Who will win the midterm elections? Analysts, candidates, and the public have long-tried to satisfy this curiosity through polls. Asking citizens would seem the most obvious way to find out what is going to happen on election day.
However, the surveys have certain flaws. One of them is that they have, over time, developed a clear bias toward the Democratic Party. Although some survey companies try to correct their older mistakes, it is easier to try than to succeed. One cause of this bias is partisanship, but another is technical: it is not easy to collect a representative sample, as many Republican voters live in rural areas or do not have cell phones.
Real Clear Politics, the elder statesman of polling analysis, is trying to overcome this bias by adjusting it in its forecasts. However, some analysts are looking for alternative ways to get ahead of the election results. One such analyst is economist Ray C. Fair.
The Ray C. Fair Model
James Piereson, a Senior Fellow at the Manhattan Institute, has written an article in the New Criterion detailing the methodology used by this economist:
Ray C. Fair, a professor of economics at Yale, has developed simple models for predicting presidential and congressional elections based on these parameters and contrasted them with election results since 1918. In his view, the performance of the economy is the main factor driving election results; the rest - personalities, polls, press and TV coverage, etc. - is mostly "noise" that obscures the main causes. Their equations are usually more accurate than polls in predicting election results, and are easier and much less expensive to use.
Although this model has been able to predict the results with greater accuracy and does not have the biases that generally condition surveys, it also has its limitations. Chief among them, as Piereson explains, is that
It only forecasts national results and does not reach down to the states or individual districts to predict which ones will pass from one party to another. In addition, it is not easily adaptable to U.S. Senate elections. Nor does it take into account non-economic issues such as war, corruption, or crime.
Inflation and GDP
Bearing this in mind, what does Ray C. Fair's model predict for these elections? James Piereson helps us understand it a little better. The model looks at the rate of inflation and GDP per capita growth in real terms, for the seven quarters prior to the election, plus a penalty to the incumbent party of 3 points.
According to data published by the Federal Reserve Bank of St. Louis and the Bureau of Economic Analysis, the inflation rate grew on average in the seven quarters -ending with Q3 of this year- at 7.2%.
Regarding GDP per capita, Fair's model takes into account the number of quarters in which annualized growth exceeds 3.2%. For the Biden Administration's time in office, there have been only three quarters exceeding this threshold, all in 2021. In 2022, GDP has entered negative territory.
7.2 points difference
According to the model, the Democratic Party should obtain a percentage of votes for the House of Representatives of around 46.7%, while the Republican Party should win around 53.3%. If these results are corrected for the 3% of the vote that goes to third parties, we see that the model predicts a percentage for the Democrats of 44.9% and 52.1% for the Republicans.
This represents a GOP advantage of 7.2 points. The Real Clear Politics average of polls conducted by Real Clear Politics on overall voting intention shows a 2.7-point Republican lead; that's just a little over one-third.
250 representatives
The difficulty now lies in knowing how this difference of more than seven points can be translated into the distribution of seats in the House of Representatives. Piereson takes into account that normally in presidential elections, the party of the winning candidate regains seats in the House of Representatives, as well as in the Senate. But the 2020 elections were atypical: the Republican Party regained 13 seats, which limits its ability to regain more representatives in these elections.
Piereson then compares that seven-plus point GOP leads over the Democrats to other midterm elections with similar results. In 1994 and 2010, they had a 6.8-point lead over the Democrats. And then they obtained respectively 230 and 247 seats in the House of Representatives. Most are from 218. In those elections, the Republicans started from 176 (1994) and 179 (2010), so the number of seats gained was much greater than the number of seats that the Republicans can win now, starting from a more favorable position: 212 representatives, just six short of a majority.
In any case, applying the Fair model, they could potentially reach 250 representatives. The "record for the entire period dates back to the 1930s. This would give them a margin of approximately sixty-five votes over the Democrats in next year's Congressional session, a much more impressive majority than the Democrats have had in the current legislature."
The Senate
Translating a difference in the overall vote count to the distribution of Senate seats is more complicated. The reasons are: 1) that there are not elections in all states, but in approximately two-thirds of them. 2) that the candidate weighs much more in the case of senators so that economic conditions are less relevant. 3) The Senate's role in managing the economy, although important, is less than that of the House of Representatives.
Candidates from 34 states are running in this election. Republicans defend 21, to 13 for Democrats, so Republicans have to defend more senators, and have fewer options to capture Democratic senatorial seats. But if the outcome for 2022 is comparable to 1994 and 2010, then the GOP will be able to regain one-third of the 13 seats held by Democrats. That would mean gaining three or four of the seats now held by the Democrats.
Such an outcome seems unlikely. If true, the Republican Party would have to win a victory in an election in which a Democratic victory seems assured. That is to say, looking at the polls, it is foreseeable that they could obtain victories in Nevada, Georgia, and Arizona. But getting a fourth senator would require them to win Colorado or perhaps New Hampshire, which until recently, were considered clear Democratic victories.
The case of New Hampshire
But that possibility, which seemed distant just a couple of weeks ago, today seems possible. Real Clear Politics predicts that the 50-way tie in the Senate will be broken in favor of the Republicans, with a 54-46 Senate advantage.
RCP predicts a GOP victory in New Hampshire, where it previously seemed impossible. Don Bolduc, a retired general, beat the powerful Republican establishment in New Hampshire thanks to Donald Trump's support. He was also endorsed by Democrats in the Republican primary, hoping that a candidate so far to the right would facilitate the re-election of Maggie Hassan (D), but they may live to regret this in the end. The average of the polls conducted by RCP shows a 0.7-point advantage for the Democrats. But since there is a pro-democrat bias in the polls, nothing assures his victory.
