HS2: how to decide

Last night John van Reenen tweeted that the controversy about HS2 shows that we need a better way of making decisions about big infrastructure projects. He's right. Economists have known the answer here for years. We should use a demand-revealing referendum, as proposed by Ed Clarke.

The idea here is simple. We hold a referendum in which everyone in the country is asked to state the net benefits to them of HS2; as the scheme is likely to cost each voter at least £700, these benefits will be negative for some/many voters. Then we add up all the answers, and if we get a positive number we go ahead, and if it's a negative one we do not. Finally, we check the net total benefit against individual voters' stated benefits. If any voters stated so large a sum that their vote was pivotal, they must pay a tax equal to the difference.

This method has some huge virtues.

One is that it's decentralized. Instead of the HS2 decision being taken by a few special interest groups after what John Kay calls a "dismal" cost-benefit exercise, the decision is taken by everyone. Sure, voters' stated benefits will be wholly subjective. But this is a feature not a bug; as James Buchanan argued, costs (and benefits) are subjective.

Secondly, it's incentive-compatible. If you state too small a sum, you risk not getting your preferred option. If you state too big a sum, you risk being taxed. A demand-revealing refernedum thus forces voters to be honest.

Thirdly, if the use of such referenda were to be widely adopted, it would improve out political culture. There'd be no place for blowhards and windbags who overstate their case. Instead, they would have to proportion their estimate of net benefit to the evidence. We'd get much more reasonable, grown-up politics.

Herein, though, lies a paradox. Demand-revealing referenda are in theory wholly consistent with neoliberal ideology, because they reject the top-down central planning  of Whitehall-run cost-benefit analysis in favour of a market process. And yet such referenda fall well outside the Overton window, and are proposed only by cranky extremists like me.

There is, though, a simple solution to this paradox. Adopting demand-revealing referenda requires that politicians and vested interests give up their power. And this won't happen. Power trumps ideas.

Comments

[Spink Bottle]

Is voting in the referendum compulsory? Otherwise aren't the overwhelming majority of the population going to say that they have no idea what value to place on HS2 and will not want to risk getting taxed. And if not many vote it might not be very democratic.

[Luis Enrique]

"Demand-revealing referenda are in theory wholly consistent with neoliberal ideology"

and because they ignore awkward fact rich people can afford to have more say in life?

[chris]

@ Spink - there's no reason why voting should be compulsory - though the fact that abstention is equivalent to a zero vote might encourage some to vote. People don't need to know a precise value; they can just vote the central value of their subjective probability distribution.
@ Luis - we should distinguish between efficiency and justice. If you think some folk are too rich, the solution is to tax them more, not to renounce devices which improve allocational efficiency. After all, the rich have more power in ordinary markets for goods and services but we don't shut down markets on this account, and nor should we.

[FromArseToElbow]

"Demand-revealing referenda are in theory wholly consistent with neoliberal ideology". Indeed, but ideology is not praxis.

Neoliberalism has repeatedly shown itself to prefer the "central planning of Whitehall" as a means of privilegeing selected business interests. It has only advocated markets where they could be rigged (railways being a case in point).

HS2 is best seen as an extension of the Metropolitan line that will allow well-heeled London commuters to trade up from Amersham to Solihull and the Forest of Arden.

Despite specific instances of blight, the aggregate effect will be good for housebuilders, by increasing the London commuter belt, and good for rail engineering firms, by creating additional maintenance revenue streams into the future.

That's not a lage enough constituency to win a referendum, but it one that has the government's ear.

[Luis Enrique]

Chris, agreed, I more meant that as a dig at neoliberal ideology

[Philip Walker]

It's a neat idea, but I see several flaws/details needing tightening.

Firstly, we need to assume the system can be explained to the electorate in general. Certainly it will not succeed if its proponents attempt to persuade the public that all they have to do is to "vote the central value of their subjective probability distribution"! Even once this is explained satisfactorily, people wouldn't be able to give a sensible value. I estimate that no government spending of any sort would ever happen.

Secondly, it would be far more sensible to ask the public to estimate the total benefit rather than the net benefit. I would be willing to bet that net benefit would confuse enough people to make the result infirm. We can simply subtract off the £700 afterwards, which is likely to cause less confusion. (Abstentions would count as a vote at cost.)

Thirdly, I assume that we allow unboundedly negative bids as well, and penalise sufficiently strong ones in the same way as positive ones.

Finally, I'm still not clear in my own mind how the taxation round works. "If any voters stated so large a sum that their vote was pivotal, they must pay a tax equal to the difference" between what and what? (Clarke's links are even less helpful here. They have arithmetic errors at precisely this point.)

[chris]

@ Phil. Thanks for those points. I was vague about the tax, but I say here that the "swing voters" should pay a tax equal to the net benefits others would have had, had they not voted:
http://stumblingandmumbling.typepad.com/stumbling_and_mumbling/2007/03/trident_how_to_.html
You're right that many details need sorting. But they would be if smarter folk than me were to think about it more.
Yes, there's a need for public education about how it would work. But the consultation on HS2 cost millions, and that moeny would have been better spent on such education. Over time, as familiarity with the scheme increases, this problem would disappear.

[Philip Walker]

Your optimism about the numeracy of the British public is touching! ;-) But as a working mathematician, I am far more pessimistic. I think we need better public education in mathematics more generally. But then I would say that, wouldn't I?

Note: your definition works for projects which are stopped but would have gone ahead had a certain voter abstained. For projects which go ahead but would have been stopped had a certain voter abstained, it must need tweaking somehow. Take the absolute value, perhaps?

I wonder what the system's pathologies are. I tried investigating last night but I think I used the wrong definition for the taxation round.

[Jim Vine]

This is certainly an interesting proposal, but I wanted to raise a conceptual challenge to go alongside the practical ones already noted.

The mechanism you propose looks like it is targeted on maximising the delivery of preferences. But econometric analysis using the wellbeing valuation method shows that preferences (whether stated or revealed) need not be the same thing as the options that would maximise wellbeing. See, for example, Daniel Fujiwara's recent work for HACT [1], where he found that the aspects of housing that maximised wellbeing were not the same as those that people placed most value on. In other words, people are willing to pay for things, presumably under a belief that they will make them happier, when to actually achieve that result they would sometimes be better spending on something else instead.

[1] http://www.hact.org.uk/sites/default/files/uploads/Archives/2013/02/The%20Social%20Impact%20of%20Housing%20Providers%20report2013.pdf

[Zby]

@Jim Vine - are the errors systematic - is it always overpaying? If not then maybe the errors would cancel each other (like in the law of large numbers).

[Andrew Mangles]

This is of course nonsense. The vast majority of the UK will be entirely unaffected by HS2 except in the most peripheral way and thus will vote negative if voting positive might incur a tax. Especially if it will cost them £700 if it goes ahead. I live in the west country were there will be no benefit but still think it is a good idea. Improving transport links between major cities has a proven net benefit, whether by road, rail or air and some of the new "costs" recently added to the HS2 project look pretty spurious to me. In my view it is never a good idea to ask people to give their views on a new project, they will always be negative until after it has become a great success, like the Olympics. If you want to see how it should be done go to France where there is minimal consultation and thus a high quality transport network.

[Capitalist Peeg]

I think the idea as described falls short of it's intentions.

In my case, I would benefit from less crowded trains to Euston but would find the excellent Bree Louise demolished when I got there. I suppose I would discount the net benefit over ten years.

Everyone is expected to weigh it up like this and put a price on it: a tall order.

Even then it is a declared rather than a revealed preference. Preference is only revealed when passengers pay for tickets. Estimating the revenue is part of the commercial decision.

The beauty of the idea is that it seeks to include the commercial and non-commercial elements and for that reason I quite like it.

[Markiliff]

"no place for blowhards and windbags"

Never in a million

[greg]

That is the single most stupid idea for deciding this stuff I think I've ever heard.

It would institutionally favour no campaigns for any large infrastructure project. We know people (and the population in general) often rabidly oppose infrastruce projects exactly until they are built and then they go back and say they are wrong and they approve of it.

The £700 per person is the kind of figure that is basically meaningless but would dominate the event unfairly. The actual funding, and recouping of costs and the cost benefit analysis and potential for growth etc is ridiculously complex and frankly abstract and debatable as well and most voters will not take the time to learn that (or all of them even be able to grasp it even if they wanted to). So this would mean that figures like that would dominate despite their essential meaninglessness and £700 gets linked to this is £700 taken out of my tax and a more visceral and misleading

I could go on for ages about how stupid this is but the comments to this absurd idea have covered the basics points on why it's a bad idea.

Also watch the Pater Cook film The Rise and Rise of Michael Rimmer which pillories perfectly in one part why a large amount of referendums are a bad idea.

[chris]

Gentlemen - your scepticism about the rationality of voters is reasonable. But it poses a puzzle. One virtue of demand-revealing referenda is that they will (eventually) improve people's thinking; as someone once said, making people pay is a way of making them think. If people are incapable of making a rational judgment when their money is at stake, how much more likely are they to be irrational when they bear not cost of their decisions.
If you think demand-revealing referenda are stupid, you should think that conventional democracy is even stupider.

[Scott]

You cannot tax negative impacts! If people have to move and find new jobs and houses because of the construction, they should not have to pay _more_ because that is going to cost them £20,000 instead of £100!