Bernie Sanders had a big weekend, notching victories in the Alaskan, Hawaiian, and Washingtonian caucuses. He won each state by large margins – the Alaska caucus by 63, the Hawaii caucus by 69, and the Washington caucus by 45. A boon to his momentum and fundraising (he brought in around $4 million in the days following the caucuses, ensuring he has the financial resources to invest heavily in Wisconsin and other upcoming primaries), the victories did not change the race’s trajectory, as we have documented. That said, while it remains to be seen whether Sanders can turn these victories into lasting momentum, a positive sign for his campaign emerged from the night: Sanders beat expectations, necessary in all the upcoming caucuses and primaries if he hopes to win the nomination.
First of all, we define “expectations” as our delegate estimates based on PoliticalEdu’s model. Democratic delegates are awarded proportionally, so we simply multiply vote share estimates by state delegate totals to create a benchmark to which we set our expectations. The model explains around 96 percent of the variation between actual vote shares, so it is quite accurate and continues to “learn” as more results become available (we add each result to the model to hone and refine the relationship between variables, allowing the model to adjust to new electoral inroads a candidate makes and the changing primary landscape). It is, of course, not perfect and thus our delegate targets are not perfect, but the model’s high degree of accuracy and focus on underlying electoral factors provides a strong baseline estimate; beating the model’s expectations shows a campaign that overcomes earlier trends and gains votes at the expense of the opponent.
With expectations defined, let’s turn to Saturday’s results. Our model showed Sanders taking 74 percent of the vote in the Alaska caucus, sufficient for 12 delegates. Here Sanders exceeded expectations. He took 13 delegates from the Alaska caucus with 81.6 percent of the vote. Though he clearly beat the model and its delegates expectations, doing so only led to his gaining a single extra delegate. Alaska is an incredibly small state and only had 16 delegates at stake. Beating expectations in Alaska is fairly meaningless – Sanders surpassed his vote share target by around 10 percent and got only one more delegate. A remarkable feat lost significance because Alaska had so few delegates in the first place. Sanders’ large Alaska caucus win netted him just 10 delegates over Clinton; he needs that margin of victory in large states where he can eat into the still very large delegate gap separating him from Clinton.
In the Hawaii caucus, our model again showed Sanders earning 74 percent of the vote and with it, 19 delegates. Here he underperformed, getting (“just”) 70 percent of the vote and 17 delegates. He fell two delegates off his target: already, his expectation-defying performance in Alaska is overshadowed by coming up short in Hawaii (up 1 delegate in Alaska but down 2 Hawaii, netting to being down 1 in delegate expectations). As delegate number increases, Sanders simply cannot afford to underperform expectations, otherwise he will have little chance to close Clinton’s delegate lead. Each delegate matters, especially with Sanders trailing by a substantial amount, and failing to meet expectations in any state – even a small one like Hawaii – makes his climb to the nomination a little bit harder.
Turning now to the Washington caucus, we forecasted Sanders taking 68 percent of the vote and 69 delegates. He easily defied expectations by earning a staggering 73 percent of the vote and 74 delegates, surpassing expectations and securing his largest single-state delegate haul of the primary season. Sanders was hugely favored in the state, which has small a African American subset and tends to be very liberal. That said, once again beating vote share estimates by almost 10 percent is an impressive feat, especially considering we already had high estimates and expectations for his showing. Importantly, unlike the Alaska caucus, the Washington caucus had a substantial number of delegates. Beating his expectations by around 10 percent earned him an extra 5 delegates (as compared to the single extra delegate he picked up in the Alaska caucus). Five delegates alone do not, of course, close the delegate gap, but Western Saturday showed that Sanders is able to surpass high expectations and drive up the score in states he should win. Moreover, the Washington caucus showed that Sanders’ appeal is not limited to small states – a trouble that has so far plagued the Sanders campaign. The question remains, though: can he beat expectations in large, diverse states – states through which the path to the nomination weaves?
|State||Sanders Delegate Expectation||Actual Delegates||+/-|
(A strong showing, but Sanders must make expectation surpassing a trend that carries over into the New York, Pennsylvania, New Jersey, and California primaries).
It remains to be seen whether these victories will propel Sanders in the upcoming contests. He still faces a daunting task in closing Clinton’s lead. Doing so seems highly unlikely, but if Western Saturday serves as precedent or starts a trend, Sanders may well overperform estimates in future primaries and caucuses. Well it be enough? Probably not – he’s trailing large in must-win states. But habitually overperforming delegate estimates will give Sanders and his supporters hope and may result in tighter contests than many are expecting.