Fruits and Votes » Prof. Shugart's Blog » Electoral reform in California

I am responding to popular pressures and opening up the ‘seedbed’ for ideas on electoral reform in California.

Bob points us towards a proposal from the New America Foundation for MMP* in a 360-seat unicameral legislature.

I’ll see how the comments grow and perhaps weigh in as time permits.

________
* Or “Personalized Full Representation”–ugh on the name, but hooray for the idea!

120 ideas sprouting »

Ugh, indeed.

As Ed pointed out in the BC STV comment thread, California’s county boards of supervisors would be a fine place to use STV. Indeed, the very large majority of California elections are (at least nominally) non-partisan and would benefit from STV: city councils, school boards, local water and sanitary districts–we’ve got plenty.

While I’d prefer STV for the state legislature, MMP would certainly be a big step forward. And expanding the California legislature is long overdue.

Seed planted by Jonathan Lundell — 15 May 2009 @ 18:03

I have possibly seen less unlovely acronyms than PFR21 but I do not at the moment recall where.

I suspect electoral reform would do better if it was decoupled from other constitutional changes such as unicameralism.

I will not be tedious by repeating my arguments for STV except to say that an electoral reform should really address the governorship as well and the only real option for electing a single executive is STV. There could be tactical advantages in first arguing for an STV governor and senate without also trying for unicameralism and an expansion of the legislature.

An STV senate could have manageable ballot size if the regions proposed for PFR21 were used as separate multimember electorates. That would also create a genuine tension between the two houses.

Seed planted by Alan — 15 May 2009 @ 20:21

Increasing the size of the legislature would be very unpopular, voters would not want to have more politicians. They dislike them already.

There have been propositions on the ballot to increase the size of the L.A County board of Supervisors from 5 to 7 members, and every time that proposition has been defeated.

I think the California State Assembly could change that by law, if it wanted, but I’m not sure about that one. I think with general law counties, but not charter counties.

Perhaps country government could use some sort of reform or overhaul in California. If one looks at a map of the counties in California, there are so many counties in Northern California, but so few in Southern California, but that is where most of the people in California live. The last new county to have been formed was in 1907 which is Imperial County.

Unincorporated Areas in L.A County is the L.A County Board of Supervisors, but have the same vote as those that are part of incorporated cities. It’s a complicated system that never would have been designed by scratch. I really admire how Hawaii forms it’s local government. All local government is provided by the county level, there are no lower levels of local government. So most U.S states have a county/city government layer that overlaps, and varies from city to city. It’s flexible, which is good. One wonders if it is efficient and would consolidation save more money.

Some cities have there own separate fire departments and police forces whereas other contract those services out to the county. The city of Lakewood is one such example, it pioneered the contract city plan.

Also your forgetting San Francisco Country which is the only consolidated county/city in California, which elects 11 members.

Seed planted by Suaprazzodi — 15 May 2009 @ 20:31

The New America proposal is, I think, an effort to influence — specifically to broaden — widespread discussion about (1) calling a state constitutional convention, and (2) getting rid of the super-majority requirement in the Legislature for passing budgets and tax measures. These discussions are in turn brought about by a belief that state government is hopelessly gridlocked and even dangerously close to collapse.

I will comment later on both the choice between STV and MMP and the various points raised so far about city and county government. I’d like to start, though, by asking for reactions to a specific element of the New America approach, one that would be equally interesting had they chosen to recommend STV with similar district magnitudes. This is the attempt to link large multi-member districts with regional interests and identities.

The authors start from the facts that (1) California’s several regions are very different from one another, both economically and socially, as well as fairly cohesive; and (2) there seems to be no realistic prospect of governmental institutions on a regional level (previous points about the multiplicity of overlapping and nested jurisdictions are relevant here). They argue that having delegations to the legislature representative of each of the regions would address this problem.

For those of us who “know the drill” about PR, this aspect of their presentation may seem like an opportunistic way to frame an argument for PR to people who have never thought about it or who have thought about it and are, at best, indifferent. At the very least, it provides an understandable rationale for regional lists as opposed to one statewide list that doesn’t involve wonkish stuff about controlling the threshold. But I have a feeling there may be more to it.

I’ve never thought about STV or PR with regional lists in terms of promoting regional identity and/or regional interests. Is this a well-rehearsed topic in countries with list PR? Does it work? Are they on to something that I don’t know about?

Seed planted by Bob Richard — 15 May 2009 @ 21:43

One of the downsides of the regional approach (in California, anyway) is that it pretty much requires MMDs with different seat counts, if the regions are to be at all meaningful.

That’s problematical. A reasonable threshold for party representation might be in the 3-5% range, (PRF21 says 5%; certainly no higher than that), but the proposal includes a district with 11 members.

Pure STV, of course, would have to address a related issue: you’d need 20 (well, 19) seats per district to get the threshold down to 5%. That means either very large districts or a bigger legislature.

Seed planted by Jonathan Lundell — 15 May 2009 @ 22:12

The approach is interesting. I don’t know how you promote “regional interests” if the regions are just constituencies for party-list legislators. If they were also part of a reorganization of the layer of government between the state and the municipalities, then I might say that they are on to something. (Abolish counties as we know them, and replace them with “regional” governments corresponding to the list constituencies? That might be an idea…)

Bob asks, “Is this a well-rehearsed topic in countries with list PR? Does it work?” Well, to the extent that most PR systems have regional districts in which PR allocation takes place and that are also administrative and political units, we could say, yes. But the “and also” is key.

I have not read the full proposal yet. Only the summary.

Suaprazzodi, I wish I knew of a way to sell the idea of more representatives as, well, more representation (where more=better). But, in the absence of such a sales approach, I fear your opening paragraph (at #3) aptly summarizes a major stumbling block.

Seed planted by MSS — 15 May 2009 @ 22:17

And I rather like the idea of combining cameral and electoral reform. In fact, I have thought that way for many years, except that I have always imagined starting a bit more cautiously: a 120-seat single chamber with 80 single-seat districts (the current number in the Assembly) and 40 list seats (the number of current Senators), elected from regional “top-up” districts of about 5 seats each.

But if the New America Foundation thinks it can sell 360 (The Great Turnaround for California!), I’m all for it.

Seed planted by MSS — 15 May 2009 @ 22:27

The STV quota has nothing like the significance of the MMP threshold because in STV there are no wasted votes, where in MPP the votes for parties that do not reach the threshold are simply binned. The legislative council of Victoria has regions. Each comprises 11 contiguous legislative assembly seats and returns 5 members by STV. The current council has 19 government MLCs, 15 opposition MLCs, and 6 minor party MLCs.

There is no inherent reason why STV regions should not return different numbers of representatives, so long as the ratio of electors to representatives is uniform. The ACT legislative assembly has 3 regions, 2 returning 5 each and 1 returning 7. STV does work better with districts that return odd numbers of representatives.

There would be no difficulty using the PPF21 regions as STV electorates.

The greatest difficulty with preferentialising PPF21 s that STVPPF21 is even uglier than the current acronym.

Seed planted by Alan — 16 May 2009 @ 00:03

I’m a little confused here… What is the point of using the regions as an intermediary? Also, why are we going half and half here with list/electorate members? If i recall my New Zealand politics correctly significantly less than half the parliament are list MPs.

On the increasing of the number of politicians, how much are members of the state legislature paid at the moment? If it’s anything remotely significant, it could be packaged to the public as follows: Cut the salary of each to stop the gross wage bill for them all rising too much.

Seed planted by Luke — 16 May 2009 @ 17:07

Alan @8: There is no inherent reason why STV regions should not return different numbers of representatives, so long as the ratio of electors to representatives is uniform.

Yes, I agree. But in the U.S. the courts might not agree. The argument would be that in a PR system the right to equal voting power implies a right to the same threshold of representation as everyone else. Why should the 5% of those in Sierra foothills who vote for the Purple Party be denied representation when the same percentage in Los Angeles gets the Purple Party a seat?

Litigation, like war, is the continuation of politics by other means. So I fully expect this argument to be used — by people who have no real concern for equality of thresholds — if any form of PR is adopted at the state level. I doubt that it would get much traction in the courts of most countries. But in the U.S. voting rights is a complex subject that has already led the Supreme Court into constitutional cul-de-sacs. We need to be ready to argue this in terms of constitutional law as well as governance, and I’m not sure we are.

Alan’s point, and my caveat, apply equally to MMP and STV.

Seed planted by Bob Richard — 16 May 2009 @ 18:42

Luke, NZ MMP (like German) aims for 50-50 but may end up with a few more list seats if one party wins more than twice its percentage share of district seats. It’s not like they can create new districts after the election! (Although never underestimate the lengths to which the Plant Commission would have bent over backwards to avoid recommending STV…)

Speaking of Plant, I suspect one difficulty with using AV-IRO for the district seats is, that it clouds which voters are “un[der]-represented” and so “deserve” topping up with seats from their preferred party’s list.

Suppose a Labo[u]r MP with 42% defeats a conservative candidate with 47%, because the 11% who voted Green overwhelmingly favoured Labo[u]r with their second preferences.

Should the Green voters then have their list-votes counted (at full 100% value?). One could say their votes have already been represented, by their second-choice district candidate; on the other hand, of course, their first-choice party isn’t (yet) in the legislature. One advantage of FPTP (gods, did I really just type those words?!) is that votes and party are co-terminous.

One could imagine a “tight” pre-electoral coalition – say, the Australian Liberals and Nationals – exploiting this feature, or being accused of exploiting it. Imagine the last few Queensland elections under MMP, if substantial numbers of conservative voters vote (say) Liberal with their Erstestimme but National with their Zweitestimme (yes, that’d be the common parlance, at least on the Darling Downs…).

My occasional attempts to come up with a top-up version of STV (no lists) run into an analogous problem. My last draft, frex, would have (say) 11 at-large seats (even if there were more than 11 mutli-seat districts, as in Ireland). You’d re-count the ballots for at-large candidates after all the local seats were filled, reducing every ballot to its transfer value as at the last stage of the local count. Again, though, this would reward minor-party supporters for not voting for local seats, if it turns out in hindsight that their party would not have reached the district quota.

Seed planted by Tom Round — 16 May 2009 @ 18:51

The problem with districts with different number of seats is that smaller parties end up getting treated differently in the various districts. A party that could win a seat in a nine-member district might well not be able to win one in a five-member district. I’m not saying it’s a fatal flaw, but it’s certainly undesirable. (I see that this is the point of BR#10.)

Odd number of seats? Not that important, in my view. Yes, it guarantees that a majority of voters can elect a majority of seats, but with many parties and many districts, how important is that, really?

WRT names, I assume that “personalized” refers to the constituency seats, and “full representation” to the regional top-up seats, yes?

Seed planted by Jonathan Lundell — 16 May 2009 @ 21:18

Jonathan,

Yes, you understand perfectly the reason for rebranding MMP as “personalized full representation”. Like everyone else who has commented on this, I’m not that thrilled with the name. I’m also not thrilled with rebranding in general. Instant runoff voting (instead of the alternative vote) is OK, but that’s because it was widely adopted for a while, until folks in some places decided they liked ranked choice voting (RCV) instead. The attempt to replace STV with “choice voting” has been less successful.

I really wish all of us who support reform could agree on one nomenclature and then actually stick with it. I know people who are turned off to reform because we can’t agree on what to call the things we favor.

On the other hand, I’m not sure you understood the thrust of my comment on unequal district magnitudes. While I agree that they are not a good thing, I don’t think they should be an obstacle to reform. But I’m afraid they may well be an obstacle in the courts. Look at the current lawsuit in Minnesota, which is trying to have the adoption of the alternative vote (oops, IRV) and STV (oops, choice voting) by the city of Minneapolis invalidated. The arguments in that suit are silly, in both legal terms and political science terms. The opponents are grasping at straws (and will, in that case, clearly lose).

Unequal district magnitudes will provide opponents with such a straw. I don’t know how to construct the constitutional argument to break that straw, which makes me afraid of it. If such a challenge were successful in the courts, it would create a huge obstacle to the adoption of any form of PR — except statewide lists — on the state, never mind the Congressional, level. The reason for this has already been discussed in the thread on B.C. — folks won’t accept rural regions over some number of square miles.

That’s what I was trying to get across in my post. We had better be ready on the constitutional questions. They will get presented to the courts.

Seed planted by Bob Richard — 16 May 2009 @ 22:02

The effect of the STV quota is that in an election for 2 seats. as in in the Australian territories, you need to take 2/3 of the vote to win both seats. If the territories returned 3 senators each you would need just over 1/2 the votes to take 2 of the 3 seats. No election of territory senators has ever produced any result except 1 government, 1 opposition.

Irish governments have been known to create even-numbered seats where the opposition is strong and odd-numbered seats where the opposition is weak, ensuring that the government gets significantly over-represented in parliament.

The odd number does matter in STV.

Seed planted by Alan — 16 May 2009 @ 23:49

Tom @11 NZ MMP (like German) aims for 50-50 but may end up with a few more list seats if one party wins more than twice its percentage share of district seats.

Not quite. When NZ MMP started, there were 60 general district seats, 5 Maori district seats, and enough list seats to take the total to 120. As the South island always has 16 general districts and their proportion of population has been dropping, we are now up to 63 general districts. As the (optional membership) Maori Roll has increased, the number of Maori seats has grown to 7. Notional total is still 120 (with 2 additional ‘overhang’ seats due to the Maori Party getting too many districts for their list votes).

Seed planted by Errol — 17 May 2009 @ 00:00

Aha. Sorry, I missed that the 5 Maori districts are separate to the 60 General districts.

Seed planted by Tom Round — 17 May 2009 @ 04:43

Can we agree on one name for MMP? The Germans would say, sure: “personalized proportional representation” is what they call it, and they pioneered it, so why are we using a wonky name like MMP that doesn’t even translate into French well (a problem in Canada)? Scotland calls it AMS. However, English-speaking Canada is now stuck with MMP as the name, thanks to New Zealand, oddly enough. If you aren’t into copying New Zealand, call it what you want. (The Germans called it personalized PR to distinguish it from their previous list PR.)

Do the regions have to correspond with another level of government? They don’t in Scotland, but they certainly represent real regions.

Do they all have to be the same size? They aren’t in Bavaria. If you use “highest remainder” a 5% party may win a seat even in an 11-seat district if they’re lucky.

But based in Canadian experience, don’t use closed lists.

Seed planted by Wilf Day — 17 May 2009 @ 11:40

Alan (#14): The odd number does matter in STV.

At the state/provincial and national levels, yes. I doubt that it’s worth getting upset about for city councils and the like. Let’s not let the perfect be the enemy of the good.

Even at the national level, the Irish form of gerrymandering — which is indeed a problem — can be mitigated by establishing boundary/redistricting commissions that are independent of the officials whose jobs are at stake.

Wilf (#17): Can we agree on one name for MMP?

At least in the U.S., we don’t seem to be able to agree on one name for anything.

Wilf again: But based in Canadian experience, don’t use closed lists.

I am not a fan of closed lists anywhere — with or without a “personalized” [sic] component.

Seed planted by Bob Richard — 17 May 2009 @ 12:18

I suspect that Bob’s concerns about legal challenges ought to be taken seriously.

Tom, at #11, perfectly outlines my concerns about using AV (IRV) with MMP.

Wilf, I doubt that the regions the New America Foundation has defined for California are as ‘natural’ as those of Scotland. But I will admit that I don’t know much about Scottish regionalism. I’m just skeptical that the NAF intention to invigorate regionalism via this new concept of party lists can do much if the regions do not have an independent existence.

As for naming, “personalized PR” never made much sense to me, as it would seem to describe open lists better than MMP (or whatever we want to call it).

Seed planted by MSS — 17 May 2009 @ 15:22

Local government in California

For background on this, see this brief survey. My reponses below will probably make more sense if you look at that document. I’d love to get feedback on it, since I’m thinking about expanding it for use elsewhere.

Suaprazzodi (#4 in the “BC: FPTP Forever?” thread) is right that the state legislature cannot unilaterally create or destroy cities and counties. But part of that limitation is statuatory rather than constitutional and could be changed by the legislature itself. So the contrast with local government in Canada, while very relevant to our discussion, is not as stark as he might think.

Suaprazzodi (#11 in the B.C. thread): What 10 jurisdictions are going to try it?

Jurisdictions would have to apply. Because this is a pilot project, the bill requires that they be stratified by both size and region. This might become a real constraint on participation; if two small towns in the northern counties of Humboldt and Mendocino applies (say, Arcata and Fort Bragg, where there are already many ranked voting supporters), I believe that only one could be accepted. Within strata selection would be first-come-first-served.

Loc. cit.: Multimember districts would be a hard sell in California.

And everywhere else in the U.S. except at the local level. Many small towns and school districts get sued under our voting rights laws because they won’t give up at-large plurality in favor of districts, so that a “safe” district can be created for an ethnic minority. There’s research that tries to show that shifts back and forth between at-large and district elections are motivated largely by the white power structure’s desire to minimize the effectiveness of minority votes. And there are other forces at work. The belief that city council members should work for the whole city and not just their own neighborhood is widespread.

The part of this that has to do with representation of ethnic minorities is a complicated topic, which deserves a long thread of it’s own.

Ed (#12 in the B.C. thread): Los Angeles has five supervisors elected from single member districts each containing almost two million people.

This is only one of many truly odd things about California government that don’t seem to attract a lot of attention or concern. All counties have top-two runoff elections (except San Francisco, which has IRV), while all but a dozen or so cities have plurality elections (and cities lacking home rule charters must have plurality elections). And so on.

Loc. cit.: it strikes me that five supervisors is the perfect number for an STV election. Of course electing five supervisors from each of the existing districts in large counties would be better.

All, or almost all, local bodies are elected for staggered terms, and voters seem to believe in staggered terms devoutly. If your proposal for STV also proposes to do away with staggered terms, it will almost certainly fail. So the effective district magnitude would alternate between two and three.

Combining an STV proposal with a proposal to have five members per district (25 in all) is an excellent idea that would also fail to win any support. As several others have already noted, expanding the number of legislators seems to be a harder sell than PR itself. Reformers in San Francisco solved this problem in their 1996 ballot measure by proposing STV in two districts with five and six members respectively, thus keeping the total number at 11. This got about 45% of the vote, while a proposal to change from city-wide at-large to 11 single-member districts passed at the same election.

Jonathan (#1 in this thread) points out that local government is a very good place for STV. This true not only as a matter of good government — something which has been recognized since the Progressive era — but also strategically. Campaigns on the local level are winnable in some places. Campaigns on the state level are not — yet.

Suaprazzodi (#3 above): I think the California State Assembly could change [the size of the Los Angeles County board] by law, if it wanted, but I’m not sure about that one. I think with general law counties, but not charter counties.

Correct with respect to “general law” cities and counties. Los Angeles County has a charter, and I’m pretty sure the Legislature could not override a home rule charter on this issue (although state law can preempt charters on a lot of issues — another complicated subject).

MSS (#7 above): But if the New America Foundation thinks it can sell 360 (The Great Turnaround for California!), I’m all for it.

This is a good place to comment on why I think the New America document is important. It is not, at least in my view, meant to gain immediate support or adoption in the near term. It is an effort to seize an opportunity to make opinion leaders more aware of PR. The opportunity is presented by widespread talk of a state constitutional convention. In this light, proposing a unicameral legislature roughly three times the size of the current Assembly and Senate combined makes some sense. These changes do belong together. Since this is not going to be adopted anyway why not put everything on the table? After all, this is an effort to broaden a relatively narrow dialog.

Day-to-day activism is directed entirely at local government. But we also try to seize every opportunity to address opinion leaders with proposals like this one.

Seed planted by Bob Richard — 17 May 2009 @ 15:36

The reason I don’t see an odd number of STV seats as important is a) that 50% something of an arbitrary point (as long as we satisfy the Droop Proportionality Criterion I’m happy), and b) with multiple multi-member districts, if we care about a majority of seats anywhere it’s in the main body, not in each district, and having an odd number of seats in each districts does not guarantee a party with a national majority a majority of total seats.

Moreover, if the choice is between (say) 5 and 6 seats, I’d prefer 6, on the grounds that (within reason) the more seats, the better the opportunity for proportionality.

WRT differing seat counts, I don’t agree with the implication that a potential legal challenge is merely a political-technical difficulty. There really is an equal-representation issue to be addressed. It may be that the advantages of varying seat counts outweighs the disadvantages, but that’s not clear on its face.

(I see heterogeneous seat counts as more of a problem with lists than STV, but I’m not certain that I could defend that position.)

Seed planted by Jonathan Lundell — 17 May 2009 @ 16:10

Jonathan: It may be that the advantages of varying seat counts outweighs the disadvantages, but that’s not clear on its face.

It’s clear to me simply because the advantage of varying seat counts is that electoral reform is pretty much impossible without them. The practical choice is between varying magnitudes and nothing.

Seed planted by Bob Richard — 17 May 2009 @ 16:35

Any change to the California Constitution would require an initiative measure with valid signatures equal to 8% of the vote in the last gubernatorial election. Many signatures are rejected so organizers should plan on gathering a million signatures. Another issue to consider is that each initiative may only have one major issue.

A major purpose of the legislature is to represent the various differences in a society. The best approach to obtaining a variety of viewpoints is through proportional representation. California elections for state legislature and above are partisan. As such I would suggest that a list system of PR be used. Each of the six parties; American Independent, Democratic, Green, Libertarian, Peace and Freedom, and Republican; could use their direct primary elections to arange their list in the order of preferrence among their voters.

A unicameral legislature would be be preferable if both houses are elected in the same mannor. However, that would require a separate initiative. In the meantime, instead of MMP, that only the State Assembly be reformed by electing it’s members by PR list. With the State Senate still being elected from single member districts, this two house approach would have the same effect as MMP.

California’s State Legislature really should be larger, much larger. Again that would require a separate ballot initiative. In the meantime, I would suggest that California have either, eight districts electing 10 members, or four districts electing 20 members each.

While the type of list system is debatable, I prefer the largest remainder. It is easier to explain to people. Too much confusion could be our downfall.

Seed planted by C. T. Weber — 17 May 2009 @ 16:51

STV is better suited in California for non-partisan races like city council and county boards of supervisors.

IRV is better suited in those races that elect only one person such as a mayor, governor and the other statewide positions such as constitutional officers. I would include U. S. Senator in this list because of its unique status.

Except for U.S. Senate, legislative races should be run using PR. It is unfortunate that several people are promoting IRV for electing members to their city councils. I understand the arguements for this but do not share the conclusion.

Seed planted by C. T. Weber — 17 May 2009 @ 17:03

C.T.: With the State Senate still being elected from single member districts, this two house approach would have the same effect as MMP.

I don’t think so. The two houses have equal power (legislation has to pass both), so having PR in one and FPTP in the other is less proportional than a single house elected by list PR or MMP or STV. That said, getting PR in one house of a bicameral legislature would be a very big step. For example, Australia (except that there the upper house is proportional).

C.T. again: It is unfortunate that several people are promoting IRV for electing members to their city councils. I understand the arguments for this but do not share the conclusion.

This is only true in places where the council is already elected in single-member districts. I have not heard anyone propose changing from at-large to districts in order to adopt AV/IRV. Those of us who disagree with C.T. on this argue that moving those districts from plurality to majority voting is a positive step even though it might not lead (directly or immediately) to PR.

Seed planted by Bob Richard — 17 May 2009 @ 17:17

Regarding even district magnitudes, it’s true that with four seats and two parties each party will win half the seats in perpetuity. But even M matters much less when there are more parties, or higher M. With 136 seats in Greater Los Angeles, even a small change in support would change the results. Similarly, if you have just two seats, but four parties competing for them, nobody is guaranteed a seat.

There’s nothing really special about even numbers, except that two-party systems are fairly common. Generally, with P parties of similar vote share, one would expect similar problems with any M = nP for integer n. It might be interesting to quantify “electoral stasis” and measure it as a function of M, but you’d want to have some distribution of “reasonable electoral results”, and that’s a whole ‘nother can of worms.

Seed planted by Vasi — 17 May 2009 @ 17:40

Two houses with different rules are certainly not the same as one with a combination of those rules.

It matters how you aggregate majorities, and where you place vetoes.

Bicameral (i.e. two institutions, each with vetoes over legislative output) with different electoral systems, is even less like a unicameral mixed-member system when the form of the latter that we are talking about is one that is compensatory, rather than majoritarian.

Seed planted by MSS — 17 May 2009 @ 17:47

@Vasi

The Australian territories, where my examples focused, are contested by many more than 2 parties. The same logic applies in electing state senators, where each state elects 6 senators at each half-Senate election, but for obvious reasons working out the examples would be more complicated.

Seed planted by Alan — 17 May 2009 @ 20:09

Even numbers do matter when they are small (below 7). With 2- or 4-seaters, any majority below 66.67% or 60% leads to a deadlock.

To put this in perspective, only three US Presidents in the past century have polled 60% (1936, 1964 and 1984). Admittedly it is statistically easier to find a huge majority in a district than in a whole nation, but it is still difficult, especially in a multi-seater.

On the other hand, it would probably make almost no discernible difference if the NSW Legislative Council had 20 or 22 MLCs, instead of 21, elected statewide by STV at each election; firstly because it would raise the threshold needed to win a majority from 50.0001% to “only” 52.38% or 52.17% of the final-count-preferred votes — and there will be proportionately fewer occasions when a party exceeds 50.0001% but falls short of 52.38% or 52.17% — and secondly, because the lower quota per seat means it is unlikely any party will get 50.0001% of preferred votes anyway. Whereas with 5- or 7-seaters, it is not unknown for 47-48% of first preferences to translate into a majority of preferred votes.

“Tullymandering” in Ireland has been used to discredit STV-PR there, especially (but not solely) back when the boundaries were drawn by a Cabinet Minister’s office and not by an independent commission. It makes a big difference whether a city entitled to 20 reps is divided into four 5-seaters or five 4-seaters. The former would give the dominant party 12 seats out of 20, the latter only 10. Given that numbers in a PR-elected chamber are going to be tight, that could be quite significant.

Seed planted by Tom Round — 17 May 2009 @ 21:56

TR#29: That’s why we need to build coalitions after an election. Even with an odd number of seats, there’s no guarantee of a majority party.

But what we’re talking about (in the case of California) is *many* multi-seat districts. No single district controls the majority.

Seed planted by Jonathan Lundell — 18 May 2009 @ 02:18

BR#22 It’s clear to me simply because the advantage of varying seat counts is that electoral reform is pretty much impossible without them. The practical choice is between varying magnitudes and nothing.

If that were the choice, I’d agree in a heartbeat. But I don’t follow the argument that it’s the case.

If single-seat districts have given us nothing else, they’ve taught us how to create districts of equal size. How about a legislature of 50 7-seats districts elected by STV? The districts would be geographically smaller than the present CA Senate districts, and with seven seats, proportionality would be pretty good. Or at least hardly nothing.

Seed planted by Jonathan Lundell — 18 May 2009 @ 02:23

Creating regions of different size but the same ratio is no more difficult than creating districts of equal size. Victoria provides that each 5-member legislative council district is to comprise 11 contiguous legislative assembly districts. There is no reason those council electorates could not comprise different numbers of contiguous assembly electorates. Once you have districts of equal size, creating superdistricts of equal ratio is not a challenge.

As Tom shows in his last post, you cannot necessarily assume that Tullymandering will affect only 1 seat within 1 only multimember district. Consistent Tullymandering can turn an election just as can consistent gerrymandering.

Seed planted by Alan — 18 May 2009 @ 06:32

Jonathan (#31) is right in principle. It would be great if he could be right as a practical matter as well. In some far-off future time, events may make him right. But voters will have to get rid of the idea that the solution to politicians is to have fewer of them.

As his example Jonathan offers a 350 seat legislature. I personally would be fine with that (although applying the cube root rule to California suggests about 260, somewhat more if you count those eligible to vote but not registered). The voters we need to appeal to are not going to be fine with it.

If we add together the current Assembly and the current Senate, that’s 120 seats. Let’s say we put them in 24 5-seat districts. I personally would be fine with that too. But the opposition campaign is going to advertise maps showing the northern-most district running all the way from the Oregon border to beyond Chico. And voters are not going to be fine with it either.

The United States Congress is a completely separate matter. Giving each voter the same threshold would require a major rewrite of the Constitution, not just an amendment. On the other hand (something I didn’t think about earlier), it’s possible that the courts would treat equality of thresholds among the states differently from equality of thresholds within a state. At least the argument could be made (somewhat disingenuously, I think) that federalism allows the former even if one-person-one-vote were held to forbid the latter.

Seed planted by Bob Richard — 18 May 2009 @ 11:30

Bob,

If we kept 120 members, then districts twice the size of current senate districts could have six seats each. If we doubled the legislature to 240 members, then a six-member district would be the same size as a current senate district, and five-member districts would be smaller. None of that seems all that unreasonable.

Congress, it’s true, is way different, especially considering that we have states with only a single representative. One can imagine schemes that could work, but only with radical restructuring. Unicameral MMP for Congress? Not any time soon.

These are all fun to talk about, but in California, STV for non-partisan local elections seems to me to have a whole lot more promise.

Seed planted by Jonathan Lundell — 18 May 2009 @ 12:27

[A note to our hosts: the ten-minute comment-editing window is very nice. Thanks.]

Seed planted by Jonathan Lundell — 18 May 2009 @ 12:29

Jonathan (#34): in California, STV for non-partisan local elections seems to me to have a whole lot more promise.

My main point, exactly. (See #20 above, along with these comments, which folks may have missed because it was delayed for moderation.) Thanks for bringing local government back into the discussion. Once folks have some experience voting in STV elections for city council, it will be much easier to talk to them about the state legislature.

Seed planted by Bob Richard — 18 May 2009 @ 14:00

The survey on local government in California linked to in comment # 20 was very informative. I didn’t know that elections to the county boards of supervisors were staggered.

The Economist also had a decent article recently on California politics that mentioned the proposed constitutional convention. The article contained the mindboggling information that you needed a 2/3 vote in both houses of the state legislature to pass a budget, and to pass taxes. Add that with the apparent ability to commandeer parts of the state budgets through initiatives and you can get a state government headed towards fiscal disaster.

Unfortunately, its hard to see how you can reform the system without persuading voters to give up some of the more gimmicky populist elements.

Someone noted the existence of all the small counties in northern California and the large counties in southern California. I really like the large SoCal counties, it makes it theoretically possible to have metropolitan area wide governments, something I think the US needs more of. I don’t know how much the actual system of local government in California takes advantage of this.

California may be too big to be a state, it has a larger population than most countries and I don’t know if the average California voter has more leverage in Sacramento than in Washington. One way around this would be to consolidate the counties in the north into a few larger counties, then delegate most state powers and revenue away from Sacramento and to the County Boards of Supervisors. But then you would have to look at reforming the county governments.

To nitpick a comment in #29, the four elections where a US presidential candidate got 60% of the vote were 1920, 1936, 1964, and 1972. On the overall point, there is enough polarization in US politics that Democrats would win three or even four seats in four member STV districts in the cities, and the Republicans have their strongholds as well.

Seed planted by Ed — 18 May 2009 @ 15:24

On the distribution of county sizes, the large counties in southern California emphatically do not allow for metropolitan-wide governance. The metro area of Greater Los Angeles for example, encompasses LA, Orange, and parts of Ventura, Riverside, and San Bernardino Counties. In some cases, the county lines go right down the middle of major arterial streets or follow (mostly dry and concrete-lined) river beds right through communities that otherwise would be governed together, sensibly.

Meanwhile, Los Angeles and, especially, Riverside and San Bernardino Counties extend well into relatively unpopulated desert and mountain areas that have little in common with what are, geographically, their fringe areas where the bulk of the people live. (Some of these counties, notably Riverside, have more distant population centers of their own, such as the Palm Springs area, but this does not change the general point; in fact, it strengthens it, as it points up the absurdity of Indio and Corona being part of the same county.)

Meanwhile, the San Francisco and Sacramento areas are divided into an even greater number of separate counties, although these counties typically do not have the ‘hinterland’ of some of their southern counterparts.

I am not sure what this has to do with electoral reform, other than that it buttresses, I think, the notion of exploiting regionalism as a new strategy for intermediate layers of governance as well as for permitting proportional representation of multiple partisan interests within the various regions.

Seed planted by MSS — 18 May 2009 @ 15:33

And I might add that, increasingly, the San Diego metro area has expanded to include part of southern Riverside County (and even somewhat into southern Orange County). I know all too well how much this expansion has accelerated, as Ladera Frutal is uncomfortably near country roads that have become part of new arterial commute routs. And the noise from I-15 (just east of us) has gotten much more noticeable over the years, as more and more people live in places like Temecula but still work somewhere in SD.

There is also some spillover of the SD metropolis into Imperial County (and into Tijuana and Tecate, but that’s obviously a separate matter).

Seed planted by MSS — 18 May 2009 @ 15:42

Someone has to have the 40th comment in this thread, and it might as well be me…

Despite how little counties ‘objectively’ mean, county identity remains strangely firm, as has been discussed in the past in Ladera Frutal’s Ballyard.

Seed planted by MSS — 18 May 2009 @ 15:47

The mention of staggered elections (electing 2 or 3 members of a 5-member board every other year) reminds me that there are a few practical difficulties that need to be worked out in a move to STV; I’d be curious to hear how others have done it.

The rationale for staggered elections is to provide continuity, which is fine as far as it goes. But we don’t get much PR with 2- and 3-seat elections. The alternatives seem to be to expand the board size, or to throw staggering overboard.

(STV with staggered elections would make an even number of seats desirable, btw.)

The other difficulties include recall and filling vacancies.

Seed planted by Jonathan Lundell — 18 May 2009 @ 16:02

Alan:
Indeed, more than two parties contest elections in Australia, but this is also true about elections in the US. That doesn’t stop me from calling the US a two-party system! What matters is the electoral clout of the parties.

Looking specifically at the Australian Territories, in the past five elections the effective number of parties has averaged 2.86 and has never exceeded 3.22. The third party has never been within 25,000 votes of one of the top two. So I feel pretty justified considering these to be two-party elections, at least at the present moment.

Bob:
I’m interested in why the voters wouldn’t be alright with geographically extended districts. Judging by the BC-STV referenda, I’d expect the reason is that they imagine themselves trekking hundreds of miles to see a representative. This doesn’t need a fix at the level of the electoral system. Just require that each district have some number of offices in different areas–and mandate that reps spend time in each office. Then when they look at the map, they’ll see a district office not so far away.

Seed planted by Vasi — 18 May 2009 @ 16:09

Vasi (#42) suggests countermeasures to offset the (perceived) downside of geographically large districts. I think this could be a very helpful approach. On the other hand, district size resonates with voters not just as a practical matter of travel and accessibility. It’s also a symbolic issue that opponents can use to reinforce their claims that you, as a citizen, need exactly one “representative” (in the sense of agent) who is “accountable” to you. I suspect that the argument works with voters in part because the practical and symbolic elements reinforce each other.

Along the same lines, has any jurisdiction ever tried to separate the legislative and constituent service/ombudsperson roles into separate offices? Would it work? Or are there valid reasons why effective constituent service depends on also being a legislative decision-maker?

Seed planted by Bob Richard — 18 May 2009 @ 16:25

Ed (#37): … theoretically possible to have metropolitan area wide governments, something I think the US needs more of. I don’t know how much the actual system of local government in California takes advantage of this.

Almost not at all. In fact, that’s the starting point for the many of the specifics of the New America plan. But they don’t explain how their form of PR would actually lead to regional structures.

In the San Francisco Bay Area, the several bridge districts have managed to established a common toll collection system, and there is some degree of schedule coordination between transit systems. I think that’s about it.

Seed planted by Bob Richard — 18 May 2009 @ 16:35

Would it be fair to say that the Scandinavian model of an ombudsman (-person) is a substitute for legislative constituent service? I do not know much about the institution, or whether the model could be generalized.

Seed planted by MSS — 18 May 2009 @ 16:54

Most democracies now have ombudsmen (ombusdperson just has not taken wing for some reason) and to a degree an ombudsman can take up the slack for an under-performing MP or legislature, but that is not their main function. Their main function is resolving citizen complaints by investigating and discussing them with the responsible agency.

In extreme cases the ombudsman will report agency failures to parliament. In some jurisdictions the ombudsman has much wider powers like enforcing the bill of rights. All jurisdictions make the ombudsman an independent agency.

The various agency inspectors-general are a close, but not exact, analogue, in the US.

Seed planted by Alan — 18 May 2009 @ 17:54

Also, in Australia (eg), an Ombo (yes, that’s what the NSW Ombudsman’s office has on its website – gender-neutral, easily pluralised, and magnificently Ocker) is barred from commenting upon, or even investigating, matters that are primarily “policy” rather than “administration”. So it helps to have a legislator as a backup.

A stricter Hayekian than I might get annoyed that, these days – as Prof John Hart Ely has noted – most actual legislation gets drafted by the executive, via regulations (Ministers in UK, Aust, NZ, Canada, etc: tenured Commissions in the USA), while legislators focus on executive matters, ie chasing “pensions and passports” for their constituents. As it is, I consider it an example of Hayekian spontaneous order…

One of the excellent Butler & Ranney ed books pub by the American Enterprise Institute 2-3 decades ago – sorry, but I have forgotten the title – contained a chapter whose author made exactly the point MSS does. In Scandianavia, s/he wrote, what is considered normal constituency service by a British MP would be criticised as improper politicised interference in the administration. Hence the “surgery” matters devolve to the Ombudsman instead. Water finds a path.

Seed planted by Tom Round — 18 May 2009 @ 18:42

Alan and MSS,

In an effort to repair my ignorance, I visited the Wikipedia article, “Ombudsman”. (Yes … I know … but it was quick.) This turns out to be a complicated subject. It appears that the existing models are mostly based in (although nominally independent of) agencies or departments, rather than based in constituencies (geographical or otherwise). But it seems possible that their existence might reduce the extent to which the legislator’s job consists of trouble-shooting constituents’ problems with the executive branch.

Wikipedia quotes an APSR article from 1977 to the effect that legislators can be effective in this role because collectively they hold the agency’s purse strings and therefore have some clout. That seems responsive to my question about whether it would be easy to separate the roles of legislator and ombudsman. Might be a hard sell, after all.

Seed planted by Bob Richard — 18 May 2009 @ 19:09

In a desperate effort to avoid the infelicity of Ombudsperson or Ombudscreature the South Africans went for Public Protector. I am not sure they helped matters much. ‘parliamentary commissioner’ was also quite popular for a time.

Seed planted by Alan — 18 May 2009 @ 20:05

Wow, is this the longest thread on this blog? This is the 50th post.

It looks like I started a long discussion. Perhaps Southern California needs more countries, perhaps these new counties should have a minimum of 25 members. It’s hard to know if the people would vote for new counties.

An expanded legislature is not likely to happen. I think we are stuck with an insanely small sized legislature. I wonder why the it wasn’t allowed to grow in size.

People do not want more politicians, they hate them already. It would be better to have more politicians and a large legislature, I agree, but persuading people would be a hard task.

Does any one have any ideas why should be have a large state legislature and county board of supervisors?

I think one reason is that members can specialize in two committees instead of being on 14 committees and know more rather than being spread thin and depending on the civil service advice on drafting bills. There would be more checks and balances and these people would be closer to the people.

Any concrete ideas to sell more elected politicians to people?

I think STV at the local level with a small district magnitude is better than nothing. Some cities elected 7 members and other elected 9 members, and so those elections would be more proportionate even if staggered. One has got to take these things slowly.

Californians are very picky, and it looks like all the propositions are all going to fail except 1F unless a miracle happens and 1A and 1B passes. *

* A miracle depending if you are Governator.

Hopefully the independent redistricting commission is going to change things, but I doubt it.

The question about California is why Arnold failed. The two thirds majority rule is to blame in many cases.

When he did that Special Election in 2005, should those propositions passed? I think they failed because he was being arrogant. The oppositions framed it as a power grab. One has to be careful not to fall in that trap.

Maybe California is just feed up having too many special elections, the voters are just fed up with politicians and want to tell them to solve it themselves, not keep asking voters to rubber stamp politicians proposals. This is the fifth election in 2 years. The problems is that the politicians don’t have much leeway in solving these problems perhaps. We have more elections than ‘supposedly’ unstable parliamentary democracies.

Will public services suffer if these propositions are defeated?

Even if they all pass, we are still in a budget deficit albeit not as bad.

Has any suggested that the state constitution be amended to repeal the balance budget proposition or require a balance budget to pass by simple majority and a unbalanced budget to pass by a two-thirds majority? This is an emergency, and thank goodness that such an amendment wasn’t passed at the federal level. Some state level experiments were made in a mad scientists laboratory.

Some on say it is prop 13. I heard other states are having budget crises, and this is just because of the condition of the economy.

Some say it is lavish public employee pensions which are defined benefits and need to be overhauled, because they are unsustainable.

It’s probably a combination of the two, but any centrist politician wants to change both third rails at the same time even just a little bit, you get the extremes clobbering them. I really think that California has no center, only extremes that refuse to budge no matter what. I think the problem is the lack of creativity is the problem.

It looks like a state Constitutional Convention is coming soon.

Maybe California is too big and needs to be broken up into a couple of states, at least we get more Senators. LA County has the population equivalent to the state of Michigan.

Seed planted by Suaprazzodi — 18 May 2009 @ 20:12

Suaprazzodi,

California has a political center. It’s just not represented in the legislature. Some of those who supported the redistricting measure thought it could cure that. It can’t.

I don’t have a good theory about why the state is so polarized, but demographic shifts contribute (definitely), as do closed primaries (maybe), and the two-party system.

What can cure the paucity of centrist politicians is, of course, PR. One of my leitmotifs on blogs and over a beer is the extent to which we talk about PR on behalf of small parties on the left and/or right, forgetting that the largest group to benefit is those in the middle, many of them non-partisan or anti-party — Henry Droop’s “moderate non-partisan section”. They are the main constituency of any electoral reform campaign. But we activists are, by and large, out of touch with them.

Many people are proposing the repeal of the anti-democratic two-thirds requirement. Like a bunch of other good ideas mentioned in this thread, it will be hard sell. But things are getting so bad now that it might happen.

Ironically, perhaps, people cling to the two-thirds requirement in part because it operates as a kind of (very poor) substitute for PR. It protects the minority — only at the expense of gridlock and paralysis, but it does protect the minority. In a legislature elected by PR people might not be so passionate about super-majorities.

Seed planted by Bob Richard — 18 May 2009 @ 22:01

Of course, Bob’s point in #51 is one I make regularly, as well (see some of the ‘Preserved Fruit’ on the left sidebar).

For that matter, it is Henry Droop’s argument. He made the case in 1869. (A key excerpt graces this blog’s banner.)

Seed planted by MSS — 18 May 2009 @ 22:06

@Bob

Alaska and Arizona have ombudsmen with reasonably comprehensive websites. I’m a tad surprised the institution is not better used or known in the US. If you want a look at an unusually strong version of the institution, see the agency variously known as the Justice Chancellor or Legal Chancellor of Estonia.

Seed planted by Alan — 18 May 2009 @ 22:14

I’m told “ombudsman” is grammatically neutral in the original Swedish. However, words don’t always stay neutral in translation (compare “negro” in Spanish and then English!). My own approach is “Ombudsman” for singular and “Ombuds” for plural. A but clunky, but it works.

Relating this Calif thread to the “Why do Canadians keep voting down STV…” thread… Most PR advocates would be opposed to partial PR, ie, single-seaters in sparsely-populated rural areas but multi-seat PR in the larger cities. The British Conservatives proposed something similar (with AV and STV) a century ago but Churchill, for one (who was Liberal at that time, and not wholly anti-PR, but who did dislike AV) pointed out that this meant proportionality for the Conservatives where they were weak but landslides for them where they were strong. A problem for all PR systems with varying district magnitudes, but much less intense when the range is 5-7, as in the ACT and WA, than if it’s 5-1 or even 3-1.

In fact, rather than having single-seaters for remote rural districts, I’d rather just give them each a second MP – “gratis”, as it were: so that, eg, an urban district might have 250,000 voters and 5 MPs, but a rural district would have 2 MPs – not 1 – for 50,000 voters. In most cases, their votes would cancel out on party lines – unless the majority ignored that district so much that 66.67% of its voters elected (or prodded) MPs who would unite 2-0 for one side or the other.

There is a partial precedent in the post-1990 compromise that ended Queensland’s former four-zone “Bjelkemander”; districts over 100,000 km2 in area (ie, larger than many a European nation) get additional “phantom voters”, equal to 1% of said area. An “honest gerrymander”, if you will. The 2-seaters could be reserved for the truly huge rural hinterlands (with, say, to 1.0 to 1.5 times the average in voter population) while 5- or 7-seaters would obtain everywhere else.

Seed planted by Tom Round — 19 May 2009 @ 01:50

I’m fuzzy on the details (after all, it is 6:00am and I face a 15 hour day as a pollworker in today’s election), but I believe that a mix of STV in cities and AV in rural areas was actually used at one time in either B.C. or Alberta or both.

As Tom himself points out, if there really were a problem of equity with STV districts of different magnitudes, that problem would be much worse if you added single-seat districts to the mix. As he also says, it’s a compromise I don’t think many electoral reform activists could swallow.

As for giving rural areas extra members, that might (or might not) be good governance, and it might be acceptable under other constitutions, but it’s not acceptable to courts in the U.S.

Seed planted by Bob Richard — 19 May 2009 @ 10:04

Multimember districts for cities and single member districts for rural areas might work with an added MMP component at the state or national level. If one party gets overrepresented by sweeping the single member rural districts, and grabbing its share in the cities, than the other parties get to appoint additional members.

Of course that brings in the component I hate with these systems, members representing the parties and even theoretically elected (even with straight party list systems each member is elected from very large multimember districts). But one way this might work would be to effectively raise the quota in the urban multimember seats for a party that has swept enough single member districts to be overrepresented.

The mathematics would be awful, but the idea would be that no party would be allowed more seats in the legislature than it was entitled too proportionally given its statewide showing. If too many victories in single member seats put it over that threshold, seats won in the multimember proportional areas would be deducted and redistributed to the competing parties in those areas.

Seed planted by Ed — 19 May 2009 @ 13:10

I’m not convinced that any system is absolutely proof against reversed results except a singe undifferentiated district that returns all representatives. An electoral system should therefore always include adjustment seats to ensure that the largest party in terms of votes has the proportion of seats to which its proportion of votes entitles it.

Seed planted by Alan — 19 May 2009 @ 13:22

I just took a walk and worked out the math for my idea in # 56. Its not as bad as I thought. Each victory in a single member district would raise the threshold for a party to win seats in multimember districts.

Take a legislature with 30 single member seats and 70 seats elected proportionally from multimember districts. A party that wins none of the single member seats would need 1/70th of the vote for each multimember seats. When that party wins a single member seat, it now needs 1/69th of the vote. For a party that wins 25 of the single member seats, its threshold would become 1/45th of the vote for the multimember seats.

Seed planted by Ed — 19 May 2009 @ 13:45

Ed,

In your proposal (#58), what happens to independent candidates (and voters)?

Seed planted by Bob Richard — 19 May 2009 @ 13:56

To answer the question in #59, in the single member district component, independent candidates are treated no differently than in single member districts elsewhere.

In the multimember districts, if STV is used I don’t see why independent candidates can’t run. Independents serve in the Irish Dail. If a party list system is used, then anyone running as an independent will have to form their own niche party. In some low threshold party list systems, candidates have been elected on effectively some version of the “Elect Bob!” party ticket. The Connecticut for Lieberman Party would be an example from US politics.

The central idea to my proposal is that the threshold needed to elect a member in the multi-member districts would be different for different parties, and would be adjusted upward for each single member seat a party won. This would preserve overall proportionality, though it would be a problem for urban supporters of parties that are mainly based in rural areas. But rural supporters of urban based parties would be disadvantaged as well compared to purely proportional systems.

Seed planted by Ed — 19 May 2009 @ 15:13

These systems of PR seem way too complicated, let’s just keep it simple. Everyone’s ideas are really good, but they are too complicated with people that will just be learning about a possible system if they ever get a chance to vote on it.

The problem with PR in the US, how do we accommodate it with the primary system.

The primary system is unique to America, and most of the world doesn’t have I guess what one would call “government regulated internal party democracy elections.”

Most countries are probably like Canada where the party runs the party elections. A closed party list system in the U.S would be a no-no.

Would open party list be declared unconstitutional if used in California because non-party members can vote within the party?

Maybe that why many European countries that use open party list PR don’t have primary systems?

This could be declared unconstitutional an open list PR system.

Why not just base it on permanent regions or communities of geographic interests, and then when redistricting is done, just increase the number of elected politicians in the districts, and decrease it in areas where the districts are small.

An example would be no multimember district shall elected fewer than 3 members, and there would be no upper limit in the largest size multimember district. And then an at large tier can be added to ensure maximum proportionality.

But then opponents would say there is no accountability because of the multimember districts, a bit strange to think that because the present system has no accountability too. Getting the people use to a new system is going to be hard, people fight against change, especially politicians who are use to old habits.

One could make the party primary determine the order of the lists, and then when the general elections come, voters can vote just for the party and also have the choice voting for a candidate.

One can also sell this system on the basis of cost savings, in that it requires no special elections are held if a member dies or resigns for higher or lower office, because the next person on the list succeeds him, and no districts are gerrymandered because of the at large tier ensures proportionality which is the whole state. For that tier whoever is the best looser in the personal vote with in that party gets elected to the state tier.

Could a system be devised where non-party members vote for a party and registered party members vote for there party and the candidates with in it? That would be too complicated to administer.

Also last but not least, the threshold for the Multimember districts can be set at 15% and the at large tier at 5% percent to protect against excessive fragmentation.

I was thinking of the Swedish system of PR for California might be a good fit.

Seed planted by Suaprazzodi — 19 May 2009 @ 18:14

If apparent simplicity is an important criterion of electoral systems then California should retain FPTP.

Russia and Turkey are usually cited as having thresholds that are too high and that create huge numbers of wasted votes. At 7% and 10% their thresholds are lower than the Suaprazzodi proposal. The question that obviously follows is what the rate of wasted votes would be under that proposal and what is the relationship between the threshold numbers and the district magnitudes?

California has a long history of duopoly by the established parties. Why advocate a system designed to entrench that duopoly?

Seed planted by Alan — 20 May 2009 @ 16:36

I think simplicity is a value, and I certainly do not think FPTP is the only system that satisfies the criterion.

And I also think that sometimes one must sacrifice simplicity for other values. If the aims are complex, the system may have to be. I rather like the way the Hungarian system operates, but it most certainly defies the ‘simplicity’ criterion.

Seed planted by MSS — 20 May 2009 @ 17:19

I should have been clearer that (1) apparent simplicity should not be the only criterion and (2) high threshold systems can have devastating effects on fairness. I also, just quietly, think it would be fun to watch someone trying to explain modified Sainte-Laguë to the average US journalist. ‘Why do you divide by 1.4?’

Seed planted by Alan — 20 May 2009 @ 18:27

Suaprazzodi (#61): The problem with PR in the US, how do we accommodate it with the primary system.

That is indeed a problem, but not the most fundamental one. We usually try to fudge it when arguing the merits of PR to the public (well, to whoever will listen). Advocating reform for local elections where these are nominally non-partisan has the advantage of sweeping the problem of party nominations under the carpet. For a while.

U.S.-style primary elections add a lot of confusion to many discussions of political reform. This is not limited to discussions of PR. Many voters don’t understand what a primary actually is. This is compounded by the existence of closed, open and several flavors of in-between primaries. Currently, California’s political parties have the option of allowing non-partisan (so called “decline to state”) voters to participate in their primaries, but they are not required to do so. The confusion is also compounded by calling the first round of a non-partisan top-two runoff a “primary”.

Thirteen months from now California will vote on a ballot measure that would (1) eliminate the partisan primary (except for Presidential elections), and (2) replace plurality with top two runoff. If I understand the measure correctly (can someone help me out here?), political parties would be able to choose whether all candidates registered with a party have party labels on the ballot, or none of them, but would not be able to select one candidate in each race to be its nominee, and would have to choose between these options globally rather than contest by contest. This formula is obviously intended to get around the Supreme Court’s decision in California Democratic Party v. Jones.

As a member of a small (I refuse to say “minor”) party, I should note that the June 2010 ballot measure also makes it extremely hard for small parties to survive by drastically reducing ballot access for their candidates.

The proponents of this referendum are calling it an “open primary” and few in the media or punditocracy are willing to call them on their misuse of terminology. We should brace ourselves for a very muddled public discussion.

Suaprazzodi asks, Would open party list be declared unconstitutional if used in California because non-party members can vote within the party?

No, because when we talk about PR, we are talking about general elections rather than primaries. In a general election, any voter can vote for any candidate.

Seed planted by Bob Richard — 20 May 2009 @ 21:17

Has anyone done any comparative work on California and Switzerland? The Swiss seem to vote directly on as many policy issues as Californians but the outcomes are a tad more sober and realistic. And their budgets make sense.

Seed planted by Alan — 21 May 2009 @ 03:06

The open primary is basically eliminating parties, and establishing a two round system. I think it would have been better if it was like the French two round system for lower house elections, where only candidates getting more than 15% of the vote proceed on to the next round or if no candidate is above 15% then the top 5 candidates move on to the second round. Is that how they do it in France?

Albeit no candidate is going to win a majority in the second round. Better yet, it might be a good idea to require a candidate get 40% of the vote to win outright, or move to IRV voting, Australian style preferential voting. Should the preferences be optional, limited to 1, 2, 3, or compulsory till they exhaust?

What about write in candidates? How many spaces should voters have to write in a candidate and rank them?

Sorry about asking all these questions, I really want electoral reform to be done right and sold well.

I think eliminating the primary system is a bad idea. At least with it, party members get some sort of say what happens with in their party, and it is up to party to decide to allow non-party members to vote with in their parties. Even though I am not a member of any party.

At least with the government regulating it, at least the election procedures are transparent within means and reason. It’s better than the local politburo rubbing stamping who is going to be their party candidate. The progressive movement did one thing right and that is primary elections, moving away from smoke filled rooms, and letting the people within the party decide by secret ballot who is going to be there party candidate.

It could have problems because the party could become lazy and not trying to appeal to centrist voters like the Republican Party in California which is not providing an alternative letting the Democrats win for over 40 years in the state legislature, with a brief internship in 1996-1998.

I mean party activists are more extreme than the electorate is at large especially in California. The Democrat Party is dominate the the San Francisco Bay area, and Republicans in the Central Valley and Rural areas.

In the U.S South, the only competitive election was the Democrat primary in the 50′s and 60′s during Jim Crow. The general election was just a formality. They are problems with primaries. It’s not a perfect system, no system is.

Suaprazzodi asks, Would open party list be declared unconstitutional if used in California because non-party members can vote within the party?

‘No, because when we talk about PR, we are talking about general elections rather than primaries. In a general election, any voter can vote for any candidate.’

That may be true, but I disagree, would we eliminated the primary for the Legislature, and in the general election, people vote for a party and candidates with in that party despite not being members of that party.

Voters are going to be confused, and think that the primary and general election are being consolidated in one round. I’m trying to think in the mind of people that know nothing about PR, and finding out about it for the first time. People that are oppose to change are going to use fear tactics to get their way across.

I think MMP using FPTP for Single Member Districts with regional open party lists with a 5% for each region would be the best and most sellable system to Californians. And this is a two vote system as well. It wouldn’t require fancy new expensive electoral system equipment like IRV and STV does, but then it can be used at the local level.

Different levels of government can use different electoral systems, and nothing is wrong with it. Of course, Scotland did, and it was a mess. I think voter education is key if any new system passes, and voters need to be prepared for it.

Seed planted by Suaprazzodi — 21 May 2009 @ 03:07

Scotland tried mixing different systems at different levels. They had an extensive public education campaign and an electorate that had several experiences with MMP and long experience with STV for school boards. They still had a mess. Advocating one electoral reform is difficult. Two would be impossible.

Moreover the technical difficulties of mechanising electoral systems are grossly exaggerated and I suggest that the exaggerations flow largely from the PR of outsourced election companies insisting that it cannot be done. Several Australian jurisdictions now offer electronic voting. None have found it necessary to limit preferences or restrict voter choice to accommodate mechanical difficulties, although obviously the number of positions filled at each election can be far greater in the US. It is perhaps a non-trivial fact that Australian elections do not use outsourced hardware or software. If a machine cannot adequately implement an electoral system the solution is a new machine, not a different electoral system.

The best answer to the number of preferences issue is to require a single preference. That rule has always been followed in Ireland and a number of electoral reformers expressed bewilderment that Australia, by contrast, used to require a preference for every candidate. Most Australian jurisdiction no longer do that, although it s a bit random how many preferences are required. The labor party tends to say one and the coalition parties, which need to exchange preferences, tend to say enough to ensure we get the preference flow. The minor parties tend to have the same line as the coalition. Best practice STV is probably that defined by the ACT Proportional Representation (Hare-Clark) Entrenchment Act 1994.

Australia once had an electoral system that was worse, in almost every aspect, than the US system. That changed between 1968 and 2005, not because academics and intellectuals defined the ‘perfect’ electoral system, but because one party adopted a policy of electoral reform and stuck to it.

Seed planted by Alan — 21 May 2009 @ 03:54

Suaprazzodi (#67): The progressive movement did one thing right and that is primary elections, moving away from smoke filled rooms …

According to Alan Ware in The American Direct Primary, this is at least partly a myth. In fact, the major parties themselves needed to institutionalize the candidate selection process and were pushing government-run nominating elections even before the Progressives took up the issue. I don’t know whether anyone — party bosses or reformers — understood at the time that they were transforming political parties into quasi-public entities. Perhaps Ware will address that question later in the book, which I have just started.

Seed planted by Bob Richard — 21 May 2009 @ 10:35

Alan said, “Australia once had an electoral system that was worse, in almost every aspect, than the US system. That changed between 1968 and 2005, not because academics and intellectuals defined the ‘perfect’ electoral system, but because one party adopted a policy of electoral reform and stuck to it.”

How was the Australian system the worse from 1968 to 2005? Which was the party that adopt a policy of electoral reform?

Do you mean the Australian state of Victoria?

Seed planted by Suaprazzodi — 23 May 2009 @ 00:15

@ Suaprazzodi

The administration of elections was deeply politicised at all levels of government. Redistribution (redistricting) was done directly by parliament at all levels. Malapportionment and gerrymandering were both common. Reversed eelctions were quite common, so much so that the result fo one Queensland general election, where the incumbenet govenrment had lost votes and gained seats, was described as ‘the government was returned with a reduced minority’.

Campaign financing grossly favoured the coalition parties. Public money was commonly spent by governments on their own re-election, although that problem has recently resurfaced in the form of pubic information campaigns that serve the political purposes of the government of the day.

Most state upper houses were grossly biased in terms of rural over-representation. Many lower houses were also biased in terms of rural over-representation. The legislative council of New South Wales was not even elected by the people. The territories had limited representation in the house of representatives and no representation in the senate.

The problem was nation-wide, but worst in the states of Queensland and Western Australia. Most of the reforms were introduced by the labor party, usually, although not always, over coalition objections.

Seed planted by Alan — 23 May 2009 @ 02:17

MMS at #19: Tom, at #11, perfectly outlines my concerns about using AV (IRV) with MMP.

Tom at #11: … one difficulty with using AV-IRO for the district seats is, that it clouds which voters are “un[der]-represented” and so “deserve” topping up with seats from their preferred party’s list.

I think I understand Tom’s point, even if I don’t agree that the Green voters in his example are getting double representation. I can also see that this is related to a statement in MSS’s 1999 article on the Jenkins Commission: Because of the use of the alternative vote in the single-seat constituencies, the system will tend to start off less proportional before PR compensation is performed than if constituencies were decided by plurality.

But I’m afraid I don’t understand why this is true. What is the argument? Is there empirical data or a Monte Carlo simulation to back it up?

This question is related to my response to C.T. at #25. Is he right, and should PR advocates oppose the introduction of AV/IRV for legislative bodies (I’m not talking about executive offices here) that are already elected from single member districts?

Thanks!

Seed planted by Bob Richard — 01 June 2009 @ 21:49

I see MMP with AV constituencies as being particularly bad when there are overhangs.

Suppose there are three parties, Labour, Conservative and Green; and Labour is set to win overhang seats. In a normal MMP system, if you vote Green or Conservative you may be the deciding vote in tipping a compensatory seat their way. Or if you vote for Labour, you may be the deciding vote in your constituency, contributing to the overhang–but since Labour will receive no top-up seats, your vote doesn’t change the allocation of those. Note that either way, your vote changes the party balance by at most one seat.

With MMP/AV, you could have your cake and eat it too. If your preference is Green/Labour/Conservative, your vote could be decisive in both adding to Labour’s overhang, and winning the Greens a compensatory seat. This doesn’t seem very fair!

Seed planted by Vasi — 01 June 2009 @ 23:51

Others can and will explain the dual voting problem in AV/MMP better than I can. My problem is more radical. We know how MMP was developed and how it was designed to address specific issues in the Weimar constitution and electoral system. AV Plus has a similar history.

Jenkins took STV, a proportional system, divided it up into single member districts to try not to frighten the horses, and then superimposed a compensatory tier to try and undo the anti-proportional effects of dividing his system up into SMDs. Sadly the Blair government was so terrified of the whole thing they promptly bolted into the far distance, dumping an election promise along the way.

AV Plus makes sense only if you believe the constituency link is the most important feature of an electoral system. Even then something like the circuits idea would be a better solution than Jenkins’ bungled mishmash.

Seed planted by Alan — 02 June 2009 @ 08:40

The comments in #73 and #74 are very helpful to me with respect to the effects of AV when used as part of MMP. My question — and it is an urgent one right now — is different. It has strictly to do with two single-tier, majoritarian systems, AV and simple plurality.

Let me ask it this way. Has anybody ever analyzed the returns from a series of Australian lower house elections to compare the actual results with the results had the counting rule been simple plurality? Is there a meaningful difference in the relationships between seats and votes?

Seed planted by Bob Richard — 02 June 2009 @ 13:51

Thanks, Bob, for rehabilitating my Representation article on the Jenkins proposal in the UK for “AV plus.”

I recently learned (via Tom Lundberg) that the AV+ is being dusted off for a possible referendum. I hope to address this and the related questions raised here in the near future. (But not till some time after the spring quarter ends.)

Of course there are no data on how a form of MMP with AV (IRV) works, because there are no such systems. But I think the objections I raised years ago apply, as do objections raised here recently. (More later–promise.)

Should PR advocates oppose the introduction of AV for legislative bodies? I would tend to think so, unless they think PR is simply not in the cards anyway.

Seed planted by MSS — 02 June 2009 @ 14:21

The majority opinion among PR activists has three parts. (1) PR is, indeed, not in the cards right now. (2) Experience using AV will help open the minds of voters and (maybe) politicians to broader electoral reform, i.e. PR. (3) Even if we’re wrong about (2), AV is a fairer system than FPTP and a therefore a positive step.

There is a minority point of view which says, in effect, that (2) is a high stakes bet that is probably not worth making. Or, alternatively, that the need for PR is so urgent that we can’t rely on long-range strategies. I’ve never heard anyone in the minority really question (3), though.

The statement I quoted from the Representation article directly questions (3). If it’s true that AV has the effect of increasing the size of the majority bonus relative to plurality, then I need to rethink my position, which has been mostly in line with the majority of activists.

Seed planted by Bob Richard — 02 June 2009 @ 15:24

Is there another example of AV being used than in Australia? For lower house elections, Australia has the most extreme two party system out side the US (and yes, I’m ignoring the Nationals’ claim to be a separate party). I think no minor party has ever won any House of Representatives election, excluding a by-election where one of the two major parties stood down, in Australia. Even in the US the Socialists were able to elect a few Congressmen.

I’ve seen the argument that most Australian voters will do anything to avoid a situation where a minor party winds up as a “finalist” in the AV count. I don’t know if this is a result of AV or something particular to Australian culture. But in FPTP systems minor parties do come up the middle, and there is minor party representation in all FPTP parliaments, except the US House.

Even in France, which uses runoffs, true minor parties (parties that have not made some sort of electoral deal with the Socialists and the conservative party of the day) really struggle to win seats.

I also used to favor AV over FPTP but have been reconsidering, since I think that in parliaments which have minor party representation their representatives have played a beneficial role, if only in holding the two major parties account.

Seed planted by Ed — 02 June 2009 @ 18:09

Ed, don’t ignore the Nationals’ claims to be a separate party. Their sole lower house MP in South Australia sits in the Labor Cabinet, and they have also allied with Labor in Victoria. In Queensland, the Liberals have withi my memory (1983-1992) seriously considered allying with Labor against the dominant Nationals. In Western Australia, coalition relations have been rocky at times. Only in NSW and federal level is the coalition marriage generally smooth. And federal Nationals like Barnaby Joyce and Bob Katter Jr have crossed the floor to vote against the Liberals (this being a much bigger headline-grabber in Aust than in the USA, or even the UK).

Seconly, while minor parties do have trouble winning single-member seats (other than the special cases of [1] the Tasmanian Upper House, where the Liberals don’t endorse candidates so all the conservative MLCs are “Independents”, and [2] One Nation in Qld, especially their shock 1998 win of 11 of 89 MLAs with 22% of the votes), independents can and do win these frequently. There were 6 out of 148 at one point in the 1996-98 House of Reps. Admittedly, the majority of Inds are former party MPs who lose endorsement; these often last only one further election (eg, Graeme Campbell, Pauline Hanson, Paul Filing, Alan Rocher… Ronan Lee didn’t even survive one), although others like Tony Windsor and Katter do get re-elected on the Ind label. And there have been a few who got elected solely as Independents: John Hatton, Ted Mack, Liz Cunningham and Peter Andren.

It is true though that if the “outsider” is framed as a threat to the democratic system, all three major parties will be expected to combine to put them last in their preference order – One Nation since 1997 being Exhibit A – in some cases, the Liberals or Nationals refused, and this hurt them among swinging voters. Illustrates nicely Domenico Fisischella’s thesis about runoffs being used to keep out “anti-system” parties. The British Big Three have no similar way to gang up on the BNP.

Apart from Aust, AV is used for presidential and by-elections in Ireland, and has been used in the past in some Canadian provinces.

Seed planted by Tom Round — 02 June 2009 @ 19:09

As far as I know Tom is right that the Australian House is the only legislative body with enough AV elections to be useful for my purpose. On the other hand, while the presence of more than two significant parties is one reason for the discrepancy between votes and seats, it is not the only reason. Precisely because of the (relative) similarity of the party systems of Australia and the U.S., the data would be helpful.

Before I launch into the project of putting Australian election results into a spreadsheet and figuring out what the seat allocations would have been under plurality, I’d like to be sure that this hasn’t already been published somewhere.

Seed planted by Bob Richard — 02 June 2009 @ 20:18

Bob, you’d need to include the legislative assemblies of NSW, Victoria, etc etc NSW in particular has extremely high numbers of independents, including the speaker of the assembly himself. You also have to remember that AV gives minor parties considerable power in preference bargaining, even if they d not win seats. I would always advocate STV over AV for legislative elections, but AV is not completely without advantages.

Seed planted by Alan — 02 June 2009 @ 21:33

AV has been used in Papua New Guinea, both before independence and (in a limited fashion) again since 2002. (The limitation is that voters may give up to three preferences, similar to San Francisco, right?)

Aside from recent US reform cases, I can’t name another non-Australian case (and, before it was independent, PNG was an Australian jurisdiction).

As for Bob’s three points:

The majority opinion among PR activists has three parts. (1) PR is, indeed, not in the cards right now. (2) Experience using AV will help open the minds of voters and (maybe) politicians to broader electoral reform, i.e. PR. (3) Even if we’re wrong about (2), AV is a fairer system than FPTP and a therefore a positive step.

I agree with 1 and 3, but am very skeptical about 2 (which is, in any case, trumped strategically by 3, though if my skepticism is justified, then the caveat in 1 falls by the wayside).

Seed planted by MSS — 03 June 2009 @ 22:13

MSS, if AV results in legislatures that are even less like the electorate than plurality (#72), then how can it be fairer than plurality (#82)? Fairer within each district separately, yes, since it implements majority rule. But fairer overall?

Seed planted by Bob Richard — 03 June 2009 @ 22:42

AV (and two round majority) give smaller parties bargaining power before the election (or between both rounds) even if they cannot win a seat.

Seed planted by Bancki — 04 June 2009 @ 04:23

@84 The capacity of a party to direct its preferences also becomes important. The Democratic Labor Party, a right-wing (on most but not all issues) split from the Australian Labor Party had famously tight preferences. DLP voters almost invariably followed their party’s ticket which almost always favoured the Coalition and kept the ALP out of federal power for a generation. There is a nasty underside to Australian politics where it is not unknown for parties to mislead their voters and try to persuade them that they must vote the party line to cast an effective vote.

@82 AV was a crucial reform in PNG where the electorate was extraordinarily fragmented along clan lines and FPTP was electing MPs who had as little as 10% of the electors in a district. Limiting the number of preferences to 3 was a compromise directed at an electorate with a high degree of illiteracy. Arguably the San Francisco 3 preference rule also reflects an electorate with a high degree of illiteracy about electoral system performance and design.

Seed planted by Alan — 04 June 2009 @ 09:20

#85: … the San Francisco 3 preference rule also reflects an electorate with a high degree of illiteracy about electoral system performance and design.

The San Francisco charter requires that voters be able to rank all the candidates (with provision for write-ins) unless the voting equipment cannot support that. The equipment currently certified for use in California cannot handle more than three preferences.

The state of the equipment industry in the United States does reflect a degree of both of both public indifference and opposition to reform on the part of election officials. It also reflects the power of a very small number of individuals who control four or five equipment vendors. The solution is public ownership of the design specifications of both hardware and software. Not coincidentally, that would also go a long way toward solving our problems of vote counting fraud, both actual and perceived.

Seed planted by Bob Richard — 04 June 2009 @ 12:39

Bob (#83), who said that AV was less ‘like the electorate’ than plurality? (Well, maybe Alan, at #85, when referring to its “nasty underside.”)

My main issue with AV concerns how it would work with MMP, not on its own (that nasty underside notwithstanding).

Seed planted by MSS — 04 June 2009 @ 13:01

MSS (#87): who said that AV was less ‘like the electorate’ than plurality?

Unless I totally misunderstand (very possible), you did, in the 1999 Representation article:

Because of the use of the alternative vote in the single-seat constituencies, the system will tend to start off less proportional before PR compensation is performed than if constituencies were decided by plurality.

Seed planted by Bob Richard — 04 June 2009 @ 13:17

Nobody asks the question will voters be required to rank all preferences till they exhaust or will preference ranking be optional by just marking one, or will just three preferences be required in order to make a vote valid?

What about voters that decide to skip numbers? I rank 1, 2, 5, 7, 9

What about voters that decided to duplicate numbers when voting in AV? I rank 1, 2, 3, 4, 4, 5, 5, 6, 7, 7

Would the above be consider valid in mandatory preferential voting?

Seed planted by Suaprazzodi — 04 June 2009 @ 14:10

In Australia, no. Before 1983, and since 1998, these would wholly invalidate your ballot. For the 1983-98 period, there was a savings provision whereby your ballot was valid up until the omitted or repeated number, then it exhausted. It was meant as a way of salvaging accidentally-spoiled ballots, but Albert Langer, an anarchist activist, tried to publicise this loophole to get people to deliberately vote only for left-wing minor parties (which would help them clear the 4% threshold to receive public funding) without needing to put Labor second-last, before the conservatives. The Labor govt enacted a ban on encouraging people to deliberately vote in this irregular-but-valid way. The Liberals reverted to the old rule – no restriction on political speech in this regard, but your vote goes straight into the dustbin.

Langer had a dodgy interpretation of the Cwlth Electoral Act whereby, if no candidate has an “absolute majority”, the election is void. It appears that Parliament meant “absolute majority of the unexhausted votes remaining in the count”, ie, that this applied only to a 50-50 (or 47-47 or 43-43) tie. Langer however argued that this meant 50% of all valid first-preference votes cast. Like other attempts by anarchists to import ideas from Russia, this did not take root successfully.

Seed planted by Tom Round — 04 June 2009 @ 18:42

@90 It’s not really a very important issue. Both AV and STV work just fine when the voter is directed to fill out only a single preference. The double vote idea merely has the effect of causing a vote to exhaust in the same way that attempting to vote for 2 candidates in an FPTP election would.

@86 It’s probably a nontrivial fact that the only privatised electoral system in the world is also one of the worst-performing.

Seed planted by Alan — 04 June 2009 @ 19:18

Bob (88), the quoted passage is within the context of a discussion of compensation based on first preferences (for parties under 2-vote MMP), whereas seats under AV are allocated to the final distribution of preferences (for candidates in the SSDs).

As I noted some days ago, I will address this in more detail when I get a chance (i.e. not before mid-June).

This is the first F&V seedbed to have 90+ sprouts! Will we reach 100?

Seed planted by MSS — 05 June 2009 @ 13:38

At #91, Alan says:

It’s probably a nontrivial fact that the only privatised electoral system in the world is also one of the worst-performing.

Indeed, and within the set of allegedly advanced countries, we could say the same thing regarding systems of health-care coverage.

Seed planted by MSS — 05 June 2009 @ 13:39

Only tax-farming, not vote-farming, had been invented in the time of Adam Smith, but his arguments against one run equally well against the other:

Even a bad sovereign feels more compassion for his people than can ever be expected from the farmers of his revenue. He knows that the permanent grandeur of his family depends upon the prosperity of his people, and he will never knowingly ruin that prosperity for the sake of any momentary interest of his own. It is otherwise with the farmers of his revenue, whose grandeur may frequently be the effect of the ruin, and not of the prosperity of his people.

An Inquiry into the Nature And Causes of the Wealth of Nations, Book Five, Chapter II

Seed planted by Alan — 05 June 2009 @ 17:19

The Sydney Morning Herald‘s view. Paul Sheehan is a maverick conservative of sorts:

‘… Despite having immense wealth from a variety of revenue streams, California now lurches from political crisis to political crisis. Its biggest problem is an excess of democracy, expressed in a number of ways:

* Micro-management by the electorate. The state constitution allows an unusually diverse array of direct participation by the electorate through voter initiative, referendum, recall and ratification. Over time, the result has become a web of conflicting legal requirements imposed on the state government. Most famously, in 1978, California passed Proposition 13, which placed a cap on property taxes.

* That proposition has had unintended consequences. “Proposition 13 opened up loopholes for corporate land-owners and shifted the tax burden to individual home owners, while dramatically reducing California’s tax base,” says Phil Ting, the chief assessor for the City of San Francisco. The state now depends excessively on income tax.

* The California constitution requires a two-thirds majority vote to pass the state budget. This measure, in place since 1933, gives inordinate blocking power to the opposition, and inordinate power to minority interests.

* “Term limits are killing us,” Dr Patrick Mason, the head of a non-partisan public policy think tank, the California Foundation on the Environment and the Economy, told me in San Francisco. Members of the California State Assembly are restricted to three two-year terms, and state Senators restricted to two four-year terms. “Our politicians used to learn their jobs, develop relationships, a common culture, and a social glue,” Dr Mason said. “Now they are more ideological. Term limits have created scorched earth politics. Anyone who makes compromises is killed.”

* Gerrymandering. Both the Democrats and Republicans have been able to manipulate the electoral system to create safe seats, leaving only a narrow band of genuinely marginal electorates. This, like term limits, has led to ideological rigidities because most real contests are the party primaries, dominated by party faithful, not the state elections, where moderates have a better chance.

* Public sector union power. In a stark lesson for Australia, the cost of California’s public sector is unsustainable, but the sector with the most interest in preventing reform is the group with the most political power. The sector which benefits most from California’s big-spending status quo is the public sector union movement. With the nation’s highest-paid teachers, prison warders and civil servants, these unions recycle tens of millions of dollars into political contributions and dominate the Democratic Party.

Last week the Governator, having been elected in 2003 and re-elected in 2006 to break the cycle of political gridlock, brinkmanship and intransigence – and been thwarted – implored the state to hold a constitutional convention to produce structural reform. The need is self-evident. California is proof that there is such a thing as too much democracy. Allow governments to govern, because if they fail they can be terminated.’

- Paul Sheehan, “No salvation for the Governator,” Sydney Morning Herald (2 June 2009), p 11.

Seed planted by Tom Round — 07 June 2009 @ 20:14

I seriously doubt the consequences of Proposition 13 were unintended. I do not doubt there have been consequences the promoters somehow forgot to disclose to the electorate.

Seed planted by Alan — 07 June 2009 @ 21:17

I’d call virtually all of the provisions mentioned by Mr Sheehan restrictions on, rather than excess of, democracy.

(I also do not think the consequences of Prop. 13 were at all “unintended” by the initiators.)

Seed planted by MSS — 07 June 2009 @ 21:19

We’re within striking distance of 100 comments here, so why not?

MSS (#92): … the quoted passage [on AV+] is within the context of a discussion of compensation based on first preferences (for parties under 2-vote MMP), whereas seats under AV are allocated to the final distribution of preferences (for candidates in the SSDs).

I’m afraid I didn’t understand this sentence when I first read it on June 5 and I still don’t understand it now. It sounds like you are saying that AV+ would base the allocation of list seats on who the SMD winners would have been had only first preferences been considered, rather than on the actual (AV) winners. But I can’t find anything to that effect in the Jenkins Commission report, and it sounds strange anyway.

What am I missing?

I did (finally!) find this extended discussion of AV and proportionality by Lewis Baston. He seems to conclude that the answer to my original question (#72 above) depends on circumstances and whether your main focus is small parties or the bonus given to the largest party. He also makes clear why my plan to look at Australian election results (#75 and #80) won’t work: first preferences in AV elections do not reflect the tactical voting that would shape plurality elections.

Seed planted by Bob Richard — 12 July 2009 @ 15:36

I think there would prolly be a consensus on this thread for a single legislative chamber elected by proportional representation in regional super-districts. I suggest an order of preferences as follows:

1. a single chamber elected by proportional representation

2. a senate and assembly each elected by proportional representation

3. a senate elected by STV and an assembly elected by AV

4. the current mess.

There seems to be a consensus that the governor should be elected by AV and that the governor should either be the sole elected member of the executive or have a joint ticket vice-governor. Personally I’d prefer special elections for governor and only temporary succession provisions.

As always I’d argue if you ask the voters to number preferences for one election you have to give them that right for every election.

Recall and initiative should have thresholds and the bill of rights should be entrenched.

The 2/3 rule on budget votes should go.

Local government should be consolidated on a regional basis on the same boundaries as the super-districts used in elections.

Seed planted by Alan — 13 July 2009 @ 03:20

May I suggest a new thread on reforming the New York state government? There is a lot happening in New York state politics that is interesting from a political science perspective, such as the recent discovery that the Governor can temporarily appoint a new Lieutenant Governor to fill a vacancy in that office. No one realized before that this power existed.

Seed planted by Ed — 13 July 2009 @ 11:44

Hooray, we made it to 100!

Seed planted by MSS — 13 July 2009 @ 14:00

And neatly synchronised comes:

Ilya Somin, “Should California Be Broken Up?”, The Volokh Conspiracy (14 July 2009).

Seed planted by Tom Round — 14 July 2009 @ 03:54

The idea of breaking up California has been around about as long as I can remember. And it is not an idea I am willing even to entertain. Don’t mess with California! We are messed up enough already.

Seed planted by MSS — 14 July 2009 @ 12:34

On the breaking up of California, what is the function of states in the US, other than that they are there? What do state governments do or what are the supposed to do that is not already being done by the federal government or more local levels of government?

I’m genuinely curious. For some reason its considered appropriate for Rhode Island to have its own state government, but not Ventura County, which has a somewhat similar size, population, and population density. So what is the theory whereby its terrible idea to deprive Rhode Islanders of statehood, but also a terrible idea to break up California and make Ventura a state?

Seed planted by Ed — 14 July 2009 @ 17:52

Fantastic thread. As attractive as breaking up the states into geographies that would undo the malapportionment of the Senate is, you really can’t mess with The Bear Republic. Regarding the evolution of party systems under MMP, anyone have a quick link to changes in the effective number of parties in NZ, Italy, etc.? One argument that needs to be made in CA is how multi-party politics would likely strengthen the capacity for centrist policymaking, and hypothetical examples of likely party system evolution could strengthen that. I’m looking at patterns of party affiliated voting on initiatives by geography for clues, other ideas would be welcome. Also, wouldn’t the public be more likely to buy into a larger legislature by way of a downsizing argument? Unicameral and eliminate county boards of supervisors in favor of regional MMP districts? Never underestimate the desirability of doing away with a level of govt in CA, we have plenty to spare.

Seed planted by Mike Latner — 15 July 2009 @ 15:14

Good to “see” you, Mike. Remember, Italy did not ever have MMP. But that did not answer your question. No, I am aware of no quick place to look for that information. For Italy, of course, one encounters the vexing question of which are the “parties” and which are blocs or other animals.

Speaking of animals, indeed, those who mess with the Bear Republic will be mauled.

Seed planted by MSS — 15 July 2009 @ 15:41

Understood that Italy is less than compensatory, but I think of Italy because it raises an interesting question about the allowance of alliance lists if PR were to come to CA in a mixed system. Seems like it might be part of the transition in a comparatively weak and financially privatized party system, no?

Seed planted by Mike Latner — 15 July 2009 @ 16:25

Thanks, Mike, for bringing the question on Italy back to California so quickly!

It is an interesting question, but I don’t see why we would expect alliances under any MMP system, such as the one that the New America Foundation has proposed (linked way up there at the top).

Alliances formed in Italy because (1) the electoral system created great incentives for them, being fundamentally MMM rather than MMP, and (2) for the path-dependent reason that the system was being introduced into an existing multiparty system (that was already becoming ever-more fragmented just before the reform).

Obviously, neither condition would apply to California. In that regard, clearly the NZ experience would be more relevant, although even NZ had more than two parties of some significance in votes well before the electoral reform (although less so than either Scotland or Wales, the other two main FPTP jurisdictions to adopt MMP).

In California, third parties would be so thrilled to be finally able to win seats, that I suspect alliance-formation (with the bigger parties) would be pretty far down their lists of priorities!

Seed planted by MSS — 15 July 2009 @ 16:55

Wouldn’t California have just minority governments lead by the Democratic party if MMP were used? It would be a party system similar to Sweden. The Green Party, and a Progressive Party (In Sweden, it would be the Left Party) would support Democratic lead minority governments.

I would prefer STV compared to MMP, and I would only accept MMP, if the party lists were open.

Seed planted by Suaprazzodi — 15 July 2009 @ 22:39

A closed party list would probably not be acceptable to lay over such a candidate-centered system as we have now. Wouldn’t STV be more supported along with MMP? First choices used to maintain proportionality, and we get to do away with primaries? Or does that give too much potential influence to minor party voters (I think that was suggested above)? Either way closed party lists might be “too European” for many Californians.

Seed planted by Mike Latner — 16 July 2009 @ 14:18

Mike (at #110!!), others who comment here could say better than I, but I think there is a wide consensus that closed lists would be a tough sell. (I do not necessarily agree that the common criticisms apply when the lists are part of an MMP system, but that is probably neither here nor there.)

I remind readers of one the “preserved fruit” that is “MMP and dual candidacy,” in which I suggested an open-list model of MMP that also gets around the common revulsion against “losers” (in the nominal tier) becoming winners due to the list: have voters outside the single-seat district in which a candidate is nominated, be responsible for ranking giving the candidate preference votes on the list.

At the time that I “invented” this system, I thought I had really invented it. But I recently learned that it is more or less what Bavaria uses.

Seed planted by MSS — 16 July 2009 @ 15:54

The “common revulsion” is not just against district losers being given list seats. The opponents of PR in Ontario developed a finely tuned attack on closed lists as a mechanism by which legislators are selected by the political parties rather than by the voters. (They omitted the part where the voters get to decide how many Tories, Liberals, NDP-ers, etc.) This was an extremely effective meme of the No on MMP campaign.

Seed planted by Bob Richard — 16 July 2009 @ 16:14

Bob, yes. I could have been clearer. Closed lists, whether or not in the context of a mixed-member system, are a tough sell for precisely the reason you note. That is a generic problem for closed lists. The district-losers problem is an additional one, so to speak.

I suspect that people generally (I mean, outside of the privileged few who can spend many hours a day thinking about electoral systems) over-estimate the extent to which any democratic electoral system promotes generalized competition. Whatever the system, a very high percentage of seats is going to be “safe” either because the district is not inter-party competitive, the legislator has a high closed-list rank, the legislator is popular and hence always gets a huge haul of the preference votes, etc.

Against all those safe seats in the typical FPTP system, closed lists do not actually look so bad.

Seed planted by MSS — 16 July 2009 @ 16:53

Following up on the references to the Jenkins report (#72 – 77, #88, #92, #98), Wilf Day has taken another look at AV Plus and decided, “As a compromise, it’s not as bad as I thought.”

MSS, did you ever revisit the question of using AV in the district half of MMP, as you once promised?

Seed planted by Bob Richard — 21 June 2010 @ 17:06

The other advantage of Wilf’s model is that the flow of preferences among parties is also a guide in coalition-making.

Seed planted by Alan — 22 June 2010 @ 01:37

I’ve also started to look more favorably on AV plus, but mainly because of optics -it just seems easier to explain the top up representatives’ functions to voters and opinion makers used to single member districts, than in full blown MMP. Theoretically the system is a mess.

Germany is surrounded by countries (with the exception of France) that use the party list system. Adopting MMP when the main alternative is party list will have a different feel from adopting MMP when the “natural” system is SMP. In the one case you are carving out constituency MPs from a group that would be elected off of a party list anyway, in the other you are taking constituency MPs and having them appointed by the party bosses.

I would prefer a system, which I think Hungary uses, where if a party elects an MP from a single member district, the votes needed to elect him (ie the votes won by his closest opponent) are deducted from the party’s total for party list purposes. The “wasted” votes, votes cast in failed attempts to elect representatives from constuencies or the surplus votes of the winners, are then used to elect a handful of at large representatives. The effect should be similar to MMP, except the at large representatives are more clearly and directly elected.

Of course, if the politicians become as unaccountable as late 1980s and early 1990s New Zealand politicians, the voters will be ready to switch to just about any alternative electoral system, and we may be reaching that point in many places in the U.S.

Seed planted by Ed — 22 June 2010 @ 15:05

As Ed says, U.S. voters may well be approaching the same level of frustration. But I think another key ingredient is still missing — namely, symptoms that are clearly and unmistakably caused by the voting rule rather than by gerrymandering, closed primaries, incumbency advantage, etc., etc. In the absence of a successful third party leading to clearly wrong results in legislatures, the main symptom we have is polarization (by which I mean too few centrist representatives for the number of centrist voters). It’s too hard for people to see that as a consequence of winner-take-all when there are so many other superficially plausible explanations. And when the other explanations are so obsessively harped upon by journalists and pundits.

Seed planted by Bob Richard — 22 June 2010 @ 18:04

Ed, you were probably thinking of the former Italian mixed-member system when you referred to deducting from party-list votes those votes that elected a district member. In Hungary, the votes of losers in the districts (where two-round majority-plurality is used) are added to their parties’ list votes.

Both the former Italian and the current Hungarian mixed-member systems lead to quite strong majoritarian results.

Seed planted by MSS — 22 June 2010 @ 18:12

“Closed lists, whether or not in the context of a mixed-member system, are a tough sell for precisely the reason you note. That is a generic problem for closed lists. The district-losers problem is an additional one, so to speak.”

In Japan, don’t they call these legislators that loose in a district seat, but win in a close party list seat, zombies.

These seats in the MMP system should be called safety seats for the politicians. If voters who know very little about electoral systems. They are going to be suspicious of close party list seats moving from a FPTP to MMP system.

It’s a pity that British Columbians rejected STV a second time. It’s a good thing the BC Liberals were indifferent towards PR. If they were oppose to it along with the NDP, I think the voters would have went for it. Similar to voters going for Prop 14 in California, and MMP in NZ, the big parties were oppose to it. The voters voted yes to spite them.

Seed planted by Suaprazzodi — 22 June 2010 @ 19:33

Suaprazzodi might be right about the effect of major party opposition to Prop 14. On the other hand, that opposition was remarkably muted — a few press conferences, mailers in some locations, and was funded by very few dollars. Bruce Cain is quoted somewhere as speculating that the major parties were deliberately lying low because they didn’t want to be too closely identified with opposition to Top Two. I don’t know whether that’s true, or whether, because of a variety of factors, they couldn’t raise much money for this issue.

I also note that this is post #120 and we still mention California from time to time.

Seed planted by Bob Richard — 22 June 2010 @ 23:19

TrackBacks

To graft a scion to this planting, please use the following URL:
http://fruitsandvotes.com/blog/wp-trackback.php?p=2897
(Non-MT bloggers click here to send pings.)

Grafted scions that are not compatible with this planting's stock will die or be pruned out by the Orchardist.

About the comment form

Please note that the name you enter below and the first several words of your comment will appear on the right sidebar of the blog's front page, under "Propagation." New propagators might want to look at the comment policy.

Please do not enter long URLs into the seedbed. Either mark them up using html hyperlinks or convert them to a "tiny URL." Thank you!

120 ideas sprouting »

SeedbedThe soil is ready for planting:

Seedbed
The soil is ready for planting: