Winning, despite the polls

The polls were right, big-time, and also wrong big-time. They were right about everything having to do with the breakdown of the blocs: not a single poll in recent months predicted that the center-left would win, whereas they all said the Right would. Most the polls gave the right-wing bloc a win of 10 seats or more. When all the polls are averaged, the Right was predicted to win 67.5 seats, compared to 52.5 for the left-wing bloc. The polls scored a bull's-eye.

But the pollsters working within the blocs had a mea culpa moment after some harsh blows. Almost all of them – the three main TV stations, newspapers, and internet site polls – gave the two biggest parties a combined 60 seats. But in the last three days before the election, Prime Minister Benjamin Netanyahu and Blue and White leader Benny Gantz took seats from their satellite parties – which was not wholly unexpected – and put them at a combined 70 or so seats. Now, for the first time since the 1990s, the two biggest parties hold a majority of seats in the Knesset.

The blind spots

In most of the pre-election polls, the Likud was behind Blue and White because pollsters predicted that the parties to the right of the Likud politically would win 10 to 12 seats. But voters decided to give those seats to Netanyahu. That's what happened with the New Right under Naftali Bennett and Ayelet Shaked, which the last few polls predicted would win six seats but as Thursday was uncertain to make it past the minimum electoral threshold; that's what happened to Moshe Feiglin's Zehut party, which had five to six seats in the last polls but would up shut out of the Knesset; and that's what happened with the Union of Right-Wing Parties (Habayit Hayehudi-National Union-Otzma Yehudit), which polls said would win six to seven seats but whom voters wound up giving a mandate and a half fewer.

Something similar happened on the Left: Pollsters predicted 30 seats for Blue and White; nine to 11 for Labor and five to six for Meretz. At the last minute, Gantz and his party pulled four seats away from Labor and one from Meretz wound up with 35.

Pollsters didn't predict what happened within the blocs. Very few votes, if any, moved from one camp to the other. Blue and White won 35 seats, the exact same number garnered in 2015 by Yesh Atid under Yair Lapid and the Zionist Union (11 and 24, respectively). The right-wing was frozen, too. Netanyahu's decade in power began with 65 seats for the right-wing bloc, just like this time, but 10 years ago the intra-bloc spread was different (27 for the Likud, 15 for Yisrael Beytenu, 11 for Shas, five for United Torah Judaism, four for National Union, and three for Habayit Hayehudi).

So where did the pollsters go wrong, and where did they get it right?

1. The pollsters did not predict the large number of votes that Netanyahu, and to some extent Gantz, took from their satellite parties.
2. The pollsters underestimated the strength of the haredi parties, repeating the mistake they made in the polls that came out before the Jerusalem municipal election last year. Then, and now, the haredi voters proved themselves to be disciplined soldiers who had a much higher turnout than the general public. Then, and now, the pollsters proved they have difficulty understanding that sector, which doesn't spend much time online or on Facebook and even shuns polls. In the 20^th Knesset, the haredim held 13 seats; in the 21^st Knesset they increased their strength to 16. But polls showed them at about 12 seats, four fewer than what they got.
3. The mainstream pollsters were very wrong about the strength of the religious Zionist parties. If indeed the results as they appeared Thursday are correct, we are seeing an all-time low in the organized electoral strength of the religious Zionists. The religious Zionist public, unlike the haredim, has spread its votes around several different parties for a few elections already. This time, it divided its ballots between the New Right, Likud, Feiglin, the Union of Right-Wing Parties, and Blue and White.
4. The polls were more accurate when it came to the Arab parties. On average, pollsters gave Hadash-Ta'al seven to eight seats, and the party won six. They gave the Balad-Ra'am list anywhere from 0-4 seats and it appeared that the second Arab list made it past the minimum threshold, winning four. Votes by prisoners – many of whom are Arabs – could push Balad-Ra'am into the Knesset.
5. The last polls by Channels 12 and 13, as well as the Maariv daily, were right about Avigdor Lieberman's Yisrael Beytenu party, predicting it would win four to five seats after Lieberman dropped off the radar for a while and people were predicting he wouldn't make it past the minimum threshold at all. Lieberman, it turns out, was right when he claimed repeatedly that the pollsters skipped over him and his supporters, who do not tend to cooperate with polling organizations.
6. The pollsters also managed to maintain their dignity when it came to Kulanu under Finance Minister Moshe Kahlon. For weeks, it looks as if Kahlon might not be in the next Knesset at all. Only in the final two weeks did he recover, and pollsters spotted that. The last six polls by the major news outlets, including Israel Hayom, gave Kahlon five seats. In the end, he won four.
7. However, they were very wrong about Feiglin, who was pegged as the surprise of the election but who turned out to be a "Fakelin." Some pollsters gave him as many as six to seven seats. He didn't make it past the threshold.

Protest votes, votes leeched, and droves of votes

Most of the polling failures were political ones, rather than a lack of professional acumen by the people conducting them. That could indicate that we need a new model to predict results. Every pollster should be assigned an experienced political strategist, and every political strategist should have a professional pollster. Only by combining the two can we do better work, it appears.

The pollsters, who do not publish any results in the three days leading up to the election, couldn't have predicted processes that were mainly political in nature – processes that even political analysts pointed out only in the last few days of the campaign when they couldn't be publicized.

The first such process was the lower voter turnout among Arabs, some of which was due to a call to boycott the election over the nation-state law passed by the 20^th Knesset. By 6 p.m. on election day, few Arabs had voted. Only in the last four hours, after a scare campaign in the sector, did voter turnout rise. But it was still lower than in previous elections.

The second process was one of leeching votes. Pollsters only have the material in front of them to work with, and four days before the election was still before the big vote leech. They couldn't have seen that coming, since it happened when they could no longer put out poll results.

And the last process, exactly like 2015, was the arrival of loyal Likud voters in the final four hours of the voting, which put the color back in the party's cheeks and the seats back in Netanyahu's pocket.

The pollsters weren't the only ones who got it wrong. The political establishment also made errors. The assumption as that it was dream for Netanyahu and the Likud to win 30 seats, the number they had last term in office. But like he did in 2015, Netanyahu proved them wrong. For the second time in four years, his cries of "Wolf!" proved effective. His "gevalt" campaign convinced more voters than the gevalt of the New Right, Feiglin, or the Union of Right-Wing Parties. Netanyahu's performance in the final days was nothing less than amazing: he gave dozens of interviews to mainstream, local, and even fringe TV and radio stations. The prime minister needed the hostile media, and after four years in which he gave almost no interviews. At the same time, he used Likud TV and social media so extensively that it will be studied by communications students. In the last few hours of voting, he launched a live Facebook stream, which he commanded with considerable talent.

Netanyahu ran a three-pronged campaign, two of which he won. He beat Blue and White, which is not the biggest party. He beat a long list of right-wing parties that threatened to gnaw away at him and put him at a disadvantage against Blue and White, and he believes that his legal position has improved, either allowing him to pass an immunity law or by putting together an "indictment coalition" that will allow him to remain prime minister during a trial, if he fails to get the cases closed in a pre-indictment hearing.

Aligning the stars

"All the stars aligned for Netanyahu," political analyst Amit Segal said on election night, but in truth, it was Netanyahu who took care to align them. He might have won a referendum on himself – since the election was essentially about voting for or against Netanyahu – but he also attached a list of achievements to his personal brand that convinced voters. He started the campaign in terrible conditions – three potential indictments for serious offenses, and three former chiefs of staff who created a party out of nothing in only two months.

Against the generals and Attorney-General Avichai Mendelblit, Netanyahu set up his own arsenal: foreign relations and his closeness to and influence with leaders of the two strongest world powers (U.S. President Donald Trump and Russian President Vladimir Putin.) These special ties led to major, visible achievements: American recognition of Jerusalem as the capital of Israel and the relocation of the U.S. Embassy to Jerusalem; U.S. recognition of Israeli sovereignty on the Golan Heights; the U.S. withdrawing from the 2015 nuclear deal with Iran; and the U.S. declaring the Iranian Revolutionary Guard Corps a terrorist organization. Putin gave Netanyahu active help locating the remains of fallen IDF soldier Zachary Baumel, who had been missing since the First Lebanon War, and has been coordinating with him to allow Israel to carry out consistent strikes against Iranian attempts to gain a foothold in Syria and use it to attack Israel.

All these pushed aside socioeconomic issues and the parties that backed them, whether it was health care or the economy. The socioeconomically oriented parties did particularly poorly in this election: Kahlon lost 50% of his electoral strength; MK Orly Levy-Abekasis didn't make it past the minimum threshold; and even Feiglin was out.

Netanyahu's handling of the ongoing Gaza Strip crisis, and his restrained responses to Hamas this past month, didn't appear to have much effect on voters, either. Even in southern Israel, in a city like Sderot, Netanyahu won a clear plurality of 44% of the votes, compared to only 9% for Blue and White.

We, the public, show up for the pollsters' performance, but they almost never manage to predict the final results. There are too many unknown factors involved. Two of them are especially important in the predictions that shape up in the month preceding an election: the first is the number of participants in the polls. Most people contacted by pollsters, sometimes 80%, decline to take part. The second factor is that many who do participate say they haven't decided for whom to vote, and even more say whom they'll vote for, but note that they are not certain they will vote at all.

Just like in the 2015 and 2013 elections, it turned out that this time many people made their decision at the last minute, sometimes on their way to the polling place. So maybe the polls, which not only aspire to predict reality but also help shape it, are unnecessary. Perhaps we should limit their use.