"The concept of utility in economics refers to the pleasure, or relief of pain, associated with the consumption of goods and services" writes economist John Quiggin. Another economist, Robert Frank, suggests that it is closer to the idea of satisfaction. In Luxury Fever he writes: "The analogous construct in the psychological literature is subjective well-being, a composite measure of overall life satisfaction." This kind of talk outrages the philosophically trained Will Wilkinson who writes:
Conceptually, happiness has nothing to with utility in economics, nor does subjective well-being, or subjective satisfaction. Utility is way of representing an ordering of preferences. It simply isn’t a psychological concept, nor a value concept, nor does it imply either. A utility function is just a little machine in which you can put an ordering of preferences, a pair of alternatives, and have something that somebody decided to call a “utility” assigned to each alternative, the most preferred getting the greater utility.
Wilkinson complains that economists talk about "in a most confusing and haphazard fashion". And it’s true. Some economists who are extraordinarily fastidious about their formal work become careless and sloppy when talking in words or discussing the policy implications of their work. I don’t think this is true for Quiggin, but it’s certainly true for others.
Quiggin argues that a cardinal concept of utility is necessary for the analysis of behaviour under uncertainty and in game theory. But this is different from saying that the concept of utility this analysis needs is the same thing as pleasure or subjective well-being. It’s not the same as saying that people are motivated only by the pursuit of pleasure and the avoidance of pain. And it doesn’t suggest that more utility is better in than less utility in a normative sense. Who knows, perhaps the world would be a better place if there was less utility.
But still, Quiggin is happy to link the idea of utility to the idea of pleasure. Wilkinson, on the other hand, insists that the two concepts are separate:
… economists qua rigorous appliers of utility theory, don’t think that being wealthier ought to make you happier. They think that a bigger budget gets you a more preferred bundle of goods, and more preferred means, by definition, more utility. But since utility is a not a subjective psychological state (since semantic satisfaction is not), no one should be surprised that having more utility won’t make you more anything, subjectively. The world could be exactly the way you prefer it, and you could be miserable, because you could prefer to be miserable.
If economics is to have any influence over policy, then these arguments matter. As Edward Mariyani-Squire points out, more and more economists are returning to the idea that utility is a psychological state. Economically trained happiness researchers like Richard Layard not only argue that utility is a cardinal variable, but that utility can be measured and the results compared between individuals. (apparently Edgeworth’s Hedonimeter has arrived!)
But economists like Layard leap to conclusions about the ethical and political implications of their theories. Layard argues that "happiness is the ultimate good and that all other goods are good because of the way they contribute to happiness". Will Wilkinson once suggested to Layard that people didn’t always want to pursue happiness. And it’s a good point. Why is happiness the only thing worth pursuing in life? For thousands of years philosophers have argued over the nature of the good life and grappled with problems like incommensurability, but many economically trained happiness researchers seem to think that this was all a waste of time. To them it’s obvious that maximising happiness is the only thing that’s worthwhile.
I’m curious about happiness research, but I’m far from convinced that happiness is the only thing that matters in life. And unless happiness researchers like Layard stay out of the policy recommendation game, they ought to make a stronger case for happiness as a policy goal.
Update: On reflection, I think my last paragraph is unfair to Richard Layard. He doesn’t avoid philosophical arguments for his position and has engaged with other thinkers like Amartya Sen (and Sen with Layard). What bothers me about the happiness debate is that there is so much focus on empirical issues such as the link between income and GDP and so little on normative issues like the role of the state in promoting well-being. There’s a huge difference between Layard’s concept of well-being and Martha Nussbaum’s — a difference that has implications for policy. But too often economists wade into normative disputes with the assumption that improving utility is a good thing and then refuse to explain what utility is or why we should care about it.
Wilkinson’s point is exactly the one I would have made, and that’s without any philosophical training. I’d be surprised if Frank really disagreed, and it’s hard to tell without reading the book whether Wilkinson is not taking him out of context. Frank seems to be interested in talking about happiness rather than utility, and may have just been sloppy.
Certainly the early neoclassicals conflated preferences and happiness, but your hard-nosed positive twentieth century microeconmist theorist clearly saw there was no need for the Benthamite baggage. Utility just became a short-hand for influence of tastes and prefernces on choices.
Which is not to say that Benthamite utility and happiness are not important — as you know, I’m glad these things are back on the table. But for much of the twentieth century, especially the post-war period, economics students were told that tastes and preferences were just data. It was none of our business why the sovereign consumer preferred bundle A to bundle B.
James
You say: “Frank seems to be interested in talking about happiness rather than utility, and may have just been sloppy.”
I think Frank really does want to claim that utility is identical to the mental state of satisfaction. The only thing that’s sloppy is his take on Bentham.
Here’s an extract from the textbook Principles of Microeconomics he co-authored with Ben Bernanke:
And as if I wasn’t confused enough already, here’s a strange defence of Layard’s Benthamism by Brad Delong.
1. In his intermediate textbook Microeconomics and Behaviour, Frank takes the word ‘utility’ to be synonymous with the word ‘satisfaction’. He does not say anything about the latter.
2. See page 31 of Armen Alchian’s The Meaning of Utility Measurement for an instructive comment on the modern meaning of ‘utility’. See also points 4 and 5 of his conclusion on pp.49-50 for his comments on the very limited value of ‘utility’ to normative analysis.
3. Although it is reasonable to say that happiness, whatever colour it comes in, is not necessarily the only “thing that matters”, it is prima facie unreasonable to say that it is therefore irrelevant OR that it should be ignored altogether – perhaps especially in policy formulation – until (yet another) “case” has been made for it or until someone has come up with a nice formal theory which economists can then wile away their days writing journal articles about. One doesn’t have to subscribe the everything Peter Singer says to agree that forcibly extracting a million dollars from a billionaire who was going to two additional Ferraris and using it instead to purchase vaccines for thousands of people who are painfully dying of preventable diseases is justifiable on (a) a pretty simple ethnical principle that reducing suffering is good, and (b) this would most likely reduce suffering. (We should not be so naive as to think that a Cato Institute employee, such as Wilkinson, is concerned to make sure utility never means anything other than individual interpersonally incomparable preferences just because of its theoretical elegance or just because it constitutes a judicious application of Occam’s Razor. It is that, but the fervour for it suggests something else: a neat little defence against the redistribution of wealth and income.)
Edward
1. It’s clear from comments both in Luxury Fever and the Principles of Microeconomics that Frank has endorsed a psychological interpretation of ‘satisfaction’.
2. Thanks for the Alchian paper. I’ll have a look. On a related point. If economists claim (following Robbins) that economics has nothing to say on normative issues then they should say nothing on normative issues (eg not rail against ‘distortions’ in the market place as if ‘distorting the market’ was a mortal sin).
3. The policy implications of — for example — hedonic utilitarianism, Rawlsian contractarianism, Sen’s capabilities approach and Nussbaum’s promotion of ‘human flourishing’ are different.
All of these lead in different directions from orthodox welfare economics and non-cognitivism leads nowhere.
Of course the positive contributions of economics matter for policy, but only in the context of a normative framework.
Normative debates are important and difficult to resolve. Far too many economists act as if they are unimportant and easy.
Either economists develop a normative position and accept responsibility for defending it (ie move beyond having ‘an opinion’) or they make their positive contribution to the debate and step back.
Edward Mariyani-Squire at #3 has it right. The elephant in the room here is interpersonal comparability. If you believe in cardinal, interpersonally comparable utility, you can say things like “another dollar to a poor family is worth more than another ten to James Packer”, and draw the obvious policy conclusions.
This implication of Benthamism is what the 1930s critique of people like Robbins was aimed at getting rid of, with some success as James notes at #1.
The natural requirement for cardinal utility in the context of choice under uncertainty provides a bridge from individual decision making to social choice via thought experiments like those of Rawls and Harsanyi.
As the mention of Rawls indicates, it is not really important here whether you accept every detail of utilitarianism – the implications are going to be much the same.
Economics is about efficiency, Don — trying to identify and measure potential welfare gains. That’s what motivates the concern with distortion, just as it might motivate John Quiggin’s advice to regulate a market that ia functioning poorly.
When someone like Robbins says he wants to steer clear of normative judgements, he means questions of equity and justice, not ones about how well the system is performing. It doesn’t mean he wants to investigate economic behaviour with the indifference of an entymoliogist studing a termite colony.
This is interesting.
John – In your last para are you suggesting that there’s not much difference between Rawls’ position and utilitarianism? (Not just on the isolated issue of income redistribution.)
James – If you say (or assume) that efficiency is a good thing, then you are making a normative judgement.
Perhaps what you mean to say is that it’s so obvious that efficiency is a good thing it doesn’t require an argument?
James – I’m confused. I don’t understand why you say:
Firstly, Robbins made it clear that he meant to exclude all normative judgements from economics — not just the subset you pick out. For example, he writes:
Secondly, I think Robbins expected economists to behave in the same way as scientists like entomologists. In one passage of his Essay on the nature and significance of economic science, he likens economists who have views on normative issues to botanists who have views on the lay out of gardens. He regarded both as examples of “outside interests”.
Of course Robbins had no problem with economists having outside interests. He just wanted to make sure they didn’t attach the authority of economics to their normative pronouncements.
Curiously, this is itself a normative judgement. So presumably Robbins is not making it as an economist.
I’m not picking out a subset — I just mentioned the main ones that Robbins warns us to steer clear of.
There’s really nothing to be confused about. Economic science is concerned with the efficiency of achieving given ends with given resources, not with evaluating the ends themselves. The sentence preceding the one you quoted is:
The botanist can advise on which shrubs to plant where to maximise shade, or how to allocate scarce fertiliser and water between plants to maximise overall flowering. But his science doesn’t help to determine whether these are desirable objectives.
I mentioned the entomolgist because he presumably has no interest in advising the termites even on matters of efficiency, let alone how to lead the good life.
The end-means distinction is starightforward in principle. The real problem with it is that it’s often hard to distinguish between ends and means, because many ends are only intermediate. Is economic growth an end in itself? If it’s only pursued as a means to happiness, for example, then the positive economist is entitled to dispute that growth is desirable if he can show that it doesn’t bring happiness.
Don @7. Rawls thinks there’s a big difference, but there isn’t. Starting with the theory of choice under uncertainty, it turns out that Rawls is advocating a polar case of the rank-dependent utility model which I’ve worked on for the last thirty years – the difference principle is the translation of this model to assessment of outcomes for populations. Utilitarianism is another special case. There’s a paper by Ebert ‘Rawls and Bentham reconciled’ which spells this out.
John – If I was wondering what to read over the holidays then you’ve solved that problem!
I missed you posts on utilitarianism and Rawls the first time round: here, here and here.
My first reaction is to say that there’s more to Rawls than maximin and a quite a bit of unpacking to do when Rawls uses terms like “everyone’s advantage”.
Without you pointing it out, I wouldn’t have made the link between Rawls and your work on anticipated utility (I wouldn’t pass an examination on this topic, so I can’t say more).
Intereting.
James – I should probably be more upfront about my position on Robbins.
Robbins’ position on ethics is noncognitivism — he would say that ethical statements refer to subjective preferences not objective facts.
Noncognitivism is a superficially attractive position for people who consider themselves tough minded. I agree with Hilary Putnam’s response to it (see his Reason Truth and History).
Robbins’ approach to economics reveals a preference for efficiency.
If you do not share this preference then you can safely ignore his economics. He allows himself no come back to this response.
Having a ‘preference for efficiency’ wouldn’t fit into Robbins’ scheme, unless it just means behaving rationallly — that is, choosing the least costly means to obtain a given objective.
Economists sometimes talk about ‘efficiency versus equity’ as competeing criteria for judging some state of affairs, but they are talking loosely — using efficiency to refer to GDP or some other tangible measure of productive output. That obviously isn’t what Robbins means by efficient or ‘economical’, to use his preferred term.
As for Robbins’ ethics, I can’t see anything in the book that excludes the possibility of objective morality. He just doesn’t think that deriving moral truths is the province of economic science. Does he say somewhere else that ethics are just a matter of preference?
James:
In his “Nature and Significance of Econ. Sc.” (the 1935 edition) Robbins does seem to commit himself to what today would be called non-cognitivism. This is to be found in point 4 of chapter 6 – esp. p. 150 – where he writes:
Robbins seems to be following Weber in this (he cites W. as his authority in a footnote on p.148). My mate Milton says almost exactly the same thing in that essay.
That said, although Robbins himself seems to hold to non-cognitivism it does not follow that Robbins was necessarily right in thinking this is the only position that coheres with the rest of what he says. I imagine one could provide a naturalistic account of moral values that might enable one to say there are certain objective values, while simultaneously accounting for Robbins’ claim of (seemingly) irrational fighting over ends.
Don:
This is a matter of interpretation, I think – and one that would hinge on knowing what Robbins (by his own account of values) ‘felt in his bones’. There’s obviously a methodological difficulty here in establishing whether Robbins did ‘feel in his bones’ a deep and abiding love of efficiency (or in his definition of the term, rationality, which comes to pretty much the same thing). It seems to me the safer interpretation is that Robbins was merely stipulating a definition of ‘Economics’: that it is the science of determining the least costly (most efficient) means of obtaining some given end. He doesn’t demand that the reader like this definition or demand that they adhere to anything that arises out of it. At best, on his own terms he can only offer a conditional of the following kind (quoting Robbins himself, p.149):
Applied to his definition of economics itself, putting words into Robbins’ mouth, we would have something like: “If you don’t want to waste scarce resources in obtaining your preferred ends (i.e. if you want maximally satisfy your preferred ends), then you should follow the method set out in my book, and if not, then I have nothing much to say. It’s up to you.”
All that said, I think Robbins unduly simplifies the arena of conflict over values to seemingly unproblematic ‘basic ends’. As you would know, this is something both Sen and Putnam put the boot into too. Nonetheless, that, I think, is a slightly different issue.
Edward, I agree with the part addressed to me.
As for the point addressed to Don, I’d prefer not to even flirt with the notion of a ‘taste for efficiency’. In Robbins’ scheme, if people are rational, they choose the most efficient route afforded by the information they have avaialble to them.
You could say that, given economics’ origin as a moral science, Robbins should have shown more curiosity about the meaning of welfare, the goals of economic behaviour, and the connections between those goals and social order and happiness. For Smith, Mill, Marx, Marshall and Wicksell, those questions were the starting point of their intellectual journeys.
But that doesn’t mean there’s no merit in Robbins’ attempt to separate the technical issues from the moral ones, which will always be contested ground — even if they are not merely matters of ‘opinion’.
Indeed Mill attempted to make a similar distinction; the difference is that, while he knew about gains from trade, he lacked most of the neoclassical apparatus of production functions, consumption functions and static efficiency gains that provide Robbins’ economist with a terrain to work on, so resource allocation alone would not have seemed a big enough topic for a science.
James:
Since Robbins took it to be a sheer fact that people were (for the most part) rational (as he defined the term), certainly the question of whether one should be rational does not arise in his mind; the ought-question is moot in the same way that, say, the question “Should I [choose to] be a mammal” is pointless.
However, irrespective of what Robbins himself thought, it would seem that the philosophical framework – empiricism – if followed through to the letter, only makes the above ‘pointlessness’ of the question “Should I be rational?” contingent. The sheer fact that people are currently rational is not, from an empiricist perspective, set in stone by any objective law of nature. Such lawful behaviour, at base, persists only as long as people choose to behave as such. Nothing objectively ’causes’ that choice to behave rationally. Thus, Robbins’ philosophical framework – though not Robbins himself – implies the possibility of there being a choice between being rational or irrational, thereby leaving room for ‘tastes’ for rationality or not.
As you may have noticed, something is going wrong here. As posited by Empiricist in Chief, Hume, Reason (rational use of means) and The Passions (ends) are supposed to be walled-off from each other, but empiricism’s epistemological implications seem to undermine the wall between them. One way around it is to just dump a thoroughgoing empiricism for some mild-mannered realism.
But I get the feeling this is, again, verging on the esoteric.
Edward
According to Robbins, people can choose to act irrationally. At the very end of the book he warns against those who yearn “for the deep unawareness of the unborn state” and refuse to reflect on their ends (it reads like a speech by one of Ayn Rand’s heroes).
For Robbins ‘rationality’ doesn’t just mean taking the most efficient means to the satisfaction of given ends. Taken at face value, many of our ends will be mutually inconsistent. Robbins seemed to think that by reflecting on our ends we can resolve these inconsistencies.
Since economics is a tool for doing exactly that, he acknowledges that:
I suspect that many economists believe that the population is in the grip of ‘irrational preferences’. They are unwilling to treat preferences for social justice or racial superiority in the same way as they treat preferences for chocolate or flat screen tvs. They have no idea what to do with preferences based on group identities. And they refuse to accept that some ends will remain mutually inconsistent.
Robbins’ economics is banal because it has no capacity to comprehend tragedy. Reflecting on incompatible or incommensurable ends will not always lead to a harmonious resolution.
Here’s Robbins on the rationality of ends:
Don:
Quick response for now.
(1) I can’t find any reference in Robbins’ Nature & Signif. to the words for the deep unawareness of the unborn state. Can you supply a page number? Ta.
(2) At the end of the book he refers to another notion of irrationality that is related, but not identical, to the ‘inefficient means to achieve given ends’ one. This second one – he indicates its different with the words “a further sense” and “in the sense that” – refers to a situation where people want incompatible or unattainable ends because (and only because) they are ignorant of the fact that the ends are incompatible or unattainable.
(3) RE the passage:
whoops:
In (2), delete “(and only because)” due to what he says as quoted in (4).
In (4), re “…and Economics as Robbins defines it is pretty useful”, useful should be useless.
It’s in the very last paragraph.
Don seems to be blaming Robbins for not being Friederich Nietzcshe, which strikes me as a tiny bit unfair.
Edward
I’ll do 1 now and have a think about 2 – 5.
You’ll need the first edition (1932). It’s on p141.
The Mises Institute have pdf versions of both the first and second editions.