Rational != Self-interested

I’ve said it before (along with Aaron Edlin and Noah Kaplan) and I’ll say it again. Rationality and self-interest are two dimensions of behavior. An action can be:
1. Rational and self-interested
2. Irrational and self-interested
3. Rational and altruistic
4. Irrational and altruistic.
It’s easy enough to come up with examples of all of these.

Before going on, let me just quickly deal with three issues that sometimes come up:
– Yes, these are really continuous scales, not binary.
– Sure, you can tautologically define all behavior as “rational” in that everything is done for some reason. But such an all-encompassing definition is not particularly interesting as it it drains all meaning from the term.
– Similarly, if you want you can tautologically define all behavior as self-interested, in the sense that if you do something nice for others that does not benefit yourself (for example, donate a kidney to some stranger), you must be doing it because you want to, so that’s self-interested. But, as I wrote a few years ago, the challenge in all such arguments is to avoid circularity. If selfishness means maximizing utility, and if we always maximize utility (by definition, otherwise it isn’t our utility, right?), then we’re always selfish. But then that’s like, if everything in the world is the color red, would we have a word for “red” at all? I’m using self-interested in the more usual sense of giving instrumental benefits.

To put it another way, if “selfish” means utility-maximization, which by definition is always being done (possibly to the extent of being second-order rational by rationally deciding not to spend the time to exactly optimize our utility function), then everything is selfish. Then let’s define a new term, “selfish2,” to represent behavior that benefits ourselves instrumentally without concern for the happiness of others. Then my point is that rationality is not the same as selfish2.

What’s new here?

The above is all background. It came to mind after I read this recent post by Rajiv Sethi regarding agent-based models. Sethi quotes Chris House who wrote:

The reason that economists set up their theories this way – by making assumptions about goals and then drawing conclusions about behavior – is that they are following in the central tradition of all of economics, namely that allocations and decisions and choices are guided by self-interest. This goes all the way back to Adam Smith and it’s the organizing philosophy of all economics. Decisions and actions in such an environment are all made with an eye towards achieving some goal or some objective. For consumers this is typically utility maximization – a purely subjective assessment of well-being. For firms, the objective is typically profit maximization. This is exactly where rationality enters into economics. Rationality means that the “agents” that inhabit an economic system make choices based on their own preferences.

No no no no no. Self-interest is the end, rationality is the means. You can pursue non-self-interested goals in rational or irrational ways, and you can pursue self-interested goals in rational or irrational ways.

Sethi’s post is about the relevance of agent-based models (as indicated in the above YouTube clip) to the study of economics and finance, and is worth reading on its own terms. But it also reminds me of the general point that we should not melange rationality with self-interest. I can see the appeal of such a confusion, as it seems to be associated with a seemingly hard-headed, objective view of the world. But really it’s an oversimplification that can lead to lots of confusion.

P.S. House’s blog is subtitled, “Economics, chess and anything else on my mind.” This got me interested so I entered “chess” into the search box but all that came out was this, which isn’t about chess at all. So that was a disappointment.

P.P.S. Some commenters asked for examples so I added some in comments. I’ll repeat them here.

First, the real-life example:

Some students in my class are designing and building a program to display inferences from Stan. They are focused on others’ preferences; they want to make a program that works for others, for the various populations of users out there. And they are trying to achieve this goal in a rational way.

Second, the quick examples:

Rational and self-interested: investing one’s personal money in index funds based on a judgment that this is the savvy way to long-term financial reward.

Rational and non-self-interested: donating thousands of dollars to a charity recommended by GiveWell.

Irrational and self-interested: day trading based on tips you find on sucker-oriented websites and gradually losing your assets in commission fees.

Irrational and non-self-interested: rushing into a burning building, risking your life to save your pet goldfish that was gonna die in 2 days anyway.

You could argue about the details of any of these examples but the point is that rationality is about the means and self-interest is about the ends.

69 thoughts on “Rational != Self-interested

    • For what it’s worth, the strict definition of rationality used in microeconomics and decision theory is that an individual’s behavior is consistent with a set of preferences which are:
      i) complete (meaning that the individual can compare any two options and make a choice – this does not exclude the possibility of being indifferent)
      ii) transitive (meaning that “A preferred to B” and “B preferred to C” implies “B preferred to C”)

      It’s a pretty minimal set of assumptions, but they are still violated in a variety of behavioral experiments.

      • Brad:

        Yes, my point is that this definition of rational could be applied to self-seeking or other-seeking goals. House’s mistake is to identify rationality with self-seeking goals.

        • I completely agree with your point – when people confuse rationality with selfishness it frustrates me as well! I just thought I’d answer Rahul’s question (or part of it), since I think the narrow definition of rationality can be useful and illuminating.

      • @Brad: “are still violated in a variety of behavioral experiments.”

        Which often is an indictment of the experiment not the assumption. A lot of these experiments make very strong assumptions for identification, stronger that the assumptions they are trying to test.

        I am waiting for the revisionist movement to catch up.

      • There’s something wrong here, it seems to me that “rationality” normally means something along the lines of “consistent with logic and a scientifically verifiable world view”.

        So for example, it is rational to drive to the beach if you want to hang out on the sand, it is irrational to drive to a parking lot in the middle of Los Angeles and demand that other people import a beach full of sand to your location. It is rational to take a medicines that have been reasonably well verified as effective against your illness, it is irrational to take homeopathic hocum which is just poisons diluted away by water until only the water is left….

        how does that kind of rationality enter into this discussion.

  1. Totally agree with this but I think it’s also important to note that with our without self-interest, rationality doesn’t imply equilibrium. Even common knowledge of rationality doesn’t imply equilibrium. The main strength of the agent-based approach is that it is explicit about disequilibrium dynamics, which may or may not converge to an equilibrium path. Leigh Tesfatsion’s comments on House’s two ABM posts are spot on.

    On chess, though, House has a wonderful post on Judit Polgar, complete with video of a famous win over Shirov:

    http://orderstatistic.wordpress.com/2014/09/03/end-of-an-era/

    • Rajiv:

      I still don’t understand. “Chess” is in the description of his blog but in the eight months of posting there’s only one chess post. With a description, “Economics, chess and anything else on my mind,” I’d like to see a bit more chess!

  2. I have a friend who is irritated when asked “what’s the payback time for your solar panels” or “what’s the payback time for your hybrid compared to buying a regular car.” He says nobody ever asks anyone “what’s the payback time on your hot tub” or “what’s the payback time for buying a BMW instead of a regular car.” He wants to be part of the solution rather than the problem, or perhaps simply a small part of the problem rather than a large part of the problem, and he’s willing to pay for it; he resents it when people imply that this is irrational of him, just because it’s not self-interested.

        • I believe the implicit reasoning here is that the desire to drive a hybrid car is really derivative of a desire to “help the environment” and/or “save money”.

          If hybrids don’t actually reduce total CO2 emissions, or are not cost-effective at such, or if CO2 isn’t really a problem for the environment, or […] then the desire to drive a hybrid car is “irrational” in that it doesn’t achieve the desired end.

          Similarly if you don’t end up saving enough on gas to make up for the extra purchase price, then you wouldn’t be achieving the desired end of frugality.

          You can come back and say “but no I really just have a preference for hybrids”, but that would be pretty unusual.

          Of course, the true preference is probably for the purchasers to “feel like” they are doing good, even if they aren’t actually accomplishing anything. In which case the criticism of “irrationality” is really just an accusation of hypocrisy.

        • Yes, if hybrids aren’t actually better for the environment in some way, then anyone who buys them because they prefer to spend extra money to help the environment is either acting irrationally or has been misinformed.

          That has nothing whatsoever to do with the payback time, however. If someone buys a miraculous ten-million-dollar car that is in fact much better for the environment — it was hand-built by unicorns in a factory that consumes no fossil fuels, and it requires no fossil fuels to operate, and it lasts for 20 years — it will never “pay for itself” compared to a conventional car. A rational person with a strong preference for reducing greenhouse gas emissions might buy such a car.

          I don’t understand your comments about hypocrisy. I don’t see how someone who wishes to feel like they are doing something good could attain that feeling by doing something they realize is not in fact good. That would indeed by irrational.

          Of course this is hypothetical since, considering the entire life-cycle cost of hybrids (including batteries), hybrids reduce greenhouse gas emission by something like 20-40% compared to conventional cars, according to the last report I saw about this.

        • Isn’t the miracle car still irrational? What if your guy still bought and drove a regular fuel economical car (say a toyota) for say $15,000 and then spent his other $900,085 on other activities that were far more emission reducing than wasting a million on one car. A person with a strong preference for emission reduction must just want to offer all his neighborhood some incentive (say $2000) for switching to a fuel economic car from their old ones or SUVs?

          i.e. Is it or is it not irrational when offered two alternatives, both good, but one grossly better to not choose the better one?

        • You’d need to know more about the guy’s preferences, and about the car. Those unicorns put in a totally awesome sound system, and the seats are so frickin’ comfortable you can barely believe it. And boy, talk about a chick magnet.

          If the hypothetical guy values _only_ greenhouse gas emissions then yeah, it would probably be irrational to buy the car, for reasons you suggest. (There are possible exceptions even then, if we freely create scenarios.) But anyway, a hypothetical person might be that one-dimensional, but no real person is; I don’t think anybody is motivated by only a single factor.

        • I’m always bothered that rational seems to always hinge on a short term personal balance sheet analysis

          So say an early adopter of a hybrid vehicle. From the personal balance sheet, maybe it doesn’t make much sense. Who is this guy fooling. However because the early adopters provided a market, hybrid cars are now cheap to make and they do make sense, having lower cost of ownership and a better environmental foot print.

          So over the short haul the early adopters are fools. Over the longer time frame, they performed an invaluable service. And consider most of these people are operating well away from the Malthusian limit. All the hard needs are met.

          I see examples like this all the time, bothers me.

        • It’s hypocrisy because of the context surrounding hybrid cars (especially 10 years ago when they were considered hip). You’ve got a lot of people emoting about global warming and then just doing the most visible “pro-environment” thing they can without actually figuring out if it helps. And this behavior is pervasive beyond hybrid cars. Even if any particular individual is just misinformed, it is reasonable to call out this sort of conspicuous but unreflective “charity”.

          You’re splitting hairs with the precise meaning of words like “irrational” when the general thrust of the critique is obvious if haven’t been living under a rock for 15 years.

        • Alex:

          “Emoting” is a loaded word, especially given that Phil went to the trouble to give an actual reference to a live-cycle analysis. Let’s keep it clean here, ok? Feel free to disagree but let’s show a little respect. Thanks.

        • I’m not offended by “emoting.” Or at least, not very.

          As I said in my response to Rahul above, I don’t think people are motivated by just one thing. Certainly I’m not. I am unhappy about my high environmental impact, which is one reason my family has only one car, I bike to work, we installed solar panels, we were early adopters of CFLs, we paid extra for an on-demand water heater and we keep an egg timer in the shower to help keep our showers short (partly a drought issue), I eat nearly no meat (partly an animal cruelty issue), and so on.

          But of course one car is still more than zero cars, we could take cold showers and less frequent showers, we could go to bed earlier to save electricity, etc. And the biggie is that I like to travel, so I fly an average of 15,000 miles a year, which makes that my biggest source of greenhouse gas emissions. That doesn’t mean I’m either hypocritical or irrational, it just means I’m weighing many factors against each other in making my decisions. I care about greenhouse gas emissions enough to bike to work and do those other things, but not enough to give up plane travel.

          I’m sure that some people who drive hybrid cars are hypocrites who say they’re concerned about the environment but aren’t. But I think most people who drive hybrid cars are genuinely concerned about the environment, but that’s not the _only_ think that motivates them.

          And I don’t think it’s “splitting hairs” to distinguish irrationality from other motives when commenting on a post that is about the importance of distinguishing irrationality from other motives!

        • @Phil

          Fair enough. But then I want to know an example of behavior you’d label as “irrational”. What’s everyday behavior that you think of as “irrational”?

          Because with the multiple-motivations & subjective-utility criteria it becomes very hard to label almost anything as irrational.

          Would it be irrational for me to pay a witch-doctor $10,000 / year for weekly healing visits? Maybe it offers me peace of mind!

        • Rahul, about fifteen years ago I read several books that have informed my thinking about rationality and decision-making: The Winner’s Curse, by Richard Thaler; Inevitable Illusions, by Massimo Piatelli-Palmarin; and The Armchair Economist, by Steven Landsburg. I learned a lot from these books, although I no longer remember what I learned from which, but from some combination of them I ended up with this:

          “Irrational” behavior is behavior that is not consistent with any utility function. Operationally, you can’t usually recognize irrational behavior from a single choice. Buying a new Lamborghini or a used Chevy can be rational.

          Still, there are some clean examples of irrationality. An example of irrational behavior is putting money in a Christmas Club account. I don’t know if these exist anymore, but until fairly recently many banks offered them. You could put money into your Christmas Club account any time you wanted to. The account paid very little interest, and there was a penalty for withdrawing money before December 1. For a rational person there was literally no advantage to putting money into such an account: at the same bank you could put money into a conventional account, earn higher interest, and suffer no penalty if you decided to withdraw it earlier. The point of the account was to “force” people to leave their money in the account so they could buy Christmas presents, but of course a rational person, when tempted to withdraw it early, would resist the temptation if in fact that was the higher-utility choice.

          (As one or the other of those books discusses, the accounts make sense if you imagine that a person is composed of two agents, whose dominance shifts with time; one of them prefers short-term enjoyment but not so much that he is willing to pay a large financial penalty, the other thinks long-term and is willing to enter a contract that will constrain the behavior of the short-term one).

          That said, I probably didn’t give your question about the rationality of buying the unicorn-built car quite enough respect. Your basic point is valid, that if someone wants to reduce greenhouse gas emissions and is willing to spend ten million dollars to do it, they can do a lot better than simply eliminating one car’s worth of emissions.

        • @Phil,

          I totally disagree with your example. The sort of Christmas Club you described makes tons of sense. Most people lack self control. Assuming you put a premium on having money to spend at Christmas it makes perfect sense to put it into a non-conventional account that penalizes you for spending early.

          Would you call piggy banks irrational too? How about one of those Gym memberships that offer you a rebate provided you work out a certain minimum number of times a month? The alarm clock that shreds a ten dollar bill if you don’t wake up soon enough. Or the alarm clock that hides itself in your room.

          The average person not having stellar self control is just fact. Is it rational to recognize it and design your life accordingly or rational to assume that super duper self restraint?

        • It’s not my example, it’s Thaler’s example, or maybe Piatelli-Palmarin’s example, I forget.

          Lack of self-control is not rational. Mister Spock does not say “I know I’ll regret this for the rest of my life, but I’m going to do it anyway.”

          As I mentioned in my parenthetical note above, if you recognize that you have two (or more) selves that have different utility functions, and that sometimes the current “you” will not be the one making the choices but that you’ll still have to live with his decisions, then it can make sense to open a Christmas Club account or to do those other things that you mention.

          The person who put the ten-dollar-bill into the alarm clock at night had a different tradeoff for sleep-versus-money than the person who wakes up in the morning.

          How about this: can we agree that schizophrenia is not rational? Just kidding…kinda.

        • @Phil

          Dunno. Maybe we are just arguing the semantics of “rational”?

          It seems to me that you accept, as an empirical fact, that it makes sense for many people to buy money-shredding alarm clocks or Christmas club accounts. Just that you aren’t willing to call that behavior “rational”? ( Ok, or only call it rational if I will split the person into two selves. )

          Is that right? If so, you think that sort of behavior makes sense but we shouldn’t call it rational? So sometimes people ought to behave irrationally for their own good?

        • How about this, Rahul and Phil. It is irrational to put $10,000 in each of two bank accounts, both of which are fully insured to $250,000, one of which pays 5% interest and the other of which pays 3% interest. The banks are on the same street in your town, next to one another, are identical in all other relevant respects, and you are aware of both interest rates.

          BTW, I’m inclined to agree with Rahul here. Recognition of your own limitations and investing in precommittment devices (Christmas Clubs, etc.) is not irrational. It violates some intertemporal principles of some utility functions (though not all) but as Andrew said, rationality is just the means. If the individuals who invest in Christmas Clubs find themselves, in the fulness of time, better off ceteris paribus than those who opted for the higher interest rates and couldn’t control themselves, how can you claim that they aren’t rational?

        • Try it this way. Suppose I am schizophrenic. I sometimes act rationally and sometimes act irrationally. When in my rational state, I take steps to protect against my irrationality…maybe this means I don’t own any kitchen knives, or I don’t have any credit cards, or whatever…picture your own blend of irrationality and you can think of some actions that would minimize the damage.

          Rahul and Jonathan (another one), I think you are saying that in the circumstances above I’m rational, it’s just that I sometimes act irrationally. I disagree. You could say that I’m “sometimes rational” or “conditionally rational” or whatever, but to characterize a schizophrenic who has irrational episodes as “rational” doesn’t make sense to me.

          Just because I am “rational enough” to protect against my own irrationality, that doesn’t eliminate the fact that I don’t behave rationally.

          If we’re agreed on that, and I assume we are, then we can move to the Christmas Club case. The only reason the Christmas Club would be “necessary” for me is because the person who deposits money has different preferences from the person who would otherwise withdraw it, even though both of those people use the same brain and inhabit the same body. That’s not clinical schizophrenia but it’s hard to reconcile it any notion of rationality that makes sense to me…or to either Thaler or Piatelli-Palmarin, whichever gave me the example.

        • Phil: suppose your discount rate varied over time. Sometimes it’s really high and sometimes it’s low. In neither case are you irrational… you’re just more impatient sometimes. There is no theory which specifies how heavily you have to discount the future before you’re irrational. Nor is there any theory which requires your discount rate to be constant. Now this sets up a difficult intertemporal problem, since the decisions one would make when highly impatient might well be different than the decisions one might make when patient. In practice, one or the other has to win each of these disputed decisions. Precommitment devices allow the patient decisions to win. Now, we come to the hard part… why should the patient self win? The answer is… because it can. By the way, sometimes the impatient self can win by committing to a course of action which constrains the patient self. But either way, the decision is perfectly rational at the time it is made, and the individual is rational throughout.
          Discount rates are somewhat special, but you wouldn’t call a guy who preferred apples to pears irrational if one day his tastes changed and he preferred pears to apples, would you?

        • I hope this reply shows up in the wrong place.

          No, there’s nothing irrational about your tastes changing with time. But there is something irrational about refusing to accept that your tastes change with time.

          Jonathan (another one), you say what about someone who sometimes prefers apples and sometimes prefers pears, would I call that irrational? Of course not. But what about a guy who prefers apples today but knows he’ll prefer pears tomorrow…yet he still arranges things so he has to eat apples tomorrow. That would be irrational. And that’s the Christmas Club. You join a Christmas Club in order to coerce your future self into doing something “he” doesn’t want to do.

          I’m going to bow out of this conversation. You guys can define rationality any way you want. You should know that your definition is likely to be at variance with the definitions many other people use.

      • I’m not sure what you mean, which is not to say that I disagree with you or that you disagree with me.

        There’s nothing irrational about being willing to pay for a hot tub or a BMW, nor to being willing to pay for solar panels or a hybrid car. Or at least, these are not _necessarily_ irrational. In any of these cases, if you weigh the pros and cons and decide that the expense is worth it to you, that’s fine. What irritates my friend is that people don’t tend to question others’ (or their own) decision to buy a hot tub or a BMW — everyone seems to understand that some people think those are worth paying extra for — but that people do question the decision to pay extra to reduce one’s greenhouse gas emissions, as if the one kind of choice is understandable but the other is “irrational.” It’s not irrational, it’s just not self-interested. Which is why I brought it up in the context of this post!

        • I agree with Alex Godofsky above, the point of the “payback” people is that typically that type of calculation is what’s necessary to actually determine whether the Hybrid does in fact reduce total CO2/pollution/etc. Lots of people want to believe that a Prius is a great thing for the environment, and lots of other people think that maybe it’s just self-love and that the CO2 created by the excess manufacturing process (batteries, etc) offsets the lifetime output of the equivalent non-hybrid vehicle, and that the price difference is an economic signal which hybrid buyers are ignoring.

        • One would hope that anyone who is worried about these issues would look up the research on the subject, which is available online within seconds. Perhaps I would go so far as to say that anyone who worries about these things and doesn’t do the required 3 minutes of research to get a rough answer is being irrational.

          The “price difference is an economic signal” argument doesn’t make much sense to me. The cost of the fossil fuel energy that is required to make a car is not the only thing that influences the price of the car. The costs of labor and raw materials are also significant, for example, and these costs are not the same for all cars.

        • I agree with you Phil. I personally have no dog in the hybrid fight, but I’ve heard enough of all the different sides to understand at least where they’re coming from. the “price as an economic signal” is compelling when your utility isn’t uni-variate (ie. your only interest is in reducing emissions etc). I don’t know that I’ve seen a 3 minute research summary that is highly convincing, but as I said, I don’t really care enough about it to have spent much time. I’m not in the market for a car, much less a fancy hybrid. Plus I hate the visibility in the Prius, and I think safety / visibility is a more important factor for me.

        • I have to admit, it will probably take a typical person, or even a motivated reader of this blog, much more than three minutes to come up with an answer they trust about the life cycle emissions of hybrids and plug-in vehicles. For one thing, they won’t know that the key to the question is to include the jargon “life cycle” in their search. You want to search on something like [life cycle emissions hybrid car], and even then many of the top results will be 5-10 years old and things have changed a lot in that time. As for me, I searched for [life cycle emissions hybrid car 2014] but I recognize that most people will not do the search that way and might get lost in a morass of opinion pieces, old news articles, etc.

          If you do the search I did, you’ll find (among other things) a recent paper that seems to do a decent job at looking at life cycle emissions (that is, including production and disposal, not just in-use emissions). Figure 7 is the main one of interest for this discussion. But this takes far more than 3 minutes to find and to read.

          As for your other topic, I’ve tried the Prius and found the visibility to be OK for me, but I’m quite tall. The windshield pillars are angled such that the farther forward the seat is placed, the worse the visibility. For my wife it’s appallingly bad, enough so that we eliminated the car from consideration. Which is too bad, because I know several people who have the and love them.

          By the way, I’ve seen a review of the Prius that gave pros and cons, and one of the cons was something like “Visibly symbolizes concern for the environment, which some drivers may not like.” Of course I’m aware that there are people who see concern for the environment as a bad thing — people who like “rolling coal” are an extreme example — but I wouldn’t have thought that view would be so mainstream as to make it a “con” in a review!

  3. Although there is a great deal of misunderstanding of this issue out there, some of those who always use rationality and self-interest together may really believe that most actions are of the type “1. Rational and self-interested”, and are trying to push an explanation in terms of that, if nothing else to see how far such an explanation can be taken.

    In a review of a book by Gary Becker, the philosopher Jon Elster wrote “Becker’s work on rational addiction is much more sustained and probing. Although I disagree sharply with much of it, it has raised the level of discussion enormously. Before Becker, most explanations of addiction did not involve choice at all, much less rational choice. By arguing that addiction is a form of rational behavior, Becker offers other scholars the choice between agreeing with him or trying to identify exactly where he goes wrong. Whatever option we take (I’m going to take the second), our understanding of addiction will be sharpened and focused.(http://www.reocities.com/hmelberg/elster/ar97mte.htm)”

    • I have to admit, it will probably take a typical person, or even a motivated reader of this blog, much more than three minutes to come up with an answer they trust about the life cycle emissions of hybrids and plug-in vehicles. For one thing, they won’t know that the key to the question is to include the jargon “life cycle” in their search. You want to search on something like [life cycle emissions hybrid car], and even then many of the top results will be 5-10 years old and things have changed a lot in that time. As for me, I searched for [life cycle emissions hybrid car 2014] but I recognize that most people will not do the search that way and might get lost in a morass of opinion pieces, old news articles, etc.

      If you do the search I did, you’ll find (among other things) a recent paper that seems to do a decent job at looking at life cycle emissions (that is, including production and disposal, not just in-use emissions). Figure 7 is the main one of interest for this discussion. But this takes far more than 3 minutes to find and to read.

      As for your other topic, I’ve tried the Prius and found the visibility to be OK for me, but I’m quite tall. The windshield pillars are angled such that the farther forward the seat is placed, the worse the visibility. For my wife it’s appallingly bad, enough so that we eliminated the car from consideration. Which is too bad, because I know several people who have the and love them.

      By the way, I’ve seen a review of the Prius that gave pros and cons, and one of the cons was something like “Visibly symbolizes concern for the environment, which some drivers may not like.” Of course I’m aware that there are people who see concern for the environment as a bad thing — people who like “rolling coal” are an extreme example — but I wouldn’t have thought that view would be so mainstream as to make it a “con” in a review!

      • Oops, sorry, I posted that in the wrong place! This was supposed to be a reply to Lakeland’s comment, above. Indeed, I’m going to repost it there. Sorry to muck up this thread.

  4. Here’s another way of thinking about it: that the more important distinction is whether people are *consequentialist* or not; and this is orthogonal to both rationality and self-interest.

    And consequentialism is an extremely useful modeling assumption: that decisions depend only upon objective outcomes; and usually this is identified as consumption of goods and services. This implies that decisions are insensitive to a lot of other things like sunk costs, fairness, promises, social norms, heuristics, etc.

    So a lot of economics is based on this conjecture that a large part of the R^2 of decision-making can be explained just by preferences over objective outcomes. It’s simple to extend the model so you care about the consumption of other people, however it’s much more of a challenge if you include norms, promises, reciprocity. So I think that’s partly why consequentialist behaviour is identified with selfish behaviour. (Because a lot of non-selfish behaviour looks more like rule-following than like utilitarianism)

  5. Maybe I’m missing something here, but isn’t this all nomenclatural? You’re certainly entitled to your definition of “selfish2” and I don’t know of any economists who wouldn’t say that there is behavior in all four boxes of the self-interested/rational binaries, so long as altruistic is interpreted to mean “not selfish2.” It’s just that except in special circumstances, economists rarely care whether behavior is selfish or selfish2. Going back to Friedman, we act as observed choices are rational means of implementing some selfish (NOT selfish2) set of preferences and see where it gets us. As Rajiv Sethi points out, it may or may not get us to some sort of equilibrium and that equilibrium (if it exists) might or might not have good properties. As to the examination of some putatively non-selfish2 behavior and whether it’s REALLY selfish2 or not, that’s a minor part of one subfield which looks at these things… but it really doesn’t occupy much of the literature or our thinking as economists.

    • Jonathan:

      Recall the quote from Chris House:

      Rationality means that the “agents” that inhabit an economic system make choices based on their own preferences.

      I think this quote illustrates a misconception (and Rajiv Sethi agrees with me). There are two concepts in House’s sentence: (1) rationality, and (2) agents making choices based on their own preferences. Concepts (1) and (2) are different. You can make choices based on your own preferences in a rational manner, or in an irrational manner. Similarly, you can make choices based on other people’s preferences in a rational manner, or in an irrational manner.

      To take a homely example, some students in my class are designing and building a program to display inferences from Stan. They are focused on others’ preferences; they want to make a program that works for others, for the various populations of users out there. And they are trying to achieve this goal in a rational way.

      OK, sure, you could simply define this as “their own preferences”; e.g., It is their own preference to create a software system that is useful to others. But to define preferences in that way is to make the definition empty: in that case, anything that people do—rational or otherwise—can be viewed as an expression of preferences (and, for that matter, anything that people do could be described as rational). Once we get away from such empty definitions, we can see lots of examples in which rational means are applied to other-directed ends (and vice-versa).

      • I agree the quote from Chris House is misleading. You can certainly make choices based on your own preferences and not be rational. but I disagree with you (and maybe Sethi… who knows?) that “anything that people do can be described as rational.” If I know your preferences and I see you take an action which is strictly dominated by another action that you might have taken, that you know you might have taken, and whose consequences you understood, then you have behaved irrationally. Of course, if I *don’t* know your preferences, it may be very difficult to figure out whether you have not behaved rationally or I have not understood either your goals or your information set, so in that sense there may be *some* set of preferences and/or information set under which what you did was rational, but we might be able to argue persuasively that that particular set of preferences and/or information set is belied by some other observed behavior.

        I think of these discussions a little like I think about solipsism. Philosophically, solipsism is not disprovable, so we proceed through the rest of our lives conscious in the back of our minds that this might all be a joke, but I’ll proceed as if it is not. Similarly, while I might, with some straining, be able to justify some really stupid-looking decisions as rational (“I have a preference for crashing my car into a tree and becoming paralyzed”) in practice we find people prefer more good things to fewer, the prefer a mix of things to all one thing, and they often have motives which (not-selfish2) cause them do things which benefit them only indirectly, whether through the hope for future reciprocation or simply a warm internal glow. We create our models (most of them) around that simple insight and elaborate the implications of large groups of people, with different preferences, behaving that way. As Tyler Cowen says, “Solve for the equilibrium.”

        • Jonathan:

          I think we’re in agreement here. When I say, “anything that people do can be described as rational,” I mean this only in the sense that rationality can be defined in this all-inclusive way. It’s not how I would define rationality. Similarly, anything that people do can be described as selfish, but only if the concept of selfishness is stretched so far as to become meaningless.

          I agree with your statement that we often have selfish interests and that this insight can be valuable. The thing I object to is the identification of selfishness with rationality, which you are not doing but many economists do, often I think without fully reflecting on the implications.

    • Maybe I’m missing something here, but isn’t this all nomenclatural?

      Yes and no. Yes, the problem is nomenclatural, but no, the problem is not *all* nomenclatural. The word “selfish” was coined (I assume) to talk about characteristics a person might possess in various degrees. As I understand it, in economics the related idea of self-interest is the point of departure for a mathematical model — but calling the resulting property of the model “self-interest” is wrong is senses #24 and #26 of this list of 37 ways that words can be wrong.

  6. To focus on one quoted line, “Rationality means that the “agents” that inhabit an economic system make choices based on their own preferences.” That means irrational is rational because irrational actions are choices based on preferences. It only possibly excludes incoherent actions, but generally actions appear incoherent in one frame but coherent in the frame of the actor so they are “rational”. I say this to point out no one actually means “rational” means “rational”. The use of “rational” is as a code word, a highly-loaded signifier intended to convey that the analysis is thoughtful, that the model is founded on respectable and carefully defined. What they actually mean is, depending on the model, not so much rational as autonomic, meaning a set of controlled responses in which the actors are automatons whose responses to stimuli are inhumanly predictable. If we removed the word “rational”, I think we’d be better able to judge the poor quality of many economic models because we’d more easily see they don’t translate into human rationality well.

  7. Your delineation of the two types of “selfishness”, as well as the tautologies underpinning “self interest”, reminds me of a wonderful essay by the Dutch primatologist/ethologist, Frans De Waal: http://opinionator.blogs.nytimes.com/2010/10/17/morals-without-god/

    “Nature often equips life’s essentials — sex, eating, nursing — with built-in gratification. One study found that pleasure centers in the human brain light up when we give to charity. This is of course no reason to call such behavior “selfish” as it would make the word totally meaningless. A selfish individual has no trouble walking away from another in need. Someone is drowning: let him drown. Someone cries: let her cry. These are truly selfish reactions, which are quite different from empathic ones. Yes, we experience a “warm glow,” and perhaps some other animals do as well, but since this glow reaches us via the other, and only via the other, the helping is genuinely other-oriented.”

  8. Economics is a theoretical science and abstains from any value judgment of human preferences, values, or goals.

    It is not its task to tell people what ends they should aim at… nor to characterize their ends as Selfish/Unselfish or Rational/Irrational.

    It is a science of the ‘means’ to be applied for attainment of the subjectively chosen ends — not a science of choosing the ends. That’s where rationality enters the discussion, not before.

    Science never tells people how they should act — it merely shows how humans must act if they want to attain specific ends.
    Choice and valuation of those ends are beyond the scope of any science.

  9. > if “selfish” means utility-maximization, which by definition is always being done (possibly to the extent of being second-order rational by rationally deciding not to spend the time to exactly optimize our utility function), then everything is selfish

    What definition, where? You object to a vacuously encompassing definition of selfishness, but conjure a definition of utility that is equally as pointless. As formally understood in economics utility is a function of possible states of the world
    with some constraints (transitivity, etc) such that (a) the notion of “maximizing” utility has meaning, and (b) it’s not necessarily true of any set of preferences (indeed, it’s probably NOT true of any real person’s preferences, but it may be a useful approximation.) Or in other words: for an actual set of actions and preferences, there’s not necessarily a utility function that makes these actions utility maximizing. On the other hand, if you want to ignore economics and utility theorizing and stick with ordinary language, it’s surely clear that in ordinary language utility (whatever it is) is something that one’s actions might or might not enhance – a contingent fact.

    Whether formal theory of “utility” or “ordinary language” – in neither case is there any justification to say that utility maximization is by definition always what is being done. Utility maximization is not definitionaly “what is always being done”,
    that would be both pointless (your own, correct, argument) but is furthermore false.

    • Bxg:

      You write, “You object to a vacuously encompassing definition of selfishness, but conjure a definition of utility that is equally as pointless.”

      Just to be clear, I object to a vacuously encompassing definition of selfishness, and I also object to a vacuously encompassing definition of utility, and I also object to a vacuously encompassing definition of rationality.

      As I’ve written many times, I find utility theory to be useful in many settings; indeed, I’ve taught courses in decision analysis, I’ve published papers on various aspects of (utility-based) decision analysis, and Bayesian Data Analysis has a chapter on utility-based decision analyses, including two examples from my own applied research.

      And I think that rationality and self-interest are two different things.

        • Rahul:

          I gave an example in an earlier comment here and could easily come up with a zillion more examples.

          For example:

          Rational and self-interested: investing one’s personal money in index funds based on a judgment that this is the savvy way to long-term financial reward.

          Rational and non-self-interested: donating thousands of dollars to a charity recommended by GiveWell.

          Irrational and self-interested: day trading based on tips you find on sucker-oriented websites and gradually losing your assets in commission fees.

          Irrational and non-self-interested: rushing into a burning building, risking your life to save your pet goldfish that was gonna die in 2 days anyway.

          You could argue about the details of any of these examples but the point is that rationality is about the means and self-interest is about the ends.

      • I feel that my question has annoyed you, in which case sorry, but I’m confused because you say (not for the first tiome) “utility-maximization, which by definition is always being done” yet I (a) don’t believe that there is any commonly used definition under which this is even approximately true (of course I’m sure one could cherry-pick something from somewhere that could be distorted thus), and (b) this comment-in-passing about utility seems to be subject to the same criticism that you are are criticizing re: selfishness.

        Utility-maximization isn’t what is always is being done, either in fact or by definition, unless you have a bizarre, entrirely private (and vacuous) definition of this term. Yet you seem to be saying this, not for the first time.

        • Bxg:

          I don’t believe that utility maximization is always being done!

          My statement (which I guess was not clearly phrased) was, “If “selfish” means utility-maximization, which by definition is always being done . . .” It is a conditional statement, conditioning on something that I do not believe.

          It was as if I said, “If unicorns exist, they would have sharp horns.” It’s an “if X, then Y” statement where I don’t believe X. This was so apparent to me (that is, I know very well that I do not believe that “‘selfish’ means utility-maximization, which by definition is always being done”) but it seems that I did not make this clear to others in my phrasing.

          I went in and added another “if” to that statement which I hope makes things clearer.

    • Thanks, TB. Very good piece explaining what economists actually do, emphasizing the diversity (and contextual nature) of the meaning and use of rationality.

  10. First, we would all do well to read Amartya Sen on these questions: https://www.uclouvain.be/cps/ucl/doc/cr-cridis/documents/sen_on_TCR_rational_fools.pdf

    Having got that off my chest, I think the issues can be illuminated by looking at a typical application of economic theory to a practical problem. I have done some work in the field of child labor and have spent more hours than I would like to remember poring over various economic models whose purpose is to generate hypotheses and assumptions for empirical estimation.

    All such models employ utility maximization, although the subject (the entity who maximizes) can vary. Generally it is either an individual parent or a composite parent, but it can also be a composite household whose utility is the sum of the utilities of the component individuals. A little reflection should lead us to doubt any simple representation of the subject, but that’s another matter.

    Now as to self-interest, here the models diverge. Some, where the subject is a composite parent, the function to be maximized is discounted lifetime expected income. Decisions about child labor and education enter in insofar as children contribute to current parental income by working as well as future parental income by growing up, getting a job and contributing to parents in their old age. This is represented as pure parental self-interest. In other models parents are represented as (partial) altruists, and their utility function is a composite of their own income and the utility of their children. Since these two classes of models yield different predictions, it is obvious that the strict self-interest assumption matters.

    The rationality part is the maximization structure. But that structure is tested empirically insofar as the model as a whole is tested. In other words, if the predictions of the model are falsified, researchers consider themselves free to speculate on which aspects of the model are disconfirmed. For instance, the model may predict that rational parents respond in a predictable manner to differences in child labor wage rates. Empirically, however, it may turn out that changes in the ratio of boys’ to girls’ earnings do not result in the expected parental decisions about whether to supply sons’ and daughters’ labor. In this case, researchers are likely to invoke “social norms”, adherence to which contradicts the “rationality” that was initially modeled.

    Yes, you could write a model that had the utility derived from adherence to social norms, either due to socialization or responses from other community members, as an argument, but I’ve never seen it. To their credit, economists usually see the rationality assumption as a falsifiable claim about the world. To their debit they hang onto it far more than they should given the evidence that it is frequently disconfirmed. This is changing with the rise of behavioral economics.

    To sum up, in practice economists do often distinguish between rationality and self-interest. They also have a substantive conception of rationality that goes beyond the minimal notion of completeness and transitivity of preference orderings. (Yes, completeness and transitivity are substantive in the sense that they can be and are disconfirmed, but the range of outcomes that disconfirms explicit maximization models is much wider than that which disconfirms general restrictions on preference orderings, as we saw with child labor.) The problem is that, while they do all these things, economists are often fuzzy on their own methodology, and their writings fail to distinguish between self-interest, rationality and specific implementations of them. (Cue “rational expectations”.)

  11. To add a bit to Peter Dorman’s useful post, two basic points that motivate the microeconomic analysis of choice:

    1. The utility maximization hypothesis of individual choice is simply “doing the best one can do under the circumstances”. “Doing the best” simply means maximizing a preference function, and “under the circumstances” is just in the textbook case a budget constraint. Writ larger, it is any constraint on choice. This formulation can be easily expressed as a calculus problem, constrained maximization.

    2. Incentives matter. Incentives take the form of arguments in the budget constraint, and if altered, leads to predictions of changed choices. Thus the derivation of the inverse relationship between prices and quantities. But of course this can be and has been greatly extended beyond simply market choices, as Gary Becker and many others have noted.

    All of this conforms very nicely to basic intuition. But we should keep in mind George Cox’s useful observation: “essentially, all models are wrong, but some are useful”. The economist’s choice model has proved to be immensely useful, but it is not a full description of all choice behavior, and Kahneman and Tversky have importantly shown.

  12. I think this common confusion between rationality and self-interest reflects the real reason why rational and self-interested agents are so useful to economists’ modeling : they are reflexively stable. A self-interested rational agent learning about the model wouldn’t change her behavior. It doesn’t work with an irrational agent (who may become aware of her shortcomings by reading economics), neither does it with altruistic agents (the paradox of lovers who want each other to be happy). So yeah, this is useful for “equilibrium”, in the sense of “reflexive equilibrium”. That’s what, to me, gives such a passionate and even moralistic tone to the epistemological defense of the rational and self-interested agent modeling. “Who are we to imagine others are not as rational as us/to hope others are altruistic ?” really means “without rationality and self-interest, my model would make predictions that would become impossible if people read my model”.

  13. Herein lie the dangers of inventing normative jargon to describe your models. See also: “normal distribution” and “statistical significance.”

    I’ve always been bothered by the assertion that individuals are only “rational” if they are “money-profit-maximizing.” Give me a scenario in which an individual exhibits “irrational” behavior and I will give you a utility function that the individual could have been maximizing. Chances are, their objective function is not monotonic with the size of their bank account. That was one of the first things i learned in freshman-year micro, that individuals simply do not act except out of self interest. What I can’t do is guarantee that it’s stable over time, or that it has identifiable parameters, or that it’s differentiable, or separable, or homothetic. Or even that it makes sense to an observer. Maybe it’s supermodular. But I’ll never forget the homework assignment in which I was tasked to come up with a supermodular utility function, and found it astoundingly difficult to construct a scenario that was both realistic and mathematically valid.

    We appeal to a restricted notion of utility and of rationality in our models to try to get something useful out of them. I don’t take issue with that; we all know what Box had to say about models. What I have a problem with is using terms like “rationality,” without any qualification to indicate its status as a technical term, to describe the behavior of individuals in these models. It begs a misleading normative interpretation, and that’s exactly what it attracts.

    Maybe I’m too young to understand, or maybe I need to peruse rather than skim Milton Friedman. But I think it breeds an antisocial sentiment in what is supposed to be a social science. I also think this mentality has done permanent damage to the public reputation of the study of resource allocation, and that it’s done even greater damage to at least one generation of undergraduates (e.g. my own).

  14. Pingback: Somewhere else, part 173 | Freakonometrics

  15. Pingback: Introduction to Agent-Based Models with respect to the Future of Macroeconomics | Zachary David's

Leave a Reply to Jonathan (another one) Cancel reply

Your email address will not be published. Required fields are marked *