Religion as risk management

Reader Justin Singh directs my attention to a new paper in economics: Religion, Heuristics, and Intergenerational Risk Management. In this paper, philosopher Rupert Reed and The Black Swan author Nassim Nicholas Taleb assert that religion is a unique, valuable, and perhaps indispensable method for avoiding unusual or "silent" risks (220), the "black swans" that Taleb has studied extensively. The authors explicitly focus on religion as practice; indeed, they consider the supernatural aspect of religion as "epiphenomenal," and assert that a traditional epistemic evaluation of religion is "extremely naive" (223). Instead, the authors insist that religion be viewed as "as closer to a form of trusting, as a form of action, or a willingness to take action, and, most crucially of all, as a set of interdicts upon action" (223), as well as establishing a rigid dichotomy between the sacred and the profane, a dichotomy taken on faith and immune to rationalization.

The authors claim that modern scientists and economists have far more confidence in their tools than is warranted. The models are fragile, sensitively dependent on initial assumptions. Crucially, modern scientists necessarily ignore unusual risk, i.e. rare but catastrophic events that have never occurred in the past, and thus do not form a (Bayesian) evidentiary basis for present action. The authors argue that religion, with its categorical interdicts immune to rationalization, is a unique mode that can protect us from modern scientific overconfidence.

I have read Taleb's book, The Black Swan, and he has convinced me of the value of his risk model, which focuses on rare but catastrophic events that typical scientific and economic models cannot describe. Catastrophic "black swan" events do happen, they are hard to positively predict, and people tend to ignore them in their planning. Indeed, Minsky's financial instability hypothesis seems to operate on much the same lines. (I would like to see Taleb comment on Minsky; perhaps he already has.) Taleb has also convinced me that modern scientific and especially economic practices are woefully inadequate to the task of describing, analyzing, and managing black swan events. Thus, I will not focus here on the topic of risk.

I concur that there is a problem, but I am not convinced that religion is a solution. First, the authors do not make an adequate case that religion really does effectively protect us from black swan events. More importantly, the authors fail to establish an effective definition of religion.

The authors give us only one case where religion might have protected us from a rare catastrophe. They juxtapose the 2008 global financial crisis with many religions' prohibitions or strict regulations against debt. However, this case is not particularly apt. First, it should be noted that religion is still ubiquitous, but failed to prevent the crisis. Indeed most modern organized religions, with the notable exception of Islam, permit nearly unregulated debt. Furthermore, the authors fail to look at the costs of using religion to prevent catastrophes. Credit and debt in general are crucial components of the modern capitalist global economy; eliminate them and the system would have to be rebuilt from the ground up. Not that I think rebuilding the capitalist system is itself a bad idea, but it would be enormously costly, not just in stuff but in human suffering. The authors fail to consider these costs. (And if this case is strained and artificial, it was the authors who chose it.)

Even taking the benefits of religion as granted, the authors do not provide an adequate definition of what religion actually is, and without such a definition, we cannot understand what it is about religion that affords the claimed benefits. The authors do have a definition. First, the authors claim sympathy to "true religion; to what we call faith as practice; to a genuinely spiritual orientation toward life" (219). This sentence, however, does not help; they authors nowhere define "spirituality," nor tell us how to distinguish genuine spirituality for its bogus or insincere alternative. The authors go on to assert that the important aspect of religion is its practice, not the epistemic basis of the practice. (Just this condition is controversial; many people who call themselves truly religious and genuinely spiritual would reject this definition.) Furthermore the authors assert that a religion makes some absolute, inviolable distinction between the sacred and the profane. The authors consider the inviolability crucial: "The sacred is not open to ‘rationalization’—what we don’t understand is not necessarily irrational, and it might have reasons that can be probed only across generations of experience and experimentation" (223). Finally, to establish these crucial interdictions, some reference to God is necessary: "People can understand the notion of God, not unexplained rules, interdicts, and categorical heuristics" (222, quoting Taleb from The Black Swan, p. 21).

Except for the mention of God, the definition of religion is probably necessary — a religion with no practice is at best trivial and probably no religion at all — but hardly sufficient. Again, except for the mention of God, it fails to distinguish between what is ordinarily considered religious and non-religious. For example, the United States constitution meets the definition: it establishes practices (the federal government, voting, citizenship, and a general democratic/republican political psychology and sociology). The constitution enforces a distinction between the "sacred" (free speech, property rights, due process) and the profane (what the government may regulate by law), and the United States has maintained these interdictions for almost twenty-five generations. The epistemic basis of the constitution is certainly not scientific, and it is immune to the weaker sort of "rationalization" the authors mention. Just because we don't (fully) understand the basis of one or another provision is not sufficient grounds to change it; we need strong positive legal or political reasons to textually or interpretively change provisions of the constitution.

Indeed, the purpose of almost every social institution is to preserve norms across generations without continual epistemic inquiry. We obey the law, we go to our jobs, we pay for our meals not because we have performed a full double-blind, peer-reviewed scientific inquiry each and every time we act to determine the best course of action, but simply because we have faith, trust, and the social commitment to act in that particular way, because that's just the way things are done. Thus, the only way to differentiate religion from any other social institution has something to do with God. But what? The authors are silent on this crucial point. They cannot mean just slapping the word "God" on things willy nilly; as the joke goes, "calling a tail a leg doesn't make it one." And they talk about true religion, genuine spirituality, which entails they must believe there is such a thing as false religion, i.e. illegitimate uses of the "God" label, but what distinguishes them? The authors explicitly disclaim an epistemic distinction: "it is an extremely naive interpretation to think that religious ‘beliefs’ map to the ‘justified true belief’ standards of modern epistemology . . .; it is naive to examine the supernatural aspect of religion as anything but epiphenomenal." But they do not offer any meaningful analytic distinction.

The authors argue for a pragmatic distinction, survivability. But relying on this distinction contains a fundamental contradiction. The whole point of Taleb's larger argument is that black swan events are rare, which means any social institution, religious or secular, can survive a long time without a black swan event. Indeed, Reed and Taleb explicitly condemn historicism: "[R]isk is not really in the visible past but rather in the future: the past is just a proxy. . . . Further, these silent risks [black swan events], when they hit, are produced most likely by some largely unknown class of distributions."* Moreover, their position depends on an epistemic critique of modern science and economics (i.e. in many ways, science and especially economics are making determinable epistemic errors), and depends on the correctness and power of that critique. (and not just in this paper but in Taleb's work in general, the critique is indeed correct and powerful). Because the authors admit to some true religion, we can conclude that both true and false religion have both survived; we need an analytical way to distinguish the two. And, of course, we need a way to evaluate change, which by definition has not survived at all.

*The elided sentence refers to the "recent past", but the visible past is broader, and not just the recent but also the distant past are distinct from where the authors assert risk lies: in the future.

The authors bury an interesting analytical definition in a footnote:

These heuristics belong to the class called “convex heuristics,” mathematically defined in Taleb (2014) [Silent Risk: Lectures on Fat Tails, (Anti)Fragility, Precaution, and Asymmetric Exposures]. Their aim is not to be ‘right’ and avoid errors, but to ensure that errors remain small. A convex heuristic has the following properties: (1) Compactness: It is easy to remember, implement, use, and transmit. (2) Consequences, not truth: It is about what it helps you do, not whether it is true or false. It should be judged not in ‘truth space’ but in ‘consequence space.’ (3) Antifragility: It is required to have a benefit when it is helpful larger than the loss when it is harmful. Thus it will eventually deliver gains from disorder. (4) Robustness: It satisfies the fragility-based precautionary principle. (5) Opacity: You do not need to understand how it works. (6) Survivability of populations: Such a heuristic should not be judged solely on its intelligibility (how understandable it is), but on its survivability, or on a combination of intelligibility and survivability. Thus a long-surviving heuristic is less fragile than a newly emerging one. But ultimately it should never be assessed in its survival against other ideas, rather on the survival advantage it gave the populations who used it.

These criteria are perhaps useful, but curious. If some heuristic is opaque (5) or unintelligible, how do we know it ensures errors remain small or has a desirable cost-benefit ratio? Especially if, as the authors assert, risk lies in the future and is produced by unknown distributions.

The authors, especially Taleb, have correctly identified a problem: a degree of epistemic arrogance in perhaps the sciences, definitely in economics, and blatantly in Western political economy. However, they fail to demonstrate the value of their proposed solution, religion, they fail to mention how religion might solve this problem, and they even fail to adequately define what their solution is. The authors' insistence that the epistemic and truth status of ideas underlying social institutions such a religion seems deeply anti-rational. We don't have a perfect epistemic basis for anything; we cannot know if anything is true with absolute certainty, but to argue that we can simply dispense with (analytic) knowledge and truth is just the fallacy of perfectionism. Indeed, such a position contradicts Taleb's larger work, which is explicitly epistemic: we know, or we want to know, how to improve our epistemology to prevent rare, catastrophic errors. It is important to examine the epistemology and truth underlying any social institution, religious or secular. They may not be perfect tools, but they're the best we have.