McFadden’s Pleasure

Daniel L. McFadden THE NEW SCIENCE OF PLEASURE NBER Working Paper 18687, January 2013

This provides a strong challenge to the neoclassical model of the individualistic utility-maximizing consumer, on behavioural grounds. In practice, people are not ‘rational’ utility maximizers. But the paper does not explicitly challenge the view that people ‘should’ be utility maximizers. I wish to make this challenge.

I.  Introduction

The neoclassical model of the individualistic utility-maximizing consumer [that] forms the basis of most economic analysis is largely a finished subject … .

II. Pleasure, Pain, Utility

Bentham thought about the pursuit of happiness in ways that did not fit into the later neoclassical synthesis, but which resonate with contemporary behavioral studies.

Dupuit’s idea of …  recovering utility from demand … remains today the standard approach to measuring and predicting consumer welfare. … Self-interest was defined narrowly to include only personally purchased and consumed goods; reciprocity and altruism were ignored. No allowance was made for ambiguities and uncertainties regarding tastes, budgets, the attributes of goods, or the reliability of transactions. The Hicks-Samuelson formulation was fundamentally static, with the consumer making a once-and-for-all utility-maximizing choice of market goods.

It is not necessary to deny that individual preferences may vary over life-times, only that the net effect is stable. It does not allow for the impact of advertising or fashion, for example.

The fundamental consumer sovereignty assumption of neoclassical theory requires that [consumer attributes affecting choice] not depend on opportunities or choice.

This sounds reasonable, especially for mature markets, but it is not so obvious that it is always true. It seems to preclude consumers from delaying their choice of a new type of product (e.g., tablet) to see how the market develops.

IV. The Stone Age

The utility-consistent demand systems mentioned above generally worked well to explain demand at the market level despite the representative consumer restriction.

But many products are now much more segmented than they used  to be.

V. Consumer Well-Being

[When the economy changes, it is hard to estimate] the net increase in income necessary at initial prices and utility level to offset the change in non-market attributes … . Tacit in most applications of neoclassical welfare analysis is an assumption … that changes in [the environment] can be translated into changes in effective market prices, and rolled into the consumer surplus calculation.

VI. Expansions

These systems could not deal easily with preference heterogeneity, acquired tastes, shifting hedonic attributes of commodities, non-linear budget sets, time, space, or uncertainty, and the frequent cases of zero and lumpy purchases. … This meant preserving the tenets of consumer sovereignty and preference maximization, but admitting the influence of (observed and unobserved) experience and memory on perceptions and on current preferences, leading to heterogeneity across consumers.

In Debreu’s interpretation, utility spanned the lifetime of the consumer, with a single decision-utility-maximizing choice specifying in advance the response to the realization of each contingency, and determining the entire life course. This was a complete, logically elegant, and instructive implementation of consumer theory, with utility incorporating a complete system of perceptions and subjective probabilities, and including in the life plan of the consumer full allowance for the strategic impact of choice on later options and preferences. Nevertheless, the approach has severe limitations, … fundamentally because it is clear from behavioral evidence that life plans … are subject to continual updating and revision.

However, the strongest form of the [dynamic stochastic programming] model, with a representative consumer and rational expectations, is vulnerable to behavioral rejection, because the solution of these programs involve levels of complexity and computation that fairly clearly exceed human cognitive capacity, because it is unrealistic to assume that historical experience and market information and discipline are sufficient to homogenize subjective expectations, particularly for rare events, and because the axiomatic foundations for utility jointly additively separable in time and uncertain outcomes are not persuasive.

McFadden is drawing attention to some of the limitations of ‘standard’ utility theory, which relies on ‘standard’ (Bayesian) probabilities. It may be that some of the wider concerns of Keynes also matter.

VII. New Frontiers: A Behavioral Revaluation of Consumer Decision-Making

Neoclassical consumer theory implies that with rational calculation, we cannot be harmed by choice and trade. … Yet, people … often use procrastination, rules, pre-commitment, habit, suspicion, and imitation to avoid “rational” decision-making and trade. The psychiatrists even have a word for it – agoraphobia, or fear of the market. There are two possible reasons for this behavior. First, while trade is calculated to advance our self-interest, the calculation may be burdensome, and the cost of mistakes substantial. We may simply be too lazy or timid to trade. Second, trade involves social interaction and the emotions that go with this. Choice alternatives and trades may be misrepresented in the market game, and suspicions may be justified.

McFadden appears to identify ‘rational calculation’ with utility maximization, and to suppose that it cannot lead us to harm. But it may be that people avoid “rational” decision-making in circumstances where it is unwise. For example, in mid 2008 it seemed reasonable to delay investing in stock markets.

Neoclassically trained economists think of these behavioral elements as arising from the limits of memory and cognitive capacity that bound rationality, slips or anomalies that the individual will detect and correct if they become obvious. Many psychologists and biologists think of this instead as a product of evolution, the result of a rough correspondence between generalized self-interest and survival, a hodge-podge of rules, processes, and strategies that mimic rationality in circumstances where rationality increases survival value. Day-to-day economic choices are explained by either paradigm, but perception and choice in novel situations tests the neoclassical premise, and challenges easy transitions between conventional demand analysis and the effect of novel economic policy on consumer well-being.

Again, departure from utility maximization is characterised as a ‘slip’ or ‘anomaly’, seeming to preclude deliberate, soundly-based deviations from utility maximization. ‘Deviations’ are said to follow a ‘hodge-podge’ of methods, which would seem to preclude them being soundly based. But could it simply be tjhat psychologists do not understand decision-making under uncertainty? (My view is that people are far from perfect, but that some of their ‘slips’ are actually more sensible than the assumed norm of utility maximization.)


An experiment where half a class is given pencils which are then traded, seeming to reveal that those given pencils value them more than the others. But other explanations are not precluded. For example, the class is full of economics students, who perhaps would not wish to be caught paying ‘over the odds’, or accepting ‘under the odds’. By bidding low or only accepting high bids they can avoid looking foolish.


[C]onsumers will often refuse to take any share of either side of an offered lottery, a result consistent with the observed paucity of real-world wagers. Kahneman and Tversky attribute these effects to an editing process that determines the reference point and the perception of lottery outcomes as gains or losses, and to systematic misperception of probabilities. An additional reason that individuals are ambiguous about lotteries, and often avoid them, is the superstitious belief that there are hidden causal forces at work, interventions that place the lottery in ambiguous relationship to the rest of life.

If a magician, trickster or psychology experimenter offers you a gamble, maybe it is not superstitious to think that there might be ‘hidden causal forces’ at work. Even with tricks, there may be good reason not to gamble, for example if you can’t afford to lose.

Suppose that you want to invest in the stock market to build up a pension pot. The gains this year give you more to invest next year, so the yearly returns multiply. Hence you should seek to maximize log value, not pure value. This means that you are loss-averse in terms of value, and so should pass over some investments that have a positive expected return, if their possible down-side is too great. As you get near retirement it is prudent to become even  more risk-averse, and it is not at all obvious that the best strategy is equivalent to maximizing any kind of utility.

Superstition can arise and persist even when people are consistently Bayesian. Start with a prior that admits the possibility of complex, hidden causal paths. The experiments that life offers, and selective memory of outcomes, allows these cognitive castles in the air to survive; see McFadden, 1974c; Hastie and Dawes, 2001.

McFadden , quite credibly, is claiming that flawed Bayesian reasoning will not necessarily tend to remove the possibility of ‘complex, hidden causal paths’. He seems to suppose that idealised Bayesian reasoning would, and that this would be a ‘good thing’.


Hyperbolic discounting occurs when individuals systematically underweight future consequences relative to contemporaneous ones, and make choices that gratify now and leave lasting regret, in patterns that cannot be explained by maximization of consistently discounted present value of instantaneous utility.

McFadden seems to suppose that ‘consistent discounting’ of utility is a good way to reason, and that underweighting future consequences will ‘leave lasting regret’.

VIII. The Future

New results challenge the standard assumption of maximization of individualistic utility, indicating that social networks as information sources, reciprocity, and altruism enter human behavior and cannot be ignored.

The challenge facing economic consumer theory is to utilize the disparate measurements and experimental methods that have become available to synthesize a new behavioral science of pleasure that retains the quantitative, predictive features of neoclassical theory in the economic settings where it works well, and extends these features into areas of individual sensation of well-being and choice in the context of social network information and approval, so that the theory can better predict the impact of novel economic policies on consumer well-being.

McFadden seems to have supposed that ‘consumer well-being’ is represented by an appropriate utility, whereas consumers reason irrationally. The aim, then, is presumably to give consumers ‘nudges’ or otherwise arrange things so that consumer utility tends to be maximized despite their irrationality. In particular, the aim seems to be to get consumers to gamble more when their is a positive expected return, presumably by ‘improving’ their expectations.


  • This paper seems to assume that utility maximization is always best. In VII.1.2 it even seems to suppose that there is some ‘objective’ utility function that ought to be maximized. Its assumptions are not justified, and appear wrong.
  • The assumptions make more sense in stable situations, where decision-makers have ample evidence. But is this ‘the real world’?
  • It regards conventional time-discounting as optimal, and supposes that underweighting future consequences would be regretted. But this is not justified, and it seems unlikely that the real world is always as it is conventionally conceived. (For example, the financial difficulties of 2007/8 seem to be a counter-example, where ‘rational’ investment strategies were regretted.
  • It is regarded as ‘a bad thing’ to suppose that ‘hidden causal forces’ might be at work. yet the alternative seems to be to suppose that we have full knowledge of all the actual and latent causal forces.

Most worryingly, the trend of the conclusions would seem to support the view that one should act to improve people’s expectations, so that they become more ‘risk neutral’. Yet isn’t this what financiers and governments were doing before the crash? Might it not have been better to uncover the ‘hidden causal forces’?

Dave Marsay

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: