A study of the voting records on conscription of U.S. legislators with draft-age sons in the 20th century shows how self-interest ruled votes
Men register for the military draft in New York City in 1917. “We’re showing that politicians are malleable, that if you change their private incentives, they will change how they vote,” Eoin McGuirk says. Photo: Shutterstock
How much do politicians vote based on their own interests, even if that conflicts with the ideology of their party or public opinion? Suggesting that this might be the case is one thing, but finding a way to test the hypothesis is entirely another. It’s not like politicians can be asked about these things, even long after the fact.
But Eoin McGuirk, an assistant professor of economics, and two of his colleagues figured a way around that hurdle. They looked at the voting records of U.S. representatives and senators in very specific circumstances—when the vote was on issues about conscription, drafting young men to fight in wars.
There were four wars in the 20th century in the U.S. in which troops were conscripted to fight, and many votes in Congress related to conscription—its expansion and duration, for example. The key variable for McGuirk’s study was if a politician had a male child who was draft age—say, a 23-year-old son. (The draft was eliminated in 1973, as the U.S. moved to an all-volunteer military.)
Luckily there was a control group the researchers could compare their test group with—politicians who had daughters of draft age, but not sons. “The only difference between the two groups of politicians is the gender of a given child, which is essentially as good as random,” says McGuirk. “It’s like we were doing a clinical trial.”
“On average, these legislative roll call votes on conscription would be pro-war about 58% of the time,” says Eoin McGuirk. “We found that having a draft-age son reduces your likelihood of voting for the draft by between 7 and 11 percentage points.” Photo: Alonso Nichols
The researchers gathered data on legislative voting in the 20th century, linking it to individual politicians and their census records, which showed their children’s age and gender. “We put it all together to see if politicians who have sons of draftable age at a given time are less likely to vote for conscription than otherwise identical politicians who have daughters of the same age,” says McGuirk.
The results were unmistakable. Having a draft-age son made a clear difference in the voting of a sizeable percentage of legislators. “On average, these legislative roll call votes on conscription would be pro-war about 58% of the time,” says McGuirk. “We found that having a draft-age son reduces your likelihood of voting for the draft by between 7 and 11 percentage points.” In close votes, that’s a very significant amount—“it has a meaningful effect,” he says.
Because the gender of a draft-age child is as good as random, “it’s unlikely that this is going to be correlated with ideology, with background, with any other sort of variable that we think could influence one’s proclivity to vote for war,” he adds.
And while it might be argued that having a son who could face military action could change one’s ideology about war, the researchers found that as soon as sons of legislators aged out of the draft window, the propensity to vote for conscription went back to where it was beforehand.
“It’s unlikely that a politician’s ideology would change over that very short period, but we know for sure that their own personal cost of conflict did change, so we can put that down to private interests rather than ideology,” McGuirk says.
Why It Happens
The study, published in the Journal of Political Economy, grew out of a larger body of work looking at the causes of war. From an economist’s point of view, war shouldn’t happen. “Conflict is extraordinarily costly. It destroys physical capital, it destroys human capital, it destroys social capital,” McGuirk says. “The puzzle is, why does it happen? Surely both groups would be better off settling their differences before conflict happens rather than fighting. You need to come up with a theory that allows for the failure of a settlement.”
But say a political leader stands to benefit from the conflict, without paying a price, he says. Then it is clear how war becomes more likely in that situation.
The paper provides evidence that at least one determinant of legislative voting is a politician’s private interests. “We’re showing that politicians are malleable, that if you change their private incentives, they will change how they vote,” McGuirk says.
“It’s a bit of a smoking gun,” he adds. “I think a lot of us probably assume that some politicians are malleable to financial transfers, that lobbying can influence how they vote. That’s a really hard thing to prove and we don’t prove it here, but we do prove the politicians are malleable and that their own private incentives that are completely orthogonal to political incentives matter. It’s an interesting area for future research.”
