The concept of preference plays a central role in economic theory. Surprisingly however, a monograph dedicated to precisely this, and trying to give us a coherent view on how this concept functions or should function as a building block in various branches of the discipline, has been missing. Daniel Hausman's recent book is a welcome addition.

The book consists of three parts. Part I, Preferences in Positive Economics, is the largest; it contains six chapters, devoted to the role preferences play in descriptive economics. Hausman argues for a specific conception, preferences as total comparative evaluations, which is of crucial importance throughout the book. He also defends it against other notions of preference, which are misconceived in his view. There are clear expositions of utility and how the axioms of utility theory work. This part also contains nice discussions of consequentialism and game theory, and how these relate. Part II, Preferences, Welfare, and Normative Economics, contains two chapters on the question of how preference satisfaction relates to welfare. Inter alia, Hausman is critical of the ‘laundered preference’ approach. Part III, Psychology, Rational Evaluation, and Preference Formation, also contains two chapters. Here Hausman discusses recent research done by psychologists and behavioural economists on choice behaviour and how these findings might inform empirical work on preference formation, but also the normative question of how one ought to form one's preferences. Overall, it is a short and compact book, and a pleasurable read.

A central theme in Preference, Value, Choice, and Welfare is how to conceive of preferences in the discipline of economics. Hausman argues for preferences as total comparative evaluations. Together with corresponding beliefs, they determine choice. By being comparative, preferences differ from desires. I may desire both chocolate and strawberry ice-cream but I cannot prefer both. The evaluations are total in the sense that they take everything that is relevant into account. For example if a person wants to act on a promise, this should not be seen as countering his preferences, as a constraint, but as something that influences or shapes his preferences. Ordinary language usage can differ from this conception, Hausman admits, but this is how we should interpret preferences in economics – it is the most adequate conception of a technical term. Hausman is explicitly legislative on this issue:

Regimentation is necessary because, apart from the fact that ordinary usage varies, the common practice among economists is not as nearly unequivocal as Hausman wants it to be. Amartya Sen for example allows for different uses of the term, and also gives moral commitments a distinct role in determining choice, outside of a preferences structure. Sen's idea is that moral objectives function as what Robert Nozick has called ‘side constraints’.

Both Hausman and Sen agree that economists must be aware of the complexity of human motivation and make room for behaviour that is not self-regarding. Both find it important to provide a place for commitments and other-regarding motives. However, they disagree where to put them in the theory, where exactly their place should be. For Sen, moral commitments and motives are constraints, they counter one's preferences, while preferences are self-regarding. For Hausman, on the other hand, a moral commitment is a consideration that feeds into one's final comparative evaluations, i.e. preferences as he conceives of them, while these can also be other-regarding. He finds it more ‘straightforward to model [them] as features of preferences’ (p. 62).

Now it can be asked whether these two conceptions are substantially different after all. Hausman gives several reasons ‘why economists should reject Sen's advice and instead adopt a single notion of preference as a total comparative evaluation’ (p. 64). According to him, this notion best accords with general practice, and it would also clear away some confusion. Perhaps most importantly, he argues that this conception is more fruitful: it ‘permits game theory and expected utility theory to serve their predictive and explanatory roles’. Hausman then illustrates this using a particular model of social preferences. Regardless of the merit of the illustration, I found it a pity that Hausman does not explain this also contrastively and has not made clear that a preference–constraint–choice model would have been hopeless for such a case.

My impression is that Hausman seems more wedded than Sen to the idea of economics as a science in the sense of Thomas Kuhn. A science, Kuhn famously argued, is characterized by a commonly accepted paradigm that includes concepts that are relatively precise for all the researchers who work within its boundaries. You do not want scholars to waste their time disputing each other's interpretations of the paradigmatic concepts. When this happens, it usually means that the paradigm has entered a state of crisis or that a discipline is stuck at a pre-paradigmatic (or pre-scientific) stage.

Another important contribution of the book consists in a discussion of the distinction between monetary or physical outcomes on the one hand and comprehensive outcomes on the other. (Again Hausman is debating with Sen here.) This distinction is important, Hausman says, ‘because people may have reasons for preferring paths through game trees, including their own strategies, that do not derive from preferences among the results’ (p. 52).

At the same time Hausman says that economists standardly employ a consequentialist model of choice:

More specifically, a model of choice is consequentialist if and only if:

1. 1. An agent's final preferences derive from

   1. a. the agent's beliefs about the properties and consequences of the alternatives, and

   2. b. the agent's preferences over these properties and consequences.

2. 2. An agent's choices causally depend on the agent's beliefs and the agent's final preferences among the alternatives, given the constraints, can be chosen. (p. 43)

This consequentialist approach would for example allow the inclusion of a social preference like ‘inequity aversion’ (a term introduced by Fehr and Schmidt) in the utility functions of players involved in mixed motives games. Inequity-averse agents in fact have a clear-cut consequentialist goal in view, namely to reduce a difference in monetary pay offs. Consequentialism rules out, however, typical cases of sensitivity to the history of play, e.g. when people act on trust and display ‘chooser dependence’. As for trust, in a sequential prisoner's dilemma it may make an important difference for the second player whether the first player has intentionally made a cooperative (trustful) move or whether this move was produced by a randomizing device or by trembling hand. ‘Chooser dependence’ occurs when you don't pick an apple from a bowl because it is the last one. According to standard economics any apple is just an apple, but for a polite person it can make a difference whether somebody else leaves the last tasty apple in a bowl for her to take, or whether she herself is in the position to make that choice, thereby possibly excluding it as an alternative option for somebody else. Hausman acknowledges such cases and does not treat them as atypical. So on the use of the consequentialist model of choice in economics Hausman seems less legislative than on other issues discussed in the book.

From Hausman's conception of preferences as total comparative evaluations and his nuanced view on the complex of human motivations follows that he must reserve an important role for preference formation. This he does, throughout the book and in two special chapters. As an example from consumer research Hausman refers to the well-known case of Corona beer. Corona spectacularly increased its market share by building an association in consumer's minds between the attributes and consequences of drinking a beer and relaxing on a beach. Hausman also mentions Robert Frank's work on relative income and status goods, and many other examples. These cases demonstrate that economists’ excursions into the territory of psychologists are likely to bear fruit. Economists can usefully broaden their models of decision making and choice behaviour by accommodating findings on what shapes and influences preferences. Taken together, much of the presented material in these chapters appears not very systematic for a general theory about preference formation, but that is not Hausman's fault. It is just the state of the art.

Where he is dealing with more developed research, however, Hausman is able to delve deeper in the issue of preference formation. The heuristics and biases research in the tradition of Tversky and Kahneman, for example, is discussed at length in the book. As is well known, loss aversion, framing, preference reversals and adaptation violate the standard axioms of rational choice theory, perhaps most notably the axiom of context independence. Hausman proposes to treat these robust anomalies as stemming from factors that influence people's preferences, e.g. context sensitivity, and suggests indeed that economists drop the context independence axiom. Heuristics and biases must be incorporated in models of preference formation.

The orthodox view in economics is of course that preferences are exogenously given. However, it seems fair to say that many empirically minded economists tend to see this as just a provisional stance. Hausman also makes clear that more data and models on preference formation would enrich the field. Moreover, his consequentialist model of choice can give one a handle on the normative question how one ought to form preferences. For example, in the sub-discipline of health economics people can be asked to compare health states that are quite alien to them.

It is a stressful thought experiment, but for a rational comparison one indeed ought to consider the values of the various attributes and consequences of these health states, Hausman explains. One would have to gather information about how paraplegia and deafness affect people, try to imagine how these deficiencies impact on herself, given technological resources, her personal circumstances and what she generally value in life.

Overall, the prospects for a general theory of preference formation are bleak, says Hausman in his concluding chapter. ‘But useful theories can have narrow scope. Narrowly focused models of preference formation and modification are already in use’ (p. 135).

There are two things that I find particularly likeable about this book. One is that the debates that Hausman engages in are alluring and germane to both philosophers and economists. Hausman does not just speak to an inner circle of philosophical experts. Secondly, and relatedly, Hausman's methodological appraisal is well rooted in the field of economics. He has for example a keen sense where to strive for more generality and where it seems wise to applaud partial models.

The book intends to be interesting for different audiences: economists, philosophers of economics and philosophers working in other areas of the discipline (especially ethics). I believe it succeeds in this. The book will interest economists who like to think about conceptual topics and frontiers issues in microeconomics, game theory, and welfare economics. Philosophers of economics will see that the book makes various substantial contributions to the literature. Last but not least, Hausman's writing style allows other philosophers to follow these discussions without effort. Where necessary, the book contains brief and non-technical introductions, for example to the axioms of rational choice theory, to expected utility and welfare economics. It is commendable reading for philosophers who work in the areas of practical rationality, social theory, ethics and human well-being.