Annie Lowrey speculates:
Based on Census Bureau data, five senators would represent Americans earning between $100,000 and $1 million individually per year, with [2/10 of a senator] working on behalf of the millionaires. Eight senators would represent Americans with no income. Sixteen would represent Americans who make less than $10,000 a year, an amount well below the federal poverty line for families. The bulk of the senators would work on behalf of the middle class, with 34 representing Americans making $30,000 to $80,000 per year. . . . Or how about if senators represented particular demographic groups, based on gender and race? White women would elect the biggest group of senators — 37 of them, though only 38 women have ever served in the Senate.
I don’t know how well all of this would work in practice–for one thing, I wouldn’t want the senator who represents two-year-olds to be anywhere near the nuclear button–but I agree that ideas of fairness and political representation are subtle.
Along similar lines, here is my response to economists who complained that there were not enough economists in elective office:
I was curious about this so I looked up some statistics–not on Congress but on the workforce. According to the 2001 Statistical Abstract of the United States, there were 139,000 economists employed in the United States, which represented 0.1% of the employed population. 1% of 535 is about 1/2, so with at least two economists in Congress, the profession is hardly unrepresented.
139,000 is a crude estimate because it presumably represents the people whose job title is “economist” (and thus wouldn’t include, for example, Matt Kahn, who originally raised the “not enough economists in Congress” issue and whose job title is “professor”). But, even throwing in all these economics professors and various other practicing economists, I still don’t think it would add up to the half-million that would be necessary to reach 2/535 of the employed population.
This is not to debate the merits of the argument–perhaps Congress would indeed be better if it included more economists–but rather to note that people with this sort of job are a small minority in the U.S. (In contrast, there were 720,000 physicians, 170,000 dentists, and 2.1 million nurses, and 1.7 million health technicians in the U.S.)
To put it another way, without reference to economists (or to the 2.1 million “mathematical and computer scientists” out there): the Statistical Abstract has 260,000 psychologists. Certainly Congress would be better off with a few psychologists, who might understand how citizens might be expected to react to various policies.
I’m willing to believe that the country’s 890,000 lawyers are being overrepresented, but what about the 114,000 biologists? A few of these in Congress might advance the understanding of public health. And then there are the 290,000 civil engineers–I’d like to have a few of them around also. I’d also like some of the 280,000 child care workers and 620,000 pre-K and kindergarten teachers to give their insight on deliberations on family policy. And the 1.1 million police officers and 340,000 prison guards will have their own perspectives on justice issues.
So I think that representation is a tricky issue. Most of us would probably like more “people like us” in Congress, but that’s tough with only 535 seats to go around, and given that there are a lot of politicians already out there (many of whom are lawyers) who you’d be competing with.
Does it make a difference? Maybe so. In his article, “Does the Numerical Over-Representation of the Upper Class in U.S. Legislatures Matter?”, Nicholas Carnes finds:
Throughout America’s history, most political decisionmakers have been highly-educated, wealthy individuals from prestigious, high-paying occupations. . . . On balance, this study’s findings suggest that the numerical over-representation of affluent Americans in elected offices promotes more conservative economic policy outcomes, although not for the straightforward reasons that political observers have often suggested.
P.S. One suggestion that’s come up from time to time is to form the Senate from a national random sample of adults. This would give you all the representativeness you’d want.
P.P.S. Apparently the House has four (former) blue-collar workers: Phil Hare of Illinois, Stephen Lynch of Massachusetts, Mike Michaud of Maine, and Bob Brady of Pennsylvania. I don’t know how many former nurses, police officers, or child care workers they have.
P.P.P.S. Some data here.
I wrote a little about a random sample as the means of selecting representatives and senators et c:
http://michaelkenny.blogspot.com/2009/10/randomoc…
I have been interested in this question for a long time. The use of geography to proxy for interest made a lot more sense when all of the voters were homogeneous property-owning white guys but is now not a particularly good way to cluster opinion. (Geographical clustering was also the only technologically feasible alternative as well at the time.) Contra both Lowrie and Gelman, I wouldn't have any requirement for the relevant member to actually be a member of the group… thus, if the wealthiest 1/435 of the voting population wanted to elect a homeless guy, they could. In any case, once the proper clusters are created (and what are the criteria?) gerrymandering is eliminated. I'm not saying income is the best stratification possible, but it has to be better than geography… this is a simple implication of Red State, Blue State, no?
Take all the 2010 census information. Throw that in a big k-means clustering with k=435. Take the 2 people closest to the centroids who would be agree to be in Congress. These are your nominees.
Voters get to vote in all 435 races, but their vote is weighted inversely to their distance from the centroid.
This is not a totally outrageous idea; in Lebanon, for compelling historical reasons, representatives represent religious groups rather than geographical ones.
What about proportional representation? In principle that might result in all groups being represented in proportion to their demographic fraction. In practice it seems to result in minority governments, which can be a problem. Perhaps the current system works by grouping fundamentally dissimilar people so that the majority-rules elections pick a unified policy rather than representing their diverse interests?
There's also an important distinction here: who the representatives represent is distinct from who they actually are. A senator representing two-year-olds need not actually be two years old; she must simply represent their interests (which might have more to do with foster care and child protective services than, say, coloring on walls). That said, it's perfectly reasonable to wonder about the ability of someone to properly represent a group they do not belong to; certainly I would find it difficult to represent our First Nations population since I'm not a member and don't have any personal experience with their problems and concerns.
Just to take the thought experiment a bit too far:
If Congress represented income groups, they'd have a bizarre time planning their re-election strategies… If they've done a great job and helped their constituents grow wealthier, then those folks won't be voting for them anymore in the next election. In fact, the longer you have the same constituents, the less happy they'd be that they're not moving up, and thus the lower your chances of re-election (which I assume is the opposite of what happens in the current system). There'd also be less of a penalty for doing a terrible job — if many of your constituents dramatically lose wealth/income, their unhappiness becomes someone else's problem. It seems like accountability would not be a major focus under a system like this.
There's also the fact that *every single congressperson* would have to run a national ad campaign around their election season — making elections that much more expensive, and confusing the heck out of viewers who'd be subject to ads from candidates irrelevant to them.
It's not exactly overrepresentation or underrepresentation that matters, but rather the consequences of these things. My book, Personal Roots of Representation, offers a theory of the precise conditions under which the traits of members of Congress are consequential for policy-making.
We really only need one person: the median voter. Make him or her king, and voila, problem solved.
Regarding your title question: What problem are you trying to solve?
One problem with democracy is that the Senator from West Virginia strives to divert resources from the people of New York while the Senator from New York strives to divert resources from the people of West Virginia. Along the way a lot of resources get wasted.
To solve that problem, we'd want to give Senators constituencies with as few common interests as possible. By that criterion, this proposal seems to go in exactly the wrong direction.
By contrast, I argued in my book "More Sex is Safer Sex" that Senators ought to have alphabetical constituencies: One for everyone whose name begins with AA through AE, etc. It's easy to cook up earmarks that will benefit everyone in northern Michigan, but not so easy to cook up earmarks that will benefit everyone whose name happens to begin with Q.
Steven: The problem, I think, is not so much that senators have distinct constituencies but that the Senate represents some groups of people so much more than others. That said, maybe your idea has merit (practical considerations aside).