When I run a quantitative analysis, I typically have some sort of expectation as to what I will find. Sometimes I’m surprised by the results. For example, in “Take 2: Could Employing Tax Cuts Tax Employment?” I discovered that, while higher taxes on the top 20 percent of earners corresponds to increased employment, higher taxes on the next-highest quintile corresponds to lower employment. It’s not an outcome I would have predicted.
Today is another such time. In this case, I was looking at the relative representation of each American. To explain what I mean by this requires a bit of background.
As you’re surely aware, each state receives the representation of two Senators in Congress, regardless of population. Additionally, each state receives the representation of at least one Representative in Congress, while additional Representatives are provided at a rate proportional to the state’s population. This, of course, means that the democratic concept of “one man, one vote” doesn’t apply to our representative democracy, even though our national colloquial rhetoric regularly implies otherwise.
Our congressional structure should mean in theory is that there is a wide disparity in the number of people represented by each Senator, but a small disparity in the number of people represented by each Representative. Moreover, since the red states tend to be sparsely populated compared to the blue states, one would expect to find that Republican Senators represent fewer people than Democratic Senators. We shouldn’t expect to see any such disparity in the House.
I put these hypotheses to the test.
This is a slightly awkward time to run such an analysis, because we’re in the midst of reapportionment. That is, some states are gaining Representatives this year, while others are losing them. My calculations are based on the 112th Congress, i.e., the current one. They will shift somewhat in the 113th.
I collected the population of each state, using the same numbers used for apportionment. Naturally, the actual population may be different for a host of reasons. First, let’s look at the populations of the states, along with the number of Representatives they currently have, and the number of people covered by each Representative and Senator. In this table, I’m treating it as if each Senator represents half of the state, although they actually both represent the whole state. It doesn’t actually matter in terms of the final relative representation calculations, though.
A few things may jump out at you right away. For example, unsurprisingly, California’s Senators cover about 60 times as many people as do Wyoming’s. But there are some other less expected results; Montana’s at-large Representative covers nearly twice as many people as does Wyoming’s.
From these, I calculated what I’m calling the Representative Factor and the Senator Factor. If the factor is one, it means that the Representative or Senator covers exactly the national average. If the factor is greater than one, it means that the Representative or Senator covers fewer people than the national average, meaning that each constituent has more power. For example, a factor of two means that each constituent has twice the power of the national average, while a factor of 0.5 means that each constituent has half the power of the national average. I also created a Combined Factor, which covers the constituents’ relative power in both houses together.
This is merely another way of looking at the first table. Even in the House, the relative power of each constituent varies by quite a bit.
What does this mean in terms of the relative power of each party’s constituents? To answer this, I took a slight shortcut, assuming that, within a given state, the number of constituents is the same across congressional districts. For each state, I broke down the number of Representatives and Senators of each party, and calculated the relative factor weight based on how many members of each party are representing the constituents of that state. I did this by taking the number of people represented by each party, and then compared that to the national average of all Representatives and Senators.
The average Democratic Representative has 703,682 constituents (Representative Factor of 1.01), while the average Republican Representative has 716,312 constituents (Representative Factor of 0.99). This means that the constituents in Democratic congressional districts have about 1.8 percent more power than constituents in Republican congressional districts. It’s not a lot, but it’s there.
In the Senate, the average Democratic Senator has 3,282,771 constituents (Senator Factor of 0.94), while the average Republican Senator has 2,876,524 constituents (Senator Factor of 1.07). This, you may recall, assumes that each Senator represents only half of the state. You can double those numbers if you wish to assume that each Senator represents the entire state. In either case, the Senator Factors are the same, and the result is that the average constituent of a Republican Senator has 14 percent more power than the average constituent of a Democratic Senator. The impact in the Senate, then, is an order of magnitude greater than in the House, ignoring the cloture rules. If one includes cloture, the relative power of each Republican constituent is that much greater in the Senate, at least to the extent that one wishes to prevent the passage of legislation.
As I said at the beginning, the results weren’t exactly what I expected. Yes, I expected Republican constituents to have significantly more influence on the Senate than would their Democratic counterparts. But I thought it would be more than 14 percent. And I certainly did not expect to find that Democratic constituents have more influence on the House than their Republican counterparts, slight though it is.
What do you think of all of this?