No, I don’t have a cite, and I’ll concede that “almost certainly” is too strong a term. However, constitutional law experts (and judge watchers) who I have asked about this feel that, barring constitutional amendments, these ideas are unlikely to fly to and through SCOTUS (and, of course, the losing side at each level almost certainly will appeal – although whether the appeal will be heard is annother set of good questions).
A lot of perfectly viable alternative to the status quo exist and are used all the time in other countries.
Non-partisan redistricting commissions. Britain also uses single member districts, but they are drawn up non-partisan Boundary Commissions. I’m sure lots of cynical Americans will say that the Boundary Commissions are secretly partisan, but they probably won’t hazard a guess as to which party they are favoring (and yes, I’m talking about drawing the districts, not about how many members each district would elect).
Proportional representation. PR makes district shape a lot less important. PR systems do not necessarily require the vote to choose a list as a whole, as in Israel. Open list systems also exist, and allow the voter to pick a preferred candidate from the list, with candidates seated in the order of the number of votes they won as individuals. There are also single-transferable vote systems, like in Ireland, where there are no lists or slates of candidates at all, but these are too complicated for me to want to explain them.
PR systems do not necessarily allow parties to win seats with extremely low portions of the vote; if the constituencies average ten seats then it will take around 10% of the votes to win seats in them. One problem would be, a lot of states have very few House members, and wouldn’t have enough to make for a very close proportion. This bothers some people a lot and me not at all. One- and two-seat states could contine to elect their people by first-past-the-post; medium-sized states would be a single constituency; the most populous states could be divided into a two to five constituencies.
A simpler alternative might just be to have a computer draw the districts. A mathematical program could be written to generate the most compact districts possible of equal population. “Compact” doesn’t mean “small” of course, it just means, not looking like a salamander, more or less. Gerrymandering schemes usually involve a lot of “nooks and crannies” where borders are drawn to divide or unite certain neighborhoods. I think another way to perceive of the most-compact-districts model is, the redistricting plan wherein the totalled length of all the district boundaries is the shortest (i.e., add up the perimiters of all the districts), but I’m on shaky mathematical ground here.
Was. Check out http://www.ncga.state.nc.us/html1999/geography/html4Trans/Thematic/thematichome.html, and click on the map to see a pdf of the current and two former incarnations of NC District 12, currently served by Congressman Mel Watt, I think.
It’s kind of funny. The Supreme Court case that got it created showed that African-Americans and Republicans were systematically gerrymandered out of proportional representation in North Carolina. I don’t have the cite, though.
Later Supreme Court cases, however, whittled it back. The decision was that the racial makeup of districts should be one aspect considered, but that it must not be the overriding basis for the construction.
Oh, and it’s interesting to note that, yes, the original district did closely follow parts of I-77 and I-85, in part because when those highways were built, they often were put through the black communities in the area.
My apologies for not noticing that you had dealt with the problem of states with a lot of districts. But what number is viable for people to vote on?
Los Angeles County has all or part of 17 Congressional districts inside it (24-39 and 41). Would the County have to be divided up arbitrarily to make everything manageable? It seems that any division of the state into smaller areas is still going to leave open the opportunity for gerrymandering.
I believe it’s one of those things that would have to be challenged, after the first election after Admission, to determine whether or not it still applies. The current commonwealth constitution was taken to Congress for approval in 1952 and the legislative setup sailed thru with no trouble (and the concept of the at-large seats was created by Fed statute in 1917). However the constitution just says e.g. “The Senate shall be composed of at least 27 senators: 2 elected in each of 8 districts, plus eleven at-large, provided each voter may cast his ballot for only one at-large candidate” and the way we apportion the a-l’s is left to local statute.
On top of this, we have a provision that unless the majority party actually got more than 67% of the overall votes in the General Election, the minority parties may not get fewer legislators than 1/3 of the base size of each house (thus, the Senate in January will be 28 instead of 27, since the majority party elected 19 senators).
jrd
In response to your first question, I don’t know. My guess is that, if and when this first gets tried, there’s going to be a bit of experimentation involved to see what range of numbers works.
I have to disagree with you, though, on how much gerrymandering opportunity this still leaves. Take LA County, for instance. If it has sufficient pop to be represented by 17 Congresscritters, we chop it up into pieces of, say, 6,6, and 5; twist and turn them as you like. If the Dems control the legislature and try to stuff all the Republicans into one of these three mega-districts. If MegaDist 1 (with 5 reps) is drawn 80-20 GOP, with the other two 60-40 Dem, then worst case for the GOP is probably that they get 4 of MD1’s 5 seats, and get 2 of 6 seats in each of MD2 and MD3. So they get 8 seats out of 17, when they might’ve gotten 9 by population share - and still might, under this system.
That suggests that, in by far the most populous state in the Union, the majority party in the legislature might be able to draw lines so as to get an extra two or three seats in Congress for their party. That’s small compared to the current system, and more iffy too, I think. And given that the effects would be somewhat balanced across the country (Dems getting this slight advantage in CA and NY, GOP in TX and FL), I think this has the desired effect of minimizing the effect that one state legislature can have on what legislation stands a chance in the U.S. Congress.
The 12th district in NC is the long narrow district. In some places it is the width of I-85 only.
Back in 1992 the legislature in NC created 2 districts that are over 50% black to get black members in Congress. The legislature did this to satisfy a Fed. Court order. This did result in 2 new black congressmen from NC. But it also ended up helping to elect more Republicans since other districts then became more conservative. Right now there are 7 Reps to 5 Dems in Congress. For senators it is 1 Dem and 1 Rep. There is still a court case now over the 12th district to get it thrown out but it has not been decided.
As someone pointed out earlier in the thread, the Constitution doesn’t mandate how Congresscritters are elected. Given that, I find it somewhat surprising that PR hasn’t been implemented, even briefly, in any state for electing congressmen. I mean, we’ve experimented with most other things. (TMK, PR hasn’t been used in any elections in any state for any office. Anyone know if this is right or not?) Are there any court cases out there specifically dealing with PR and its constitutionality?
As for your idea of shortest total length of district boundaries, Boris, assuming that is valid mathematically (I’m as innumerate as you on this point), I think that’s probably the best idea. It combines geographical closeness and nonpartisanship.
Sua
No longer. See my post above, and follow the link.
How about each party gets to create a number of districts proportional to the number of seats it won? It would still cause unfairness, but not quite as much.
What was the justifiation for that rejection?
That’s funny. I don’t recall the phrase “one man, one vote” appearing in the Constitution anywhere.
RTFirefly:
Your plan definitely sounds good from a mathematical point of view, (in fact, it’s the best plan I’ve heard from a theoretical point of view) but from a numerical methods point of view it sounds horrible. I mean, it makes the traveling salesman problem look trivial. I’m not an expert on this. Would anyone who is care to take a guess how many flops this would take?
According to an article I dug up in a newspaper database, Congressional districts could be chosen at large until 1842, when Congress passed a law saying that districts needed to be used (with the exception of one member delegations). The idea was that a Representative should be more responsive to a small consitutency than an entire state.
As for “One man, one vote” it’s not in the Constitution, but it was the subject of a Supreme Court case that I referenced above Baker v. Carr. It specifically dealt with State legislatures.
As for PR, most articles I read on the subject say that it could be instituted in the US with a simple act of Congress.
I don’t know any cases of it either. I think there may have been some use of it in city council elections, but I don’t have any examples of it. I think PR has never been tried because it would end the dominance of the two parties.
In Illinois at one point they used three-member districts to elect the state House of Representatives. I don’t know how many votes each voter got to cast, but I seem to remember the number was two.
Just to elaborate on multi-member constituencies, let me describe the effects of various systems (in each case assuming categorical (non-transferable) votes and no PR of lists):
Each voter having one vote This systems is called Single Non-Transferable Vote (SNTV). This is the systems currently used in Japan (both houses) and the Phillipines (upper house). If each party correctly estimates the number of votes it can win, it can run the right number of candidates (too many and its candidates will lose; too few and its small number of candidates will be elected with plenty of wasted votes). If each party’s votes are evenly distributed between its candidates, then this system will result in proportional representation. A couple of big ifs, but it is a fairly simple system and pretty much guarantees representation of minorities.
Each voter having as many votes as seats, but not being allowed to vote more than once for the same candidate (this system is thus different from Guinier’s - pardon my spelling?). Back in the olden days, some states used to elect presidential electors this way. If your state had 10 electors, you’d cast 10 votes for no fewer than 10 elector candidates. Needless to say, this usually resulted in a single slate of electors being elected; ticket splitting could happen but was rare (small parties would sometimes run incomplete slates and this might means some of their Presidential votes could slip off to the major parties when voters to chose to vote their full 10).
To answer The Ryan’s question, this system has been rejected since it totally crushes the losing folks. That doesn’t just mean political parties, it can also mean, among other things, racial minorities. If you had an “At Large” city council of seven people, and each voter had 7 votes, that would let all your white voters vote for 7 white candidates apiece; the unfortunate black voters could vote for however many black candidates they wanted - it wouldn’t matter a bit. If your guy comes in 8th place it’s no better than 100th place. So the courts have said - you have to use districts in city council elections, so that the black neighborhoods won’t be snowed under. It isn’t a universal rule - Portland, OR has an At Large City Council (it’s actually a commission form of goverment, which might make it an exception).
Each voter having a small number of votes This puppy was used in Illinois, IIRC. It gives an intermediate result between “near PR” SNTV and “crush the minority” many-vote systems.
Each voter having multiple votes, and being able to vote as many times as they please (until they run out). This is Lani Guinier’s system, and has been amply described on this thread. One thing I’d like to point out, is that voter behavior is everything here. If everybody concentrates their votes on one candidate, it becomes the de facto equivalent of SNTV (q.v.) If everybody votes no more than once for their candidates, it’s just like the above. So it’s anybody’s guess what the result would be.
Municipal governments in California aren’t required to elect their city councils by district I believe. Most cities elect their city councils here at large. However, the larger cities all elect their councils by district.
The tiny gated city of Bradbury elects its five-member council by district. There are only about 200 people in each district.