Inside and Outside Perspectives on Institutions

Transcript below. This is a presentation of “Inside and Outside Perspectives on Institutions: An Economic Theory of the Noble Lie” for the Entangled Political Economy Research Network.

The screenshot here is from a movie called They Live, and the guy gets a magical set of glasses that when he puts them on he can see through all of his society’s institutions and this is not a wonderful thing; in fact this is kind of a horrifying thing. And the movie goes on from there.

First of all what i want to start with is what we mean by inside and outside perspectives on an institution, and then I’m going to come back and build back up to where those perspectives come from and why they’re always going to exist for any social institution.

A lot of you have probably read Pete Leeson’s Ordeals paper. And that i think is a good illustration of what we mean by inside and outside perspectives on institutions even though I don’t think it explains where those come from adequately. So in the ordeals paper, basically what he argues is that in medieval Europe, and in a lot of places actually, you had these kinds of rituals where if there was ambiguous evidence, then there was a process where you would put the accused through an ordeal. You see if she weighs more than a duck, or more realistically you’d put put his hand in boiling water and see if it’s scalded. This was designed to reveal the judgment of God, or the judgment of the spirits as the case may be.

Now that’s an inside perspective on institutions: if you ask somebody in that institution how does this work, why does this work, what are you doing. The inside perspective is how they would answer. Now the point of the Ordeals paper is to say that, well we don’t really believe that that’s a good way of determining the judgment of God anymore. So what he argues is that the reason this institution worked is that it results in a separating equilibrium: conditional on whether the accused feels reluctant to undergo the ordeal, that provides valuable information about the accused’s innocence. So this is going to be a way of eliciting information about the accused’s guilt or innocence.

That’s what we’re going to call an outside perspective. The outside perspective is going to be kind of a functionalist perspective on how these institutions work. Now these two explanations the inside and the outside perspective are mutually exclusive not just in the basic sense of being different explanations for the same kind of institution, but also more fundamentally if you accept the outside perspective it kind of spoils the inside perspective. If Leeson’s paper were to be transported into the hands of the accused in an ordeal situation and he read this paper, that would be bad news right? And then he could exploit the system. And if the operators of the system figured out that people are figuring out how this works, it’s going to break down. This requires that people believe in the institution in order for it to work. It’s self-confirming in that sort of sense. And if you take the functionalist perspective, if you take the outside perspective, that spoils the possibility of the institution working.

And second what i want to show throughout the course of this paper is that every social institution is going to have both of these: every social institution is going to have an inside perspective and an outside perspective in a similar sort of way. And the reason for that is that social organization, broadly speaking, is not incentive compatible, and there are no strategies to make it totally incentive compatible. Now that might seem like a strong and surprising claim. Why can’t we just take the ordeal story at face value, why can’t we just leave it where the ordeals paper left it?

What the ordeals paper does not explain is why belief can be taken for granted. The ordeal story works if we can take belief for granted. Does that make sense to do? Why would humans be the type of people to be prone to believing that kind of thing?

So what I want to do is take a step back and think about a strategic model of belief. We don’t want to take that for granted anymore. Why can’t we take that for granted? Because belief is individually sub-optimal, but collectively optimal. So think about the accused in the case of the ordeal again if the accused doesn’t believe, but professes belief, he can do better for himself. He can commit crimes with impunity and then pretend like, well yeah I’m willing to undergo the ordeal. And then it loses its informational value. But then that makes the society worse off because then you lose this sort of separating equilibrium.

So when we think of belief being individually sub-optimal but collectively optimal, that should ring our economist’s bells: belief is a collective action problem! Whether or not to believe in any given inside perspective is a collective action problem, or a public good, essentially the same kind of thing.

Structurally now when we think of public goods problems or collective action problems as economists, there are a few things that our minds auto-complete to. One of those things is going to be beliefs, norms, common culture. And these kinds of things are in practice very helpful for solving particular collective action problems. That’s not going to solve our problem because beliefs norms and common culture are themselves collective action problems. They are themselves public goods.

Same thing with economists who autocomplete to the state. When we have a collective action problem, we get the state to coordinate people. And we have a public good, we get the state to tax people and provide it. That’s not going to do either, because the state’s a big collective action problem too! Where does the state come from? – and this i think is what makes this an Entangled Political Economy paper, is that we’re thinking of the state as a not the solution to a public goods problem, but as a public goods problem itself, just like any other kind of public good problem.

So we’ve got to take these easy answers off the table when we’re interested in the more fundamental issue. We’ve got to go deeper.

This is what I’m going to call the Incentive Gap. There’s a big literature on prisoner’s dilemmas, and the variety of strategies that can support cooperation in a prisoner’s dilemma. Now one important feature of prisoners dilemmas is that they involve two people. Two people is not a good prototype for a theory of institutions in general. So a lot of economists look at things like the Axelrod game. The Axelrod game is pairwise matching, and two people play a prisoner’s dilemma against each other, and you get cooperation with tit for tat.

Fine in a two-person game but this doesn’t generalize to collective action problems or public goods games where you have more than two people. Why not? The public goods game, like the prisoner’s dilemma, ends in defection if you have a one-shot game. You can fix this in a prisoner’s dilemma by repeating. you can’t do that in a public goods game. Why not? Because repeating a a prisoner’s dilemma lets you punish the other party for defection, and the benefits of future cooperation are going to outweigh the the losses of both punishing and the losses of defecting. So tit for tat is a strategy that stabilizes cooperation in prisoner’s dilemmas. This doesn’t work for public goods games when there’s lots of people. Why not? Because this takes away your ability to punish. If you try to punish somebody else who defects by defecting yourself, by not contributing to the collective action or not contributing to the public good, you’re punishing everybody. And that person that you’re trying to punish, the benefits of them cooperating accrue to everybody, not just to you.

So you say, all right well what about targeted punishment? What if you find a way to punish the defector specifically? – and that solves the problem at one level by making it in the interest of the defector to cooperate. But then it’s not in your interest to punish. Why would you ever take a costly action to punish a defector if the benefits of doing so are going to accrue to everybody? This is what Yamagishi called the second order punishment problem. Punishment does not actually solve public goods games. In fact it just superimposes another one on top of the first. And so you say okay, well you punish the people who fail to punish. That’s your third order punishment problem. And as a colleague of mine summarized this argument when i presented this paper to the department, he says it’s collective action problems all the way down.

That’s our basic problem: there’s no way of eliminating the incentive to defect in n-person social dilemmas. So we’ve become maybe too optimistic by just looking at prisoners dilemmas. When we generalize to the n-person case, the problem gets a lot harder. And fundamentally governance is about these kinds of n-person social dilemmas, in any kind of institution of consequence.

So let’s look at a few places where these problems have been treated before. As far as the theory the firm goes, Miller has this book where he talks about the firm as a locus of joint production. And under joint production, effort is a public good: if your payoffs depend on everybody else’s payoffs as well, the question then is how do you make it incentive compatible for nobody to shirk? How do you get everybody to that pareto optimum?

Alchian and Demsetz’s answer was to get a residual claimant who is the monitor. The monitor makes sure that nobody shirks, and then gets the leftover surplus. But what a series of further papers showed is that there’s no strategy by the residual claimant that will work to disincentivize shirking among the employees without creating an incentive for himself to shirk. That’s our second order punishment problem. And so what Miller shows is that firms have to rely on something like company culture right to close the incentive gap. You can’t do this all with incentives; you need something like company culture. We’ll come back to that and think about the significance of company culture afterward.

Bowles and Gintis in their book A Cooperative Species show that in society more broadly when we’re thinking about governance, your homo economicus is not going to be able to arrange for the joint provision of public goods they’re not going to be able to engage in collective action. What’s the significance of that? People look at this sometimes and say well did Ostrom show that, you know that that’s possible, that we can do that. The difference is Ostrom’s not dealing with homo economicus. And I’ll show you exactly where they diverge. She’s not exactly clear on this, but Ostromian agents are are not homo economicus in key points.

So our question is how is it possible for non-homo economicus to survive and thrive. So we’re going to bring in some game theory here. And there are two ways that the argument from the previous section can be interpreted. First of all in classical game theory, we understand the payoffs to be in terms of utility. Classical game theory is a theory of choices being made on the basis of a utility function, and that utility function referring to subjective preferences. And the argument of the previous section is that if we are utility maximizers, then nobody will choose to cooperate in n-person public goods games.

The second interpretation of game theory is going to be one where the payoffs represent not subjective utility, but reproductive fitness; the propensity to reproduce. Evolutionary game theory is not a theory of choice; in fact choice isn’t really involved at all. And what we end up with is not choices selecting strategies, but competition between strategies. Individuals aren’t going to change their strategies, but more more fit strategies are going to outcompete less fit strategies.

So the argument of the previous section in this frame would be that cooperators are going to do worse, and cooperators are going to get outcompeted by non-cooperators. You can think of the accused in the ordeal example here: if he sees this outside perspective, professes belief, and spoils the informational value of that signal, he’s gonna do better than the other people in his society because now he can commit crimes with relative impunity. That’s bad news.

Both of these interpretations so far on their own are going to agree: cooperative strategies aren’t going to be chosen because they’re not utility maximizing, and cooperative strategies aren’t going to survive because they’re going to be less fit. So economists have sometimes taken this agreement to be an evolutionary foundation for homo economicus. Binmore did this and this is kind of in the background of the Alchian and Becker papers “Uncertainty, Evolutionary Theory, and Economics”, the idea being that agents have to be assumed to prefer their own objective payoffs. In the Alchian and Becker papers they’re preferring income. In Binmore’s case they’re preferring things that lead to their increased fitness because those are the preferences that will allow agents to compete more effectively.

Vlad asks: do you mean income maximizing? They are utility maximizing because in binmore’s case we’re talking about fitness maximizing. In Alchian and Becker’s case they’re talking about income maximizing and therefore we can assume that utility corresponds to income.

Vlad: If altruism is part of their utility function, then it is utility maximizing. Like you care about another person’s welfare then it becomes utility altruism becomes utility maximizing.

But then the question is how can altruism be a stable strategy? How would an agent who prefers something other than his own income or his own reproductive fitness, how would such an agent ever out compete agents who don’t prefer such things? So our question is, where do these altruistic preferences come from. Because they clearly do exist. So how is it possible that these agents who are altruistic, that do prefer things other than their own selfish maximization, whether their income or their fitness, how do they survive?

And it turns out that though if you take evolutionary game theory or classical game theory on their own, they agree. But if you take both on their own terms then it’s going to be possible for subjective utility to diverge from objective fitness. right And that’s going to be possible if individuals are organized into a group structure. And it turns out that this is exactly true of humans. So this is David Sloan Wilson who’s been banging this drum for a long time, and his pithy way of putting this is that selfish individuals are going to out compete altruistic individuals, but altruistic groups can out compete selfish groups.

This is a simulation I ran in a group structured situation, altruistic preferences can survive. And if you just have a matching game, if you just have a non-group structured setting then you start with 99 altruists and one selfish person and altruism just collapses immediately. But if you have groups that are competing against each other, then you get a stable proportion of altruists. So that can be viable if we think of group structured populations.

And there’s two ways we can think of that, not mutually exclusive and not exhaustive, but if you think of assortativity where cooperators can choose to work with each other and they can exclude selfish individuals, this is going to be a public good. the exclusion is going to be a public good because it’s costly to exclude people. But if they do that, or if there’s competition between teams, where groups will fight with each other and less cooperative groups will do worse in these competitions, then it’s going to be possible for the between-group selection to balance the within-group selection against these altruistic preferences.

And as it turns out both of these situations, both assortativity and this this interdemic competition do seem to characterize the early human environment. Humans did develop our moral senses in environments where assortativity and interdemic competition were the rule, not the exception.

So here’s the upshot: this much is not original. This is just summarizing the background to what I want to do here. The relevance of all of this to how we think about institutions is that there’s going to be a mismatch between the level at which cooperative strategies are selected, and the level at which choices are made. So if cooperation is selected for at the group level, but choices are made by individuals, then individuals are going to have to employ strategies against their own narrow interests. So there’s gonna have to be some way of representing to the individual interests other than his or her own. And this is going to appear to the individual making the choice as as a divergence between objective fitness and subjective payoffs. This is where the altruistic preferences that Vlad mentioned, this is where they come from, and this is why it matters that they come from there. And they’re not something we can just take for granted; we have to know where they come from because if it’s true that human social life is such that altruistic preferences are necessary, then you’re always going to have to rely on people acting in ways contrary to their own narrow interests. Their subjective preferences are going to have to diverge from their objective fitness, whether that’s income or whether that’s fitness or something else.

Two things this can look like: this can look like false beliefs which Leeson’s written a number of papers on. This is gonna be people who are convinced that it’s not in their interests to employ a strategy that it actually is, or vice versa; convinced that it is in their interest to employ a strategy that it’s not in fact in their interests to do. Folks in the ordeal situation are convinced that it’s not in their interest to do certain things that maybe the guy could get away with it, but he’s convinced that he can’t, and that’s a good thing.

The other way this can happen is through maladaptive preferences. If you just prefer doing the right thing for its own sake, then that’s gonna do the same kind of work now. In practice any kind of social institution is going to rely on some combination of these things. And there’s going to be a diminishing marginal rate of substitution between these: if you’re trying to do all of one versus all of the other, that’s going to be more fragile than if you can rely partly on one partly on the other, though there are varying mixes in various places.

So now we’ve built back up to where the inside and outside perspectives come from – the inside perspective again is going to be explaining an institution in terms of the preferences and beliefs of the people in that institution, so you ask them what’s the meaning of this to you what are you intending to do with this, that’s going to be your explanation of ordeals, that rely on the judgment of god or the judgment of the spirits or something to that effect. Whereas the outside perspective is going to be looking at this in terms of objective fitness, whether that’s income or something else. That’s going to be our separating equilibrium story. And now we know having gone through all that background, why exactly the one is not reducible into the other. Why do you need an inside perspective? Why can’t you just get by with an outside perspective? And it’s not just that people in the medieval era were more superstitious or something to that effect, but rather that this is how they crossed the incentive gap, by relying on that particular sort of that particular sort of divergence. And in our own societies we have other things that are going to generate that divergence between subjective preferences and objective fitness and so as long as we’re dealing with an institution that involves a social dilemma, a collective action problem, or a public goods problem. These are never going to correspond you’re never going to be able to get rid of an inside perspective.

Now there’s two different ways of approaching norms. Oftentimes you think of these sometimes as left and right respectively, but maybe it’s better to think of these as iconoclasm versus conservatism. The iconoclast looks at norms and says these restrict freedom. The individual’s interests are all that matters, so let’s get rid of restrictive norms. They’re bad because they’re irrational. Or, you might look at this paper and take a more conservative approach and say, no, actually norms are good because norms encode information that that you might not want to get rid of.

I don’t want to take either of those positions. As a matter of fact there’s a middle ground, and the middle ground is that the inside and outside perspectives make it very difficult for a culture to critique its own norms, or for an institution to reform itself. Because the inside perspective has to remain kind of invisible, lest it devolve and break down into that war of all against all where everybody’s interested in their own objective fitness. If the inside perspective is invisible, it’s going to be possible for maladaptive norms or beliefs to hitchhike on this general capacity for internalizing altruistic norms. And so it’s not the case that everything is efficient. In fact highly Pareto sub-optimal norms can persist for a very long time because the inside perspective is invisible. It’s difficult to critique them from any outside vantage point, and especially in isolated societies or hegemonic societies, there’s not that intercultural competition to ensure that your norms and beliefs are pro-social rather than anti-social.

So what does this look like? Abstractly this is going to look like anti-social punishment. There’s a superior norm available, but we can’t get there because there are people who will punish the people who try. And there’s a number of examples from for example Edgerton and Bicchieri; have both written books about perverse norms. And in the Edgerton book in particular, he talks about several interesting cases where you have these isolated island societies, and they have some aberrant norm, some self-destructive norm like a food taboo. Not that food taboos are always non-functional, but in some cases there have been a rich food source on the island, maybe the only rich food source, that nobody will touch. And so they’re all malnourished. And if you try to eat the food, you’re going to get punished in some way or another. Or you could think of things like human sacrifice, ritual cannibalism – these are situations where European colonizers come in and say, guys what are you doing, is this is terrible. And everybody in the society says, you know, you’re right, this is terrible. this sucks. And they’re quite happy to get rid of it, and no ill effects.

So this argument should not be taken to be relativist in the sense of “we can’t judge anybody else’s norms”, or conservative in the sense that “nobody else can judge our norms”, or iconoclastic in the sense that “norms should always be judged”, but rather, we have a problem that it’s difficult to judge norms. It’s not that they’re always going to be good or beneficial or efficient; many times they won’t be. But it’s systematically hard to know, because if we did know, then it’d be hard to sustain that institution in the first place.

So again it’s difficult to reform institutions both from the inside and from the outside. So we probably all know a lot of examples of the failure to transplant Western style institutions to developing countries. That has mostly not worked and not had really good outcomes. Why not? Acemoglu has this nice paper called “Why No Political Coase Theorem” where he argues that when you’re designing political institutions, there’s no commitment power at the very top. Again it’s collective action problems all the way down. So if you know enough to design an institution, if you have the authority to design it, you also have the authority and the know-how to exploit that. And so if you clear away a coordinating mechanism it’s very possible that it’s not gonna reconverge on anything more beneficial. You’ve just cleared away the old without introducing something to substitute.

So this points to the necessity for legitimation. it’s not enough to design institutions just with a view to incentives. The incentives are important and you don’t want to rely more than you have to on maladaptive preferences or false beliefs, but incentives are never going to get you all the way. Which is why legitimation is so important. Brennan and Buchanan were quite self-aware about this: they talked about political economy or institution building as the process of creating a mythology. And they didn’t mean this in a disparaging way. There’s always going to be some necessity of an inside perspective, of some something that looks like a mythology in order for any institution to work.

My reading of Hayek’s “Intellectuals and Socialism” is that he was getting at the same problem when he talks about the courage to be utopian. I think this is what he’s referring to, that there’s this necessity for excitement for culture to overcome that incentive gap and to make liberal institutions work. And of course Plato called this the Noble Lie, that you have to convince people that it’s in their interests to do things that are perhaps not objectively in their interests. Their subjective preferences have to diverge from those objective payoffs.

One of the most interesting consequences of this kind of analysis, I think, is that it flips the logic of rules versus discretion on its head. The conventional wisdom in economics is that rules are preferable to discretion. If you know the North and Weingast paper “Constitutions and Commitment” in the case of the Bank of England and the Glorious Revolution, or if you know Kydland and Prescott’s paper in terms of monetary policy, both of these are about the fact that constraints, explicit rules, can generate better outcomes than if you left it up to discretion in certain cases, because maximizing can lead to commitment problems.

So suppose that you wanted to take this to the limit and say that rules are good, let’s make everything a rule. Let’s hire Robocop, and Robocop is going to commit us all to cooperation and enforce the rules that he has set. What’s the problem with delegating enforcement to Robocop? We’ve solved the second order punishment problem, because Robocop just obeys programming. So why couldn’t Robocop enforce cooperation? Why would this not solve the commitment problem?

The basic problem is that the fact that the incentive gap exists, the fact that it’s never going to be possible to align objective preferences and subjective utility and make the institution work, means that it’s not going to be ever possible to fully articulate the decision rules governing an institution. Because if you could articulate them, and if you committed to a rigid interpretation of them, then it’s going to be vulnerable to defection. There’s always going to be some way to take a rigid set of rules and exploit them. This is essentially Boyd and Lorberbaum’s problem that no strategy is the best response to all other strategies. And so because of the fact that social institutions are always going to need ways of identifying and fighting against novel forms of defection, it’s never going to be able to be exhaustively pre-specified. That means it’s going to be necessary to switch strategies at key points without realizing that you’re doing it, because if you realize that’s what you were doing, that would be an explicit meta rule that could itself be exploited.

This gives you kind of an arms race between defectors and cooperators who try to exclude them. And there’s an argument that this was a major factor in the development of human intelligence. Why are humans smart? Because we needed to develop this kind of inarticulate gut sense of when we’re being exploited, when a rule’s being exploited. So the capacity to articulate, in this view, is not the fundamental feature of human intelligence.The fundamental feature of human intelligence is the ability to sense inarticulately that hey, somebody’s trying to screw me here.

So what’s the upshot? in Jacobellis versus Ohio, there was an obscenity case. And the defense lawyers wanted to know, what’s the definition of pornography? Why are you charging our client here, what about these National Geographic pictures? And the judge said, no those don’t count. And the lawyer said why not? What’s the rule? And judge Potter Stewart did not give a rule. He refused to specify a rule. In fact his response was “I know it when i see it.” That’s the kind of discretionary enforcement that allows you to switch strategies to deal with these kinds of novel forms of defection in a public goods game. And any legal system is going to rely on this. If you think about it, basic legal concepts like negligence – what specifically counts as negligence? – that’s a very discretionary sort of question. It requires judgment, and the reason it requires judgment is precisely because of this incentive gap problem.

The blockchain paper that Marta mentioned at the very beginning, that’s an application of this that Jim Caton and I had just published. What we show is that it’s not going to be possible to govern a firm or a society entirely algorithmically through smart contracts or things like that, because a smart contract is a great commitment mechanism, but it’s entirely explicit. You have to say “if x then y.” That’s not going to be able to get you very far in cases where you have social dilemmas like like finance. So why have blockchain applications been so fruitful everywhere but in finance? That’s why you can’t borrow and lend on the blockchain still.

So ultimately some kind of non-articulable decision that you can’t explain is going to be necessary in order to protect against novel forms of defection. And the inability to explain that is precisely what generates its commitment power.

I want to end by identifying inside and outside perspectives in the discipline of economics and political philosophy. Because it would be a mistake to think that we as economists are all on the outside perspective. A lot of what we do is, but a lot of what we do is also an inside perspective. I mentioned already Buchanan talked about the role of the political economist being to establish a mythology or a civic religion. And I don’t think he was saying that metaphorically. So what he’s doing in for example The Calculus of Consent is he’s connecting moral intuitions to existing institutions. He’s engaged in a project of legitimacy.

Doug North I think is more on the outside perspective, also a Nobel Prize here, but he’s taking more of this functionalist view, thinking of things in terms of objective payoffs. You could also think of Lock and Hobbes as doing different things. So Locke is on the inside perspective; he’s talking about natural rights, legitimating this idea of English liberty. Whereas Hobbes is worried about this war of all against all, and people’s incentives to cooperate in a society. Now i don’t think Hobbes’ solution to his dilemma is very fundamental, because again the overall leviathan is itself a public goods problem. So he hasn’t solved that, but the fact that things like Locke’s work exist – Locke’s work doesn’t respond to Hobbes, but the fact that Locke’s work exists, the fact that we are prone as people to generating these kinds of legitimating structures – that’s where we get the answer to Hobbes’ dilemma.

So political economy has feet in both perspectives, and that’s okay. As a lot of radicals are very fond of pointing out, economics plays a somewhat minor role in legitimating the role of markets in modern life. And it relies on simplifying assumptions that are not strictly true, or it relies on tautologies in precisely those places where the danger of defection would be greatest. For example I really like the the 1993 Krugman paper about the Narrow and Broad Arguments for Free Trade where he argues that classical trade theory is kind of a noble lie. It’s true in principle that maybe we could tell an optimal tariff story that would do better than free trade. But if we think about the incentives here, the point of classical trade theory and its assumption of perfect competition is to obscure opportunities for rent seeking. Because if we don’t obscure those opportunities for rent seeking, we open ourselves up to the possibility of doing better with an optimal tariff, but more likely the Nash equilibrium is worse. And so it’s good that we pretend these opportunities for rent seeking don’t exist.

Professor Wagner, your book with James Buchanan on Democracy in Deficit, same sort of argument, that maybe it’s true that we can do better for ourselves by active aggregate demand management. But the point of classical public finance is that it obscures these opportunities for rent seeking that make it more likely that the Nash equilibrium is going to be worse if we open up the door for active aggregate demand management.

And so another implication here is how do we think about iconoclasm. And we can think about critical theory for example as kind of systematic iconoclasm. Foucault, for example, had a very expansive understanding of power. And he would say that if there is a group of people that is convinced to act in a way against its own narrow interests, that is an exercise of power against those groups. What’s he doing there, he’s trying to attack the inside perspective and outside perspective distinction. He’s saying that these supposedly marginalized groups could do better for themselves if they were to decline to buy into these noble lies.

Now whether or not the critical theorists have actually identified the right groups, and whether they in fact are the ones that have the possibility of productively defecting (productively for themselves), it is true that if you understand social relations as power relations, that’s a valid analytical thing to do. But it’s not going to be possible to sustain a society that way. And so what makes this especially potent is the combination with liberationism. Foucault himself, you can read him as having some appreciation for the functional value of exercises of power in this sense. But the literature the subsequent literature has been mainly liberationist in the sense of trying to inform these people, hey, there’s an opportunity here for you to do better for yourselves by minding your objective fitness.

Now i don’t want to pick on critical theory too much, because the game theory that we’ve been doing throughout this paper points the same possibility. There’s no liberationist tradition of game theory, which is probably a good thing, but at least in terms of the kind of analysis it’s doing. But it points to the same possibility. If we’ve used game theory to identify this outside perspective and distinguish it from the inside perspective, we’ve also eaten from the tree of the knowledge of good and evil. And we’re gonna be liable to get ourselves kicked out of the garden of good institutions.

So what do we do with that? One thing that i think we all appreciate at some level is the importance of tacit knowledge in institutional design. But I want to think of this tacit knowledge not as just information that we’d like to incorporate but it just happens to be inarticulate. There’s a reason that that knowledge is inarticulate, and it pertains precisely to those areas of social life where the incentive gap is the most salient. That’s why it’s hard to transplant institutions, because you can’t do this instrumentally. If you try to convince people you should believe this because there are good consequences, who’s going to do that? So it has to be an organic process because it has to be invisible. You can’t know that that’s what you’re doing, and you can only get there by pretending you’re doing something else.

It’s probably not necessary for social scientists to censor themselves, for reasons i talk about in the paper. Even so, even when economists run afoul of sacred values, we shouldn’t disparage the impulse that takes offense at us. One example I use is the value of the statistical life. When we put a dollar value on human life in order to get at the fact that there are real tradeoffs, there are real opportunity costs of a lot of things in terms of human life and we have to be able to navigate those tradeoffs, people get offended at this. They say you’re putting a dollar value on human life, that’s terrible. Why why are they offended at that? Because putting a dollar value on human life opens you up to all kinds of avenues of defection from good society. For example, assassinations, paying to murder somebody, things like that. That’s the kind of thing that people jump to. And even though we know that we’re really trying to save lives here, we’re really trying to take account of opportunity costs, we shouldn’t pooh-pooh the impulse that leads people to be wary of those kinds of things. And again we shouldn’t not criticize or examine norms and beliefs, at least particular ones. But it’s also important not to reject the idea of sacredness in general, because that’s an important way that the incentive gap is crossed in one way or another.

So the structure of social life is such that it’s never going to be possible to come up with a mechanism or a strategy to eliminate the incentive to defect, and therefore that’s going to require separate inside and outside perspectives, that’s going to require something that sounds like a noble lie.

Q&A

Abby: Do both perspectives fail on the axis of open-ended deduction/induction, given that deduction is the neoclassical theory, induction is the evolutionary neoclassical theory. So deduction is game theory proper, where individuals deduce solutions in analytical fashion by deriving their best response mathematically, and induction is the evolutionary strategy where we drive things like grim trigger strategies and tit-for-tat kind of mechanics. So in either of these cases we still are looking at closed systems right? So for either strategy to be meaningfully informative, we are necessarily in the world of closed systems. But i don’t think that your analysis necessarily lives in the world of closed systems. I think it necessarily lives in the world of open-ended systems.

So that last remark is interesting and I might have you elaborate on that. The way I’m thinking of this is that especially when we talked about rules versus discretion for example, the conventional wisdom works only in an open system when there’s something external to the system that imposes upon the system. And I’m thinking of this as a closed system analysis because even if you have two systems, one of which can impose on the other, the system as a whole is going to be closed. There’s going to be some level at which outside of that system, there’s nothing else to impose. And so if you haven’t closed the incentive gap at that level, well you have a closed system. And what are you going to do with that?

Abby: I’m not sure I buy that. So i do get the idea that we can layer these external worlds, by which we have some kind of closed system in the limit. But this is not something anyone has access to. And so this is where discretion lives: discretion lives in the external world that is far enough out that it cannot be accessed. It has to, or else someone would access it.

Right, so you can think of the that discretion as simulating an external—

Abby: Okay, so the belief in that sense creates that distance. So okay, maybe that is another way of understanding it yeah.

Pavel: I’m generally unsatisfied about altruism-based explanations, because it seems like an exposed kind of a move where you simply shove whatever you want into the utility function to explain whatever happened in your backward induction. I think the work by Vernon Smith, who builds on Adam Smith’s idea of people striving to be loved and lovely and avoid more than anything else to be blamed and blameworthy, gives you a different kind of motivation, and the puzzle largely disappears. So then you can say well how do I know I’m loved and lovely, and Smith has four sources of moral approval which are both consequentialist and deontological, meaning you want to act according to the rules of the game in the community where you are, but you also are mindful of the impact of what you’re doing on the person that receives whatever is at the other end of the exchange. And you also want to have some kind of an understanding of the overall impact of what you’re doing, which is much more kind of a complicated exercise than just maximizing pleasure. But i think it has evolutionary connotations to it that you are interested in. So Vernon Smith might be something to consider for you. I’ll share the paper in the chat.

Thank you. And i do actually criticize also the the same kinds of things that you opened your criticism with, is the idea of just assuming altruism in the utility function. That’s not very satisfying, because we want to know where does that come from? But I would also ask the same question of propriety. And in this evolutionary literature, altruism has the specific meaning of anything that you prefer that benefits others rather than yourself. So I would call propriety altruistic in this sense as well, noting that this isn’t something that I’m just assuming and going from there, but rather that I’m looking for the foundations of both altruism or propriety or however you understand that. So i do think it’s definitely compatible with that kind of analysis too.

Roger: This is hugely ambitious, which I think is good. And an ambitious project is one that people will struggle with. So all of that is by way of two thumbs up. So apologies if I’m just behind here, but it seems to me you’ve got three distinctions running: subjective/objective, altruistic/selfish, insider/outsider. And I’m not really getting how they relate. I thought maybe it was outsider equals selfish equals subjective. But I think Jim Caton said no that ain’t right. Help me please.

So objective and subjective refer to just preferences, versus something like fitness in an evolutionary model, or income in a Alchian-Becker model, or some objective interpersonally comparable criterion that allows you to go into the next period in a better position. So that by itself—

Roger: I’m sorry, who is you – is you the organism, or–

The organism, yes. So by selfish and altruistic I would not say that’s the same distinction. Selfish refers to a situation where your subjective preferences correspond to your objective fitness. That would be selfish, and that’s what I’m calling homo economicus here. And altruistic would be a situation where your subjective preferences include some other element than your own objective payoffs.

Roger: So objective would be something like fitness or income. What would subjective be?

Subjective would be the basis on which you personally make your decisions. This would be like the standard conception of utility in econ.

Roger: Ok. Altruistic versus selfish i think we know.

And then the inside and the outside perspectives are approaches to institution. The outside perspective is in terms of objective fitness, and so takes that as given. The inside perspective takes subjective beliefs and preferences as given and doesn’t ask where they come from.

Roger: So insider/outsider the same as subjective/objective.

Yeah, I would be happy with that.

Marta: Would it be practical to give up the the objective/subjective going forward and just stick with inside/outside? I don’t know if that’s a semantic point.

Jim: Thanks for the paper Cam. So the insider outsider/perspective distinction reminds me of the exegesis/eisegesis issue But the problem is – and i think this was Abby’s point – is that exegesis is itself a kind of eisegesis. Can we really use this word objective when we’re engaged in in this analysis? Does that have a proper place?

Can you elaborate on what you mean – this has to do with textual interpretation right?

Jim: Yeah right so you talk about textual interpretation, but it’s also in some sense like a cultural interpretation. It comes from religious studies largely, we talk about exegesis and eisegesis, so the perspective within a group, and of course that was one of the kinds of groups that you had identified. But then the exegesis is something more objective, or the intention of it is to be more objective. Which, intention’s all good, but at the end of the day I don’t think we escape the subjectivity. And we’re maybe we’re trying to approximate objectivity, which would be again the exegetical side of the thing. but it’s it’s not clear to me that we can we can make that leap even analytically.

So i think that actually gets partly at the necessity of the inside and outside perspective distinction because you can ask from an outside perspective, where do subjective values come from, where do subjective interpretations come from, and why are they what they are? And you can tell a story like that in terms of the outside world and the forces operating in it. On the other hand you can also ask the question, where does the objective world come from in terms of our perceptions? Why do we perceive something as an objective external world? And ultimately we’re going to have to tell that story basically in terms of subjective elements. And so I do think both of those are valid approaches it’s a question of what are you telling the story in terms of. Like what are you taking as your fundamental? And i guess one thing that the argument here implies is that it’s valid to go either way, and that you can’t go one way and then expect to obviate anybody who’s going the other way.

Jim: So my one stick is the idea of objective payoff. I’m hearing something more like material resources and power when you mean objective. And maybe we could more particularly define that.

So this this is actually something that one of the reviewers complained about too, is that i was trying to be general. And like I mentioned to Roger I’m trying to tell a story that applies both to reproductive fitness in a an evolutionary story and to income in an Alchian-Becker type story. And so I think the fundamental characteristic that makes something an objective payoff is the interpersonal comparability. It’s something that you can compare between people. You can compare the number of offspring somebody has, you can compare the income that people have, and then that is going to affect the viability of the strategy going into the next period, the more children you have, the more those strategies proliferate, the more income you have, the more able you are to effectuate that strategy. And so that strategy is going to do better the more income you make. So it’s really that interpersonal comparison and affecting the next period, that’s what i mean by objective.

Richard: I hope you can help me out Cameron because i have the distinct impression that you’re trying to square the circle with the same results as everyone else has obtained who’s trying to do that. And there I think with both your presentation and the paper and your work you’re trying to do with the inside outside kind of distinction in which I think what you’re wanting to say is that trying to build up and explain a well-working society starting with a single individual and assembling a whole set of individuals through that, you’re never going to get there without invoking some kind of noble lie that’s going to somehow bind all those people, because you need some binding agent. And yet, I don’t think that can be done. And also besides not necessary I think, because we have no historical record in the slightest about that kind of process. Everything we have suggests that however we got there, we originated in groups. Even Eve was fashioned from Adam’s rib, and it went from there. And then you get to things like Norbert Elias – I think he has the Civilizing Process book – who was fundamentally raised in this question of how do adults become adults starting from infancy. They all grow up in some kind of group setting where values are instilled, modes of conduct are created. There’s various kinds of anthropological stuff on this, also archaeological about tribal life, was generally pretty peaceful, that the idea of outlaws came about because tribes wouldn’t kill tribal members. They’d banish them. And that could be a scary thing to have happen. But only as territory became filled up we got what we call more civilized orderly procedures that when you get to that point, you get in necessary conflicts because part of our nature i think is to prefer to govern someone else. And to govern ourselves I don’t think there is any kind of generally cooperative way that that’s going to get handled. And I would remind everyone that I think for instance pugilistics is a social process. So ‘social’ i think is a very badly misconstrued adjective if you think it just means harmony peacefulness and dancing and stuff like that. It also means fighting. Because one person can’t fight with his or herself. And so then when you get to this, the inside/outside and subjective/objective, I think there is a very important distinction there, but I think it’s the distinction between participants and spectators. Like game theory is of no use to playing a game. Game theory might be of use perhaps to observing the game. I doubt it, but the perspective of spectators is a wholly different thing than the perspective of people on the ground. Take a relatively simple game like football – it’s not so simple actually – but the feel of how a play unfolds to the participants has a whole lot of impromptu improvisation going on. But you can’t put that improvisation into the format of game theory. And yet there are those different perspectives for sure, but I think in terms of social studies, it’s more a difference in orientation between some people who have arrived at a position, whether by accident or by choice, where they like the position they have, and it includes a position of domination generally. And there are others i think who aren’t in that position, perhaps would rather not be dominated, but even perhaps don’t care as much as those who want to dominate. And so I think this insider/outsider perspective is an important one, but it’s not – which I think your paper verges on – is that in a way it’s it’s all individuals as as actors, but individuals occupying different positions within some kind of social structure that leads them to want to act in in different ways, and not necessarily all compatible or congruent ways. I get the impression you’re going down a kind of placid social equilibrium path, and you want to say that’s that’s not going to work. Where I myself am very happy to start off with recognition going back to Elias then that societies are fundamentally arenas of ongoing civil warfare. And the basic social challenge is, you hope that it can be kept more or less simmering rather than boiling. But historically we know, on the one hand, I don’t think there’s any evidence anywhere historically of a placid peaceful society, and there’s plenty of evidence of modest variable degrees of simmering, but there’s also plenty of historical evidences of downright fierce boiling. And so I like the basic kind of structure where you’re trying to go with it, but I had the impression that it might be better if the metaphors were aligned more more consistently, is how I’ve reacted to what I’ve read and what i heard today.

Yeah i think i agree with most of what you’ve said here and I like the idea of maybe instead of inside/outside perspectives, talking about participant perspective and spectator perspective. Maybe that makes things clearer with respect to an earlier question. One thing that I think is the next step of this kind of analysis is to think about what kinds of differences there are here. So you mentioned you know clearly this problem must have already been solved way early in prehistory since we’ve had tribes for a very long time. On the other hand something qualitatively different happened at the Neolithic revolution. The kinds of noble lies that we were telling ourselves changed, and then therefore the structure of our social cooperation changed. And talking about societies as boiling, or taking a turbulent versus an equilibrium perspective, that’s part of what i was struggling with while writing this paper, is that i don’t think it’s satisfying to just make a decision like that, to stake out a claim and say I’m approaching this as a social equilibrium theorist, or I’m approaching this as doing kaleidics or something like that. I am in some sense trying to square that circle and showing you know why do we always have to deal with that boiling, why there are going to be some respects in which things are never going to settle down into a stable equilibrium. Why do we always get this pervasive conflict, and despite the fact that there are lots of arenas where we do find reasonable equilibria, clearly there is a lot of turbulence, a lot of boiling. I don’t want to dismiss that as somebody who just kind of assumes social equilibrium. But I also want to show specifically where that comes from rather than assuming it.

Richard: Well I think we all face this kind of question, what are the kinds of what you might call stylized facts with which we’re trying to deal or explain. And if we’re trying to explain I think in this case, the material here is always something about the “what it’s like” or what it means to live within particular societies, what its features are, just like the DSGE model assumes a basic flat line equilibrium in societies unless disturbed by exogenous shocks. Where if you want a framework that says that there is continual novelty that takes place within society, you can’t use a DSGE type model, because you need some place from inside the society for the new energy to be inserted. Which means you can’t have everyone playing Nash equilibrium and so on and so forth. And so the whole constellation of assumptions that pair with different kinds of explanatory frameworks. I think the right orientation is that there’s no point in theorizing how to get Kant’s eternal peace. He wrote a little monograph called internal peace once upon a time. And there’s no way we’re going to get there. But there might be a way that we can get to that the world is relatively friendly. And it’s not the world we live in now, and there are some places that are less friendly than others, but it’s that question of what we’re trying to get out of our theories. And I think that is what needs to direct our questions. And maybe it’s just showing my own prejudices or biases or whatever, but I think there’s an awful lot of economic theory over the last century or more that has taken the form of some people want to run the world for good. Of course they’ll always tell you it’s for good. They want to run the world for good, and they want to use economic theory as some of the constructs to propel them into that position of heading some kind of crusade. The American Economic Association was founded on this kind of progressivist orientation in its very beginnings. So I think it’s those kinds of asking questions about what you think our theoretical efforts or conceptual efforts should help us to do is going to depend on what we think it is we want to do as well as what we think are the undesirable influences that are in play there. And I think there are a lot of undesirable influences that are in play in economics.

Vlad: I want to push back a little bit against the reluctance to accept that generosity is real, or that there’s something wrong with using preferences like that. On purely empirical grounds, I think it’s very bizarre to deny that we care about other people. I genuinely care about them. So then why would you deliberately want to have a model that is empirically false? So then if you just insist on having purely selfish preferences, then it creates all sorts of problems that you’re highlighting. It’s very difficult to make sense of social order if you just assume that everyone is purely selfish. But that’s just a a fake problem that you’ve created arbitrarily by having an unrealistic model. So now you raise the good question: if you’re assuming altruism or some level of generosity and also distinctly from that of various norms, which is a different type of preference like say fairness, that’s different than generosity. So if you assume these kind of things, where do they come from? So my question to you is, what’s wrong with the typical evolutionary psychology story which says that we evolved in small groups, so we still have that mind from small groups, despite the fact that now we’re in giant groups? For example the evolutionary story is that you care about others simply because in those small groups they were related to you, so there’s that Selfish Gene kind of argument. So of course that’s no longer the case in big societies. But our mind doesn’t really know that. The intuitions are still there as if we are in a group in a small group. So in a sense it’s a misfiring of an intuition that was developed in a different evolutionary environment, but we still have those intuitions. And those help us to some extent to build the social order in our society. And of course there’s also problems that are created by those intuitions like say xenophobia or stuff like that. So it seems like part of the answer here is to look into this kind of other-regarding preferences, rather than just insist that we must assume selfishness.

I very much don’t want to deny the reality of generosity and altruism. As a matter of fact that’s exactly what the the group story allows me not to do. Because if you’re a Dawkins or a Pinker telling the Selfish Gene story, then you do have to assume that generosity is in some sense fake, because you have to imagine it as something like a misfiring that’s intended for one purpose, but not accomplishing that purpose. You think you’re being selfish but really it’s not. What the group story allows me to do here is to say that, no, generosity altruism are real. We really do care about each other as ends in themselves. And my use of homo economicus in this paper is not to say that we should use a model of selfish behavior, but to use that as kind of a benchmark. Homo economicus is the benchmark of dynamic stability, because an agent that prefers its own fitness is going to be the one that survives. So that leads us to the question, well clearly that’s not the case for humans. Clearly we are generous people and altruistic people. So what kinds of mechanisms would we have to imagine that would lead humans to be not the kind of homo economicus that sometimes economists would assume. And by telling a dynamic story, it is one thing to say, put altruism in the utility function. Yeah you can do that, and that probably gets you a more empirically accurate model in some respects than your standard homo economicus model. But if you just throw altruism in the utility function, you’re telling a static story, and you’re not asking a question of, first of all where did altruism come from? And second of all is that going to survive? Have the conditions today perhaps changed such that altruism is not a good strategy? Maybe we really are misfiring, and if that’s true we’d want to know that, because that to me is an extremely nihilistic kind of story. If it’s a misfire, then we’re not going to expect that to survive, if it is in fact maladaptive on the individual and the group levels. So part of what I’m interested in doing here is thinking about the future conditions as well. Not just the development, but the development sheds light on what we can expect in the future. And one of the things i mentioned in the conclusion is that we have two mutually reinforcing dangers that we got to watch out for. First of all the dominance of individual level selection, which would make it a true misfire, and which would decay the altruistic tendency in humans. And second is anti-social punishment, the use of altruism to prevent mutually beneficial norm changes, and to prevent certain forms of cooperation. And i think we’re seeing that a lot with – i could point to specific examples; i don’t want to get too in the weeds of current events or anything – but i think you can see some of that happening. And if that’s the case we’ve got to know where that impulse came from, and how to adequately deal with that. And knowing why it came about, and maybe some obvious strategies that wouldn’t work for the same reason that it’s hard to transplant institutions, it’s going to be hard to fight back against this kind of anti-social punishment. So we’re going to have to be creative here.

Vlad: Yeah, I think that maybe we would disagree far less than I thought.