Improving Germany’s model democracy

Germany has one of the best electoral systems in the world. If I were writing a constitution, I think only New Zealand could serve as a better model. I’ve briefly discussed the advantages of mixed member proportional representation (MMPR) before, but here are some of the main benefits:

1. The question of which person you want to represent your local area and the question of which party you want to govern the country are separated.
2. Each voter has equal influence over the ultimate distribution of seats – unlike systems which give rise to '’swing states’ or ‘marginal constituencies’, where voters have more power
3. There is less incentive for tactical voting, as proportional representation ensures that votes for smaller parties are not ‘wasted’

I think these three features – less ambiguity over what people are voting for, equalising the influence of each voter, and ensuring that voters’ genuine preferences are considered – mean that MMPR is more democratic than other systems.

However, recent events in Germany show that it can still fall short of these lofty democratic principles. In the January 20th Lower Saxony election, it has been suggested that a number of CDU supporters voted for the FDP. Why would they do this? Because the FDP is more likely to cooperate with the CDU than other parties, and it was in danger of falling short of the 5% share of the vote necessary to win any seats in parliament. It has been suggested that the general election later this year might witness a similar phenomenon.

The election was analysed mostly in terms of its implications for the fortunes of the CDU, but more troubling are its implications for democracy as a whole. CDU supporters voting FDP violates principle 3) above, that democracy should aggregate authentic, and not falsely stated preferences. Less obviously, it also violates 2) – that each vote should have equal bearing on ultimate electoral result. The CDU supporters misrepresented their preference because they believed that voting FDP would give them more influence on the distribution of seats. Voting FDP gave them more power because they had the chance not only to be the marginal voter who decided the destination of a single seat, as all other voters did, but additionally gave them the chance of being the marginal voter whose vote carried their party over the threshold, and decided a hatful of extra seats.

The flipside of the extra influence wielded by voters whose party is close to the parliamentary threshold is the disenfranchisement of those whose favoured party fails to meet the threshold – their votes effectively count for less than the rest of the electorate, raising the spectre of the ‘wasted vote’, another phenomenon MMPR is supposed to avoid.

These problems are minor niggles compared to the inequity of voting systems like the US or the UK, but the undignified scenes in Lower Saxony let the German system down. If Germany is to live up to the high democratic standards it has set itself, it would do well to lower the threshold for entry to parliament (at 5% it is amongst the highest in the world), or better yet, abolish it entirely.

Should votes be for sale?

Ahead of Tuesday’s U.S. Presidential election, Stephen Levitt, citing the work of Glen Weyl, has made a provocative suggestion for improving the electoral process – people should be permitted to vote multiple times, paying increasingly higher fees for each additional vote. I’m attracted to this idea because it addresses one of the major drawbacks of ‘one man, one vote’, that every vote counts the same, regardless of how strong the preference expressed is. It is problematic (to me, at least – others may have different intuitions) that a vote cast by a fervent anti-Republican, who fears that a Romney victory would be as calamitous as a zombie apocalypse should be cancelled out by the vote of someone who thinks the candidates are much the same, but that Romney just has the edge. To be clear, the objection is not that the former voter is better informed or qualified to vote, just that they feel stronger about the issues – the opinionated but ignorant would still be likely to buy plenty of votes.
 
A trivial example should make my intuition clear. Suppose a group of people are trying to decide what topping to get on their pizza. Four people have a weak desire to get prawn, while two have a weak desire not to get prawn. Meanwhile, one person is strongly against prawn because it will cause them to have a severe allergic reaction, and require them to go to hospital. If the decision is put to a simple majority vote, the group will opt for prawn, 4-3. Yet it seems wrong to me that the preferences of those who gain so little from having the prawn should be counted equally with the person who is allergic. For a more clearly political example, perhaps the views of those enjoying a modest tax cut should not be counted equally alongside the person who loses vital services to pay for the cuts.

The obvious objection to the vote buying proposal, which Levitt discusses, is that it gives disproportionate political power to the rich. His response is unconvincing – he points out that such a transparent system would be better than the status quo, where the rich exert greater influence through campaign finance. But that is just an argument to reform campaign finance, not to create an alternative way to perpetuate the inequity.

Moreover, the objection is deeper than Levitt appreciates - insofar as the vote buying scheme gives greater power to the rich, it fails in its attempt to reflect the strength of voters’ views, because it fails to account for the law of diminishing marginal utility. The logic behind the vote buying scheme, as I understand it, is that it forces people to put their money where their mouth is. In buying $50 worth of votes for the Republicans, I am showing that a Republican government is worth $50 to me. The trouble is that that $50 is a much more powerful statement when it comes from a poorer person, involving a more significant sacrifice on their part. If $50 pledged to the Republicans means giving up a meal at a nice restaurant, and person A dines at nice restaurants regularly, while for B it is an annual treat, then it is clear that B wants the Republicans to win more.

The obvious way to reflect this would be to make voting fees progressive i.e. proportionate to income. So the richer you are, the more you have to pay to vote.

Even with this amendment, there are two types of concern I have about the scheme. On one hand, I am sceptical as to whether this measure would be successful in achieving its own stated goal - to successfully reflect public opinion. On the other hand, I wonder whether this goal is the only desirable one to have.

The first reason for scepticism about vote buying accurately capturing people’s real views is that it depends on people being able to provide an accurate and meaningful valuation of how much an electoral outcome is worth to them. Yet as behavioural economists such as Dan Ariely have shown, this is something most people struggle to do without a set of reference prices for context (do you have a clear idea, right now, how much you would be willing to pay for a certain number of votes in the next election?) This is further complicated by the fact that people wouldn’t be buying a Romney or Obama victory, but just the mere possibility of being decisive one way or the other.

Even if people could accurately gauge how much their vote is worth, the market for votes might still be distorted. Just because a person is willing to  pay $100 to vote for Obama, this might not fully be a reflection of their endorsement of the president. People might pay a voting fee for all sorts of reasons, just as people vote for all sorts of reasons just now – out of a sense of civic obligation, to express gratitude for the historical sacrifices of their forebears, out of a sense of power. Each of these motivations means that the amount they pay for their vote will fail to track their enthusiasm for the candidates.

A more fundamental objection to the vote buying scheme is that it fails to understand hat voting is really about. It could be argued that it is not just about taking a snapshot of public opinion, but that it has some other function, which is subverted by vote buying. For example, political theorists have long argued that democracy has an educative value – by giving citizens the responsibility of a say in how the country is run, they are encouraged to live up to that faith by engaging with politics, developing and trusting their own judgement. If engagement with politics becomes a prerequisite of voting, then those who are not inclined to involve themselves in politics will simply shrug off this responsibility, and opt not to develop these faculties.

Whatever you think about the vote buying proposal, it offers an illuminating challenge to the way democracy proceeds just now – does it really expose an important problem with the status quo? And if so, what is wrong with its proposed solution? In answering these questions, we get a clearer idea of the strengths and weaknesses of democracy, and why (if at all) we value it.

When is campaign finance undemocratic?

That the recent cash for access scandal that has hit Prime Minister David Cameron and his Conservative party has put campaign finance back on the agenda is certainly a good thing. Yet I worry about the current focus of the debate. Most people seem to be under the impression that the major problem with the current system is that single individuals or organisations contribute too much, and therefore wield too much influence. However, I think it is at least as worrying that there are such great disparities between the resources of the different parties. In the 2010 British general election, the Conservatives spent twice as much money Labour, and enjoyed over three times the budget of the Liberal Democrats. Unsurprisingly, there is a strong correlation between how much a party spends and its electoral performance. It has been estimated that an extra £2.25 of electoral spending corresponds to an additional vote. If this is correct, then the £8 million more that the Conservatives spent than Labour were worth over 3 million votes to them. Since the Conservatives received just over 2 million more votes than Labour at the election, this implies that their vote share could even have been less than Labour’s had there been a level financial playing field.

These calculations are a bit sketchy, but the point I am trying to make should be clear enough: there is significant inequality of wealth between parties, and this almost certainly influences electoral results. This seems fundamentally undemocratic. The idea of ‘one person, one vote’ is often taken to be essential to democracy. Yet for a donation of as little as £2.25, a person can effectively double their political influence – voting once themselves, and basically buying a second vote.

Obviously this ‘multiple voting’ is more heinous in the case of larger donations where more votes are ‘bought’. But the argument applies in the same way to the ordinary trade union members whose support Labour keep stressing. The implication of the argument is that all privately funding of political parties is undemocratic.

Yet funding electoral campaigns is clearly not the same as straightforwardly buying votes. The voters who are influenced by it still get to freely choose who to vote for. Campaign spending may be a way of deploying one’s resources so as to alter people’s voting intentions. But the same could be said of rational persuasion, and it is very hard to argue that that is undemocratic. We do not think that it is undemocratic ever to try to influence others, that democracy requires us to leave voters to make up their minds alone. On the contrary, better informing fellow citizens of their options is seen as a democratic virtue.

So how, if at all, is spending money different from other forms of political campaigning? I think the intuition which explains the unease that people like me have with the role of money in politics is expressed by Jurgen Habermas’ claim that democracy requires the “force of the better argument” to prevail. This is doubtless idealistic, since elections are influenced by any number of factors which have nothing to do with the arguments: personalities, tribal loyalties etc. But it is a regulative ideal: the more that elections are decided by arguments, the more democratic the political system is; the more that other influences intrude, the less democratic it becomes. If campaign spending becomes a more significant determinant of electoral success, this moves us away from democracy.

A certain level of campaign spending brings us closer to the ideal of elections decided by arguments. If parties could not spend, they would be unable to publicise their message, and the electorate would be relatively uninformed. However, this beneficial effect is counterbalanced by the malign effect of uneven electoral spending: if some parties have far more money than others, then they are likely to do better regardless of the success of their arguments. This implies that it might be better to have a relatively ignorant electorate, as long as they are equally ignorant about all parties.

This would seem to be a case for the public funding of parties, since this would ensure that voters are (a) better informed, but also (b) not much more informed about some parties than others. However, it would also seem to have other radical implications. If parties are to triumph only by the force of their better arguments, then this means that they cannot base their success on more enthusiastic grassroots volunteers. It seems like the use of volunteers by political parties for canvassing and leafleting is subject to the same trade-off as their spending of money: more is better, but only if it is evenly spread. Yet the implication of this thought is that there should be limits on the use of volunteers. This seems like an even more puzzling paradox than campaign spending. On the one hand, it seems undemocratic to prevent people volunteering their time and efforts to promote a political cause. Yet on the other hand, it seems undemocratic for the party with the most loyal volunteers to win on that basis alone. Have I gone wrong somewhere?

Some Bad Arguments in the AV Debate

One of the features of the Alternative Vote debate – a common element of many modern political arguments – is that politicians (on both sides) keep making the same bad arguments. Despite the fact that these are regularly debunked, they keep resurfacing as if nothing had happened, a clear sign that certain messages are intended to be hammered home, regardless of their veracity. That these counterarguments are so regularly ignored, despite the fact that anybody involved in these debates must have heard them numerous times suggests flagrant cynicism and duplicity. Anyway, these are the arguments that really irritate me:

A. Arguments from the No Camp

AV will help extremist parties like the BNP

Any sort of claim about what will happen under AV is suspect because it involves trying to guess people’s second preferences and how these preferences would affect results in different constituencies. It is possible that AV could help the BNP, I suppose because BNP sympathisers would no longer see supporting them as a wasted vote. However, as numerous people have pointed out it is extremely unlikely that AV would gain the BNP seats in parliament. Under AV, as I have already argued, the winning candidate is the one that fewest hate. In most constituencies, there are more voters who would prefer any alternative to a BNP MP, such is the hatred for the party. As long as this is the case, the BNP cannot win under AV. If it were not true, the BNP would win even under FPTP. Yet under FPTP, the BNP can capitalise on a split vote, and take a seat based on a strong core vote of a minority of constituents. So AV appears to make BNP success less likely.

Even if it were the case that the BNP would profit under AV, this still remains the worst argument in the history of arguments about electoral systems. To show this is the case, just replace ‘BNP’ with any other party name. Arguments like ‘X is bad because it benefits Labour’ or ‘Y would help the Conservatives get in’ are unacceptable because it is fundamentally undemocratic to rig the system to produce a desired outcome. If people vote for the BNP, the BNP should do better. If you don’t agree, you clearly don’t like democracy and have no place debating democratic systems.

AV leads to hung parliaments

This argument is much like the one about the BNP, though slightly less clear cut. Once again it depends on predicting how the AV system will affect electoral results, which involves a lot of guesswork. Once again, the No campaign seem to have got their guesswork wrong. Their prophecy is fairly unambiguous: “AV leads to more hung parliaments, backroom deals and broken promises like the Lib Dem tuition fees U-turn. Instead of the voters choosing the government, politicians would hold power. Under AV, the only vote that really counts is Nick Clegg's. We can't afford to let the politicians decide who runs our country.”

However, as this article points out, it is more complex than that. The Political Studies Association has estimated the results of past elections had they been held under AV: “Contrary to some claims, AV would not lead to permanent hung parliaments and coalition governments. Of the last seven elections, only the most recent would have delivered a hung parliament – just as under FPTP. By boosting the Liberal Democrats, however, AV does increase the likelihood of hung parliaments a little. Some simulations suggest a hung parliament in the close election of 1992.”

Though the link between AV and hung parliaments is more plausible than the link to the BNP, both arguments are invalid for the same reason. It is fundamentally undemocratic to rig the system to produce a desired outcome, whether that outcome is success for a particular party or ensuring a majority.

AV gives some people more influence than others

The argument seems to be that since some people’s lower preferences are counted but not others, those people have more effect on the ultimate result than others, violating the sacred dictum of ‘one person, one vote’. This is clearly nonsense. For a start, if the preferred candidate of those voters whose second preferences weren’t counted had been eliminated they would have supposedly have more influence, so there is no systematic bias.

But the truth is that nobody’s preferences count for any more than anyone else’s. If your first preference wins, that preference is taken into account at every round of voting. In effect, it is as if there is a fresh round of voting every time a candidate is eliminated, and it is presumed (surely uncontroversially) that people would vote for the same candidate if they can. If we consider these distinct rounds as different elections, we see that every person has one vote and one vote only in each round.

B. Arguments from the Yes Camp

AV means all MPs will have the support of over 50% of the electorate

The first thing to point out is that getting a vote from 50% of the electorate is not the same as having their support or confidence. This is obvious when you consider that some of these votes could be people’s fourth or fifth preference. Getting 50% of the final vote under AV means nothing more than that 50% thought that at least one person would make a worse MP than you.

Worse still, there is no guarantee that winners under AV will in fact get over 50% of the voters. If a substantial number vote only for candidates knocked out in early rounds of the count and fail to express lower preferences, then the threshold of victory will be reduced. As an illustration, twelve of 31 Scottish local authority elections carried out under AV have produced minority winners.

AV would have made the expenses scandal less likely

This argument appears to have two premises:
1. There will be fewer safe seats under AV
2. MPs in safe seats are more likely to be corrupt
As far as premise 1 goes, it is certainly the case that MPs who rely on a strong core support of less than 50% of the electorate will have to woo new voters to keep their positions. However, certain MPs may be more secure if they can attract many second preferences. So, for example, many Lib Dem MPs with relatively small majorities may have their position strengthened if they know both Conservative and Labour voters will put them second. Moreover, over 200 MPs already get over 50% of the vote, so AV is unlikely to have any effect on them. On balance, it seems plausible that AV will cut the number of safe seats, but hardly dramatically.

As for the link between safe seats and expenses, Nick Clegg’s claim that “Research suggests that MPs with the safest seats were much more likely to have been exposed by The Daily Telegraph” is certainly true. However, it is misleading because ‘exposed by The Daily Telegraph’ is not the same ‘made excessive expenses claims’. The study that Clegg cites was carried out before the full extent of expenses repayment were made public. More recent research suggests no statistically significant relationship between repaid expenses and the size of an MP’s majority.

Some Thoughts on the Alternative Vote

Much has been said and written already about the 5th May referendum, when British voters will decide whether UK Parliamentary elections should be carried out under the First Past the Post (FPTP) or Alternative Vote (AV) system. David Cameron and Nick Clegg have already summarised the case for and against the change surprisingly well. Here’s my perspective, though, and possibly a different way of looking at the matter.

The first thing to get clear (at least for the purposes of my discussion) is what the AV vote is not about. One of the problems of the British electoral system is that voters are (or at least think they are) being asked three questions, but are allowed to give only one answer. Under the current system, we vote for an MP, an individual to represent our local area in parliament. We do not, in fact, vote on which party should be in government. Nor do we vote on which person should be Prime Minister and run the country. The three questions are constantly conflated, but under FPTP we only have a say on the first.

Certain reforms would address this problem. Proportional Representation would mean that voters are asked at the polls which party they want in government. Mixed Member P.R. would split the questions of constituency MP and national government and ask both separately. A separate Presidential election would allow voters to choose the leader of the country.

Nothing this radical is involved in the switch to AV. AV asks the same question as FPTP: ‘Who do you want to be the MP for your constituency?’ This means that all the discussion about the effects of AV on the proportionality of parliament or the possibility of coalition government seems to me to miss the point. John Prescott has claimed that AV is a good system for electing individuals (it was the system under which he was elected deputy leader of the Labour party), but a bad one for electing governments. But neither FPTP nor AV are about electing governments, they are systems for electing individual MPs. It is on their merits for this purpose that I’ll try and compare them here.

Comparison is difficult, not least because the economist Kenneth Arrow has long since demonstrated that there is no such thing as a perfect electoral system. Under different circumstances, different systems best reflect the popular will. This can be illustrated with a couple of examples.

Example A: Socialist Split

FPTP AV

A. Black (Conservative)              40%               B. Brown     60%

B. Brown (Democratic Socialist) 35%               A. Black      40%

C. Green (Social Democrat)        25%

The left, true to form, has split, dividing almost equally between the Social Democrats and the Democratic Socialists. Their division stems from some obscure point of doctrine, so there is little substantive difference between them. However, the rift is too bitter to be patched up. Nevertheless, the socialists are all adamant that they would prefer a socialist representative (of any stripe) to a dreaded Conservative. In this case, the FPTP seems clearly to misrepresent the views of the constituency, producing a Conservative MP to stand for a majority of socialists. AV, by contrast would allow the social democrats to use their second preferences to keep the Conservative out.

Example B: Genuine Third Party

FPTP AV

A. Butcher (Socialist)             45%                  B. Baker      55%

B. Baker (Conservative)          35%                  A. Butcher    45%

C. Smith (Liberal)                   20%

Suppose that the Liberal supporters are genuinely liberal, with no real sympathy either for the Conservative or the Socialist. On balance, most of the liberals are more scared of the Socialists, and so their second preferences give the election to the Conservative candidate. In this case, it seems that the FPTP result is better, fired by the positive vision of the Socialist supporters rather than the ambivalence of the liberals.

This example is not straightforward, and many people may have a different intuition that the AV result is preferable. However, these conflicting intuitions seem to get to the heart of the dispute. To put it crudely, FPTP gives victory to the candidate that most people like; AV to the one that fewest hate. AV is for those who fear that a vocal minority can sneak someone despised by the others into power. FPTP is for those who reject this conservative line of thought.

FPTP is best when people vote for a candidate they want to get in; AV when they vote against a candidate they want to keep out. The trouble is that we can never know in advance whether more people are casting their votes for or against politicians.