Rationality, Psychology and Capitalism

When asked to review Arkadiusz Sieron’s book, Rationality, Psychology and Capitalism, I was hesitant, but not for lack of interest in the subject matter (as suggested, albeit obliquely, by the book’s title). The author seemed to speak to themes in my decades-long writings on the methodological intersections of psychology and behavioral economics (which I shorten to “behaviorism”), and neoclassical economics.

Rather, my hesitation was founded on my practical concern over the book’s high price, $125 for a hard copy. I had to figure that the book’s price would make it attractive only to university libraries, while scaring off many of this journal’s readers. I would not have paid the listed price—before
immersing myself in it!

Once I began reading, however, I was more than pleased at the book’s readability and its steady flow of insightful arguments, largely driven home by an alternative (but still familiar) economic perspective to mine, Austrian economics. My only problem reading the book emerged from my dedication to underlining key points, as well as adding notes in all margins. Early on, I found my underlining was so dense that it lost its purpose, which is my way of saying Sieron’s book is a bargain in terms of insights per dollar.

I’ve read much of the behavioral literature critical of mainstream economics, only to become progressively unimpressed with the likely reliability of behaviorists’ inductive experimental findings and with their infusion of oft-repeated misunderstandings of economics’ methodology. Most fail to understand one of Sieron’s most basic points: Rationality is “a core principle that enables economics inquiry, as people’s behavior would be inexplicable without it” (p. 3).

Behaviorists really do need to read economics with greater care and take their critics, including Sieron, more seriously. They also need to give greater credence to psychologist Daniel Kahneman’s warning that the mounting criticisms of behaviorism research represent a major threat to behaviorism’s professional credibility.

Sieron has convinced me that I should have been more critical of behaviorism, even though I’ve not always been kind. In his book, he carefully documents behaviorists’ many methodological misunderstandings and flaws and adds damning arguments. He has greatly expanded the array of problems that behaviorists must address, drawing on insights of the Austrian greats—including, most notably, Friedrich Hayek, Ludwig von Mises, and James Buchanan—relating to exclusive reliance on empirical findings and inductive thinking.

If nothing else, today’s Austrians should be interested in Sieron’s book for his inclusion of a library of relevant Austrian references. Neoclassical economists, especially those unfamiliar with behaviorists’ anti-economics arguments, should be interested in this book for Sieron’s short summaries of behaviorists’ challenges to deductive neoclassical economics, along with the many counter arguments posited by a wide array of neoclassical and Austrian scholars.

Behaviorists should read Sieron’s book because in it he throws down a professional gauntlet, effectively daring behaviorists to rethink their claimed deficiencies of economics, as well as their claimed methodological superiority, despite their own defective social science methods. If for no other reason, readers of this review should consider Sieron’s book for reference purposes. He has provided an impressive, and perhaps a close-to-complete, 25-page bibliography. The book will have an honored place in my bookshelves among my collection of Austrian notables.

Couple Sieron’s powerful methodological critique of behaviorism with behaviorism’s snowballing “replication crisis” (a major challenge to behaviorism’s empirical findings that, unfortunately, Sieron skips), one must wonder how psychology and behavioral economics can avoid a loss of professional standing over coming years. “Replication” of empirical findings is the gold standard for studies to be called “science” (even “social science”).

To worsen their credibility problems, behavioral policy advocates have put their reputations at risk by proposing, with unchecked enthusiasm, the imposition of regulatory “nudges,” which they have presented as “virtually harmless,” claiming they are founded on principles of “libertarian paternalism” (surely an oxymoronic phrase). To me, “libertarian paternalism” is nothing short of a semantic con, especially if their intent is to manage in detail all other’s personal choices. My sense is that behaviorists have concocted many experiments to justify, in Sieron’s words, their “strong pro-authority psychological bias” (p. 110).

What credible scientist would support imposing nudges (and other regulations) if their foundational evidence cannot be reliably replicated, especially when not all subjects in their studies make the choice errors, which nudges are, supposedly, designed to correct? Why would people want to take directions from (politically appointed) human beings who must be drawn from, behaviorists attest, a population of grossly fallible decision makers? The ironies in behaviorism, which Sieron exposes, are breathtaking.

Understandably, some behavioral researchers have reacted defensively (even with hostility) to would-be-replicators’ findings, arguing that researchers have not exactly reproduced their original laboratory settings. Perhaps they should take to heart Sieron’s repeated observation: Behavioral findings are necessarily “context dependent,” with their contexts unavoidably complex—so complex, in fact, that exact replication of their original settings is very difficult, if not impossible, given the large number of uncontrolled, and uncontrollable, variables at work—especially when studies are done at different times with different subjects guided by different researchers.

Behaviorists can’t claim that all known and unknown, small and large, conscious and unconscious changes in variables matter; confess that not all variables can be controlled; and then not expect to have serious replication problems. In effect, Sieron reminds behaviorists of Hayek’s Nobel-worthy point, that behaviorists’ “choice architects” (like Hayek’s central planners), don’t and can’t learn as much from their studies as they presume. Much knowledge that affects decisions is out of reach to science, a major reason decision errors, biases, and irrationalities must be accommodated in a free society that honors individual welfare improvement via their choices, not behaviorists’ choices.

Behaviorists’ intellectual problems are compounded by the fact that they have itemized a lengthy list of human decision biases, mistakes, and irrationalities, yet seem to think that they can, in concocting nudges, escape all others’ decision failures. They also seem to presume that they, blessed with behavioral insights, will be anointed, by enlightened political operatives, as “choice architects”—certainly not Sieron, nor me. Moreover, they should consider a criticism of behavioral methods that Sieron makes with force: In the behaviorist paradigm, people “cannot do anything about them [‘mistaken choices’]; they cannot learn and improve their performance” (p. 5). Such may be true in behaviorists’ one-shot experiment settings, but it’s hardly the case in real-world social and market settings in which learning is a major purpose.

Sieron in effect asks why people, whose progenitors have met evolutionary tests, would be expected to waste resources on improving their welfare yet get so many decisions so “systematically wrong”?

Also, behaviorists must come to understand that, as Sieron stresses, quoting another Austrian, the fact that “in some experiments people violate the axioms of normative decision theory, does not necessarily prove that people are not as rational as axioms assume” (p. 6). Sieron fortifies this incisive point by quoting von Mises, “He who shoots wants, as a rule, to hit the mark. If he misses, he is not ‘irrational,’ he is a poor marksman” (pp. 22–23).

Sieron argues repeatedly that behaviorism lacks the type of general theory of human decisions that neoclassical and Austrian economists have long honed. Granted, behaviorists have a “general” theory, sort of, which is that almost everything in any setting can affect almost everything else. But that conceptual foundation hardly does what theory needs to do, which is to make deductive thinking manageable within the biological and economic limits of the human brain. Behaviorists skirt that problem by choosing the choices of subjects.

However, their “theory” does ease behaviorists’ difficulty in coming up with “hypotheses” (many of which, in my view, border on being ad hoc speculations-of-the-moment guides to laboratory experiments). For example: behaviorists have hypothesized that

• “Dinner plate size can affect diners’ consumed calories,” and
• “A few words on aging problems conveyed to a couple dozen undergraduates in a laboratory study can ‘prime’ them to walk slower.”

Such hypotheses have been supported in studies, leading many behaviorists to generalize, with confidence, that “larger plates lead to more calories consumed” and “people primed with words suggesting aging walk more slowly than subjects not primed.” These kinds of studies have led Kahneman to insist that “primes work,” a claim he insisted is beyond dispute—just as primes came under repeated replication challenges.

With such a loosey-goosey conceptual foundation, no one should be surprised if some studies (indeed, up to half of studies reconsidered) have failed replication tests. Sieron’s best explanation: behavioral studies are “context dependent,” with studies’ circumstances ever-changing across locations, subjects, time and other details.

Indeed, my two examples of behavioral hypotheses above were chosen because they undergird “classic” (widely cited) psychology studies that have failed replication tests. The first study might have involved academic fraud. The second might have failed replication because the original study involved the use of stopwatches to time subjects’ walking speeds while the replication study used a more precise timing device, using a laser beam.

The central theme of Sieron’s book is one that can’t be dismissed: Psychology and economics are distinct disciplines with substantially different intellectual interests, goals, and methods. Psychologists are mainly interested in inducing, through experiments, the “content” of people’s decisions, especially “errant” ones they might be able to “correct.”

Economists, on the other hand, mainly assume, as their conceptual starting point, what people want and then use general theories to deduce predicted “patterns” (Hayek’s favored word) of people’s behaviors in aggregated settings, for example, commercial and political markets, with the predictions required to pass statistical tests.

Many behaviorists advise economists to become more like them, which means substituting “real world” findings for “unreal” theoretical deductions. Sieron advises economists, in effect, to stay their methodological course, which has been highly successful in terms of generating insightful hypotheses. I can add another reason: Be wary of behaviorists’ advice, especially when founded on misunderstandings of economic methodology in the middle of a “looming,” slow-motion professional “train wreck.”