G. Harrison, & D. Ross,  The Gambling Animal, London: Profile, 2025

According to its authors, The Gambling Animal (TGA) is a natural history of risk management, written by two economists. Risk is thus predictably defined, from an economic point of view, as “the entire range of uncertain possible outcomes associated with an action or circumstance” (12). Perhaps less predictable is that Harrison and Ross appraise risk management mostly “at the scale of the general environment” (8), i.e., they analyse how two species “manage” the risks their members share in an ecological niche in which they need to survive. These two species are humans and elephants. The natural history TGA presents is a comparative analysis of how humans and elephants have managed their environmental risks (e.g., climate change) with outstanding success for the former, so far, and less success for the latter. Both species are highly social, highly communicative species that evolved under similar pressures and developed a comparable number of neurons for their respective size. Their biological brains “implement a close approximation to Bayesian learning” (16), allowing economist to study their choices with their standard tools and explain why humans have been comparatively so successful at environmental risk management.

At this point, some readers may be surprised at such an expansive view of what economists qua scientists can do. The answer lies elsewhere: as a disciple of Daniel Dennett, to whose memory he dedicates the book, Don Ross is not only an economist, but has made a parallel career in philosophy defending that microeconomics is about capturing real decision patterns in the world, whether the decision maker was a human or a bug — see his Economic Theory and Cognitive Science (2005). From an economic perspective, so broadly understood, the sort of brain humans and elephants developed is expensive, it requires lots of energy to maintain. Having all those extra neurons would have been an evolutionary gamble that paid off when environmental changes were too quick for any species to adapt rearranging its body structure. The rapid cooling that happened in the Neogene period (23 to 5.3 millions of years ago) gave an advantage to those expensive brains. In the Pleistocene, the forerunners of contemporary elephants and humans diverged even further, with the former would have evolved a high-fidelity memory and the latter a super-expensive brain (“a uniquely large battery of neurons in a cerebral cortex”, 158) that flourished in the then climatically volatile Rift valley. The high-fidelity memory of elephants made them very conservative with risk, because every new experience for them is a tiny increment against a steady growing sample. Hominine brains became instead capable of creating abstractions and imaginary realities, from which a distinctive style of risk management would emerge.

For this, a second building block was necessary: social norms of cooperation in small groups, which ultimately led to moral universalism via proto-market exchanges. These social norms allowed humans to jointly manage risks to a point beyond any other species’ reach. In chapter 7, we finally witness the rise of several risk management technologies: domesticated dogs, agriculture, writing and, ultimately, capitalism — the authors’ definition of technology is quite expansive, as you may have guessed. All of them bring about new risks, but competition between groups to reap their benefits will make them prosper. The eighth and final chapter is about climate change, how our species generated these risks and how it could face them:

This pattern we have observed through the whole history of our species is what we should expect to see continuing into the future. The main elements of that pattern are clear. First, we will not confront our major new risks, of which climate change is now the largest, by devolving them to individuals; new institutional structures are required. Second, the new institutions will not be deliberately built by some universal bureaucracy but will be assembled piecemeal through cultural-evolutionary processes (which, history shows, need not always unfold slowly). Third, the mechanisms by which the risk is ultimately managed will give rise to new, still more complex risks for management by our descendants. Is this a hopeful or a discouraging vision? That is a question to which there can be no objective answer. (344)

In a final coda, the authors present the experiments they are making to develop new technologies for elephant conservation, providing the grounds for an (economically coherent) ecological conservation policy.

The achievement of TGA is thus to explain the ecological dominance of our species and our struggle to contain its consequences “in terms of a single concept: risk management” (2). When I found this claim after opening the book, I could not help wondering if there was some particular epistemic merit in deriving an explanation for such a complex phenomenon in terms of just one concept, instead of many. Harrison and Ross seem to bet on the persuasive power of a coherent big picture. The risk is, of course, to deliver instead a just-so story: plausible, but impossible to test. As the authors acknowledge at several points, the topics they address refer to phenomena about which the evidence is often scarce and open to speculation. I, at least, would have appreciated a more systematic and explicit acknowledgement of the uncertainty about each step in their argument. The footnotes could have also considered alternative interpretations.

Leaving aside these formal qualms, my main objection is that Harrison and Ross are offering a general explanatory framework that, on its own, will not elucidate the actual risk managements techniques we observe in our species. The framework would help us mostly to understand how our risk management style compares with other species. But about this latter, TGA says surprisingly little. So what have we learnt?

Let me illustrate first the problem with the explanation of risk management techniques. On pp. 273-277, for instance, we find a short history of insurance, the canonical risk management technique. Yet, the account Harrison and Ross summarise here is entirely independent of the first principles laid out in TGA: surely, ecological considerations, Bayesian brains and social norms of cooperation did a lot of work to get insurance started, but the story Harrison and Ross tell about the development of insurance contract provides no particular evidence to test their central claims about how our species manages environmental risks. The story is equally compatible with many different explanatory frameworks. The same happens with the other risk management technologies presented in the book. In my view, this argumentative strategy increases the risk of TGA being a just-so story.

However, the TGA framework will surely be fruitful at the very abstract level of inter-species comparison, where the differences in brain abilities and social norms may be significant. The paradox is that, despite the recurrent invocation of elephants, TGA is mostly about humans. We learn very little about the risk management style of elephants: for instance, the authors do not say much about their experiments with them, and how those experiments support their claims. It would have been interesting, in general to learn more about what risk management in other animals could be, if only to make sense of the authors’ central point: in what sense human ecological dominance is owed to our risk management style as compared to those of other species. Among other things, because, as the authors themselves acknowledge, human ecological dominance may be transient. If somebody were to write about animal risk management in a few thousand years after our species’ extinction, would they be so focused on human skills? Perhaps two economists writing about animals and risk can’t help to be somewhat anthropocentric, but they are, nonetheless, fascinating to read.

 

{April, 2025}