Friday, 15 April 2011

The Alternative Vote #4: likely to lead to better or worse governments being elected?

 Charles Roddie is a Postdoctoral Fellow in Economics at Nuffield College, Oxford. In this essay he provides the most honest account I have read of why one might vote NO to AV: to avoid hung parliaments and left wing policies.

Is the Alternative Vote system a good idea?

What makes a voting system a good one?
AV is a proposed voting system for UK national elections, an alternative to the current first past the post (FPTP) system. The history is that Liberal Democrats have had a smaller share of representatives (MPs) than their vote share, and they would do better under other systems than first past the post, which favours the main two parties. Understandably, they feel very strongly about this, so when at the last election they were in the position of kingmaker, a referendum on AV was the main concession they required from the Conservatives to form a coalition government with them.

Now we are naturally getting a lot of arguments from the main parties rationalizing why they support or reject AV. They rhetorically attach to AV terms like fair/unfair, or simple/complex, that sound pleasant or unpleasant to most people. Perhaps if they argued more honestly, for example “vote for AV because it helps us, and we are better than those other parties for these reasons”, supporters of those other parties would take the point and vote against. The result is people think that voting for an electoral change is not about policy and governments, but about the particular aesthetics of voting.

What does a voting system do? It produces governments. So the question is: 
Is AV likely to lead to better or worse governments being elected? 
In reality the AV vote is about exactly the same issues that you vote for in a normal election: policies and governments. But it’s more difficult because we have to try and guess what the effects will be on policies and governments, which is not easy to work out. 
So if you think submitting a vote under AV is complex, or tactics in FPTP are difficult, the vote on whether to vote for AV is much harder. But we’ll try to make some sense of it.

The effects of AV
A voting change produces two effects:

  1. A direct effect: Given existing parties and their positions, how will the voting scheme change which party gets chosen?
  2. An indirect effect: this incorporates how parties adjust to the new system.
(This is a tautology, but it can split up the problem usefully.)

The indirect effect is much harder to gauge, and I don’t know that anyone has much idea of what it will be. Would new parties form and succeed under AV that would not under the FPTP? How will parties adjust their policies and tactics if AV comes in? 
I don’t have any strong view that these effects will change policy in any particular direction. So I’m going to look at the direct effect. The main parties and the published analysis seem to be taking the same line and assuming the direct effect is the important one here.

The direct effect
      1) Who will gain and lose MPs?
In the short term, probably parties won’t adjust to the change immediately. So the direct effect is likely to dominate.
  • To estimate this, some people have done simulations based on survey data. These seem to indicate Labour and the Liberal democrats will gain, and the Conservatives will lose from AV.
    • See
    • See also 
  • This is consonant with the actual positions of parties: Liberal democrats strongly for AV, Labor mostly for, including the leadership, Conservatives solidly against. Presumably, they have paid consultants to analyze the question and we can rely on their supporting the position that is in their self-interest.
So who gains and loses? Liberal democrats gain almost certainly; conservatives lose probably; Labour gains (slightly less) probably.
Overall there is probably a shift to the left with AV.
      2) Majority governments
This is an easier observation, and is a consensus about AV. AV increases the likelihood of a hung parliament, where no party has a majority; so it increases the chance of coalition governments.

Is this good or bad?
  • Minority and coalition governments
These are much less likely than majority governments to have the stability and internal relationships needed to govern effectively. This is a common informed opinion about political systems in the world. But importantly, and this is the key argument for me, the people who stake their money on this question are confident that this is the answer. 

The stock markets move on the assumption that this is true. Stocks fall if a hung parliament becomes likely. This is reported in the media whenever an indecisive outcome is likely. The effect is independent of policy positioning, with markets typically preferring a center-right majority to a center-left majority to a hung parliament. 

As measured by competence of a new government to manage the economy, the lack of a straight majority is bad (probabilistically speaking). The risk that ideologically very different groups will not be able to work together is a significant problem here.
  • A move to the left
It’s harder to persuade people on a big question like this if they don’t already agree, but I believe this is the wrong direction to move and here are my reasons. Let me concentrate on the economics here. Left wing economic policy:
    • Reduces incentives to work and invest. This is the consequence of the large taxation and redistribution system that is in place. It should be reduced in scope and not increased.
      • The aim is to reduce relative poverty, but it is a burden on the activity of the private sector, including the many aims that people set for themselves, and also reduces total economic output. It also reduces the link between actions and their consequences, and leads to a dehumanizing dependence on the state.
    • Overregulates business. Labour markets are particularly overregulated, with restrictions on firing, discrimination law, health and safety regulation, etc.. Firms will be risk-averse in hiring and unable to contract with completely flexible wages, causing frictions between supply and demand, i.e. unemployment. Some regulation is necessary for trust but current regulation goes far beyond that.
    • Distrusts market solutions, believing instead in direct control of certain services. But markets can price goods and allocate them efficiently. Government is limited in knowledge and constrained by ideological and institutional factors.
      • Suppose a government were in charge of the automobile industry and distribution of cars. Then it would ensure that everyone’s car is of the same quality. Anyone can see that is nonsense, particularly for the poor, who would want to sell theirs and buy a cheaper one, and rich, who would want to sell theirs and buy a more expensive one.
      • But this is exactly what the government does in healthcare and education.
    • Will become increasingly costly in a competitive and globalizing world. This world is not dangerous; people will still be able to work and make a living. But the ability to survive by transfers from wealthier people will be reduced as people and businesses can relocate if they face high taxes.
Moving from FPTP to AV increases the risk of hung parliaments, with their likelihood of producing governments that are less able to govern effectively. It also seems to favour left-wing parties, which I believe will lead to inferior economic policy.


  1. Charles - a fascinating analysis. Immediate responses: (1) Electoral systems are hard to change, so if you want to take a consequentialist approach I think you need to look further into the future. For example, suppose we ended up with a third party that was centre-right rather than centre-left. Then the tables would be turned the other way, and AV would likely favour the right rather than the left. (2) Even if AV leads to more consensus (coalition) governments, I don't see that as necessarily a bad thing for markets. Consensus governments can give more stability, avoiding extreme swings to the left or to the right. For example, speaking largely from ignorance, I wouldn't say Germany has the weakest economy in the world, but they have had coalition governments for years. (3) The indirect effects could well be significant. First, see Ed Miliband 40 seconds into Second, the indirect effects of empowering voters are likely to be very good for society as a whole. Finally (4) I do sympathise with much of what you say about left-wing economic policy. But I think right-wing politics tends to ignore the consequences of the current market system, in terms of international relations (the arms trade etc.), the plight of the poor (climate change etc.) and sustainability (enshrining the principle that unfettered greed is always the best way to act). I don't know what the answer is, but I think an electoral system that empowers voters to express themselves is not likely to be a bad thing, long term.

  2. Unfortunately, Charles Roddie's article is based on the fallacy that AV will lead to more balanced parliaments than FPTP does. There is no reason for this to be the case; the nature of the two systems is too similar to suggest that one is more likely to lead to the HORROR of coalition than the other (although, incidentally: what do the stock markets do in Germany when a coalition is formed?).

    This is because both systems are majoritarian and constituency based and so do not take into account national vote share. Neither system benefits parties whose vote is distributed thinly but widely over the country; it is mathematically possible under AV for a party to get 49% of the national vote and not gain a single seat.

    In fact, in Australia, there has been one balanced parliament in the last century compared to our six under FPTP.

    But ultimately the problem with the article is that it takes a consequentialist approach which is inimical to democracy - either we believe in the rule of the people or we do not. In which case, the only pertinent question is certainly not "Is AV likely to lead to better or worse governments being elected?" but "Does AV allow people to better express their political views?"

    To which, of course, the answer is yes.

  3. Hi Anthony. Thanks for your comments.

    1. & 3. Yes in the longer term the indirect effects may become more important. It's just hard to see what these might be. Maybe the main two parties will be nicer to smaller parties. What effect this has I'm not sure; some influence of the smaller parties on the policies of the main parties?

    2. A priori this argument makes sense. But the behavior of the markets seems to indicate that any advantage of moderation that consensus government brings is outweighted by the negative effects I mentioned.

    Not sure what has been done with cross-country comparisons. I hear them used more often against coalitions but cause and effect are hard to deterimine.

    Yes, eventually there will be some tradeoffs. Maybe there will be a chance to discuss one or two of these at some other point.

    Overall I'm not sure what you mean by empowering voters. AV would change the way they use their power but I don't see why it would be "more power".
    Or "more expression" unless it's the fact that people are submitting more information. But I would think the extra information, insofar as it doesn't determine the outcome, would have similar importance to a poll, i.e. not much since polling is done all the time.

  4. Charles - thanks for the responses. On "empowering voters", I mean AV makes it easier for an individual voter to cast a vote that has (what they perceive to be) a positive effect on the outcome, by removing the need for agonising tactical voting. And by requiring some kind of majority support for the winning candidate, AV empowers the voters corporately to get an MP who is generally acceptable to the majority, rather than the vote-splitting under FPTP that can allow a generally disliked candidate to win.