Skin in the Game

by Nassim Nicholas Taleb

  It may not be ethically required, but the most effective, shame-free policy is maximal transparency, even transparency of intentions.

  However, the story doesn’t tell us whether the purchaser was a “Swiss,” those outsiders our ethical rules don’t apply to. I suspect that there would be a species for which our ethical rules would be relaxed or possibly lifted. Recall our discussion of Kant: theory is too theoretical for humans. The more confined our ethics, the less abstract, the better it works. Otherwise, as we will see with Elinor Ostrom’s result later in this chapter, the system cannot function properly. And, before Ostrom, our old friend Friedrich Nietzsche got the point:

  Sympathy for all would be tyranny for thee, my good neighbor.

  Nietzsche, by the way, is the one person Fat Tony (upon hearing his quotes) said he would never debate.

  MEMBERS AND NON-MEMBERS

  For the exclusion of the “Swiss” from our ethical realm is not trivial. Things don’t “scale” and generalize, which is why I have trouble with intellectuals talking about abstract notions. A country is not a large city, a city is not a large family, and, sorry, the world is not a large village. There are scale transformations we will discuss here, and in the appendix of Book 3.

  When Athenians treat all opinions equally and discuss “democracy,” they only apply it to their citizens, not slaves or metics (the equivalent of green card or H-1B visa holders). Effectively, Theodosius’s code deprived Roman citizens who married “barbarians” of their legal rights—hence ethical parity with others. They lost their club membership. As to Jewish ethics: it distinguishes between thick blood and thin blood: we are all brothers, but some are more brothers than others.

  Free citizens, in ancient and post-classical societies, were traditionally part of clubs, with rules and member behavior similar to those in today’s country clubs, with an inside and an outside. As club members know, the very purpose of a club is exclusion and size limitation. Spartans could hunt and kill Helots, those noncitizens with a status of slaves, for training, but they were otherwise equal to other Spartans and expected to die for the sake of Sparta. The large cities in the pre-Christian ancient world, particularly in the Levant and Asia Minor, were full of fraternities and clubs, open and (often) secret societies—there was even such a thing as funeral clubs, where members shared the costs, and participated in the ceremonials, of funerals.

  Today’s Roma people (aka Gypsies) have tons of strict rules of behavior toward Gypsies, and others toward the unclean non-Gypsies called payos. And, as the anthropologist David Graeber has observed, even the investment bank Goldman Sachs, known for its aggressive cupidity, acts like a communist community from within, thanks to the partnership system of governance.

  So we exercise our ethical rules, but there is a limit—from scaling—beyond which the rules cease to apply. It is unfortunate, but the general kills the particular. The question we will reexamine later, after deeper discussion of complexity theory, is whether it is possible to be both ethical and universalist. In theory, yes, but, sadly, not in practice. For whenever the “we” becomes too large a club, things degrade, and each one starts fighting for his own interest. The abstract is way too abstract for us. This is the main reason I advocate political systems that start with the municipality, and work their way up (ironically, as in Switzerland, those “Swiss”), rather than the reverse, which has failed with larger states. Being somewhat tribal is not a bad thing—and we have to work in a fractal way in the organized harmonious relations between tribes, rather than merge all tribes in one large soup. In that sense, an American-style federalism is the ideal system.

  This scale transformation from the particular to the general is behind my skepticism about unfettered globalization and large centralized multiethnic states. The physicist and complexity researcher Yaneer Bar-Yam showed quite convincingly that “better fences make better neighbors”—something both “policymakers” and local governments fail to get about the Near East. Scaling matters, I will keep repeating until I get hoarse. Putting Shiites, Christians, and Sunnis in one pot and asking them to sing “Kumbaya” around the campfire while holding hands in the name of unity and fraternity of mankind has failed. (Interventionistas aren’t yet aware that “should” is not a sufficiently empirically valid statement to “build nations.”) Blaming people for being “sectarian”—instead of making the best of such a natural tendency—is one of the stupidities of interventionistas. Separate tribes for administrative purposes (as the Ottomans did), or just put some markers somewhere, and they suddenly become friendly to one another.*3 The Levant has suffered (and keeps suffering) from Western (usually Anglo-Saxon) Arabists enamored with their subject, with no skin in the game in the place, who somehow have a vicious mission to destroy local indigenous cultures and languages, and separate the Levant from its Mediterranean roots.*4

  But we don’t have to go very far to get the importance of scaling. You know instinctively that people get along better as neighbors than roommates.

  When you think about this, it is obvious, even trite, from the well-known behavior of crowds in the “anonymity” of big cities compared to groups in small villages. I spend some time in my ancestral village, where it feels like a family. People attend others’ funerals (funeral clubs were mostly for large cities), help out, and care about the neighbor, even if they hate his dog. There is no way you can get the same cohesion in a larger city when the “other” is a theoretical entity, and our behavior toward him or her is governed by some general ethical rule, not someone in flesh and blood. We get it easily when seen that way, but fail to generalize that ethics is something fundamentally local.

  Now what’s the reason? Modernity put it in our heads that there are two units: the individual and the universal collective—in that sense, skin in the game for you would be just for you, as a unit. In reality, my skin lies in a broader set of people, one that includes a family, a community, a tribe, a fraternity. But it cannot possibly be the universal.

  NON MIHI NON TIBI, SED NOBIS (NEITHER MINE NOR YOURS, BUT OURS)

  Let us get into the gut of Ostrom’s idea. The “tragedy of the commons,” as exposed by economists, is as follows—the commons being a collective property, say, a forest or fishing waters or your local public park. Collectively, farmers as a community prefer to avoid overgrazing, and fishermen overfishing—the entire resource becomes thus degraded. But every single individual farmer would personally gain from his own overgrazing or overfishing under, of course, the condition that others don’t. And that is what plagues socialism: people’s individual interests do not quite work well under collectivism. But it is a critical mistake to think that people can function only under a private property system.

  What Ostrom found empirically is that there exists a certain community size below which people act as collectivists, protecting the commons, as if the entire unit became rational. Such a commons cannot be too large. It is like a club. Groups behave differently at a different scale. This explains why the municipal is different from the national. It also explains how tribes operate: you are part of a specific group that is larger than the narrow you, but narrower than humanity in general. Critically, people share some things but not others within a specified group. And there is a protocol for dealing with the outside. Arab pastoral tribes have firm rules of hospitality toward nonhostile strangers who don’t threaten their commons, but get violent when the stranger is a threat.

  The skin-in-the-game definition of a commons: a space in which you are treated by others the way you treat them, where everyone exercises the Silver Rule.

  The “public good” is something abstract, taken out of a textbook. We will see further in Chapter 19 that the “individual” is an ill-defined entity. “Me” is more likely to be a group than a single person.

  ARE YOU ON THE DIAGONAL?

  A saying by the brothers Geoff and Vince Graham summarizes the
ludicrousness of scale-free political universalism.

  I am, at the Fed level, libertarian;

  at the state level, Republican;

  at the local level, Democrat;

  and at the family and friends level, a socialist.

  If that saying doesn’t convince you of the fatuousness of left vs. right labels, nothing will.

  The Swiss are obsessive about governance—and indeed their political system is neither “left” nor “right,” but governance-based. The thoughtful mathematician Hans Gersbach once organized a workshop on skin in the game in Zurich on how to properly reward (and punish) politicians whose interests are not lined up with those of the people they represent. It struck me that if things worked well in Switzerland and other Germanic countries, it is not because of accountability so much as scaling, which makes them very prone to accountability: Germany is a federation.

  Let us next generalize to risk sharing.

  ALL (LITERALLY) IN THE SAME BOAT

  Greek is a language of precision; it has a word describing the opposite of risk transfer: risk sharing. Synkyndineo means “taking risks together,” which was a requirement in maritime transactions.*5

  The Acts of the Apostles describes a voyage of St. Paul on a cargo ship from Sidon to Crete to Malta. As they hit a storm: “When they had eaten what they wanted they lightened the ship by throwing the corn overboard into the sea.”

  Now while they jettisoned particular goods, all owners were to be proportioned the costs of the lost merchandise, not just the specific owners of the lost merchandise. For it turned out that they were following a practice that dates to at least 800 B.C., codified in Lex Rhodia, Rhodian Law, after the mercantile Aegean island of Rhodes; the code is no longer extant but has been cited since antiquity. It stipulates that the risks and costs for contingencies are to be incurred equally, with no concern for responsibility. Justinian’s code summarizes it:

  It is provided by the Rhodian Law that where merchandise is thrown overboard for the purpose of lightening a ship, what has been lost for the benefit of all must be made up by the contribution of all.

  And the same mechanism for risk-sharing took place with caravans along desert routes. If merchandise was stolen or lost, all merchants had to split the costs, not just its owner.

  Synkyndineo has been translated into Latin by maestro classicist Armand D’Angour as compericlitor, hence, if it ever makes it into English, it should be compericlity, and its opposite, the Bob Rubin risk transfer, will be incompericlity. But I guess risk sharing will do in the meanwhile.

  Next, we discuss some distortions from the introduction of skin in the game.

  TALKING ONE’S BOOK

  I went on television once to announce a newly published book and got stuck in the studio, drafted to become part of a roundtable with two journalists plus the anchor. The topic of the day was Microsoft, a company that was in existence at the time. Everyone, including the anchor, chipped in. My turn came: “I own no Microsoft stock, I am short no Microsoft stock [i.e., would benefit from its decline], hence I can’t talk about it.” I repeated my dictum of Prologue 1: Don’t tell me what you think, tell me what you have in your portfolio. There was immeasurable confusion in the faces: a journalist is typically not supposed to talk about stocks he owns—and, what is worse, is supposed to always, always make pronouncements about stuff he can barely find on a map. A journalist is meant to be an impartial “judge,” yet, unlike Sisamnes in the Judgment of Cambyses, there is no threat of a secondary use of his skin.

  There are two types of “talking one’s book.” One consists of buying a stock because you like it, then commenting on it (and disclosing such ownership)—the most reliable advocate for a product is its user.*6 Another is buying a stock so you can advertise the qualities of the company, then selling it, benefiting from the trumpeting—this is called market manipulation, and it is certainly a conflict of interest. We removed the skin in the game of journalists in order to prevent market manipulation, thinking that it would be a net gain to society. The arguments in this book are that the former (market manipulation) and conflicts of interest are more benign than impunity for bad advice. The main reason, we will see, is that in the absence of skin in the game, journalists will imitate, to be safe, the opinion of other journalists, thus creating monoculture and collective mirages.

  In general, skin in the game comes with conflict of interest. What I hope this book will do is show that the former is more important than the latter. There is no problem if people have a conflict of interest if it is congruous with downside risk for themselves.

  A SHORT VISIT TO THE DOCTOR’S OFFICE

  The doctor doesn’t have the Antaeus problem: medicine, while wrapping the garment of science around it, is fundamentally apprenticeship-based and, like engineering, grounded in experience, not just experimentation and theories. While economists say “assume that…” and produce some weird theory, doctors have none of that. So there is skin in the game at many degrees, except perhaps not fully in the agency effect separating customer from provider. And attempts at putting skin in the game there have brought a certain class of adverse effects, in shifting uncertainty from the doctor to the patient.

  The legal system and regulatory measures are likely to put the skin of the doctor in the wrong game.

  How? The problem resides in the reliance on metrics. Every metric is gameable—the cholesterol lowering we mentioned in Prologue 1 is a metric-gaming technique taken to its limit. More realistically, say a cancer doctor or hospital is judged by the five-year survival rates of patients, and needs to face a variety of modalities for a new patient: what choice of treatment would they elect to do? There is a tradeoff between laser surgery (a precise surgical procedure) and radiation therapy, which is toxic to both patient and cancer. Statistically, laser surgery may have worse five-year outcomes than radiation therapy, but the latter tends to create second tumors in the longer run and offers comparatively reduced twenty-year disease-specific survival. Given that the window used for the calculation of patient survival is five years, not twenty, the incentive is to shoot for radiation.

  So the doctor is likely to be in the process of shifting uncertainty away from him or her by electing the second-best option.

  A doctor is pushed by the system to transfer risk from himself to you, and from the present into the future, or from the immediate future into a more distant future.

  You need to remember that, when you visit a medical office, you will be facing someone who, in spite of his authoritative demeanor, is in a fragile situation. He is not you, not a member of your family, so he has no direct emotional loss should your health experience a degradation. His objective is, naturally, to avoid a lawsuit, something that can prove disastrous to his career.

  Some metrics can actually kill you. Now, say you happen to visit a cardiologist and turn out to be in the mild risk category, something that doesn’t really raise your risk of a cardiovascular event, but precedes the stage of a possibly worrisome condition. (There is a strong nonlinearity: a person classified as prediabetic or prehypertensive is, in probability space, 90 percent closer to a normal person than to one with the condition.) But the doctor is pressured to treat you to protect himself. Should you drop dead a few weeks after the visit, a low probability event, the doctor can be sued for negligence, for not having prescribed the right medicine that is temporarily believed to be useful (as in the case of statins), but that we now know has been backed up by suspicious or incomplete studies. Deep down, he may know that statins are harmful, as they will lead to long-term side effects. But the pharmaceutical companies have managed to convince everyone that these unseen consequences are harmless, when the right precautionary approach is to consider the unseen as potentially harmful. In fact for most people except those that are very ill, the risks outweigh the benefits. Except that the long-term medical risks are h
idden; they will play out in the long run, whereas the legal risk is immediate. This is no different from the Bob Rubin risk-transfer trade, of delaying risks and making them look invisible.

  Now can one make medicine less asymmetric? Not directly; the solution, as I have argued in Antifragile and more technically elsewhere, is for the patient to avoid treatment when he or she is mildly ill, but use medicine for the “tail events,” that is, for rarely encountered severe conditions. The problem is that the mildly ill represent a much larger pool of people than the severely ill—and are people who are expected to live longer and consume drugs for longer—hence pharmaceutical companies have an incentive to focus on them. (Dead people, I am told, stop taking drugs.)

  In sum, both the doctor and the patient have skin in the game, though not perfectly, but administrators don’t—and they seem to be the cause of the troubling malfunctioning of the system. Administrators everywhere on the planet, in all businesses and pursuits, and at all times in history, have been the plague.

  NEXT

  This chapter introduced us to the agency problem and risk sharing, seen from both a commercial and an ethical viewpoint, assuming the two can be disentangled. We also introduced the problem of scale. Next, we will try to get deeper into the hidden asymmetries that make aggregates strange animals.