bq. I heard this morning for about the hundredth time a very smart person give a presentation in which she (or more usually he, but this time it was a she) massively overstates the role of gerrymandering in creating problems for federal politics and policy.
bq. There’s actually a really easy way to think about this: Look at the Senate.
bq. State boundaries for better or for worse can’t be changed. So there’s no gerrymandering. So whatever problems gerrymandering is causing don’t exist in the Senate. Whatever problems gerrymandering is exacerbating should be much better in the Senate. If you want to believe gerrymandering is at the root of our problems, you ought to be able to produce a healthy list along these lines.
bq. Personally, I have a hard time coming up with one. Conversely, since every Senator has a “district twin,” the Senate offers a good test of how constrained (or not) legislators are by their constituents. It seems to me that Judd Gregg is closer to Jeff Sessions than to Jeanne Shaheen and that Kay Hagan is closer to Pat Leahy than to Richard Burr.
He is correct on this last point, according to Simon Jackman’s data (.csv). Gregg’s ideal point (0.81) is closer to Sessions’ (1.44) than Shaheen’s (-.59). Kagan (-.39) is closer to Leahy (-1.18) than Burr (1.42).
On how much Senators are _not_ constrained by their constituents, see this piece by Grofman, Griffin, and Glazer (pdf). A brief quote summarizes their argument:
bq. We make use of a ‘natural experiment’, comparisons of senators from the same state but of opposite parties. This allows us to inviestigate the magnitude of the differences in ideology caused by party differences when the potentially confounding effects of constituency differences have been completely controlled for…Our key findings about the US Senate fly in the face of simple Downsian [JMS: median voter] model:
bq. 1) US Senators from the same state but of opposite parties vote quite different from one another, while differences between senators from the same party and same state are minimal. Moreover, in states where one senator is a Democrat and the other a Republican, the Democrat is almost always to the left of the Republican from that state.
bq. 2) For most of the past 25 years [JMS: 1960-84] the mean ideological difference between senators of the same state who are of opposite parties exceeds that between two randomly chosen senators of opposite parties in the nation as a whole.