After four years of continuous campaigning and five election cycles, finally an election has ended in a knockout. While coalition negotiations are yet to formally begin and more surprises could theoretically be in store, it is now highly likely that Netanyahu will get his wish and be able to form a right-wing coalition.
Much has been written about the election and its implications at home and abroad. This column does not seek to add to this debate.
Instead, as we have done throughout the campaign, we will focus on the data, answering two key questions: What happened to shift a very close election to a relative blowout? And what does this mean for polling in Israel moving forward?
Throughout the four-month campaign, the polling was neck and neck, with both the Netanyahu and anti-Netanyahu blocs hovering around the 60-seat mark. In our final pre-election column, we noted that two things could change this and create a markedly different result: differential turnout and the threshold.
That is exactly how it played out. Simply put, differential turnout secured the victory for Netanyahu’s bloc, and the threshold ensured the relatively wide margin.
Let us consider them one by one.
This was never going to be a persuasion-based election. With so much water under the bridge, the likelihood of getting large numbers of voters to switch sides always looked unlikely. And so, much of the campaign was about generating differential turnout, mobilizing the persuaded rather than convincing the undecided: bringing more supporters to the polls than the other side could.
The first thing to note is that overall turnout actually went up, from 67% in 2021 to 71%. Irrespective of everything else, it is quite remarkable that the Israeli public’s response to election after election is actually to go to vote in greater numbers.
What is more significant, though, is who went to vote in greater numbers, and who did not. Center- and left-leaning cities in the center of the country – such as Tel Aviv, Herzliya, Kfar Saba, Hod Hasharon and Raanana – saw turnout figures virtually unchanged from 2021. In contrast, cities that lean rightward, such as Beersheba, Jerusalem, Ashkelon, Ashdod, Tiberias, Kiryat Gat and Afula, all saw turnout rates increase by between three and seven percentage points.
The four percentage point increase in turnout, therefore, came largely from areas that lean right.
Looking at the raw vote figures further illustrates this point. In 2021, 2.22 million Israelis cast their votes for parties opposed to Netanyahu, as opposed to 2.13 million for pro-Netanyahu parties. This time, while the non-Bibi bloc grew its vote by 5 percent, to 2.33 million, the Netanyahu bloc grew by 11 percent, to 2.36 million. All in all, Netanyahu added 230,000 votes to his bloc from the previous round, while the other side added half as many.
What happened is even clearer when we break down the blocs into their component parts:
On the right, the biggest shift was an unprecedented turnout among supporters of Haredi parties, who managed to increase their number of voters by 19 percent. The gain among the non-Haredi right (Likud, Religious Zionism and Jewish Home) was a more modest, but nonetheless significant, 8 percent.
In contrast, the Zionist center and left (Yesh Atid, National Unity, Labor and Meretz) saw a one percent decrease in their number of voters despite the enlarged electorate, showing how their attempts to motivate and mobilize their voters fell short.
After weeks of talk about Arab turnout, in the end the Arab citizens turned out in large numbers, with the vote for Arab parties increasing by 35 percent. This, however, was ultimately unable to make up for the drop-off among the center-left.
In total, Netanyahu’s bloc very narrowly “won” the “popular vote” (a heavily disputed and questionable concept in this context, but interesting nonetheless) for the first time since April 2019.
While Netanyahu’s narrow 30,000 vote “victory” explains how his bloc would have reached 61 or 62 seats, it does not explain his significant majority. For that we need to look at the threshold.
As noted throughout the campaign, Netanyahu built his bloc in the optimal way: Four parties, each with a surplus-vote-sharing agreement, with one other smaller party that he attacked mercilessly to ensure it hardly got enough votes to matter. In total, his bloc “wasted” around 56,000 votes, or 1.5 seats.
On the other side of the map, the situation was very different. With four parties hovering just above the threshold and another just below, there was always a strong chance for wasted votes that would prove costly. In the end, Meretz, the veteran left-wing party, fell about 4,000 votes short of the threshold, while Labor just about crossed. The outcome meant that despite the two left-wing parties winning enough votes for around eight seats, they ended up with just four.
This has triggered a wave of mutual recriminations on the left and center.
Without wading into the blame game, we would highlight the point made frequently in these pages throughout the campaign: when elections develop a head-to-head dynamic, with two big parties competing to be the biggest, voters tend to get drawn from smaller parties as the last minute. While Meretz crossed the threshold in every poll over the past two months (albeit just about), this dynamic means that its falling underneath should not be a surprise.
The other party not to cross the threshold was the Arab Balad party, which ended up with 2.9 percent or about 15,000 votes short. This means that while the Arab parties won a combined 511,000 votes, virtually the same as Religious Zionism, they ended up with just ten seats. Ultimately, the 130,000 increase in the Arab vote since the last election, and the successful effort to increase their turnout, was largely meaningless.
In total, the anti-Bibi bloc wasted 289,000 votes, worth around seven seats – the difference between a very close election and a resounding defeat.
What about the polling?
As ever, when election results are out of keeping with the polling, questions are asked about the quality – and even utility – of public opinion polling. This has become a familiar argument across the world, and as those who have noted the accuracy of Israeli polling in recent cycles, this is an important question for us to address.
The biggest challenge for pollsters in Israel surrounds the threshold. Meretz dropped under the threshold by around 0.1 percent, which is far beyond the accuracy of any poll. This in turn boosted the other parties, as the Meretz vote gets allocated elsewhere, all adding to the perception of polling “error.”
If we are being extremely critical, we could point to the fact that Meretz averaged 4.6 seats in our final average, and did not drop under in even one poll, but nonetheless, the polls showed that it was consistently in the danger area – always within the margin of error – and the fact that it dropped under therefore should not be a major surprise.
Had Meretz gained those extra 4,000 votes, the Netanyahu bloc would have won 61, or more likely, 62, seats, which is not far off the polling average of 60.3. Further analysis down the line will be able to expand on this further, but it is likely that the late surge in turnout among the right, and in particular among the Haredim, was responsible for this discrepancy.
On the issue of the Haredi parties, however, there is much room for improvement. Despite polling consistently at a combined 15 seats throughout the campaign, the Haredi parties ultimately won 19 seats, the fifth time in a row they have exceeded their polling. Twice – in April 2019 and this time – this has been by a very large margin.
Given the particular nature of the Haredi community, Israeli polling companies will need to invest further in new ways of reaching this community more accurately moving forward so as to ensure this is not repeated.
On the whole, however, the polling was pretty accurate. With the exception of Meretz and the Haredi parties, every other party ended up within a seat of their average, with several right on the nose. In an extremely fluid multiparty field, that is not to be sniffed at.
Furthermore, even with Meretz narrowly missing the threshold, the average “error” by party in this election was 1.2 seats, well in line with the average over the past four rounds.
In the end the polling fulfilled its purpose of giving us an accurate picture of the state of the race. It showed us that the election would go right to the wire, a seat or two either side of 60-60, unless a party on the left dropped under the threshold. It also showed us that four parties were extremely close to that threshold. While it was not able to identify exactly which party would in the end drop under, it did its job.
Which brings us back to where we started with this column, four long months ago, with our borrowed and misappropriated Winston Churchill quote: After five election campaigns in quick succession, the reality is that polling remains the worst way of measuring public sentiment, except for all the others that have been tried.
Simon Davies and Joshua Hantman are partners at Number 10 Strategies, an international strategic, research and communications consultancy, who have polled and run campaigns for presidents, prime ministers, political parties and major corporations across dozens of countries in four continents.