Dear Economists, Please Stop Talking About Human Nature

So, after reading a recent post on the Monkey Cage, I started browsing the most highly downloaded papers on the Social Science Research Network (gated access to the top list, sorry). Unsurprisingly, papers by the top agency theorists – Fama, Jensen and Meckling – rank very highly, as they are of broad interest to folks in business, law, economics and organizational theory, which make up a good chunk of SSRN’s coverage. What was surprising, to me at least, was a boldly titled paper by Jensen and Meckling that ranks number 15 on the all time top downloads list: The Nature of Man.

Last semester I had the opportunity to take an interesting class called “Human Natures”, taught by historian of science John Carson, so I was quite intrigued by the paper and wondered where it might fit into that literature. Here’s the abstract:

Understanding human behavior is fundamental to understanding how organizations function, whether they are profit-making firms, non-profit enterprises, or government agencies. Much disagreement among managers, scientists, policy makers, and citizens arises from substantial differences in the way we think about human nature–about their strengths, frailties, intelligence, ignorance, honesty, selfishness, and generosity. In this paper we discuss five alternative models of human behavior that are commonly used (though usually implicitly). They are the Resourceful, Evaluative, Maximizing Model (REMM), Economic (or Money Maximizing) Model, Psychological (or Hierarchy of Needs) Model, Sociological (or Social Victim) Model, and the Political (or Perfect Agent) Model. We argue that REMM best describes the systematically rational part of human behavior. It serves as the foundation for the agency model of financial, organizational, and governance structure of firms.

Ok, some disciplinary methodological individualism aside, not a terrible start, right? But the paper rapidly goes into the land of demonization and heavy ideology of individualism land. I’ll just briefly outline their take on the “sociological model” and some of my issues with it. Here we go:

In the sociological model, individuals are viewed as the product of their cultural environment. Humans are not evaluators any more than ants, bees, and termites are evaluators. They are conventional and conformist, and their behavior is determined by the taboos, customs, mores, and traditions of the society in which they were born and raised. In this model individuals are also often viewed as social victims, a concept that has gained widespread acceptance in many quarters. (See Sykes, 1992).

Basically, J&M take the extreme sort of “oversocialized”, culturally scripted model criticized by Granovetterr (1985) as the general sociological model. For sociologists, then, humans have no agency, their lives are completely determined by rules which they passively accept. J&M contrast this with their nuance “resourceful, evaluative, maximizing model”:

By contrast, REMM is an evaluator. The REMM model recognizes that customs and mores serve as important constraints on human behavior, and that individuals who violate them incur costs in many forms. But REMMs compare the consequences of alternative courses of action, including those that involve the flouting of social norms, and consciously choose actions that lead to their preferred outcome. Moreover, if the costs or benefits of alternative courses of action change, REMMs change their behavior. In the sociological model, individuals do not.

Now, I readily admit I am just a budding Sociologist, but who exactly are Jensen and Meckling talking about? What empirical or theoretical literature really treats humans as that rule-governed? Is this some sort of reaction to Parsons (whose work I am not at all familiar with) or Durkheim? When I read, say, Selznick’s 1949 account of the TVA, I see a whole lot about individuals, constrained and enabled by certain social structures and institutions, making do. The TVA members seeks out and coopts the grassroots elites (a phrase that makes sense in the context Selznick uses it in, even if now it feels anachronistic) in order to accomplish its more primary mission. So, clearly these are not the overdetermined actors J&M rail against. Alas, J&M don’t bother to cite any examples of this model in action, so I am left to speculate.

Next, J&M minimize one of the primary (potential or actual) contributions of cultural and norm-based investigations:

To be sure, social practices, customs, and mores play an important role in determining the attitudes and actions of individuals at any point in time. They serve as an external memory device that aids in the storage of knowledge about optimal behavior. In addition, they represent a major force for teaching, learning, disciplining, and rewarding members of a group, organization, or society. But if the group or organization is to prosper—and, indeed, if the society itself is to survive—these cultural practices or values must adapt to approximate optimal behavior given the costs and benefits implied by the opportunity set faced by individuals in the society.

No, no, no! Culture, values, norms, etc. do not simply serve as an “external memory device”. Those cultural features define what is optimal, and indeed, even the whole framework of seeking optimal outcomes. Optimality, especially when that concept is broadened to include non-monetary outcomes (as J&M do by contrasting their REMM model with a narrow economic model) has no fixed meaning. Weber’s story of rationalization is precisely the story of the move towards intended rationality, of trying to behave as economists believe people should. But how means and ends were connected and, most importantly, the set of ends themselves (be it religious purity or large bank accounts) are not features of the rational system itself, but part of the larger cultural framework. Culture (writ large, and obviously with lots of heterogeneity and space for change and conflict) tells us what to want and how to want it.

Ok, so why is it such a problem that Jensen and Meckling have such a distorted view of sociology, and ignore the importance of culture in determining preferences and such? Because it informs the policies they prescribe for society:

As another illustration of the workings of the sociological model, consider the current debate over the causes of homelessness. The very use of the term “homeless” suggests no choice on the part of street people (who are therefore victims of the system); it also carries little or none of the social disapprobation of “vagrant,” a now unfashionable label. This change in language and attitudes reduces the decentralized sources of social or cultural punishment for being a street person—again something the REMM model predicts would result in an increase in this socially undesirable behavior. New York City now spends in excess of a half billion dollars a year on subsidies for the homeless, and the problem shows no signs of going away even with the improvement in the economy. (And the “de-institutionalizing” of the mentally ill, a common explanation, by no means accounts for the vast increase in the numbers of street people.)

The sociological model suggests that if an individual’s income and wealth are small, it is entirely due to cultural factors, environmental adversity, or bad luck—not to conscious effort, the choice of leisure over work, the choice of a particular type of work, or the failure to invest in learning. Therefore, “justice” requires that we confiscate the wealth of the more fortunate to recompense the unfortunate. Of course, the higher the recompense, the more attractive it is to be poor, and REMMs will respond by taking more leisure, by choosing occupations in which employment is more unstable, and by investing less in learning. The REMM model predicts that if we make the payoff high enough, we can attract an arbitrarily large number of people to become poor or unemployed—or at least to meet the established criteria for those programs. This describes important aspects of our welfare and unemployment systems.

Yes, that’s right folks, homelessness and poverty are entirely choices made by free individuals who simply didn’t like to work much. Don’t take our money away!

Now, I agree that policies that ignore how individuals respond to incentives are doomed to fail. And in this, thinking in terms of people as resourceful and evaluative is essential. But that’s very different from trying to understand the system as a whole or, as R.E.M. might say it, “how the west was won and where it got us”.* The REMM model, on its own, predicts nothing because it requires exogenously determined preferences which are more or less constant, e.g. culture. The two together might make some predictions, but on its own, an assertion that individuals actively try to achieve what they want is tautological and predictively worthless.

And it goes on, as Jensen and Meckling move to tackle the health care crisis:

There is a U.S. health care problem, to be sure; but it does not stem from too little regulation and too few subsidies. Rather it comes from our third-party insurance system that effectively removes responsibility for the costs from the most important decisionmaker, that is, the patient. The key to solving this problem is to impose the financial consequences of their medical decisions on patients through greater use of co-pay insurance with larger deductibles that place first dollar costs on patients while protecting them against catastrophic illness.

Clearly, the system we have in place now, about 14 years after this paper was published, approximately more closely this solution, and people hate it – partly because it’s wildly inefficient, and partly because we simply don’t want to have to worry about some things. We don’t want to have to make choices about our health care, we want people who know a lot more than us to do so on our behalf. By so readily dismissing Herbert Simon’s insights about the costs of rationality (as Jensen and Meckling do in a footnote on page 4), J&M ignore the value of expertise and of systems that help us make the right choices, sometimes by giving us very few choices at all. I don’t want to have to choose what exact illnesses to be insured against, for example. I wouldn’t be very good at it. I want experts to figure it out and make it work for society as a whole. And many societies do just that – indeed, most other wealthy nations, and at a fraction of the price we pay here. And yet, J&M argue with no facts or figures, but rather pure faith in free markets and maximizing individuals, that health care centralization will inevitably be more wasteful.

I’m sorry Jensen and Meckling, but you can’t impose your representation on reality without reality fighting back, and you can’t keep reality out of your representation either. The REMM model is just that, a model, and a valuable one at that – for situations where preferences are fairly clear and straightforward, like firms trying to maximize profits. Even there, the “maximizing” part remains very contentious, as the flow of recent work in the institutionalist literature would attest to. But stay away from human nature, ok? You are economists, not philosophers.

* Sorry, after reading “REMM” over and over again, I had to toss in an R.E.M. reference somewhere.



  1. The idea of sociology as “victimology” is pretty widespread from what I can tell (among outsiders, anyway). I’ve encountered it with my students, too. The logic is easy to follow, even if it is flawed: if sociology is the study of how the social world influences people, then people are pawns of the social world.

    If only we could discover the laws which guide our actions, we could harness them for the powers of evil and TAKE OVER THE WORLD!!!!! MUAHAHAHAHAHA!!!

    That’s why I got into sociology, anyway 😉 .

  2. So, the “wicked” in “wicked anomie” is like the wicked in wicked stepmother? 🙂
    Of course, if we do discover the laws that bind us, we’d be just as much trapped, no? There’s nowhere to stand, in Aristotle’s terms, for us to move society with.

    I mean, until I build my space lab in space.

  3. haha, that may be my new favorite song.

  4. Secretly, Jonathan Coulton may be the best combination of music + nerdiness ever. I highly recommend “Re: Your Brains”, “Mandelbrot Set” and “Ikea” amongst many others. All are available via youtube, and all are creative commons’d, because Coulton is cool like that.

  5. Bennett Holman

     /  June 15, 2008

    As regards to human nature, after 6 years of studying psychology (and given my new found career in philosophy of science (since you gave the task of specifying what people are to philosophers)) I can tell you definitively a few things regarding their nature:

    1) Claim: People are not rationalizing agents
    Proof: I have better things to do than continue this post

    2) Claim: People are not like termites, ants or bees
    Proof: Termites don’t have health care, ants don’t pay taxes, and I lack the ability to sting people that upset me.

    3) Claim: Human action is not exclusively motivated by need fulfillment
    Proof: If you flash something in the corner of someone’s peripheral vision they will orient their gaze, this is governed by a separate subcortical neural pathway than the main visual pathway that enters awareness (though if you asked someone to explain their behaviour they tend to come up with a story that involves things like I was curious to see what was over there)

    4)Claim: When it comes to predicting the aggregate behaviour you need not have a realistic theory about the constituent parts.
    Proof: Nobody does physics to explain the stock markets.

    5) Claim: Jonathan Coulton is the best combination of nerdiness + music
    Proof: See code monkey

    Conclusion: I argue if you are concerned with understanding social behaviour on the group level that you may employ some sort of theory about individuals and that while this theory may be informed by psychology, it is not required. This is a level of analysis issue and for the forseeable future, all theories used in sociology will be vast oversimplifications and have very little import as to what people “ARE”. The main concern is what assumptions allow for the best understanding of the social phenomena in general, or perhaps the one that is of particular interest to termites.

  6. Bennett Holman

     /  June 15, 2008

    CORRECTION: Claim 1 should read “People are not rationally maximizing agents”
    Further proof: I corrected claim 1, when I continue to have better things to do…

    Additionally, pretty much anything that serves as a mechanism for decision is going to end up being tautological, like “people are motivated by motives” or “people influenced to action by things that influence them to action.” This is precisely the dilemma that faced behaviourism: People’s actions are governed by reward and punishment. An action that is rewarded is more likely to repeated in the future, an action that is punished is less likely to be repeated in the future.
    This raises the question: What constitutes a reward (or a punishment)?
    Well you see, a reward (punishment) is something that increases (decreases) the frequency of the action that preceded it. Thus, the circle is complete. But that is fine, provided we understand what we up to! Definitions are neither true nor false. Of course we want our theories to be self-consistent so hopefully something like this will occur. The hard part is cashing out the coordinative definitions, viz. what things in the world our definitions link up to.

    Behaviourism served as a useful model in psychology for some years and lead to some very effective therapies (and some pretty terrible ones). However, it fell not because it contained patently false assumptions (like inner mental life is inconsequential), but because (amongst other reasons) it was shown to be an inadequate account of certain phenomena (e.g. language), ceased to be considered a fruitful paradigm for further research and because it was augmented by a paradigm that served the psychological community better for understanding human behaviour (namely cognitive behaviourism).

    But even in it’s heyday there is no reason to think that the success of behaviourism actually implied that human being have no inner mental life (or that inner mental life is of no consequence), and it was a mistake for people to interpret entities in scientific models as ontological. Further, Skinner was making a methodological point concerning inner life, not an ontological one: “A great deal goes on inside the organism, and physiology will one day tell us more about it… At the moment neither introspection nor physiology provides very adequate information what is going on inside a man as he behaves.” (Beyond Freedom and Dignity (pg 195)) The point is merely this: researchers are faced with the questions what are the best avenues currently available to advance our understanding about the issues in question and how best can we go about achieving our goals. Now of course, opinions will differ on the matter especially when there are differences in opinion concerning appropriate goals, but this is not a scientific question, but rather a question that must be answered before science can begin.

    It seems that the goal of sociology is to capture phenomena that are occurring at a higher level of complexity than individual action and to approach these issues researchers will make simplifying assumptions about individuals (or the interactions between individuals and surrounding culture if you wish). But if these assumptions violate our notions of what human beings are, this does not seem to me to be an argument against these assumptions, because these “models of individuals” should never have been taken to imply ontological claims about “real people” to begin with. They are methodological claims about what models capture the emergent social phenomena the best and as such can only be arbitrated with reference to the efficacy of how certain models do or do not capture the phenomena of interest.

  7. Geronimo

     /  December 12, 2008

    You’re reading J&M as if they would have any intention to produce decent social theory where their only goal is to introduce the tenets of agency theory into social theory. Their article is a prime example of poor scholarship: contentious and epochalist approaches to theory that they clearly have not read and a substitution with a purely personal system of understanding the world with no references whatsoever.
    Thanks for you blog on this.

  8. Kenneth Curmi

     /  September 6, 2009

    The article is not claiming that people are like termites or ants. It is claiming that they are no more of an evaluating agent than them, which is a considerably different claim.

    People, according to the article, are shaped by the culture they live in. We don’t become Christians or Muslims, we don’t choose to adopt Western philosophy instead of oriental. We are born Christians with a Western mind, or Taoists with an oriental philosophy, etc.

    The question is how much this is true (how much of our nature is, for instance, shaped by evolutionary factors rather than cultural ones) and how much say (freedom) do we have in determining our own lives.

  9. Lisa-Anne

     /  February 18, 2014

    Was given this 15 page document as part of a management course. I peeked ahead in the course, and noticed a section on something along the lines of “All talk, no action” – after about 2 pages of reading, I assumed that this document was fake, and planted as an example of someone “windbagging”. I very seriously thought that there would be an eventual test to see who would be the first to reveal that this article was an incoherent rambling mess. I read through all 15 pages and conclude: this was probably one of the most memorably painful reading experiences of my life.