Crack and pack didn't get them back

A commonly held view is that the extraordinarily high incumbent re-election rate in the US is due to gerrymandering (colloquially known to political operatives as “crack and pack”). My friends John Friedman (a Harvard PhD student) and Richard Holden (an Aussie who’s just started at MIT Sloan) beg to differ. In a piece in the New Republic, they argue that money, political polarisation, and changes in the media are to blame.

In fact, while the technology for gerrymandering has become more sophisticated, Friedman and Holden argue that the constraints on gerrymanderers (eg. minority protections in the 1982 Voting Rights Act) have gotten tougher. As a result, there’s actually less egregious gerrymandering going on now than in the past.

If you have a TNR subscription, click here. Otherwise, a draft version is on John’s website. 

This entry was posted in From the Frontiers, US Politics. Bookmark the permalink.

12 Responses to Crack and pack didn't get them back

  1. Patrick says:

    If they wanted a useful study, it would have helped to cross-check with Senate and Governor’s elections, which are obviously immune to gerrymandering (although still susceptible to Kennedy-style plain cheating), and with Presidential voting, too.

    Without that their data isn’t worth a pile of beans – who knows what effect gerrymandering has just because incumbency goes down or up if there is no benchmark?

  2. Sacha says:

    It’s amazing that state legislatures can still draw the electoral districts (within constraints like the act you mention), except for one or two in which an independent commission draws them. I don’t understand how this isn’t a focus for US reformers.

  3. the view that incumbency advantage is driven by favorable redistricting is hardly the dominant view in American political science. so I’m not sure that the finding is “news” to anyone who has looked at the issue seriously. all the estimates I’ve seen on the redistricting components of incumbency advantage are on the small side.

    the question about partisan redistricting targets of reform is an interesting one. yes, it is a scandal that many US state legislatures continue to control redistricting of their state’s congressional districts (and their own districts), but it doesn’t end there: *all* aspects of election administration, save for the relatively limited set of matters covered by the (Federal) Voting Rights Act, are run state-by-state or even county-by-county, which means there is an awful lot of scope for incompetence (via under-funding, hack appointments, etc) and good old fashioned corruption. Incidentally, this is why I think compulsory voting is such a good idea: it brings into being a Federal bureaucracy like Australia’s Electoral Commission, that oversees all aspects of Federal elections (everything from allocation and staffing of booths on election day, to ballot design, to redistricting, absentee procedures, etc) and is generally non-partisan and professional. Help American Votes was supposed to fix some of this, but don’t hold your breath.

    on the up-side, the U.S. Supreme Court ruled malapportionment unconstitutional eons ago (approx 1960, Baker v Carr etc), and that went away immediately (save for the malapportionment inherent in the 2 senators per state provision of the U.S. Constitution). The Australian High Court didn’t/wouldn’t/won’t touch malapportionment in the Australian states…

  4. Sacha says:

    “The Australian High Court didn’t/wouldn’t/won’t touch malapportionment in the Australian states… ”

    I think there’s scope for the High Court to interpret the relevant constitutions to find malapportionment unconstitutional. I doubt it’ll happen in the next 10 years though, unless state constitutions are amended to include those principles (think WA and Qld as the states in which malapportionment will continue to exist for the forseeable future).

  5. Sacha says:

    It’s almost as if the US electoral system was frozen a long time ago and is now stuck: as Simon says, the basic administration of electoral laws plus the drawing of electoral boundaries seems light years ahead here compared to the US.

  6. Matt Cowgill says:

    Malapportionment in WA was largely brought to a close during the term of the previous Gallop government. Incidentally, this did involve a High Court challenge, though it was concerned with the manner of the one-vote-one-value legislation’s passage (‘constitutional majority’ vs ‘simple majority’ and the voting rights of the speaker) rather than the legislation itself.

  7. Sacha says:

    Yes – largely brought to a close, just as in the electoral redistribution in Qld in the early 90s. In the Qld one, weighted seats ended up with approximately 16000 (or so) voters per seat (there were 5 weighted seats) while the unweighted ones had about 24000 per seat (these are from memory, so the numbers are plus minus 15 years of memory).
    Currently, the average seems about 15000-17000 per weighted seats versus 28000 per unweighted seat (from a quick look on the website). No doubt someone can post the correct figures.

  8. Geoff R says:

    The party split in Congress doesn’t seem that different from what you would expect in a ‘normal’ single-member system, although the Democrats seem to do a bit better than you would expect on the basis of their overall vote. It would be simple to calcualte the correlation between the Congressional and the presidential vote which might provide evidence of incumbency. The current view that the Democrats will struggle to win the Reps in November logically implies that the likley pro-Democrat swing will be concentrated in Democratic seats. More than 1994 the 2006 mid-term it will provide an interesting test of the impact of incumbency.

  9. Mork says:

    It’s almost as if the US electoral system was frozen a long time ago and is now stuck: as Simon says, the basic administration of electoral laws plus the drawing of electoral boundaries seems light years ahead here compared to the US.

    Well, that’s kind of the facts. The U.S Constitution was drafted in the 1780s and is damned hard to amend.

  10. Sacha says:

    Mork, it should be fairly easy for the state legislatures to legislate for independent commissions (as independent as you can get in the US system) to draw up electoral boundaries. No doubt it should be possible for other elements of it to be altered: state voters could amend their state constitutions to institute state-wide constistent, at least, voting systems.

    The Australian federal constitution is perhaps even harder to amend than is the US constitution, and the Australian electoral system continues to evolve.

  11. Sacha says:

    Of course, state constitutions in Australia are relatively easy to amend.

  12. Sacha: the Australian parliament controls its own method of election, while the U.S. state legislatures control their method of election (modulo court challenges). Hence, provided you get an amendment to the Commonwealth Electoral Act through the parliament, you’re done in Australia; contrast the state-by-state tinkering needed for comprehensive change in the United States. An interesting development is this call for states to enter into a pact regarding how they allocate Electoral College voters: i.e., irrespective of how your own state votes, give your Electoral College votes to the winner of the national “popular” vote.

    Geoff R: yes, presidential vote at the level of congressional district is the usual measure of basline partisanship in that district. A simple back-of-the-envelope calculation of the “personal vote” for an incumbent is to look at the difference between his/her vote share and that of the pres candidate from their party. Of course, you can only do this in the “on-year” elections, when both president and Congress are up. Other estimators include looking at the “sophomore surge” (the difference between the incumbent’s vote share in the 1st election they contest as an incumbent, and the vote share they obtained in the previous election as a challenger, with some adjustment for national swing between the two elections) or “retirement slump” (the decline in a party’s vote share when they go from having an incumbent in the district to the “open-seat” race you get after an incumbent’s retirement). All of these estimators are subject to some biases since you’ve got the “strategic entry” problem: an ambitious challenger will not run against a safe incumbent, but wait until the seat opens up, or shop around to find a seat that is opening up etc, which has the consequence of making incumbents look safer than they really might be. Kind of a neat problem, as these puzzles go…

Comments are closed.