« Chesterton, & inflation | Main | Rennard's lesson »

February 27, 2013


Luis Enrique

it's a common idea that angst is a luxury only the rich can afford. I don't see why a weak link between wealth and happiness, or choice and happiness, should come as a surprise to anyone. And, as I expect you agree, there's more to life than happiness. Even if wealth might not lead to happiness, that does't mean poor people won't benefit from being richer.

I think there's a lot of be said in favour of the basic econ approach supposing we have preference orderings over consumption items etc. but it unlikely to have much to say about the unhappiness we experience when we realise we're never going to marry Cate Blanchett. I'm puzzled by the quote from Guthrie and Sokolow. Presumably they've got one definition of utility, and then they refer to another thing, "discontentment", which isn't in the utility function, so they can say that you might increase your utility and get unhappier in the process. I think it would make more sense to say that the utility function need redefining/expanding so that the discontentment arising from the introduction of a cinema decreases your utility.

Luis Enrique

on second thought, if people want to define utility as something different from happiness, or overall welfare, well fine, that's just semantics.

I think you raise an interesting possibility here. Standard econ utility max, where utility is based on consumption baskets, is usually criticised on basis of being unrealistic (which of course it is, having nothing to say about many dimensions of behaviour. This is a feature not a bug.) But if I understand you correctly, you are suggesting that people actually do behave in something like a u-max fashion, yet this does not lead to happiness. They would rather live in a town with a cinema than one without, because utility is higher with a cinema than without, but they'd be more miserable because the existence of choice makes people less content with whatever they actually choose.

This would also explain why people do things like work long hours in miserable jobs (those that have any choice in the matter, I mean). It's as if people are maxing only consumption and are thus unhappy, whereas maxing happiness would encompass family, leisure, "flow" etc. So crude consumption based u-max then becomes a more realistic description of how people behave than more general forms of utility function. I always presumed that a more realistic utility function would incorporate things like dignity, fairness, "flow" etc. but perhaps not.


There is an obvious logical error from these Guthrie and Sokolowky people.

Economists, as you demonstrate, have no idea what "utility" means.

From Wikipedia: "Utility is usually applied by economists in such constructs as the indifference curve, which plot the combination of commodities that an individual or a society would accept to maintain a given level of satisfaction."

G&S: "An expanding choice set off ers opportunity to increase one's utility, but it also evokes rising discontentment."

Spot the thinking mistake? Utility as applied to indifference curves (by morons, who are the only ones who can tolerate this nonsense, I hasten to add) is defined by EQUIVALENCE BETWEEN SETS OF CHOICES. WHERE INDIVIDUAL PREFERENCES ARE CONTINUOUS, COMPLETE AND TRANSITIVE.

G&S have then applied the above definition of utility to derive that transition between choice set A, and larger choice set A+B increases utility.

They have then thought up ways that this utility increase might lack a commensurate increase in happiness.

BUT THIS IS INCORRECT. If the agents would overall be happier or no different in choice set A than choice set A+B then there is no increase in utility by adding choice B. Rational agents (!) would NOT prefer the more miserable set of choices.

So, to derive this apparent paradox between utility and happiness, G&S in fact have to violate one of the follow assumptions:

1. Rational agents.
2. Transitivity of preferences.


Luis - It is not a feature, the model is internally inconsistent and also inconsistent with empirical facts.

Classical U-max microeconomics makes no sense whatsoever. There is no reason, in theory, for a complete, transitive and continuous indifference function to be assumed to exist in a model of individual differing rational agents.

Furthermore there is no way to derive welfare maximization from any such function.

So the whole thing is just a waste of time.


To put this more concisely, if the addition of the cinema option makes people who choose to stay at home less happy (because of new opportunity cost)then it decreases the utility of staying at home.

Yes, the utility, not just the contentment.

If for some reason you can't understand that this is entailed by the definition of utility in terms of preferences, then imagine instead of staying at home a therapeutic de-stressing meditation session. The utility of that would obviously be reduced if - during it - you are stressing about all the other choices you might have made.


I've read the paper now, and as expected it is drivel.

"Since Jack's value of work and his
time allocation remain unchanged, the utility he derives from work should be unaltered. For
Jill, on the other hand, to choose leisure over work at the margin, it must be the case that
additional leisure yields higher additional utility than what she gives up in forgone work."

Wrong. Preferring new option B over option A does not entail that choice set A+B has higher utility than choice set A.

Luis Enrique

Andrew I am reluctant to engage with your because you seem solely concerned with calling people like me morons, but I am intrigued.

My micro lectures are a distant memory, but I'd have thought if I faced choice of two from orange, apple banana and chose orange and banana, then if I was offered expanded choice set with addition of kiwi fruit and chose that, my utility must be higher (unless for some reason I am indifferent, an uninteresting possibility)


Luis, I'm not *solely* concerned with that. Thanks for responding.

The whole thrust of this S&M post is that expanding choice may have costs. These costs may be trivial if the choice is limited to three fruits, but non-trivial if it is between uncountable options.

My point is that either:

1. this idea is true, in which case the utility (not just the contentment) of the pre-existing option is decreased by the new choice (e.g. added stress whilst doing a meditation class). In that the paper seeks to contrast increasing utility with decreasing happiness it is wrong. The utility goes down too.

2. It isn't true, in which case there is no contentment penalty from increasing choice and the paper is still wrong.

The authors and Chris equivocate between the additional positive utility of a single new "commodity", and the *total* utility of the set of choices.

Classical micro (and therefore macro) is based on indifference curves between *entire* baskets of commodities or choices sets.

You can't claim that a rational consumer would opt for basket "miserable" over basket "happy", without doing violence to the assumptions of classical utility maximising micro/macro.

Of course I believe that classical micro/macro is entirely deserving of extreme violence for crimes against reason, but that is a wider issue.


Chris, thanks for posting on our paper – I love your take on it. Luis and Andrew, thanks for engaging in this discussion. You’ve brought to light some potential misunderstandings, and I’d like to take this opportunity to clear them up.

(i) Andrew: you are focusing on the choice over choice sets. In contrast, in the paper we treat the choice set as exogenously given. In this setting, it doesn’t matter whether one maximizes utility or minimizes discontentment – both yield the same optimal choices from within the given choice set, and consequently, there is no need for utility to subsume discontentment to be a valid representation of preferences over choices within the choice set. I’ll expand on this below.

(ii) Luis: as you state, we could use different words to capture what we call utility and discontentment (e.g., standard economic consumption utility and über-utility). But redefining the terms does not change the message of the paper. Regardless of which terminology we use, the paper explicitly characterizes a novel dimension of happiness/über-utility.


In fact, had we used the alternative terminology, Andrew’s concern would not have arisen in the first place. So why did we choose not use that terminology in the paper? We wanted to draw attention to a neglected dimension of utility and to highlight a novel perspective on the causes of unhappiness. I think this is better achieved by differentiating between the common interpretation of utility and the novel contributing factor we term discontentment. And with the choice set fixed, there is no need to be technically precise, because utility maximization and discontentment minimization lead to the same choices.

So what’s the common interpretation of utility we’re trying to get away from? To most economists, students of economics, and policy makers, what comes to mind are the benefits from the consummation of choices (i.e., benefits flow from the things we choose to have or do). Yet, one can predict the same choices by looking at the opposite side of the same coin, namely by looking at the value of the alternatives not chosen. Since the value of the alternatives not chosen is based on the their “standard economic consumption utility”, the concepts of utility and discontentment are inextricably linked, yet distinct. While the two theories of the determinants of über-utility are observationally equivalent, they are psychologically distinct. If consumption utility is the yin, discontentment is the yang.


So when the choice set expands due to technological progress or life cycle effects or changing societal norms (all of which are outside an individual’s control), the effect on “standard economic consumption utility” is different from the effect on discontentment. This is the important insight of the paper. If, on the other hand, we were concerned with choices over choice sets, then it is of course important to recognize that preferences over the choice set are validly represented by happiness/ über-utility, not just “standard economic consumption utility”. Note that in the setting of choices over choice sets, discontentment would feature at two different levels: first, it is present within each choice set under consideration; second, it is present in the decision of having to choose one choice set over others (e.g., think of choosing between the bundles of living on the beach vs. living in the mountains). As you can see, the concept of discontentment can be fully characterized without resorting to choices over choice sets, which just makes things more complicated without offering additional insights.


Unfortunately, the implication of our theory is a bit depressing – discontentment arises from being constrained, the severity of which is largely beyond our control. I would very much appreciate your suggestions on how to wrestle greater happiness out of this insight.


Jan - Thank you very much for chippping in from the horses mouth, so to speak.

I still don't see how you have separated contentment and utility.

In the end you are still making a point that compares entire choice sets, i.e. the effect "when the choice set expands" as you say above. If you want to look at utility effects, then you are indeed talking about preference between choice sets, whether this is purely hypothetical or not. Utility is as I understand it defined by and only meaningful in terms of consumer preference.

If you wish to restrict the discussion to a single choice set (and I don't see where you have separated contentment and utility here either) then you forfeit the ability to make any inferences about what happens when choice sets expand.

See my point?

Luis Enrique


thanks for your response. If I understand you correctly, are you (intentionally) violating the standard "irrelevance of unchosen alternatives" assumption in decision theory? You are saying that our utility should be seen as depending on what's not chosen as well as what is chosen?


`Independence of irrelevant alternatives’ means that adding more elements to the choice set does not reverse the preference ordering of the elements that were already in the choice set. Alternatively, removing alternatives that were not chosen does not affect the preference ranking of the remaining elements.

But our theory is not about choices: Linking happiness to the value of foregone alternatives does not reverse the preference ordering of the elements. So we are not violating it in the conventional sense.

Instead, our theory is about how individuals feel about their choices, and we make explicit that choices do not materialize in a vacuum. As Luis puts it, “utility should be seen as depending on what's not chosen as well as what is chosen.” We posit that discontentment counteracts the consumption utility, and thus also affects the overall experience of choice.

It is in the sense of magnitude that we violate the independence of irrelevant alternatives axiom: The magnitude of happiness does depend on `irrelevant’ alternatives, but choices do not.


Andrew: We are not saying that discontentment is separate from utility. In fact, discontentment cannot exist without consumption utility (its absence causes discontentment).

Discontentment and consumption utility are two sides of the same coin. A choice for something implies a choice against something else. This perspective is able to reconcile the puzzle of rising incomes with flat happiness and many others, while retaining the rational choice paradigm.


Your place is valueble for me. Thanks!

cheap jordan shoes

The comments to this entry are closed.

blogs I like

Blog powered by Typepad