In Defense of Selfish Rationalism

by Henry on September 9, 2010

(as an analytic approach mind you, not as a way to behave).

Chris and Lenin argue back against my defense of rational choice theory for lefties (and John Quiggin has some criticisms scattered through comments too). Below the fold is my clarification of my argument, at Holbovian length, with some responses folded in.

First – some concessions (most of which I’ve made somewhere or another already, I think – but not in one place). (1) Selfish rationalism is not the only possible set of microfoundations for theories of human behaviour by any means. (2) There are large and important swathes of behavior that it explains poorly, and some forms of behavior that it barely explains at all. (3) As John Quiggin has argued repeatedly on this blog, rationality is often used by economists and political scientists in tautologous ways. (4) The specific form of egoistic rational choice analysis that I am going to defend here – non-cooperative game theory – makes ridiculously far-reaching assumptions about e.g. actors’ cognitive capacities, degree of common knowledge etc. (5) It furthermore has (as one of its foremost practitioners, David Kreps forthrightly acknowledges), become dominated by ever-odder and more demanding equilibrium refinements that seek to solve a fundamentally insoluble problem of social indeterminacy through manifestly inadequate technical fixes.

Yet despite this all, I think that it is highly valuable – and especially valuable for lefties – for two reasons. First – it provides a very useful testing bed for arguments about putative social change, and the mechanisms through which they can be achieved. Second – it highlights fundamental incoherencies in the standard economic accounts of politics that economists themselves usually glide over serenely (see Brad DeLong’s aptly titled elementary mistake post for a fine example of this in action). In all the below, I lean heavily on discussions I’ve had with Jack Knight and Jim Johnson over the years.

I’ll start with a quote from Jon Elster’s The Cement of Society) (written back in the days when he used to like rational choice).

[The above] does not add up to a claim that the social usefulness of norms is irrelevant to their explanation. I find it as hard as the next person to believe that the existence of norms of reciprocity and cooperation has nothing to do with the fact that without them civilization as we know it would not exist. Yet it is at least a useful intellectual exercise to take the austere view and to entertain the idea that civilization owes its existence to a fortunate coincidence. (p.149)

This rather Gellnerian formulation nicely captures my personal take on rational choice theory – it is valuable not because it is right in some universal sense, but because its particular form of intellectual austerity is useful. But this of course begs the question: useful, how? In two main senses.

First – because non-cooperative game theory is astringent in ways that are particularly useful for the left. To quote myself from a few years back:

Starting from assumptions of rationality, narrow self interest and so on is an excellent way of assaying the worth of left wing arguments under unfavorable conditions. Lots of lefty arguments assume that humans are more or less benign and sociable. They may sometimes be right – but making assumptions of this sort also allows lefties to load the dice. If, in contrast, lefty arguments still work when the dice are loaded against them, say, by assuming the kinds of stark clashes of interest that are often modelled in game theory, it suggests that these arguments are likely to be robust under a variety of assumptions and conditions.

I’ll unpack this a little. We live in a political world which is characterized by stark disagreements, many (although not all) of which are the result of clashes of interest between different groups of actors, with different preferences over how the benefits of social production ought to be distributed (I am borrowing this from Knight and Johnson’s excellent forthcoming book with Princeton ). Non-cooperative game theory – if it is used properly – can capture this world nicely. It too portrays a world where interests clash starkly in ways that political discussion of the kind proposed by e.g. deliberation theorists may be inherently incapable of soothing. If we think in these terms, we are obliged to pay proper attention to the possibility of distributional conflicts, clashes of interest, problems of collective action etc that often get elided in leftwing accounts of how we can create a fairer polity.

NB - this is not to say that all accounts of social and political change which fail to take proper account of such conflicts will be unrealistic or unworkable. Because – as Chris stresses – human beings have innate dispositions towards sociability and fairness, they will be able to accomplish forms of cooperation that we might consider impossible if we considered their selfish material interests alone. Even so, human beings are very often self-interested too, so that accounts which focus on self-interest will capture important aspects of social interaction. Furthermore, leftwing proposals for social change which are feasible under assumptions of irreconcilable clashes of interest etc will, by that token, be particularly robust ones, since they can succeed even under circumstances that seem inhospitable.

Second – just the same kind of robustness tests can be applied to many right-wing and/or market liberal assertions about social order, with equally astringent consequences. Lenin sees market-based reasoning and rational choice as being impossible to separate from each other, arguing that:

Henry’s most basic assertion is that rational choice models don’t contain inbuilt normative assumptions, or theses about the social world that would lead to them having a right-wing bias. It is certainly possible for rational choice theorists to resist right-wing interpretations, but to do so it would be necessary to acknowledge a few things that Henry does not. First of all, to return to one of the previous quibbles, the origin of rational choice theory is precisely in the Cold War state and the right-wing free market theology it produced, which Henry claims are somehow neatly separable from it.
Secondly, the founding assumptions of rational choice theory are directly imported from classical liberalism. The idea that social action ultimately takes place at the level of the individual, that people are utility maximisers, and that such maximisation optimally takes place in the process of exchange is obviously derived from liberal, contractarian political economy. …
It is obvious enough that rational choice theory is not necessarily right-wing, since there are plenty of left-wing exponents of it, but nor is it avalent. And insofar as there have been attempts, largely in the retreat from Althusserian structuralism, to reconcile rational choice with marxism, the tendency has been for one to be adulterated rather than complemented and strengthened by the other.

Here, I strongly disagree – and I think I have good reasons for disagreeing. First – the social origins of ideas does not necessarily tell us about their implications and how they may be developed by others (e.g. while Marx disagreed with the classical economists on many issues, he was also profoundly and explicitly indebted to their way of thinking about the world). Second and by far more important – the ‘valency’ of the underlying theories of selfish rationality can cut against the broader claims of e.g. liberal claims about the benefits of markets as a form of governance and the social efficiency of market supporting institutions.

Brad DeLong’s rather odd post attacking Chris provides a nice example of how economists commonly mistake what might be described as the ideology of economics for its microfoundations. Brad cites Adam Smith on “the propensity to truck, barter, and exchange one thing for another” as microfoundational to economics (here, he seems to agree with Lenin). But – as several of Brad’s commentators point out – the actual microfoundations of economics – as it is taught – consist of mathematically derived assumptions about the behavior of self-interested individuals. To put it a different way – without agreeing with Lenin’s suggestion that rational choice is ineluctably shaped by its origins, he is entirely correct to note that it provides the microfoundations both for modern economic theory of market behavior and for the kinds of Cold War military strategic analysis pioneered by Thomas Schelling. Brad would prefer to concentrate on the market behavior stuff and forget about the military strategizing. But this is to mistake a set of ideological claims (Albert Hirschman’s ‘doux-commerce’ thesis) which owe more to sociology than economic theory, for the actual microfoundations that are employed by most economists when they try to come up with models.

Borrowing from Knight and Johnson again – the specific claims that Brad and many other economists make (that ‘truck and barter’ prevail over ‘pillage and dominate’) at best apply under a highly restrictive set of conditions, given assumptions of selfish rationality. It is only when actors completely lack strategic power, or when they are indifferent among available alternatives, that they will not seek to use what power they have (which may include e.g. forms of violent coercion) to secure distributionally favorable arrangements. These conditions clearly apply under perfectly competitive markets with full information – but such markets are a theoretical curiosity rather than an empirical fact. Even rough approximations of them are vanishingly rare in real life.

This matters – because economists have indeed sought to make claims that implicitly extend perfect-competition arguments to other social domains, as Lenin notes. But contra Lenin, these are exercises in market-imperialism rather than rational choice imperialism, and there is a very important difference between the two. Indeed, if one actually systematically applies rational choice logic to these arguments, their logic usually collapses.

Debates over institutional economics provide an important example of this. Roughly speaking – institutional economics began when economists began belatedly to realize that institutions did indeed play an important role in structuring economic behaviour. But economists – being economists – argued that the ‘institutions of capitalism’ were produced by actors seeking to lower transaction costs and increase economic efficiency (see e.g. Oliver Williamson). The problem of course is that actual rational self-interested actors are, as the description suggests, only going to prioritize achieving social benefits such as lowering transaction costs under highly unusual conditions (Knight). Most of the time, they may be expected to prefer pushing for arrangements that give them a preponderance of the goodies – even if such arrangements are relatively inefficient. In short – rational choice – if you take its arguments seriously – suggests that economically rational actors are not usually going to behave in the ways that institutional economics claims that they are.1

This critique generalizes quite nicely. If you take arguments about rational self interested actors seriously, then you will only find them behaving in ways that maximizes social efficiency (however you want to define that rather fuzzy term) under quite unusual and stringent conditions. Under conditions where they can exercise power strategically, through threats, coercion or other means, and where such exercise will be to their material advantage, they will do so, even if this hurts social efficiency very badly.

In other words, there is a serious mismatch – perhaps even a contradiction – between the valence of rational choice and the ideological uses that are made of it by economists. Unless you pay attention to the microfoundations of these arguments, these contradictions are likely to slip by. To put it another way – there is nothing within the underlying concepts of rational choice which requires, or even implies, that the terms of exchange between actors be equal. This is a good thing, since such a requirement would be in gross contradiction to what we see over the few millennia of recorded human history. Yet just such equality is required if many of the ‘nice’ qualities of market, political and social equilibrium are to obtain.

There are some further disagreements with Lenin which I may go into in a later post. But this – such as it is – is my argument for the value of rational choice theory plus egoism for leftwing analysis.

1I note that Douglass North is a partial exception – he wavers inconsistently between a framework in which institutional change comes about through actors seeking to lower transaction costs and in which it comes about through powerful actors seeking distributional benefits. The partial tendency towards the latter may be in part explained by his initial theoretical interests – he used to be a Marxist, and described himself a few years ago in conversation with me as still being a ‘Marxist of the right’

{ 70 comments }

1

dsquared 09.09.10 at 7:33 pm

I am going to perhaps push back a bit on the implicit assumption that if we were to depart from rationality as a modelling technique, the first departures we would make would be in the direction of “nice” qualities like solidarity and charity. Selfish rationalists aren’t racists, and they know that bygones are bygones. There are no “age-old ethnic hatreds” in the world of game theory. As with a world system in which America dominated other countries on the basis of pure national self-interest, selfish rationalism might not be a very great way to behave, but it’s probably better than what we’ve got now.

2

dsquared 09.09.10 at 7:36 pm

I’d also suggest that the problem with institutional economics as described above is basically an analytical one. The whole approach was to say that we’ve got a model of transactions costs, and we’ve got a model of principal-agent problems, and we’ve got a couple of other toolkit models, so we can model institutions. But any institution worth analysing is going to consist of literally thousands of layered and connected transaction-cost and agency problems – trying to analyse it with the game theory toolkit is like studying diagrams of logic gates and hoping you’ll learn how to play Super Mario Brothers.

3

Metamorf 09.09.10 at 7:42 pm

To put it another way – there is nothing within the underlying concepts of rational choice which requires, or even implies, that the terms of exchange between actors be equal. This is a good thing, since such a requirement would be in gross contradiction to what we see over the few millennia of recorded human history. Yet just such equality is required if many of the ‘nice’ qualities of market, political and social equilibrium are to obtain.

I’m curious what you think those “‘nice'” qualities are. Outside of mathematical models, market proponents generally recognize that:
a) the “terms of exchange” — assuming a broad enough definition of the phrase — between actors are never equal;
b) nevertheless, a trade as such must be to the perceived benefit of both or all parties, or it doesn’t happen — which is its primary “nice” quality;
c) trade, as opposed to looting, generally requires a legal framework and a legal system to support it; and
d) there are other forms of exchange than trade (and which can also be accommodated within rational choice models) — see this post, e.g. — but trade has advantages for all concerned.

For more on that last point see also this post.

4

John Quiggin 09.09.10 at 7:49 pm

A critical point about markets is that, in the standard (anonymous, perfectly competitive) market setting, there is no scope for anything except egoistic rationality: selling your labor at the going wage and buying the bundle of goods that maximises your utility. And, even when markets aren’t exactly like this, it still makes sense, most of the time, to make the market choices of an egoist and to pursue non-egoistic goals outside the market. The arguable exceptions are cases where you aim to influence market structure through your choices eg by buying Fair Trade Coffee, but this requires collective action organised outside the market.

5

chris 09.09.10 at 7:59 pm

Lots of lefty arguments assume that humans are more or less benign and sociable.

Most people are, most of the time; but any society contains some sociopaths (aka complete jerks, or as I think they’re called across the pond, right bastards), so if you don’t have some safeguards against them, they’re going to exploit your system for everything it’s worth. And ordinary people will kick *some* people, some of the time, if they think there’s a good justification or just think they can get away with it. (And I haven’t even started on stupidity and misinformation — whether species-innate, cultural, individually idiosyncratic, or deliberately induced by a manipulative other — which undermine the rationality of the relationship between goals and actions.)

Most large groups are heterogeneous — this principle should be wielded against any social, economic, or political theory with a simple model of what “people” want, how they behave, etc. A workable social, economic, or political theory has to be able to deal in some way with the actions of everyone in the society, including the selfish ones.

6

Chris Bertram 09.09.10 at 8:10 pm

Too much there for me respond to tonight, but a couple of observations ….

(1) It seems odd to take non-co-operative game theory as the exemplar of “selfish” rationality (rather than say, Buchananite public choice – a better example). The agents in game theory rank the various outcomes, but the reasons for them doing so need not be selfish at all. In fact, we never get to know what their reasons are, just what their ordinal preferences over outcomes are.

(2) It also seems odd to _assume_ that the more socially oriented the motivation, the more “lefty” it is. In some circumstances, sure. But given that many of those motivations are group-oriented (rather than based on impersonal morality), less individualistic motivation can also be very nasty and right-wing.

7

Chris Bertram 09.09.10 at 8:18 pm

Sorry, extra “methodological” point ….

I can understand a positivistic or pragmatic justification for making unrealistic assumptions about motivations. That justification is, presumably, that theories constructed on such foundations yield better predictions than theories constructed on other ones. But, other things being equal, I think it better to argue from true rather than false premises. That being so, there are good (but defeasible) reasons for social science to start with an account of motivation based on the best work in its branch devoted to such matters: empirical psychology.

8

chris 09.09.10 at 8:26 pm

b) nevertheless, a trade as such must be to the perceived benefit of both or all parties, or it doesn’t happen—which is its primary “nice” quality;

Setting aside interactions like slavery that clearly don’t satisfy this condition (but note that coercion can sometimes be in the selfish interest of the coercer), perceived benefit isn’t the same thing as benefit, so the “nice” quality still leaves the door open to fraud.

c) trade, as opposed to looting, generally requires a legal framework and a legal system to support it;

I’m not sure that this is true. It may be sufficient that both parties expect that violent conflict will be costly (i.e. dangerous) to them, and that the trade will be profitable if it succeeds. That could give them both a self-interested incentive to keep the trade on terms that won’t provoke the other to violence, which will work at least some of the time.

If one party is *certain* that violence will go their way at little cost, then in fact you’re unlikely to see trade instead of looting, I think. The fact that the state can punish the instigator of violence and so make it costly in the long term even if the robber (e.g.) wins in the short term is precisely what enables it to suppress looting and promote trade, it seems to me (and the belief of the undeterred robbers that they can escape the state’s retribution is what makes this suppression imperfect, and why robbery rises when the state is perceived as weak).

Or, in other words, you need either mutual deterrence or state deterrence.

9

Henry 09.09.10 at 8:38 pm

bq. b) nevertheless, a trade as such must be to the perceived benefit of both or all parties, or it doesn’t happen—which is its primary “nice” quality;
c) trade, as opposed to looting, generally requires a legal framework and a legal system to support it;

The problem with (b) is that an exchange under which I agree not to shoot you in the head provided that you agree to give me all your money is to the perceived benefit of both or all parties, assuming that I value your money more than your life and you vice-versa. The problem with (c ) is discussed briefly above under the heading of ‘New Institutional Economics and efforts to explain where aforementioned legal frameworks come from: general worthlessness thereof.’

Chris B. – I look forward to seeing what you say when you have had a chance to think through your response. But briefly – the reason why I focus on non-cooperative game theory is because it can plausibly be said to provide a set of intellectual foundations that subsume the arguments of public choice as a special (and, I would contend, mostly unlikely) case, while the reverse is not true. Knight and Johnson have a nice discussion in the book manuscript mentioned above of how many economic claims (including those of public choice) trade on the prestige of economics’ general equilibrium results while themselves being partial equilibrium findings (where, pace Donald/Deirdre McCloskey, you can prove more or less anything you want). And I don’t think I’m _assuming_ in any general sense that ‘more’ social equals more lefty – apologies for bad writing if it seems that way. Instead, what I’m trying to do is respond to the reasonable critiques of you and other leftwingers who suggest that rational choice is problematic because it does not take account of social motivations, altruism etc (i.e. my response on this is a limited one to a specific set of critics, rather than a general theoretical claim). On this, it is also interesting that some of the evolutionary game theory literature takes the willingness to fight (and perhaps be killed) on behalf of your conspecifics as a paradigmatic case of altruism.

10

zosima 09.09.10 at 8:40 pm

@Metamorf

Your claim reduces to the claim that economic transactions lead to pareto optimal outcomes, but this is not the same as optimal outcomes. Using an analogy from computer science, these sorts of gains by voluntary trade work like Hill Climbing Algorithms. This means they have a propensity to get stuck in local maxima, but will often miss a global maximum.

A more concrete example is the monopolist who deliberately forces up the price of food by decreasing supply. Yes you’re better off by engaging in the voluntary transaction of buying food rather than starving, but you’d be even better off if you have an institution in place to prevent monopoly from forming in the first place.

11

Henry 09.09.10 at 8:43 pm

And building on Chris – there is an extensive literature (cf Kandori and Greif) on trade without legal systems. Roughly speaking, there is no problem with such trade when the tit and tat can be simultaneous. Even when they are not simultaneous, trade can be carried out via reputational mechanisms, community responsibility systems etc (see Greif’s book on the path to the modern economy). What this literature mostly neglects is the terms on which trade takes place – as Kreps observes, a pattern in which I credibly threaten to take 95% of the benefits of trade or else refuse to cooperate is entirely as plausible an equilibrium, given appropriate parameter values etc as one in which we split the proceeds 50-50.

12

Lemuel Pitkin 09.09.10 at 8:44 pm

I agree with dsquared and Chris B.

13

zosima 09.09.10 at 8:46 pm

One additional point. Assuming diminishing returns on utility as a function of wealth(ie assuming the absence of Utility Monsters), an account of economics with utilitarian foundations would suggest that an even distribution of wealth is utility maximizing.

14

Henry 09.09.10 at 9:02 pm

Finally Chris B – my defense of rational choice would not be on the methodological grounds that you suggest I might use (which to me smack worryingly of Friedman’s pernicious claims about prediction trumping realism). Instead, they have more to do with my (perhaps erroneous) perception that empirical psychology still has important limitations – e.g. the Tversky/Kahneman stuff on framing is undoubtedly _right_ at the micro level, but figuring out how it connects systematically to broad social outcomes is extremely difficult. If people come up with a psychological account that subsumes or replaces the kinds of rational choice I am talking about, in that it provides a systematically better account of the kinds of things that rational choice purports to explain than rational choice itself does, it would be wonderful. My understanding is that it has not done so to date. Given this, my preferred approach is ‘horses for courses’ – to try to figure out on a piecemeal basis which specific set of microfoundations is best for investigating a particular problem or set of problems, while being as self-conscious as one possibly can be of its limitations.

15

Henry 09.09.10 at 9:07 pm

Or to put it more succinctly – I think that the austerity of rational choice is useful because it provides an (admittedly partial and limited) way of thinking about one set of factors that structure social outcomes. If it was completely disconnected from the underlying causal mechanisms, but still had some predictive accuracy, I don’t think it would be especially helpful for the social sciences or for political theory etc (it might be to others who don’t care where their predictions come from – but not to me or the others interested in this debate).

16

chris 09.09.10 at 9:11 pm

Assuming diminishing returns on utility as a function of wealth(ie assuming the absence of Utility Monsters), an account of economics with utilitarian foundations would suggest that an even distribution of wealth is utility maximizing.

Only ceteris paribus — the traditional counterargument is that without the incentive provided by the possibility of obtaining more than an even share of wealth, people wouldn’t work as much or as effectively, and therefore the total wealth of society would be lower. It doesn’t seem obviously stupid to me (although it may be flawed in more subtle ways).

17

Metamorf 09.09.10 at 10:06 pm

The problem with (b) is that an exchange under which I agree not to shoot you in the head provided that you agree to give me all your money is to the perceived benefit of both or all parties, assuming that I value your money more than your life and you vice-versa.

But that’s not what “trade” means, unless you’re Jack Benny. Even Godfathers recognize a macabre joke when they use one.

As for the legal framework, that’s what enforces trade as such, as opposed to the joke above. chris is right that trade can exist without such a framework, but it’s difficult, tense, and usually not very stable.

your claim reduces to the claim that economic transactions lead to pareto optimal outcomes, but this is not the same as optimal outcomes.

No, it makes no claim about “optimal” outcomes at all — this is because the notion of “optimal” assumes a god-like position that only modelers or game-players can adopt vis-a-vis their miniature, impoverished creations. My claim is only a definitional one — that “trade” means a voluntary economic transaction.

(I know, btw, that people around these parts think there is no such thing as “voluntary” anything; or they think that any transaction that can’t be quantified by some third party — usually themselves — as exactly “50-50” in payoff, however that’s measured, whatever it means, is either involuntary, or equivalent to someone holding a gun to your head, or just somehow not fair. For those who dwell in such intellectual sandboxes, I can only say that there’s a bigger world out there.)

18

Mark Field 09.09.10 at 11:27 pm

What I wonder is why, if you’re right about the double-edged nature of rational choice theory, so many people have for so many years seen it as having just one edge? It’s not that I don’t believe there are occasional hundred dollar bills on the sidewalk (having once found one), but I guess I want more evidence that it’s real currency.

19

Lemuel Pitkin 09.10.10 at 5:29 am

Lots of lefty arguments assume that humans are more or less benign and sociable.

You’ve said this several times now. Cites?

Because, it seems to me that no one on the left assumes this. Rather, we believe that the “way humans are” is fundamentally and irreducibly determined by the particular institutions that those humans are organized into. So we reject the rational-maximizing-individual microfoundation, but not to replace with it some other microfoundation, but rather to reject microfoundations as such, in favor of a historically grounded account of concrete institutions.

20

Tim Wilkinson 09.10.10 at 6:35 am

chris 16 – this may seem pedantic but I don’t think it is: ‘ceteris paribus’ only makes sense when you are comparing two situations (e.g. a before and after), and stating the effects (or implications, etc) of specific differences between them when all other relevant factors are held constant. E.g. Chris B @7. You perhaps mean something like ‘prima facie’ or ‘subject to certain assumptions’.

(Sometimes people mean ‘pro tanto’ by it, which is a stronger claim than the very weak ‘ceteris paribus’, and roughly speaking signifies a (vector-like) effect that always applies, even though it may perhaps be outweighed by countervailing influences. Just to complicate things it can also suggest a correlation between quantities, rather than an all-or-nothing effect.)

metamorph: much of what you say sounds close to: ‘stop tricking me with arguments, and accept my common sense position’. Which is notably deficient as an argument. But at least it saves the bother of trying to work out whether you are talking about distribution of benefits or about voluntariness, and how you decide which facts about the world count as ‘transactions’, who is a participant, etc.

21

Alex 09.10.10 at 8:25 am

In fact, isn’t the argument stronger than this? Marxism is, if nothing else, a philosophy based on the inevitability of conflict between economic interest groups and the vital, determining role of this conflict in shaping history. The economic base – you could almost say, the microfoundations – determines the social superstructure. If the classes aren’t struggling to maximise their selfish interests, then what the hell are they doing?

@zosima: this is a major issue in Red Plenty

@everyone – can I say how unimpressive I find the argument from DARPA grant application? Anyone who refused to use linear programming because it was invented in the Soviet Union would be rightly condemned as an obscurantist and a McCarthyite.

22

Z 09.10.10 at 9:04 am

In reference to “Your money or your life”-style trade

But that’s not what “trade” means, unless you’re Jack Benny.

How about “Your money or your life” (not trade) followed immediately by “OK, now let us start trading”? Is this trade? If not, where do you draw the line? Because the standard argument from left-libertarian, which I am sure you are familiar with as you are obviously well-versed in libertarian thought, against “economic freedom” as it is commonly understood in nowadays parlance is that most of “trade” going on in the world (or that would go on in the absence of a broad social system) is that it is roughly on the kind described above.

23

Harald Korneliussen 09.10.10 at 10:08 am

Metamorph: In Nicolaus Tideman’s taxonomy of collective decisions, he distinguishes trade from coercion by this point: At least one of the sides feel that unless they accept the deal, another side will act in an improper manner (way he has no right to).

With the gun to the head example, the godfather may indeed admit that he’s not conducting an equitable trade. But there are plenty of cases where it may not be so, where the offending side does either not realize, or accept, that they are behaving improperly.

There are also the interesting cases where all sides feel that the other side will act improperly in the absence of a deal: Tideman cites negotiaited peace treaties as an example. Those are also examples of cases where the improper behavior is probably not recognized as such by the offending side (“We are entitled to defend ourselves”, “We are entitled to fight this occupation by any means”).

24

Chris Bertram 09.10.10 at 10:22 am

Ok, some supplementary thoughts.

First, I’m not at all opposed to the use of formal methods in the social science and I’m quite happy to see people examining problems using these methodologies, just so long as they don’t overgeneralise about the real world on the basis of highly simplified models with admittedly false premises. I’m sure this is common ground. (I think, and have always thought, that much of the RCT Marxism stuff was terrific – especially Roemer’s work on class, which was the topic of my first published articel.)

Second, I’d have to concede to Henry that ideal social science, marrying economics, sociology, empirical psychology (the Gintis synthesis in aspiration, maybe not in the details) is a good way off and that pending working out how to do it, people had better proceed with the tools they’ve got. Fair point, so long as we don’t get too _mañana_ about it.

Third, though, I have some real concerns about the notion of a robustness test for leftish institutional projects, policy proposals etc, a test that starts from worst-case assumptions about human motivation. I could express this in terms of Rousseavian ideas about the interaction between institutions and moral character or by reference to recent work by Bruno Frei. The point would be essentially the same. Designing institutions on the basis that people are selfish may not simply be a prudent strategy, it may be a self-defeating one: because the institutions reinforce the motivations. We thereby deceive ourselves about the limits of the possible.

25

Henri Vieuxtemps 09.10.10 at 10:55 am

Designing institutions on the basis that people are selfish may not simply be a prudent strategy, it may be a self-defeating one: because the institutions reinforce the motivations.

I remember LP recently posted a comment that seems relevant…

26

magistra 09.10.10 at 11:05 am

Designing institutions on the basis that people are selfish may not simply be a prudent strategy, it may be a self-defeating one: because the institutions reinforce the motivations.

Yes, but designing institutions on the basis that people are (always) altruistic gets you David Cameron’s Big Society and the view that you don’t need a welfare state, just charities. And tower blocks with lifts full of urine, for that matter. I think it’s more useful to start from the basis that some people are selfish most of the time, and most people are selfish some of the time, and then try and work out how to design in factors that encourage the more positive aspects of humanity.

27

DFC 09.10.10 at 11:07 am

There are a lot of fields that confirm Rational Choice Theory does not works propelly, or saying in other words, there are a lot of interest in powerful “institutions” (but also in normal people behavior) for RCT not working at all, here are some few remarkable examples:

a) The Marketing= ALL the works in marketing departments are designed to avoid people make “rational choice”, so that is the reason for trademarks. The TV spot and all the thing that pursue the “fidelity” of the consumers are paid by the deviation from the “rational choice” based on price and quality of the products. From the point of view of the RCT: do you really thinks that has Tiger Woods something related to the choice by the people of Nike sport shoes and paying more than for others with similar characteristics?Are you sure Nike is wasting money in making such kind of expensives TV spots due to the “rationality choice” of the consumers?

b) The politics= I know much more the details of politics in my country (Spain) than in USA. In Spain a very important part of the messages from the political parties are based, for example in the “ghost” or spanish Civil War and all the manichaeism associated to that time (mainly by the left), also the right-wing use the archetype of the “old left” prone to comunism and disorder, and a machine of waste money, and so on…In both case parties are focus on try to build “emotional” links in the people, that are more difficult to break than “rational” ones, allegedly based on the management and results, and then, political parties try also to by-pass RCT to maintain fidelity which is the real base of its power

We can say that the philosphy behind those process is that behavior (selfish driven) is not rational, firstly is based on emotional terms. Everybody built “the rational” on the grounds of selfish (driven by emotions) as a justification of their choices.
Shakespearian Sir John Falstaff says that: ” everybdoy has reasons, and very powerful reasons, the reasons are as common as raspberries”

We need a theory of why people act, or, for example what really money buys? Or in general terms= Hey you, what are you “selling”?!

28

Chris Bertram 09.10.10 at 11:39 am

_Yes, but designing institutions on the basis that people are (always) altruistic gets you _

How fortunate that nobody advocated doing that then.

29

Tom Hurka 09.10.10 at 12:14 pm

Seconding dsquared at #1:

That was one of Butler’s main objections to psychological egoism, that it denies the existence of “disinterested malevolence.” He too thought the world might be better, not worse, if egoism were true of us.

30

Alex 09.10.10 at 12:51 pm

Re D^2’s point: Malevolent motivations are interesting, because they seem to have a special ability to survive when they make no sense at all from a cost/benefit viewpoint. Malevolence and perversity go together.

Arguably it’s a sort of negative altruism – I’m willing to sacrifice my own interests, so long as it hurts you.

31

DFC 09.10.10 at 12:58 pm

Probably people that defend Rational Choice Theory, thinks that Nike sells “sport shoes” and that Mercedes sells “cars”, and also that people that buy a mercedes is really buying a “car”, that is dead wrong
Nike or Mercedes are, at the end “dream factories”, because I can say that the human beings are those animals who expend 33% of his time sleeping and 67% dreaming (well I am exaggerating the argument)

IMO the children build a “mythology” about themselves, driven by the values around him, but not the “written” or “standard” social values, but the real ones that they finely feel in the others around him, they feel what really “matters” on the others’ examples, not the empty words about “ethics” or “moral” they received, but on the grounds of the peoples’ behavior and desires (those people really important to them)
This “mythology” is the groove that inspire the future men behavior, and are the “synthetic a priori judgments” (using kantian terms) that drives the future justifications to fill a “rational” frame of “reasons”

32

Metamorf 09.10.10 at 1:04 pm

metamorph: much of what you say sounds close to: ‘stop tricking me with arguments, and accept my common sense position’. Which is notably deficient as an argument.

“Close to”, but no cigar — what I’m saying is ‘stop tricking yourselves with specious arguments, and use some common sense as at least a starting point’. Which isn’t an argument either, but just (unsolicited) advice. Common sense isn’t, by any means, all you need, but it can certainly help as a check when you’ve let your political desires lead you too far from the world around you. It can give you a start, for example, in understanding terms like “voluntariness” or “transactions”, and it might even enable you to ask, when people speak about things like “the distribution of benefits” — distribution by whom?

33

tomslee 09.10.10 at 1:20 pm

Lemuel#19 finally spells out an alternative to rational choice.
[Those on the left] believe that the “way humans are” is fundamentally and irreducibly determined by the particular institutions that those humans are organized into. So we reject the rational-maximizing-individual microfoundation, but not to replace with it some other microfoundation, but rather to reject microfoundations as such, in favor of a historically grounded account of concrete institutions.

You divide the world into two types of people – those who can’t get past their “determination” by the institutions they are organized into, and those enlightened few who can see the bigger picture, and who can discuss hegemony, ideology, false consciousness, and point out why capitalism sucks. In other words, it replaces rational individuals with slightly stupid individuals.

So there seems to be a choice, either say capitalism sucks and the masses are dupes, or that capitalism is fine and/because people are rational.

It misses the third (and obviously even more enlightened) group, to which Henry and I belong. These are people who accept a role for selfish rationalism but also see that bad equilibria are everywhere, meaning that we can say that capitalism sucks without calling other people stupid.

34

Metamorf 09.10.10 at 1:22 pm

How about “Your money or your life” (not trade) followed immediately by “OK, now let us start trading”? Is this trade?

I’m not getting the difference here — why do you say the first is not trade? What is the second saying that’s different from the first?

Let’s accept that at some point transactional options aren’t real options, and at that point the transaction is not a trade. Where that point comes is certainly debatable. If you’re saying that it comes whenever the payoffs in a transaction are not 50-50 — as measured how? by whom? — then of course you’re saying that there’s no such thing as a trade, and in that case I’m saying that you’ve let your political belief system take you as far from reality as any other religious fundamentalism.

35

Chris Bertram 09.10.10 at 1:27 pm

Maybe Lemuel and Tom Slee should glance at the third of Marx’s theses on Feuerbach.

36

chris 09.10.10 at 2:30 pm

Designing institutions on the basis that people are selfish may not simply be a prudent strategy, it may be a self-defeating one: because the institutions reinforce the motivations. We thereby deceive ourselves about the limits of the possible.

ISTM that deception about the limits of the possible primarily runs the other way: well-intentioned people try to set up societies that assume that everyone can be encouraged to support the common good, and then they get taken over by people like Stalin.

If even a small number of people are determined to be power-hungry selfish manipulators regardless of how hard society tries to encourage them to be otherwise (and some psychological research does seem to show this, although of course it’s conducted on people already socialized by the societies they grew up in), then any society *must* be capable of defending itself against coups and power grabs by such people or it will fall to them.

the “way humans are” is fundamentally and irreducibly determined by the particular institutions that those humans are organized into.

I would say “influenced” rather than “determined”, because ISTM that “determined” gives the (IMO dangerously misleading) impression that other influences can always be overcome by the right institutions, and I don’t see any reason to believe that’s true of ultraselfish individuals.

Even if it is, though, you stage a revolution with the population you have, not the population you wish you had; if there are ultraselfish people already in your nation when you start (and there are), then you need to keep them out of power at least long enough for the new generation of utopian cooperators to be socialized, *even if* that project works.

this may seem pedantic but I don’t think it is: ‘ceteris paribus’ only makes sense when you are comparing two situations

I thought that equal distribution was being tacitly compared to the status quo unequal distribution. (The phrase “utility maximizing” requires a comparison, I think.) “Ceteris paribus” was intended to stand for “assuming the change in the system of distributing wealth doesn’t change the total amount of wealth produced”, and I then went on to explain how this assumption is sometimes challenged.

Of course it does mean “subject to certain assumptions” but since the specific assumption is that changing X (distribution of wealth) will not change Y (total amount of wealth produced by the society), I thought “ceteris paribus” was appropriate. I could be wrong though.

37

Antonio Conselheiro 09.10.10 at 2:38 pm

If, in contrast, lefty arguments still work when the dice are loaded against them, say, by assuming the kinds of stark clashes of interest that are often modelled in game theory, it suggests that these arguments are likely to be robust under a variety of assumptions and conditions.

Sort of like superman in a world without kryptonite, or a drug that cures every disease, or a perpetual motion machine.

38

Zamfir 09.10.10 at 2:50 pm

Chris, a society and institutions that encourage and rely on cooperative behaviour is hardly something for after the Revolution. Every existing society would crumble if no one tried to do “the right thing” and everyone stuck to game-theoretic games.

Good systems make sure that they can’t be ripped too much off by those who try, but beyond that it’s usually a genuine cooperative ethos that achieves the positive things.

39

Lemuel Pitkin 09.10.10 at 2:50 pm

Maybe Lemuel and Tom Slee should glance at the third of Marx’s theses on Feuerbach.

I’m pretty sure Tom S. was disagreeing with me, along the same lines as you and chris.

40

Metamorf 09.10.10 at 2:51 pm

Harald: In Nicolaus Tideman’s taxonomy of collective decisions, he distinguishes trade from coercion by this point: At least one of the sides feel that unless they accept the deal, another side will act in an improper manner (way he has no right to).

This doesn’t seem relevant to trade as such, and only partially relevant to other kinds of interactions (e.g., maybe peace treaties, etc.). In general, the other side or all sides have a right to act in any way they have a right to (with apologies for the tautology, but sometimes you have to state the obvious), and all sides should understand that going in. In a trade as such both sides have to have a reasonable possibility of refusing the deal, even if it’s at some real cost to themselves, but that’s all that they have to have, and it’s quite independent of any “improper” (but not illegal) actions of the other side(s).

41

Henry 09.10.10 at 3:03 pm

Chris Bertram – In turn, let me concede that even if the Grand Unitary Theory remains aspirational, it is something worth aspiring to – so I certainly don’t think that people should remain in their comfort zones, plugging away at theories that they know to be partial, without trying to figure out how those theories could be improved (or replaced wholesale) by contact or commingling with others. On your criticism – I actually wonder if the more worrying criticism is the inverse of this. You can model an _awful lot_ of forms of social interaction using game theory, including some that are quite benign sounding. Gary Miller has a great book on this – _Managerial Dilemmas_ – which concludes that the proper game theoretic account of how organizations should be run should not focus on brute mechanical Taylorist type schemes of efficiency, harsh punishment and reward, but instead, on Barnard type efforts to create a culture of semi-diffuse reciprocity, where workers know that their additional efforts will be rewarded, even if they hard to observe and incentivize directly. As Miller notes, this can all be supported by Folk Theorem type results – as long as people’s cooperation is _conditional_, and everyone knows this, you can do quite a lot. But the problem is that the Folk Theory doesn’t tell you that this _will_ happen, or even _when_ this will happen, but that it _can_ happen. And that’s even before we get into the “Davies Corollary”:http://crookedtimber.org/2006/11/29/reputations-are-made-of/. So I worry less that schemes based on self-interest will discourage disinterested behavior (because schemes based on self-interest can in fact support quite diffuse forms of cooperation, which does not bleed into altruism, but which does resemble it, and which does not drive it out in the same way that mechanistic reward and punishment schemes would). I worry more that the account may not be quite stringent _enough._ I think that there is a defense though – which is that Folk Theorem results only apply under certain parameter values – and these parameter values tell you something at least about the possibility and impossibility of certain kinds of social arrangement (i.e. they would strongly suggest that certain kinds of diffuse cooperation may be possible among small communities, but not so great at the level of the region or nation). This would be taking some of the work of Randy Calvert and applying it to normative questions.

Lemuel – I think that just such a theory of humanity as naturally sociable lurks behind many important leftwing theorists who seem to be arguing for the determining role of social institutions – the Marx of the Economic and Philosophical Manuscripts and Rousseau for starters. But more generally – I have great difficulty in buying into your

bq. Rather, we believe that the “way humans are” is fundamentally and irreducibly determined by the particular institutions that those humans are organized into. So we reject the rational-maximizing-individual microfoundation, but not to replace with it some other microfoundation, but rather to reject microfoundations as such, in favor of a historically grounded account of concrete institutions.

I’d contend that ‘human nature’ is not fundamentally and irreducibly determined by institutions. This is something that e.g. James Scott (who is certainly a man of the left) is very good on. Various efforts to shoehorn human beings into new institutional arrangements, on the basis that their natures will be changed by them etc have been a dismal failure, starting with New Soviet Man and going on from there. Pinker’s _The Blank Slate_ is a book that I remember disliking intensely when I read it – but on this point he is right. There are certain limits that are hardwired into human beings which limit profoundly the degree to which different institutional arrangements can determine our being. We can disagree on how best to theorize these limits, but (and here, I imagine Chris B and I are in complete agreement), these limits exist, and to figure out what they are, and what kinds of consequences we might expect when we try to introduce new institutional arrangements, we need some kind of microfoundations, whatever those microfoundations might be.

42

Henry 09.10.10 at 3:05 pm

Metamorf – let me spell it out again. Game theory _is entirely indifferent_ as to whether the exchange that takes place involves a gun to your head, or some other form of leverage that you might consider more legitimate. That is the fundamental point.

43

Metamorf 09.10.10 at 3:29 pm

Henry: Game theory is entirely indifferent as to whether the exchange that takes place involves a gun to your head, or some other form of leverage that you might consider more legitimate. That is the fundamental point.

Well, first, it’s not just me that considers the gun to the head illegitimate, it’s pretty much everyone except thieves, and probably includes most of them as well. It’s true, of course, that game theory takes place on an abstract enough plane that questions of legitimacy don’t arise at all. But that just further delimits the practical application of game theory — in particular, it at the very least hobbles the attempt to enlist game theory in one’s ideological battles.

44

djw 09.10.10 at 3:47 pm

I think that just such a theory of humanity as naturally sociable lurks behind many important leftwing theorists who seem to be arguing for the determining role of social institutions – the Marx of the Economic and Philosophical Manuscripts and Rousseau for starters

I’m not persuaded of this in general, although I’m open to being persuaded otherwise on a case by case basis. But this seems like an odd reading of 1844 manuscripts to me. While there’s some flowery language in there that we might take to indicate such a view of human nature, the actual argument about human nature made therein seems to me to be that human nature is to work, that is, to creatively and consciously manipulate nature, internal and external (and, obviously, capitalism distorts this in grotesque ways). How we treat one and other is secondary to our core ‘human nature’ and is better understood as determined, or at least very strongly guided, by structural factors. I don’t quite think your reading of Rousseau is correct either in this context–it seems to me Rousseau develops two extremely malleable features of universal human nature–self-preservation and pity–and proceeds to argue they are both almost infinitely malleable.

This is something that e.g. James Scott (who is certainly a man of the left) is very good on. Various efforts to shoehorn human beings into new institutional arrangements, on the basis that their natures will be changed by them etc have been a dismal failure, starting with New Soviet Man and going on from there.

Another way of reading Scott’s narratives about state failure would be to treat them as an investigation of people who find themselves at the violent and disruptive intersection between multiple structural arrangements that would shape their lives. The frission between these structural forces creates the possibility for crafting a moderate amount of agency in negotiating the tensions.

45

Henri Vieuxtemps 09.10.10 at 4:07 pm

the “way humans are” is fundamentally and irreducibly determined by the particular institutions that those humans are organized into

Perhaps the word “institutions” is misleading; it gives the impression that humans can be fundamentally changed by some bureaucrat changing some guidelines somewhere. “The institutions” is the totality of the social environment over your whole lifetime, up to this moment.

You divide the world into two types of people – those who can’t get past their “determination” by the institutions they are organized into, and those enlightened few who can see the bigger picture, and who can discuss hegemony, ideology, false consciousness, and point out why capitalism sucks. In other words, it replaces rational individuals with slightly stupid individuals.

One can start discussing hegemony, ideology, etc. as soon as he gets introduced to these concepts. If you don’t know what “hegemony” is, you can’t discuss it, of course, but it doesn’t mean you’re stupid.

46

Chris Bertram 09.10.10 at 4:31 pm

Well it is very difficult to know how to respond to many of these claims about the malleability of human nature and institutions etc at the level of generalization at which they’re being made in this thread. But the point I was making in the final para of my #24 above doesn’t actually require any such grandiose thoughts (perhaps my reference to Rousseau was unhelpful in this respect). Rather, it just needs the thought that humans have (a) their own individual goals and (b) some dispositions to co-operate, to conform to norms of share practices, to enforce conformity on others etc. (a) and (b) can both be permanently present and, in this respect, are enduring features of our nature (I’m not asserting this, just saying for the sake of the argument). Given this picture, institutions designed on the assumption that we are all knaves without co-operative dispositions may have the effect of promoting the knavish behaviour and suppressing the co-operative tendencies. Different institutional designs – recognizing and exploiting our co-operative and normative capacities – may do better. But you won’t get to see that if, proceeding on the assumption of 100% knavishness, you build institutions which leave those other parts of our nature invisible.

Issues of scale and common knowledge are clearly important here, since people won’t be taken for mugs, and systems that rely on trust and reciprocity will break down with too many free-riders. But assuming that everyone will free ride if they can get away with it may have the effect of increasing the proportion who will free ride _when_ they can get away with it.

(Anecdote: someone I knew ran a small art gallery in South Wales. He left it unlocked and unstaffed most of the time, with an honesty box for payment. In ten years he only had one case where he wasn’t paid for a picture, and that was a cheque that bounced rather than a painting simply removed. Obviously I don’t suggest that as a model for political and social institutions, but it does show, I think, that there are can be remarkable levels of spontaneous trust and co-operation among strangers.)

47

Will Roberts 09.10.10 at 4:50 pm

Chris B said (at 24): Designing institutions on the basis that people are selfish may not simply be a prudent strategy, it may be a self-defeating one: because the institutions reinforce the motivations. We thereby deceive ourselves about the limits of the possible.

This is a thought worth thinking, but I’m not sure it speaks against Henry’s argument, even if Henry, too, talks in terms of selfishness. RTC and game theory do presuppose that people are selfish, in some sense, but they do not draw any very precise lines around this “self” — what matters is that I prefer my preferences, not that those preferences are narrowly selfish. What fuels the collective action problems that get so much attention in game theory is not this self-regard, but the uncertainty of each actor about the preferences of the other actor(s), and the dependence of each actor’s preferences upon the action of the other(s). Designing institutions on the basis of people’s mutual dependence and mutual uncertainty does not raise the same concerns you have about building in selfishness. And being concerned with our mutual dependence and uncertainty is enough to motivate Henry’s (limited) appreciation of RTC and game theory.

48

tomslee 09.10.10 at 4:54 pm

Perhaps with HF@40 and CB@45 the middle ground is sighted?

Aside: We have a local bakery that runs on the honesty-box method too. It self consciously promotes a “community” feel (ie, reciprocity), and also self-consciously promotes the virtues of small scale. The classic large-scale institution built on such virtues is the blood supply, of course, which is insulated to some extent against free riding because it only requires relatively small participation rates, and because few non-vampires just take blood for the sake of it.

49

Lemuel Pitkin 09.10.10 at 4:59 pm

djw at 43, Henri V. at 44 and Chris B. at 45 do a better job developing the argument I was trying to make at 19, than I could.

50

tomslee 09.10.10 at 5:02 pm

Henri @44.

“The institutions” is the totality of the social environment over your whole lifetime, up to this moment.

In further seeking middle ground, I don’t think this really contradicts RCT, but is complementary to it. RCT is about how one pursues tastes, not what those tastes are (Will Roberts @46 above, and others). Whether the individual is shaped by society or society by individuals is surely a case of chicken and egg: game theory does not say (or need to say, at least) “there is no such thing as society”.

If you don’t know what “hegemony” is, you can’t discuss it, of course, but it doesn’t mean you’re stupid.

I am happy to replace “stupid” with “ignorant” – I think the point stays.

51

geo 09.10.10 at 5:08 pm

Chris @45: there are can be remarkable levels of spontaneous trust and co-operation among strangers

Yes, of course. And since high levels of trust and cooperation among strangers is a necessary and practically sufficient condition of a decent society, shouldn’t the (or a) main focus of leftists be to identify what fosters and what hampers trust and cooperation? Beyond the obvious, I mean: government and corporate secrecy, gross inequalities of income or status, pervasive economic insecurity, concentrated private ownership of media and publishing, corporate influence in universities, and the ubiquity of “strategic” (ie, manipulative) communications, eg, advertising, marketing, and corporate public relations.

Once we remove these obvious obstacles to solidarity and trust, I’m sure there will still be a great deal for social theory and social research to do.

52

chris 09.10.10 at 5:40 pm

Given this picture, institutions designed on the assumption that we are all knaves without co-operative dispositions may have the effect of promoting the knavish behaviour and suppressing the co-operative tendencies.

I don’t see why. An institution designed on the assumption that we are all knaves would be one that gives knaves a self-interested motive in cooperating. (Assuming it’s not designed by an idiot and the purpose is to produce non-socially-destructive *behavior* despite the lack of a non-self-interested cooperative *motive*.) If someone is disposed to cooperate anyway, how would that stop them?

For example, fines for littering are designed to reshape the behavior of people who would litter otherwise (because they knavishly don’t want to bother looking for a proper trash can, even though that makes the park a mess for others). If I’m a person who would not litter anyway because I don’t want to screw up the park for everyone else, how would the existence of a fine break down my cooperative tendency?

I suppose you could simply argue that we shouldn’t assume 100% knavishness because it’s factually wrong, and we should instead assume that individuals will be drawn apparently at random from different points on the knave-cooperator spectrum and the knavishness of any given person may vary over time, and we need an institution that will nonetheless get them all to work together to the extent possible. But I’m finding it hard to come up with an example where that institution is actually different than the one you’d design for pure knaves — because you always have to have a way to keep the pure knaves that *do* exist from infiltrating and exploiting your system, which is complicated by the fact that they can and will appear to cooperate for an arbitrarily long time if they think they’ll get something sufficiently valuable out of it. And positions of trust can be very valuable indeed.

53

Henry 09.10.10 at 5:50 pm

bq. . And since high levels of trust and cooperation among strangers is a necessary and practically sufficient condition of a decent society, shouldn’t the (or a) main focus of leftists be to identify what fosters and what hampers trust and cooperation?

If anyone is interested in one effort to do this (which builds upon broadly rational choice foundations but extends them a bit), I’ve uploaded a “paper”:http://crookedtimber.org/wp-content/uploads/2010/09/trustcompliance.pdf by Jack Knight and I that we really need to get around to trying to resubmit somewhere one of these days soon. Perhaps this will also help focus the debate a bit since (as Chris Bertram rightly says), it’s difficult to really work through claims at this very broad level of abstraction.

54

Chris Bertram 09.10.10 at 5:53 pm

Chris @52 “I don’t see why”

– well let me just refer you to some of the motivation crowding lit. Eg.

Bruno S. Frey, “A Constitution for Knaves Crowds out Civic Virtues” _The Economic Journal_ , Vol. 107, No. 443 (Jul., 1997), pp. 1043-1053.

55

geo 09.10.10 at 5:59 pm

this very broad level of abstraction.

Is it really an abstraction to say that trust and cooperation are necessary for a decent society? And doesn’t it help focus the debate a bit to suggest getting “government and corporate secrecy, gross inequalities of income or status, pervasive economic insecurity, concentrated private ownership of media and publishing, corporate influence in universities, and the ubiquity of “strategic” (ie, manipulative) communications, eg, advertising, marketing, and corporate public relations” out of the way first, before tackling theoretical questions?

56

marwan 09.10.10 at 5:59 pm

Professor Bowles’ lectures at yale address a lot this under the term “Machiavelli’s Mistake”

http://www.santafe.edu/~bowles/castle.html

57

Henry 09.10.10 at 6:05 pm

George – the ‘broad level of abstraction’ bit was in response to Chris’s suggestion (if I got his drift) that it is hard to evaluate the relative merits of different theoretical approaches without thinking more concretely about their specific applications. Which e.g. studies of gross inequalities etc could get at (that is one of the motivations for the paper that Jack and I wrote).

58

piglet 09.10.10 at 7:02 pm

chris:

For example, fines for littering are designed to reshape the behavior of people who would litter otherwise (because they knavishly don’t want to bother looking for a proper trash can, even though that makes the park a mess for others). If I’m a person who would not litter anyway because I don’t want to screw up the park for everyone else, how would the existence of a fine break down my cooperative tendency?

True but in many cases, everybody is affected by restrictions that are placed specifically to discourage knaves. Examples: parks that close at night; rules against “loitering”; rules against alcohol consumption in public, or even against bringing food to a swimming pool. These types of rules are fairly common and they are not intended to discourage specific, asocial behaviors. Rather they are based on the assumption that people can’t be trusted and that their freedom needs to be restricted preemptively because they would only abuse that freedom.

59

chris 09.10.10 at 7:35 pm

These types of rules are fairly common and they are not intended to discourage specific, asocial behaviors.

Aren’t they? To take one example, parks closing at night, I thought that was instituted to reduce the number of people who would be assaulted, robbed, etc. at night. It doesn’t take very many knaves in the park to make it in the best interests of even conscientious people that they not go in at night, but they won’t necessarily realize that. The rule is a bit paternalistic, maybe, but you can’t really expect a state to say “If you go into this park at night, you might get robbed, but feel free to take your chances anyway if you like” — it would be an abdication of the state’s responsibility to suppress violence. The state closes the park because it would be too expensive to patrol it adequately to keep it safe at night — that’s a political judgment based on the unwillingness of voters to pay (through higher taxes) the salaries of the extra policemen you would need to do the patrolling.

Anyway, it seems like an example of my point. How would you rewrite these rules to account for a society in which only some people are knaves? Close the park only to knaves? That would work great, if only they were wearing their special knave hat that identifies them to everyone, but being knaves, they take the darn thing off!

Of course the idea of knave hats is deliberately silly, but it makes the point: If you can’t identify the trustworthy people, then effectively you *can’t* trust anyone. That’s exactly why our species has such elaborate reputation-establishing and tracking methods to identify which people it is relatively safe to trust when you need to trust someone.

P.S. If the purpose of the rule against bringing food to a pool is to prevent accidents in which the food is accidentally spilled into the pool, damaging everyone’s ability to use it, then a conscientious person wouldn’t bring any food whether there was a rule or not, since that would be taking a risk with the public good for personal convenience. So they’re not *really* affected by the restriction any more than the littering example.

60

novakant 09.10.10 at 8:15 pm

Issues of scale and common knowledge are clearly important here, since people won’t be taken for mugs, and systems that rely on trust and reciprocity will break down with too many free-riders. But assuming that everyone will free ride if they can get away with it may have the effect of increasing the proportion who will free ride when they can get away with it.

A large proportion of people do free ride when they can get away with it: before the advent of the internet very few would have thought about stealing a record, software or a book from a shop, much less actually do it – now everybody and their dog seems to do it and the people who still actually pay for such things tend to be bombarded with self-serving arguments or ridiculed. I don’t like elaborate copy protection and such, but the blame rests squarely on the large number of people who make it necessary.

61

piglet 09.10.10 at 8:16 pm

“Aren’t they? To take one example, parks closing at night, I thought that was instituted to reduce the number of people who would be assaulted, robbed, etc. at night.”

Ok to be more precise: the restrictions in question do not cover just the behaviors that are intended to be discouraged. They also cover a whole range of activities that are perfectly innocuous and would normally be legal. That is different from your example, in which a specific behavior – littering – is made illegal and punished with a fine. The park that is closed at night isn’t just closed to knaves, it is closed to anybody. Could it be that policies like that “may have the effect of promoting the knavish behaviour and suppressing the co-operative tendencies”? That requires a more complex discussion than I have the time to offer here but I think it is a possibility. As a potential park user who finds the park closed as a security measure, or who is denied a civilized glass of wine at a picnic in the park with friends, I could imagine that promoting a somewhat cynical outlook on the maintenance of public infrastructure.

62

piglet 09.10.10 at 8:27 pm

“If the purpose of the rule against bringing food to a pool is to prevent accidents in which the food is accidentally spilled into the pool, damaging everyone’s ability to use it, then a conscientious person wouldn’t bring any food whether there was a rule or not, since that would be taking a risk with the public good for personal convenience. So they’re not really affected by the restriction any more than the littering example.”

I would argue that I am perfectly capable of eating a sandwich at a pool without causing any accidents, and of course I would properly dispose of any waste, and many pools that I have visited have no problem with that. The point is whether you as the pool operator rely on trust and cooperation, assuming that your patrons will use a public resource responsibly, or whether you find it necessary to restrict any kind of behavior that might potentially cause trouble. And perhaps there are good reasons for adopting the latter approach but it certainly is distinct from the cooperative approach.

63

Alex 09.10.10 at 8:35 pm

Piglet@61: Urban spaces like your park are interesting examples. If you put up a sign saying ENTER AT YOUR OWN RISK and leave the place to the knaves during the hours of darkness, this may be fine in theory…until the knaves’ knavish activities start to affect the streets around it. Perhaps people go there to buy drugs off the knaves, or knaves bring stolen goods there to sell, and being knaves they tend to settle disputes in their own way, including in the pub that faces onto the park. Knaves need their beer too.

On the other hand, it might be better to have non-knaves about all the time. “Self-policing” is a fairly well known phenomenon, and an example of natural cooperation and norm-enforcement. Obviously there’s a balance between the two, and a complex relation between self-policing and police policing. If you expect to be surrounded by knaves, or at least that you can’t count on any support, would you be more likely to be more knavish? And people confronted with a massive police state often resort to being more knavish than the average in order to get on with life.

64

chris 09.10.10 at 8:48 pm

I would argue that I am perfectly capable of eating a sandwich at a pool without causing any accidents

Well, of course you would. And you’d be right — most of the time. And the rest of the time there’d be a really good reason why it wasn’t really your fault, after all, nobody could have expected that other person to bump into you like that just because someone’s kid made a loud noise.

It’s still a risk. That’s how risk works.

The park that is closed at night isn’t just closed to knaves, it is closed to anybody.

I thought I addressed that already: it’s impossible for it to be closed just to knaves, because knaves knavishly disguise themselves as regular people (especially when they can obtain a benefit, such as access to a park). So in this case the policy that results from assuming everyone is a knave and the policy that results from assuming some people are knaves is the same policy: you have to keep out everyone to be sure of keeping out the knaves.

If you really assumed *absolutely everyone* was a knave, though, you wouldn’t have a park at all: everyone would litter, there’d be no way to hire enough policemen to stop it, especially when they would all take bribes to look the other way. At best, you might hope to put out enough garbage cans that bribing a policeman to overlook your littering would become more inconvenient than using a garbage can. But then people would steal the garbage cans and bribe the police to overlook that. And how do you pay any policemen when your entire population is made up of tax evaders, and you yourself would rather embezzle the payroll than pay it out?

There are societies that kind of do look like that and despite the prominence of freshwater economics, the US isn’t actually one of them.

65

piglet 09.10.10 at 11:05 pm

“Urban spaces like your park are interesting examples. If you put up a sign saying ENTER AT YOUR OWN RISK and leave the place to the knaves during the hours of darkness, this may be fine in theory…until the knaves’ knavish activities start to affect the streets around it.”

Maybe you are right but as an anecdote, it happens that where I grew up, parks were open at night, you could have a drink in public as long as you don’t cause public disorder, etc. so I can attest that you don’t necessarily need heavy-handed restrictions to maintain orderliness. You, and I think chris, are arguing that asocial behavior will always win against cooperative behavior and that is precisely why you argue in favor of “institutions designed on the assumption that we are all knaves” as CB put it. I just wanted to clarify this distinction.

66

piglet 09.10.10 at 11:10 pm

“So in this case the policy that results from assuming everyone is a knave and the policy that results from assuming some people are knaves is the same policy: you have to keep out everyone to be sure of keeping out the knaves.”

Again, you argue in favor of “institutions designed on the assumption that we are all knaves”. That is all I’m saying.

67

Walt 09.10.10 at 11:21 pm

Henry, I didn’t understand your point about partial equilibrium way up above. Partial equilibrium is a special case of general equilibrium where at the end of the day when there’s nothing you want to buy, you say “fuck it” and eat the money. Thus any defect possessed by partial equilibrium models will automatically be possessed by general equilibrium models.

68

Guido Nius 09.11.10 at 11:38 am

Well, whatever else may be true, one will not be able to make any sense of economic behaviour (or any other behaviour) if one doesn’t start with a high dose of the Principle of Charity. People may be selfish but their selfishness can only be understood on the substrate of them ‘allowing’ a whole lot of stuff to be as others would have them.

This is especially true from an analytic point of view.

69

zosima 09.15.10 at 11:03 pm

@metamorf
“My claim is only a definitional one—that “trade” means a voluntary economic transaction.”

You make the claim that trade is voluntary, you also make the claim that “a trade as such must be to the perceived benefit of both or all parties.” My point is that a lot of little benefits summed together doesn’t necessarily lead to the best possible outcome or even a good one, even in absence of coercion. It has more to do with the ways in which a system can transition from one state to another.

“No, it makes no claim about “optimal” outcomes at all—this is because the notion of “optimal” assumes a god-like position that only modelers or game-players can adopt vis-a-vis their miniature, impoverished creations”

If “beneficial” means better than some alternative, “best” or “optimal” means better than all alternatives. Once you’ve defined a metric of “good” or a mechanism to order alternatives, then it necessarily follows that there is an alternative that is as good or better than all other alternatives.

Perhaps only modelers can reach the optimal, or even identify the optimal, but we can know that it exists even without our “miniature, impoverished creations.”

70

zosima 09.15.10 at 11:21 pm

@Chris
You say:
“the traditional counterargument is that without the incentive provided by the possibility of obtaining more than an even share of wealth, people wouldn’t work as much or as effectively, and therefore the total wealth of society would be lower. It doesn’t seem obviously stupid to me (although it may be flawed in more subtle ways).”

That is the traditional counter-argument but I’m not clear that it necessarily follows from the micro-foundations. This is actually a specific irrational belief that conservatives will often assume to co-opt the micro-framework. It implicitly assumes that our incentives are a social function( by comparison to the people around us rather) than a function of income or wealth.

If I live in a world where I can be little better off at T+1 than T+0 because everyone is better off, then it is still rational to work, even if it isn’t fair. Much like it is still rational for me to accept any deal in the ultimatum game, even if I’m getting totally ripped off.

I’ll agree that there is a strong case to be made that people actually are irrational in this way, but along the lines of Henry’s OP, my point is more that it is good to be aware of which claims are entailed by the model and which ones are tacked on than a claim that the model is a good one.

Comments on this entry are closed.