A framework for normative economics
This is a post that Matt really should have written, but unfortunately he’s snowed under at work at the moment. Having said that, it’s not really much more than a plug for Steven Landsburg’s fascinating new paper entitled “The Methodology of Normative Economics”. Economists tend to shy away from making normative judgments: when constructing a social welfare function they claim to be maximising efficiency. Yet the form of that welfare function determines the form of the efficient solution. Moreover, people clearly have strong feelings about what constitutes an appropriate welfare function for a regulator to maximise.
Given that people feel strongly about what form this function takes it makes sense to model people as having preferences over different forms of the welfare function. For a welfare function to be ‘consistent’ it must predict that it is itself the optimal welfare function to use. Landsburg shows that, if rich people care about equity, there is only one such function that exists. Even if people are selfish and don’t care about equitable distribution of income, there are only a limited number of consistent welfare functions.
What I really like about Landsburg’s paper is that it forces economists to confront the normative judgments that they inevitably make when they talk about welfare. Yet it forces them to do so by using an analytical framework that they are comfortable with. Rather than forcing them to make value judgments, it allows them to talk about the normative framework of their model without having to resort to moral or philosophical discourse that they are ill-equipped to deal with.
Check out another of Landsburg’s methodology papers here if this stuff dings your bell.
As it is a normative framework it is filled with value-judgements.
From a quick skim it seems to me that this paper is endogenising the choice between welfare functions? As a result, the value judgement the government makes is based on the social preference of their own welfare function.
Furthermore, if you actually wanted to get anything significant out of the model, you would have to decide on a preference ranking in order to apply an appropriate utility function, which is again a value-judgement.
Also, this idea seems sort of funny to me. We are defining social preferences over welfare function which define social preferences over outcomes. All we are doing is moving our set of assumptions from the ‘ends’ to the ‘means’ of maximising welfare. If we have preferences over certain social welfare functions, surely we are making value judgements about the final set of outcomes implicitly.
Well, what I see the paper as doing is limiting the normative judgments that need to be made in order to solve a welfare problem. If people care about social justice then it is shown that only a limited number of welfare functions will be self-justifying. The modeller then makes a value judgment about which of that remaining set is best to use. If the populace leans towards a socialist frame of mid then there is a unique welfare function that is self-justifying. That means that the modeller need not make any value judgment over the set of possible welfare functions since only one is plausible.
It is true that the modeller still needs to define a set of preferences over welfare functions. However, this is equivalent to defining preferences over bundles of goods. I think you’re asking a bit much if you expect the model to endogenously determine an appropriate utility function. What this paper does is show that the set of plausible welfare functions is limited, given a particular set of preferences. Preferences are always exogenous in microeconomics. Ideally, the function to be maximised arises endogenously from those preferences and a few other assumptions about rationality. This paper shows how that can happen in the context of welfare maximisation.
I am not disagreeing that it can be useful for economists to model the normative judgements they make, after all abstract models need not only be positive statements. However, I did disagree with the following claim you made in the blog post:
“Rather than forcing them to make value judgments, it allows them to talk about the normative framework of their model without having to resort to moral or philosophical discourse that they are ill-equipped to deal with”
This framework does force people to make value judgements. Also by endogenising the social welfare function Landsburg hasn’t reduced the set of normative judgements we have to make, he’s just rearranged them. The welfare function itself is just a means to an end, the end is the outcomes. By valuing different welfare function we are implicitly placing values on the set of outcomes that the welfare function produces.
However, if this framework is a more efficient way of solving social welfare problems, that is a good thing. But I don’t think it changes the issues associated with the value-laden sections of economics.