California's Two Electorates:

The California legislature's attempt to reverse a 5-year-old 61%-39% referendum on same-sex marriage reminded me of something that political science professor Bruce Cain pointed out a while back — states have, in a sense, two electorates: The electorate as filtered through legislative election, and the direct electorate; and the two will inherently (in many states) yield different outcomes on many issues.

The most interesting reason Cain pointed to (and I realize there are many others, but I want to focus on this one) is that legislative districts have equal population, but not equal numbers of voters. Some districts consist nearly entirely of citizens, include relatively few children, and have high voting rates among eligible voters. Those districts cast many votes in statewide elections for governor, for U.S. senator, and on ballot measures. Other districts with the same population have many more noncitizens (whether legal immigrants or illegal aliens), many more children (for instance, Mexican Hispanics, Hispanics generally, and blacks have higher birth rates than whites), and lower voting rates among the eligible voters (for instance, nonwhites are more likely not to vote even when they are eligible). Those districts cast fewer votes in statewide elections. But each district elects one state assemblyman, state senator, or U.S. representative.

Consider, for instance, a simple and oversimplified example, with three districts, each with population 100,000, with the legislators faithfully tracking their residents' views, and with the breakdowns of views being the same among voters and nonvoters in the same district:

District# (each with population 100,000)Sentiments on issue ALegislator's vote on issue APercentage of residents who votePopular vote on issue APopular vote against issue A
Total47%-53%2-1 against50%76,00074,000

The legislature votes 2-1 against issue A, and that it is indeed the sentiment among the public as a whole (including nonvoters); but as a ballot measure, issue A wins by 1%. (In other models, the legislative vote is close in one direction, and the popular vote is a landslide in another direction.) So even without gerrymandering, political horse-trading, account being taken of the intensity of preferences (which, many assume, are reflected more in legislative votes than in popular votes), and the like, the different percentages of voters in each district are enough to make the legislative-filtered results be quite different from the direct results.

Now you can decide for yourself whether this is good or bad. Is the drawing of district lines by total population bad, because it gives each of the 20,000 voters in district 3 more influence over the selection of a legislator than each of the 80,000 voters in district 1? Is it good, because it leads each legislator to represent the same number of actual people, even if not the same number of voters? Should this lead us to like initiatives and referenda (or the decisions of statewide elected officials) more than the decisions of the legislature, or vice versa? Should it lead us to like a mixed legislative-popular system, or oppose it? I express no opinion on these subjects.

But I do think that this should remind us not to be surprised when Californians consistently elect Democratic House delegations and legislatures, but often elect Republican governors and U.S. senators, and enact more conservative ballot measures. And we should expect the same in other states that have a lot of immigrants (who are more likely to be noncitizens) and a lot of nonwhites (who are more likely to be underage, and more likely not to vote when eligible).

UPDATE: One of the comments complains that the 80%-20% disparity in the table above is unrealistic (though others point out that very substantial disparities, though not fourfold, do exist). But this is a stylized example, with only 3 districts. If you have 80 districts, you can have the same effects with much less stark differences.

Just to keep the arithmetic simpler, say you have 100 districts of 100,000 voters each. 55 are 55%-45% against proposal A, but they have on average only 45% turnout. 45 are 55%-45% in favor of proposal A, but they have on average 55% turnout. In the legislature, A loses 55-45. In a ballot measure vote, we have an exact draw -- 55 districts with 45% turnout are the same for statewide election purposes as 45 districts with 55% turnout. Adjust the numbers by a hair in A's favor, and A would win statewide, though it would still lose in the legislature.

Brett Bellmore (mail):
And, let's be clear, the legislature is quite often engaged in thwarting the public will, rather than "filtering" it, on subjects where the personal interest of legislators differs from that of the general public. Term limits, for instance, which, IIRC, was adopted by every state with the initiative process, and maybe one that lacked it.
9.8.2005 2:12pm
Goober (mail):
You get that this is all specious, right?
9.8.2005 2:28pm
Jack B.:
This is yet another example of while Baker v. Carr was necessary for some states, it is not necessarily a wholly good thing for the State Assembly and Senate to be forced in to a representation model they may not fit the needs of the state.

Especially if, one state house could be apportioned on the basis of citizenship and the other on GP, or any other relevant factors.
9.8.2005 2:29pm
All true, but I wonder whether, in reality rather than theory, differing numbers of actual voters in various districts makes that much of a difference. I mean, how likely is it that in any election there is such a disparity in the percentages of person voting as in Eugene's example - I doubt VERY highly that in any real election there is a district in which 80% of the population voted and another district in which only 20% voted. More likely, to me, is that gerrymandering, the effect of interest group politics, etc. causes there to be differences between how the public votes in referenda and how the legislature votes.

One other point: if you dislike drawing of district lines by total population, how would you do it differently? By the number of registered voters? By the number of actual voters? When would these be determined ("actual" voters obviously differs depending on which election you look at). Those have pretty obvious flaws, to me.
9.8.2005 2:33pm
Arthur (mail):
Empirical data supports the theory here, at least to some extent here: There is a wide discrepancy in the number of votes. It looks like the numbre of votes for President ranged from 430,000 in Senate District 1 (Oregon border to Sacramento; 60% Bush), to 175,00 in Senate District 16 (Fresno and Kern/immigrant farm laborers and large rural families, 49% to each candidate)
9.8.2005 2:33pm
Anon1ms (mail):
Aren't both CA Senators Democrats?
9.8.2005 2:43pm
Jack B.:
Looks like as few as about 143,000 in SD 22. Wow, impressive point Eugene.
9.8.2005 2:43pm
Goober (mail):
Aren't both CA Senators Democrats?

As well as all its presidential electors for like a decade-plus, yes. And when Republicans do get elected statewide they're, you know, usually pretty liberal.

Plus, if you've ever read Robert Dahl, there's the whole pluralist take on Prof. V's numbers. For those who haven't, I'll just say that the reason one constituency might turn out to vote in greater numbers isn't necessarily derived from their legal status or socioeconomic disinclination to vote; the people in the first district might have had a much greater interest in the matter on the ballot (say, if it were a local water issue, etc.) and therefore translated their more salient interest into greater turnout. Indeed it's not obvious that we should want the other two district's casually indifferent support for a program to override the greater interest of a smaller number. The insight that counting votes isn't as simple a matter as it appears to high schoolers in civics class is, by now, about fifty years old. (Although I have no idea what Professor Volokh thinks he's shown here, I'm slightly reassured that he doesn't seem to know, either.)
9.8.2005 3:02pm
Shelby (mail):
Well, Goober, it seems to me that he's pointed out discrepancies in voting results, and a likely mechanism for those discrepancies. That need not be novel as a political science proposition to be worth posting here. Moreover, the fact that it's poorly quantified and ambiguous in meaning doesn't make it "specious".
9.8.2005 3:20pm
Goober (mail):
... doesn't make it "specious".

No, forgive me: It makes it trivial. What's specious is the intimation that distortion of California's voting process is what's responsible for California's delegation being heavily Democratic, casually ignoring that California is a reliably Democratic state.

Look, if Prof. V. had been serious about this, I would expect the following couple of observations:

1) Different methods for counting votes yield starkly different political outcomes;
2) Different methods of apportioning representation yield starkly different political outcomes;
3) Normatively, we are concerned about our political institutions reflecting the majority preference;
4) Normatively, we are concerned about our political institutions responding to the very intense preferences of at least some numerical minorities;
5) Representative legislatures have idiosyncratic structural failings;
6) Direct legislative mechanisms have idiosyncratic structural failings of their own.

But he didn't, because he wasn't serious about this. He just wanted to theorize out loud that the California Democratic party survives by subverting the will of California voters. And I think such casual slander, especially if made without any real thought, is worth pointing out.
9.8.2005 4:00pm
jgshapiro (mail):
Indeed it's not obvious that we should want the other two district's casually indifferent support for a program to override the greater interest of a smaller number.


Why is it not obvious that every vote should be counted alike, even if some support is passionate and other support is 'casually indifferent'? I have never even heard of an electoral system where a voter would get an extra vote for propositions he feels especially strongly about, or that affect him more directly.

Nor can it be said that the votes in a legislature are any different: some legislators will feel strongly about a question and others will not care or will horsetrade their vote away for support on something else.
9.8.2005 4:03pm
jgshapiro (mail):
Normatively, we are concerned about our political institutions responding to the very intense preferences of at least some numerical minorities

If this is meant to suggest that we WANT our institutions to respond to intense preferences of numericl minorities, I think this is precisely wrong. In fact, we should be concerned about institutions responding to loud and intense preferences because a silent majority will (in general) not have the time nor the willingness to be as loud in opposition as a focused and passionate minority could be in support.

This is one of the benefits of initiatives as a policy-making process: it short-circuits the ability of special interests to have policies adopted which are contrary to the will of the majority.
9.8.2005 4:11pm
I didn't mean to "complain" - I just wondered about whether the explanation you posted actually does (rather than theoretically could) explain the real-life discrepancy between the results of the referendum and the legislature's action.

I take the other commenters' data. There is more of a disparity among voting rates of various districts than I had imagined (although the extremes are, well, extreme - my very cursory examination leads me to think that most districts had ~250,000 voters in the Presidential election).

However, your second example (with 100 districts) ALSO presents problems when applied to the real world. What is the likelihood that ALL the districts with high turnout rates would be in favor of the proposition, while ALL the districts with low turnout rates would be against? I, of course, realize that you could come up with various scenarios where there is not "all" the high/low turnout districts pointing one way or another. But my point is that that, isn't it most likely that you get ~half of the high turnout districts for the proposition and ~half against (and the same for the low turnout districts)?

Again, this is not a "complaint", but merely a question about whether the theory actually applies to the real world in any significant way.
9.8.2005 4:29pm
Goober (mail):
Why is it not obvious ...

(1) Because utilitarians don't agree.
(2) Because pluralists don't agree.
(3) Because republicans don't agree.
(4) Because liberals don't agree.
(5) Because very few except committed egalitarians agree.

Just for starters.

I have never even heard of an electoral system ....

In our electoral system, if you don't care and don't go to the polls, your vote isn't counted because it isn't cast. How is that system of one or zero votes meaningfully different from a sytem of regular votes and "extra" votes? Further, I don't get to vote in Alaskan elections, but I'm a New Yorker, and Alaskan residents, whom the result will "affect ... more directly" do get to vote. I don't think this is really quite so simple as you believe.

Nor can it be said that the votes in a legislature are any different....

Actually, that's exactly what we (a) predict, and (b) observe. Funnily, for precisely the reason that you observe: Legislators only "horsetrade" their votes when they don't care very much about the issue themselves; so those votes that get bought get bought by organized interests that care a lot about the issue at hand. To pick the practical example of fuel economy, which inspires medium-level interest in the country at large, but rather fierce interest in the congressional districts in Michigan, the politics always favor Detroit---because Detroit's representatives work full time out of political necessity, whereas consumer interests that might conflict with Detroit's interests are diffuse and don't inspire any representatives to advocate on their behalf as though their political lives depended on it. End result: The representatives that care a lot about the issue tend to win, because they're willing to make more sacrifices for it.
9.8.2005 4:30pm
Paul McKaskle (mail):
The disparity in voting in various districts in California hold up over several elections, so it is unlikely that a hotly contested local issue in one district would account for the disparity. The districts in California with the lowest turnout have high Hispanic populations, likely due to a high proportion of non-citizens. In the 1990s, two congressional districts in Los Angeles which elected African American members of Congress actually had more Hispanics in the districts than African Americans, but the number of registered Hispanics was as low as four percent in one of the districts, and not much better in the other. Appendices in Wilson v. Eu (Cal Sup Ct) 823 P.2d. 545, have the statistics.

As to using other data bases, Britain uses voter registration as the basis for determining the size of Parliamentary Constituencies, and Los Angeles did at one time until the Cal. Sup. Ct. outlawed it under the state constitution. In Britain, however, county boundaries are followed rigorously and population (i.e., voter registration) equality is very loosely applied. Since voter registration is extremely high in all parts of Britain, using voter registration probably doesn't have a great deal of political effect--but for technical reasons, it probably slightly favors Labour.

The political effect of using voter registration is not clear. Obviously Hispanic areas (which are relliably Democratic but somewhat conservative on social issues) would have fewer representatives but areas such as San Francisco, Berkeley and Western Los Angeles, which have lots of voters and few children, and are very Democratic (especially on social issues) would benefit. I suspect it would only make a minor dent in the comparative strength of the two parties, but it would make a major dent in the Hispanic component of the Democratic Party shifting strength to the urban white liberal component.

One way of making the legislative result and the popular vote result on issues more congruent is to have some form of proportional representation. Some forms of p.r. do a better job of this than others. But in a state, it could be used for one of the houses of the legislature to keep the legislature more in line with "the people." For a discussion of these issues, see, for example, 35 Houston Law Rev. 1119 (1998) (an instance of blowing my own horn, I must confess).
9.8.2005 5:01pm
Goober (mail):
Just as a curiosity, Berkeley has a California voter info database here; I couldn't find voter turnout, but assuming voter registration is a useful proxy for turnout, Prof. V's hypothetical disparity is entirely plausible. The most registered district in California in 2004 (the 4th) showed registration of nearly 434,000; two other congressional districts (31st and 34th) had registrations under 170,000. By my approximation, that's not quite a four-fold disparity, but still more than two-and-a-half.
9.8.2005 5:22pm
"(for instance, Mexican Hispanics, Hispanics generally, and blacks have higher birth rates than whites), and lower voting rates among the eligible voters (for instance, nonwhites are more likely not to vote even when they are eligible)."

As a technical quiblle, wouldn't the net result of this be approximately the same overall rate of voting as demographics with higher turnout and less children?

Presumably since life isn't static, many more children become eligibile in places with higher birthrates, offsetting the lower turnout.
9.8.2005 5:42pm
Shelby (mail):
He just wanted to theorize out loud that the California Democratic party survives by subverting the will of California voters.

Well, that's ... creative, I suppose. Unfortunately it doesn't seem to be supported by anything Eugene wrote in this post, or anywhere else for that matter.
9.8.2005 6:15pm
Goober (mail):
Oh, bother, Shelby; you go read his final paragraph before the update and you tell me what you think he's saying.
9.8.2005 6:43pm
If we are looking for a different example of the same phenomenon, remember in 2000 when the political scientists explained that the reason that Bush and Gore went 50-50 in Florida while the state legislature was 2/3 Republican. (The reason why Gore desperately needed to keep the election from going to the legislature.) The reason is that voter turnout in majority-Republican districts was (probably still is) significantly lower than in majority-Democrat districts.

My 8th grade teacher was a long-time Democratic official in Cook County and the sister and campaign manager to a Chicago mayor. She taught us that voter turnout is one key to a political machine's power. Given that average voter turnout in the US is pretty low on average, there is a lot of room to gain disproportionate influence if your political alliance can boost its turnout by significant amounts.

cathy :-)
9.8.2005 7:07pm
Shelby (mail):

Yep, just did. No "subversion" to be found.
9.8.2005 7:19pm
Goober (mail):
Terrific; you can look for a word. But would you mind reading the post, and telling me what you think his point was? I suppose I could be wrong. But snarky answers don't terribly convince me.
9.8.2005 11:29pm
Eugene Volokh (www):
Folks, I'm an academic. Sometimes my point is simply to describe what I think is a really cool and interesting phenomenon.

I am not a hard-core referendum buff who thinks popular vote always beats the legislative vote. There are perfectly good arguments, as I suggested in my second-to-last paragraph before the update, for thinking that the interests of nonvoters, children, and even noncitizens should be weighed in the decisionmaking process, and the filtration through legislative elections would help do that (the first electorate). There are also perfectly good arguments for a system that's more focused on the second electorate, or for a mixed system.

I really don't have an axe to grind here; I just heard this nifty explanation, and thought I'd do a good dead and pass it on.
9.8.2005 11:37pm
Eugene Volokh (www):
Besides, isn't it more interesting to discuss the substance of the post than to speculate about my ulterior motives in posting it?
9.8.2005 11:37pm
Joel B. (mail):
But Eugene!...all us commenters are like the shadow, but we don't know what evil lurks in the hearts of men, just the evil that lurks in the heart of the blog proprietors we comment on.

"When all else fails, assume malice." - too many commenter's motto.
9.8.2005 11:52pm
Goober (mail):
Joel B. nails me like a mounted butterfly; guilty as charged.

But [ahem], pray tell, professor, what was your point in your final paragraph (before the update)? Because it sure looks like you were saying "Look, California is so conservative but somehow got gerrymandered into sending a lot of Democrats to Congress." You'll forgive me if I don't think that's a neutral example. And my point emphatically is not that you're shilling for direct democracy, but rather that you're making insinuations about California's Democratic delegation.

And I did already comment on the substance of your post; it's a trivial contribution to a field that's already been covered with a lot more rigor. No worse than that, but no better, either.
9.9.2005 11:45am
Leon J. Page (mail):
Excellent post and discussion.

Consider these real-life examples. The Speaker of the State Assembly, Fabian Nunez, received 44,570 votes in his 2004 reelection. 7,837 votes were cast for his Republican challenger.

Contrast these results with the votes received by Assembly Republican leader Kevin McCarthy. McCarthy received 129,110 votes -- almost three times the votes received by Nunez -- and 35,130 votes went to his challenger.

In the Assembly, both Nunez and McCarthy have a single vote. (Of course, because Nunez was selected as Speaker, he enjoys considerable additional power and influence.) However, because McCarthy's constituents can and do turn out to vote -- in droves -- McCarthy's consituency would have a more decisive impact on any statewide initiative measure.
9.9.2005 4:05pm
P Wallach (mail):
The supposition made in this post that the opinions of voters roughly mirror the political opinions of non-voters is really quite remarkable, and I am shocked no one commented on just how unlikely it really is. There is a significant portion of the population that doesn't vote because they are politically disaffected, and to pretend that their political opinions are likely to be similar to voters, or even in the political mainstream, is quite mysterious. I like to think that in this mass of disaffected voters there is something else altogether going on--but the truth is, nobody really knows for sure. I would venture one guess: consistent non-voters probably don't care one bit about gay marriage.
9.9.2005 6:00pm
I wish people when they do this type of analysis would actually run some statistics to show if the criticism even matters.

Well I did. I looked at 15 statewide results since 1996: President, US Senate, Governor, LT. Governor, Attorney General, Secretary State including the special election in 2003.

Democrats won 13 of 15 elections. Their two party share of the vote was 55 percent. That would be their expected vote if districts perfectly comported with the population.

In the Assembly and Senate, Democrats control 73 of the 120 or 61 percent of the legislatures.

How big a difference is this? It means that Democrats have roughly two additional seats in both the Assembly and Senate than we would expect under perfect distribution. Considering there are partisan gerrymandering and other considerations, I would say that is not a heavily biased differential.
9.9.2005 6:57pm
Anthony Watson (mail) (www):
Well this gets right to the heart of the differences between a republic and a democracy.

America is not really based upon Athenian democracy as many have said. Every citizen voted and performed civic duties in that model...albeit a model completely based on slavery. That was how every citizen had time to participate at this level. The closest thing that America ever had was, ironically, the old Confederacy.

America is a republic, a representative democracy. There are always these types of statistical anomalies related to such a system. The founding fathers tried to even this out with the bicameral legislature. Even adding the electoral college layer to prevent a dictatorship by popular vote.

That final layer did not seem to work correctly in 2000, though. Given the electoral irregularities in Florida I think that the electoral college votes should have been split. The Supreme Court should have made this recommendation or told the adversaries to take the matter up in the legislature where all previous such disagreements had landed. But I digress

I think that it is important for us to understand the differences between a republic and democracy. Without that understanding there is much discontent and apathy.

Twain summed it up best.

There are lies, there are DAMN lies and then there are statistics

9.9.2005 7:28pm
Leon J. Page (mail):
Actually, "lies, damned lies, and statistics" comes to us from Benjamin Disraeli.
9.9.2005 7:56pm
California redistricting junkie:
Is the 80-20 example realistic? Definitely.

I looked at the 2002 primary, when the districts would have been just redrawn so their populations are essentially identical. Compare Congressional District 31, where a total of 38,632 votes were cast in the 2002 Primary, against Congressional District 48, where 128,356 votes were cast in that election (neither Congressional primary was seriously contested).

At 77 to 23 percent, this almost exactly matches Prof. Volokh's hypothetical example.

If both were uncontested, what difference does it make? There were six propositions on that ballot. The voters of CD 48 had nearly four times more influence on whether those propositions passed than the voters of CD 31.
9.9.2005 11:50pm