at all. But even there they seemed to come one seat under their predicted number.
I do not think that level of accuracy is an accident. We are talking about 9 parties – each with the final poll prediction within one seat of the actual election result. Pretty damn good. Not conclusive certainly. But I think that suggests that there wasn’t some dramatic flaw in the polling. What happened was a big move within the right bloc that coincided with Netanyahu’s hard pivot to the right in the last five days and frank statements to right wing voters that he was on the verge of losing. That all comes mainly after the last polls were done.
Netanyahu did something that would likely be fatal in the binary US election system. He spent the last 5 days telling voters that he was likely going to lose and that they needed to save him. In the US that would prove wildly demoralizing and would probably accelerate a candidate’s drive toward defeat. But remember, he didn’t need votes from the center or the left. He needed them on the right. In most cases, further to his right – voters who were deeply committed to preventing a center-left, peace-bargaining government. To those voters what might be taking as Netanyahu’s admission of weakness was not demoralizing but frightening.
For those voters, Netanyahu’s basic appeal made very good sense. Voting for another right wing party meant increasing the chances of a a center-left government. Netanyahu told these voters to keep him in office they had to vote directly for his party. He was right. And that is what a lot of them did.
So how did the exit polls miss the scope of the Likud win?
Here’s an educated guess. On Tuesday, a source shared with me mid-day (Israeli time) exit polls that showed ZC with a clear lead. I believe those numbers were accurate. But look at the turnout trend over the course of the day. For most of the day, turnout levels ran just under the totals for 2013. As late as 6 PM Israel time, they were 1 percentage point under the 2013 levels. By 8 PM they were 2 points above. And by the time polls closed at 10 PM they were a whopping 5 percentage points over the 2013 turnout levels. That is a huge, huge surge of voters in the last three or four hours of the voting day.
If the midday numbers I heard were accurate, and if we take into account that the initial release of exit polls at 10 PM (Israel time) dramatically understated Likud’s margin, it seems reasonable to assume that this huge surge of voters at the end of the day voted overwhelmingly for the right and for Likud.
As painful as it is to admit, Netanyahu’s hard appeal to the most chauvinistic and let’s say it – racist – elements of Israeli politics worked. But I do not think it was only that. He was also making a very rational appeal to voters who were already going to vote for right wing parties. It was both. Why those voters turned out in such big numbers in the final hours of voting I’m not clear on. I’m curious to hear from others more familiar with Israeli politics for insight on that. But it seems a very reasonable surmise to assume that a rush of right wing voters coming in the last two or three hours of voting played a role in the exit polls being as far off as they were.