Joe Biden’s remaining two years in the White House will be much more challenging if, as experts and polls predict, the Democrats lose control of the House of Representatives and the Senate in the November 8 midterm elections. There are 470 nationwide elections to examine: all of the 435 House seats and 35 Senate seats. The outcome will determine which party controls a Congress that has been virtually deadlocked for years. An avalanche of polls, statistical models, historical and economic analyses, and prediction markets are all trying to forecast the winners and losers. As of today, most are predicting a more or less definitive victory for the Republican Party in the House of Representatives, but the Senate is still too close to call.
In the House of Representatives, the average of the several benchmark forecasts gives 216 seats to the Republicans, 190 of which are virtually guaranteed victories. The remaining 26 races are likely Republican wins or trending toward the Republican candidates. The Democrats are predicted to win 196 seats, and would have to take almost all of the seats that are trending Republican to retain control of the House.
FiveThirtyEight and PredictIt are both forecasting that the Republican Party will take control of the House. FiveThirtyEight’s statistical model is based especially on poll weights, but also incorporates data on poll bias, quality, historical data and other fundamentals. PredictIt uses a prediction market approach, where anyone can buy or sell shares that represent a bet that an event (like an election victory) will take place or not. Only one in five of FiveThirtyEight’s models resulted in a surprise victory for the Democrats, while only one in 10 of PredictIt’s markets had the Democrats coming out on top. These forecasts for the House of Representatives have remained steady for several months, which adds weight to the probability that the predictions are accurate.
But the Senate is a different story. Since senators serve six-year terms, only 35 of the 100 seats are up for election this year. Most benchmark forecasts are predicting that the Senate will remain fairly evenly split after these elections, with 48 or 49 seats for each party. Of the 35 seats in contention, 14 or 15 are likely to be won by Republicans, while the Democrats are expected to win eight or nine. That leaves four to six races trending for either side, and two high-stake races that are too close to call.
All the statistical models, prediction markets and pundits (such as those on the Metaculus platform, a massive online prediction community) are now trying to estimate uncertainty levels and probability differences for a Republican victory. FiveThirtyEight and Metaculus estimate roughly equal probabilities for Democratic or Republican control of the Senate. PredictIt has evolved from expecting a blue win (Democratic Party) to a red one (Republican Party), which is likely due to the prediction market’s greater sensitivity to rapid swings. But every pollster, model and analyst agrees that the Republican Party has a good chance of retaking the Senate.
It may seem paradoxical that there is more uncertainty where there are fewer races to accurately predict. But when a forecast is made up of the sum of many small forecasts, it may be easier to get it right. This is because the forecaster can afford more misses that will be offset by hits, and because errors in one direction can be cancelled out by errors in the opposite direction. Thus, control of the US Senate is coming down to a handful of very tight races, and two in particular – Georgia and Nevada.
In Nevada, home of the infamous city of Las Vegas, Democrat Catherine Cortez Masto is running against Adam Laxalt, the co-architect of Donald Trump’s failed 2020 presidential campaign and an unrepentant proponent of unfounded allegations of election fraud. The 2022 Senate race in Nevada is not all that different from the 2016 race won by the state’s first female senator: it’s very close and the two candidates are firmly positioned in opposing ideological camps.
The Senate race in Georgia is unexpectedly close because this southeastern state has gradually become more liberal due to urban demographic trends, enabling Democratic candidates to win more public offices than in the past.
None of the forecasts mentioned above have ventured to predict the victors in Nevada and Georgia, or even indicate a slightly higher probability of a win in either direction – they are simply calling both races a tie.
If the Senate ends up with an even 50-50 split, the Democratic Party will retain its advantage because Vice President Kamala Harris will have the tie-breaking vote per the US Constitution. But if the Republicans gain control of the Senate, President Biden’s ability to govern effectively and push his party’s agenda will certainly become even more difficult.