| 摘要 | This article appears in the November 7, 2016 issue of National Review.
With the election right around the corner, polls and betting markets agree that there will likely be a big victory for Hillary Clinton. As of mid-October, betting markets posted an 83 percent chance that she defeats Donald Trump, with the margin of victory in polls then averaging about seven percentage points. Granted, false predictions of Brexit taught us that polls may be less reliable in this strange new world. But assuming Clinton does win by margins as wide as the data suggest, what happens next?
The answer, it would seem, depends on what happens to the House and the Senate. A scenario in which President Clinton comes to power in a landslide seems like a scenario that could, at least in theory, deliver her both houses of Congress. At that point we would discover whether she is a Bill Clinton-style moderate Democrat, or the Bernie Sanders-like capitalism-hater she has at time portrayed on the campaign stage. But if Republicans maintain a grasp on Congress, there will be more pressure on the deal-making Clintons, who helped give us welfare reform and capital-gains tax cuts, to make a reappearance.
To tether this conjecture to reality, we collected data concerning the 1944 to 2012 elections from the Office of the Clerk within the U.S. House of Representatives, which reports the political divisions of the U.S. Senate and the House going back to the 40th Congress. These data detail the number of seats held by Democrats and Republicans after each bi-yearly election. We also collected data from the American Presidency Project on the percentage of the popular vote won by each Democrat and Republican presidential nominee.
To understand the relationship between the presidential-election outcome and the number of seats held in Congress by the incoming president’s party, we took the difference in the popular vote between the winner and runner-up and compared that with the change in the number of seats that the president’s party held in both the House and the Senate.
For example, in 1996, Bill Clinton ran against the Republican nominee, Robert Dole, and won with an 8.5 point margin in the popular vote. In that election, the Democrats gained three seats in the House but lost three seats in the Senate.
So how does the percentage by which a president wins affect the number of seats held by the president’s party? The graph below shows very little relationship between how much a president wins by and how many seats are gained (or lost). In fact, when running a simple regression of the president’s percentage win of the popular vote on the change of seats held by the president’s party, we find that there is no relationship. This result may appear counterintuitive, but the data we gathered do not tell any other story. Indeed, while one should not get too excited speculating about statistically insignificant results, the patterns in the data suggest that voters may well be wary of both political parties, and favor ticket splitting when the presidential outcome seems certain.
There is, then, no statistical evidence that the presidential victory margin drives pickup of seats in the House or the Senate. So if Donald Trump is trounced in the election but Republican candidates for the House or Senate “surprise” on the upside, it should be considered no surprise at all. |
