Tim Worstall at the Adam Smith Institute blog is not at all keen on Polly Toynbee’s “very patrician idea” that the government is there to provide “entertainments” and “diversions”.
“And what really grates is that at least the Romans insisted that it was the patricians that paid for the bread and circuses, Polly’s insistent that we must be charged for what she insists we must have“. [My italics]
Worstall is opposed to this neo-patricianism for a rather excellent reason: “If we want it then we’ll buy it, thank you so very much”. As I said, excellent.
Elsewhere in the post, he contrasts Polly’s idea of what the government is for with the ASI’s idea of what the government is for. According to this notion, the government
“is there to do those few things which both must be done and can only be done by government”.
“Must be done” has a familiar ring to it. What, one wonders, became of that “If we want it then we’ll buy it, thank you so very much”? It disappeared so quickly!
However that may be, is there not a whiff of the illiberal about these “things that must be done”? There must at least be a strong presumption against them given that doing them requires that we ride roughshod over the fearsome “revealed preferences” liberal economists are usually so respectful of.
But perhaps the matter is not so simple. After all these are things that not only “must be done” but that “only government can do”. That is, they are public goods.
A textbook definition of public goods would say that they are goods which are “non-excludable” (they cannot be provided at all unless they are provided for all), and “non-rival” (their being used does not leave any less for others to use). A more cynical definition might say that we call public goods those goods that it is thought unwise, or uncivilised, to make people purchase themselves. Be that as it may, we cannot derive an “ought” from an “is”, and so the brute fact that a thing “can” be done does not mean that it “should” be done – not even if the would-be doer is the “only one who can”. Something more is needed to get us to that. But what?
The traditional way of bridging the gap is to say that the people only appear not to want to buy public goods; really they do want to buy them, and would do, if only there were not powerful forces preventing it. These powerful forces are manifested in the Prisoners’ Dilemma.
In a game of Prisoners’ Dilemma players choose between Cooperate and Defect strategies. In a two person game, a player who Cooperates will get his second most preferred outcome if his opponent Cooperates, but will get his least preferred outcome if his opponent Defects; a player who Defects will get his most preferred outcome if his opponent Cooperates, and should his opponent Defect he will safely avoid his least preferred outcome. As players must make their choice ‘in the dark’, choosing Defect is each player’s best strategy. The logic of the Prisoners’ Dilemma is such that, despite each player preferring mutual Cooperation to mutual Defection, rational, self-interested play leads inexorably to the Pareto-inferior outcome. As provision of public goods is typically assumed to involve a Prisoners’ Dilemma (Cooperate/Defect describing willingness/unwillingness to pay voluntarily for the public good), and because Prisoners’ Dilemmas produce a result that is less preferred by every player than one that could be arrived at if only some kindly third-party were to lend a helping hand, we appear to have our justification for Leviathan.
But before we say something we might regret, let us pause. After all, “the Prisoners’ Dilemma is not the only game in town” (de Jasay).
Take the Assurance game, for example. As in the Prisoners’ Dilemma, in games of Assurance players choose between Cooperate and Defect strategies; however, unlike in Prisoners’ Dilemmas, Assurance games played by rational, self-interested players result in Pareto-optimal mutual Cooperation. The key point here is that there is no ‘objective’ difference between these two types of interaction, only a difference in the preference structure of the players. In Prisoners’ Dilemmas players prefer free riding off the labour of others to cooperating with them at some cost to themselves, in Assurance games players prefer cooperation to free riding; once the preference structures are posited we have our result. But there is no necessity for the preference structures to be what they are, nor is it necessary they should remain what they are. Players have only to change their minds to change their game.
Now as liberals, we are, on the one hand, strongly, even mortally opposed to the idea that because we cannot be trusted to prefer doing the “right” things to the “wrong” things, government must force us to fund transgressive artists and bad poets. On the other hand, we feel it is only right that government taxes us to provide the public goods that we must have – even though it must do this only because we might prefer free riding to chipping in. In light of this, perhaps we would do well to give some serious thought as to what is the precise difference between ‘good paternalism’ and ‘bad paternalism’, and why we support the one but not the other.
This was written with a more ‘general’ market libertarian audience in mind than some stuff on here. Readers should take the phrasing in the final paragraph cum grano salis .