Whatever the result of the Scotish referendum, almost half of all Scots will be disappointed, some very much so. Would there have been a better way of arranging things?

Possibly. I'm thinking of a demand-revealing referendum. The idea here is simple. Rather than simply put an X in the yes or no box, voters are asked to state the monetary value they would put upon independence; for nationalists this would be a positive amount, for unionists a negative one. And then, we look at whether any individual's vote was big enough to change the outcome, and levy upon him a tax equal to the net gains the other side would have had in the absence of his vote.

Take an example with three voters. Alec would vote £100 for independence, Bruce would vote £30 for no, and Charlie would vote £40 for no.

As £100 beats £40 + £30, independence wins. However, because Alec's vote is decisive, he must pay a tax of £70 - the net gains Bruce and Charlie would enjoy has Alec not voted.

The beauty of this system is that it forces people to express their true feelings. If you vote too big a sum you risk having to pay a big tax, whereas if you vote too little you risk not getting your way.

This procedure is consistent with the greatest happiness of the greatest number. In our example, independence is a potential Pareto improvement, in that Alec could in principle compensate Bruce and Charlie and still be better off*. By contrast, simple majority voting allows the weak preferences of the many to outweigh the strong preferences of the few.

In this sense, utilitarians should prefer a demand-revealing referendum to a simple one.

I'd add two further benefits:

- Demand-revealing referenda would improve the quality of debate by forcing partisans to say how much they value indepdendence or the union. This would moderate the rhetoric on both sides and militate against some forms of self-deception - though it would of course be too optimistic to think demand-revealing referenda would eliminate all cognitive biases.

- Demand-revealing referenda reduce bitterness. In my example, Bruce and Charlie can console themselves by knowing that if they had really wanted to save the union, they should have voted more. In this sense, losers have themselves to blame, not others. This should reduce the recriminations that follow the result.

There's one objection to this scheme that I don't think is adequate - that you can't put a price upon feelings of nationhood. This fails because we already put prices onto lots of things: libel laws put a price upon reputation; the criminal injuries compensation authority prices injuries; and NICE even puts a price upon life. There shouldn't, therefore, be an objection in principle to pricing national sentiment.

Another objection is that demand-revealing referenda give power to the rich; the more money you have, the more you can afford to vote a big sum.

But it's not clear how decisive this objection is. Existing political arrangements also give the rich power, not only because they can make donations to parties and causes but also because the media defer to and publicize their opinions.

Insofar as this is a decisive objection, it is a case for reducing inequality rather than for rejecting more efficient forms of public choice.

And herein lies my point. One under-appreciated cost of inequality is that it might reduce the quality of public decision-making.

* I say "in principle" because if they were to be directly compensated, they would have an incentive to overstate their opposition to independence in the hope of getting a bigger pay-cheque. This would remove a big benefit of demand-revealing referenda, namely that they encourage people to state their true preferences.