In November, 26 incumbent U.S. senators will be on election ballots. Chances are, most of them will win, according to a study by Washington University in St. Louis professor Gautam Gowrisankaran and two colleagues.
“Since 1914, incumbent U.S. senators running for reelection have won about 78 percent of the time,” said Gowrisankaran, assistant professor of economics at the John M. Olin School of Business and research fellow with the National Bureau of Economic Research. “This is sometimes referred to as an incumbency advantage. We seek to explain the causes of the incumbency advantage, using data on the outcomes of Senate elections since 1914.”
The study considered three potential explanations for the incumbency advantage: tenure, selection and challenger quality. Gowrisankaran began researching the issue after the 2000 election along with Matthew F. Mitchell of the Department of Economics at the University of Iowa and Andrea Moro, a professor in the Department of Economics at the University of Minnesota. The study encompasses every Senate election from the start of the elected Senate in 1914 through 2002.
“While many explanations have been posited to explain the incumbency advantages, including pork-barrel spending, media coverage and incumbent visibility, one important explanation is that some of the incumbency advantage is due to a selection effect: Becoming an incumbent requires winning, and winning candidates will tend to be of relative high quality,” the study states.
In other words, “Incumbents may win more often because on average, voters like them, which is why they won election the first time,” Gowrisankaran said.
“We seek to understand whether selection can explain the incumbency advantage, separately from all the benefits such as media coverage that accrue to a candidate simply by being in office,” Gowrisankaran said.
Related to this, is a third explanation, “that incumbents face candidates of weaker quality than in open seat elections,” Gowrisankaran said, “perhaps because candidates know that they will be harder to defeat.”
The researchers found that the effects of tenure — the advantage of being in office itself — are negative or small. They also found that incumbents face weaker challengers than candidates running for open seats, those that are available when the sitting senator is not running for re-election.
Further, the researchers predict that if incumbents faced challengers as strong as candidates for open seats, the incumbency advantage would be cut in half, and senators would win reelection only 64 percent of the time.
“All the advantage is in being a better candidate, not in the experience of having actually done the job,” Gowrisankaran said.
The study model implies the probability of winning reelection is influenced by the incumbent’s tenure and history of the seat following an open seat election, in other words, how many terms were served by each candidate who was later defeated by another candidate.
“It doesn’t just matter whether you won, but who you beat,” Gowrisankaran said.
Two senators with identical tenure could have important differences in their electoral history. For instance, a one-term incumbent could have beaten a two-term incumbent, or he could have beaten a five-term incumbent. “If five-term incumbents almost never lose, then the one-term incumbent who beat a five-term incumbent in his first election will have a higher probability of re-election than a one-term incumbent who previously beat a two-term incumbent,” he said.
The importance of the study is “in understanding whether we should be worried about the fact that senators win so often,” Gowrisankaran said. “We want to shed some light on this. If it’s just selection, it’s just because we’re getting senators people like.”
The study is not a predictive tool for who will win elections, although the researchers do make predictions on the 2004 election.
Moro’s website lists the 34 Senate seats up for election in November and shows the probability of the incumbent’s re-election. Interestingly, John Edwards of North Carolina, the Democratic vice-presidential nominee, would have had only a 67 percent chance of winning had he chosen to run again, less than any incumbent actually running for reelection. The formula assigns him this probability based on the fact that he is a one-term senator who beat a one-term incumbent, who had also beaten a one-term incumbent who won an open seat. On the other hand, Patrick Leahy of Vermont has a 90 percent chance of winning because he is a five-term incumbent who first won an open seat. Overall, the authors predict that the Republicans have a 75 percent chance of retaining control of at least 51 Senate seats.
The researchers caution that the probabilities are based solely on the history of the seat, and do not take into account other considerations such as party affiliation, scandals and other indicators of candidate quality.
The economics professors are interested in senate elections because the Senate is basically a labor market. “Senators are people who are selected into jobs,” Gowrisankaran said. “One of the things economists do is worry about selection. We need to think of this as a selection problem.” And senatorial elections provide much richer data than most labor markets.
“Selection will drive candidate quality,” Gowrisankaran said. “If you’re a better quality candidate, you’re more likely to get elected.”