This is the third and last post to summarize our discussions about preferences. The previous two posts presented the question about the definition of preferences and the possibility to detect what they are. This post will highlight a different aspect of the discussion. Reviewing Erik Angner’s article (2018) and his critique of Daniel Hausman’s work on preferences, this post will be devoted to the presentation of Hausman’s view on preferences in the context of his broader philosophical work. Specifically, my claim is that according to Hausman, an inadequate definition of preferences may undermine the firmness of the common rationality assumption.
In the previous post, Yam introduced ILS’s criticism of Hausman’s work. ILS criticized Hausman on two main issues: a methodological one and a normative one. First, they argued that purifying preferences is practically impossible. Second, they argued that the role of normative economics is not to maximize welfare by satisfying preferences. Instead, it maximizes welfare by increasing the possibilities one can choose from.
Angner’s arguments against Hausman takes a different path. According to Angner, Hausman’s discussion misses three crucial points: He misunderstands the economic definition of preferences; He misses the methodological use of preferences, and; He ignores empirical evidence that shows how his understanding of preferences is incorrect.
All these, Angner argues, stem from Hausman’s misconception of the role of the philosopher. But, Angner himself fails to acknowledge Hausman’s critique as only one fragment in his broader outlook on the philosophy of economics. This post attempts to fill this gap.
Hausman’s Philosophical Outlook
Angner’s view is that philosophers of science should not intervene with the scientific practice itself. This, he says, should be left to scientists. The philosopher’s role is to reflect on the scientific method and on the tools scientists choose to use. Consequently, his role is to suggest a discussion about the implications of such uses, for example of correspondence rules of models and the relevant phenomenon they represent. When dealing with concepts or scientific terms, Angner endorses the theory of contextual meaning that accepts that concepts can receive their meaning from actual implementation in the relevant scientific theory. He therefore rejects and criticizes what he sees as Hausman’s reductionist view as he insists on a pre-defined concept of preferences.
The difficulty with Angner’s account is that he misrepresents Hausman. He neglects Hausman’s broader philosophical perspective and the place his criticism of preferences takes in this more comprehensive outlook. This is a major issue in Hausman’s philosophical project, in which he asserts that the philosopher’s role is not only to observe how scientific “vehicles” are operating or to reflect on the possible consequences of their operation. Instead, they should encourage scientists (in this case – economists) to look “under the hood”:
“Economists must look under the hood of their theoretical vehicles. When they find embarrassing things there, they must not avert their eyes and claim that what they have found cannot matter.” (Hausman, 1992)
The concept of preferences, following Hausman’s terminology, is one of these embarrassing things that can be found in economic theory. Although Hausman does not say it explicitly, I will follow his lead to show how the concept of preferences is embarrassing since it threatens and undermines the methodology of rational decision-making models.
Mainstream economists’ preference for the concept of preferences arises from the difficulty of adhering to psychological considerations, considerations that researchers have minimal access to. The need to break free from subjective assumptions about the utility function brought post-war economists to embrace the concept of preferences as revealed by consumer choices. The methodological justification for this move was that applying the concept of preferences, as revealed by actual choices, maintains the fundamental axioms of economics. Preferences used this way can be formally translated to form the consumer’s utility function (See Mickey’s post). In that sense, preferences were (and in many respects still are) used as a technical concept that neither reveals nor assumes anything about the inner psychic structure of humans.
The fragility of the rational-choice model
This use of preferences imposes some difficulties as for the justification of the rationality model. If revealed preferences are actual choices, there is still a mystery as to the nature of the decision-making process. In fact, there is no methodological justification to be exclusive with a model of a rational decision-making process (as some economists have argued).
According to Hausman, the justification for the rationality model is apparent only if we treat preferences as reasons (therefore causes) for actions:
“Every theory that takes beliefs and preferences to be reasons that explain choices must incorporate some theory of rationality. Since desires cannot function as explanatory reasons if they do not induce at least a loose ranking of what is better or worse and thereby influence choices, it follows that preferences must not be radically incomplete and that they must have a large measure of consistency.” (Hausman 2006 p.62).
Following Ronald Davidson’s concept of “effective reasons,” Hausman argues that explanations in economics involve both causes and reasons. Since desires cannot be regarded as reasons, this role is reserved for preferences. But this change in terminology is immediately followed by a strong need for a clear definition of the notion of preferences. If economists wish to keep the rationality model of choice, they need a clear concept of preferences. These preferences, alongside beliefs, will fill the role of reasons, therefore causes, for choosing. Failing to do so will harm the explanatory power of economic models.
This conclusion is even more significant once we accept the role of normative economics as maximizing consumer’s welfare by satisfying their preferences.
Hausman points to an embarrassing problem that arises from behavioral economics research: that “people’s choices diverge from the predictions of rational choice theory” (Angner, 2012; p. 664). The embarrassing problem is simple: If preferences are choices, there cannot be any deviation from the prediction of models. Since we find such differences – choices and preferences must be two different things. Hausman argues that the rationality model is precisely the way to overcome this difficulty: “Agents must be rational if their beliefs and preferences are to explain their choices.” Rationality, according to his view, is the mechanism that couples preferences and choices. But it does not identify one as the other.
To keep the rational-agent model, we have to hold a coherent definition of preferences that is removed from choices.
Preferences, rationality and the inevitability of Nudge
If we accept rationality models as an explanatory tool and understand welfare economics as interested in satisfying individuals’ preferences, there is no escape from socially evaluating these preferences. This evaluation is always against social norms or rules that are, at least partially, stable over time. Indeed, individuals’ preferences are generally created in light of norms (See Bicchieri, 2010).
By insisting on the possibility of purifying preferences, Hausman highlights a methodological problem that inevitably leads (with the help of Richard Thaler) to soft paternalism:
- To hold on to a rational-model methodology, we need to have preferences and choices as distinct concepts.
- If preferences are created in light of norms they can be socially scrutinized.
- If they can be socially scrutinized they can be socially evaluated.
- Evaluation is always against a scale of values (or norms)
- Therefore, it easy to end up with paternalism.
Following this line of reasoning, if we wish to escape soft paternalism, we have to relieve the demand for rationality.
Back to Angner’s critic. Philosophers of economics, Angner says, should not direct their efforts (as Hausman does) to the abstract calculus of the theory. Rather, they should focus their work on correspondence rules and models. Angner argues that we can find this contribution in Hausman’s work as he proposes a “useful model… (that can) serve a range of scientifically valuable pedagogical, heuristic, and other purposes”. But in attributing this contribution to Hausman, Angner concentrates on a specific line of arguments and detach it from Hausman’s more general perspective of economics.
I tried to show in this post that Hausman’s work is valuable for a very different reason than the one Angner points at. My claim was that in directing his criticism to the heart of the economic theory, Hausman does not miss the proper role of the philosopher of science. All the more so, and following Angner’s advice, he illuminates the failures both in economics’ corresponding rules and its models.