
Ariel Rubinstein on Game Theory and Behavioral Economics 04/25/2011
What this covers
Ariel Rubinstein of Tel Aviv and New York University sits down with EconTalk host Russ Roberts to challenge prevailing views of what game theory and behavioral economics can accomplish. Drawing on his Afterword for the 60th anniversary edition of Von Neumann and Morgenstern's foundational text, Rubinstein—himself a game theorist—argues that the field has dramatically oversold its value. The conversation centers on a single, recurring claim: that economic models and game-theoretic solutions function as logical fables rather than predictive instruments or guides to policy, and that the profession promotes them as more powerful than they are because theorists have material incentives to do so.
The discussion ranges across several domains of concern. Rubinstein contends that game theory identifies patterns of reasoning but yields no normative conclusions and minimal empirical force, resembling logic more than physics. He pushes back against the notion that solution concepts like equilibrium provide actionable answers to real conflicts, and traces the value that game theorists have attained to their intelligence and intuition rather than to game theory itself. On behavioral economics, he faults experimental methodology: economists fixate on statistical thresholds while ignoring larger uncertainties about data collection and protocol fidelity, and they too rarely replicate findings. Roberts raises the problem of publication bias and hidden researcher degrees of freedom—how many regressions were run, how many outliers discarded—that inflate the reliability of reported results. The pair also examine why sensible people laden with political commitments come to believe their views are scientifically grounded, and they conclude that economics can sharpen thinking about incentives and unseen forces without claiming scientific certainty about outcomes or policy.
Rubinstein argues that game theory and economic models are not predictive or prescriptive tools but rather 'fables' or logical exercises that yield insight without normative or strong empirical force, and that the profession oversells their usefulness for self-interested and psychological reasons.
- Game theory is a cousin of logic: it identifies patterns of reasoning but does not tell us what action is preferable or predict behavior
- The value attributed to game theorists comes from their intelligence and intuition, not from game theory itself
- Economic models are 'fables'—sterilized, unrealistic stories that yield insight but not policy prescriptions
This asset isn't compiled yet
You're seeing its claims, ranked. Compile it to build the argument threads, weight them, and check each claim against your library — the full view.
Economic models are best understood as fables (sterilized, unrealistic stories): like fairy tales, they strip away the world's complexity to a few elements in a sterile environment, and a good model—like a good story—yields an insight about how people reason and how the world works, not a recommendation or prediction.
“my view about economic model in general is that it is a... it's a... and it's not like... modeling in the sciences... I... the model that I build or invent or other people's models I read them as stories and the good model... is a good story”
Relying on introspection to validate findings is dangerous because it is prone to confirmation bias—only believing the experiments that 'make sense to me' is a self-confirming trap.
“he said to me oh don't worry I only believe the ones that make sense to me so of course that's a dangerous road also”
Empirical work across all fields suffers because researchers present an 'expose story' of what they did rather than the truth—not revealing how many regressions or specifications they ran, how many outliers they discarded, or how many times they re-ran until they got the result they liked.
“the researchers don't reveal what they really did they tell us an expose story that we tested this and it came out yeah we were right and of course you don't know how many times they ran the regression”
Economists are insufficiently skeptical about experimental data: they fixate on statistical significance (p-values, the 5% threshold) while ignoring the far larger and unmeasured uncertainty about how the data was actually recorded and whether the reported protocol was truly followed.
“the big uncertainty about results in my opinion is not... some discussion that usually is measured by the [p-value] but the uncertainty about the way that the data is recorded the... way that the experiment is done”
Predicting behavior in the social sciences is fundamentally difficult because prediction itself is part of the game—forecasters are themselves players—so the vision of game theory as a predictive tool that will improve with more parameters is unfounded.
“we need also to deal with the fundamental difficulty of predicting behavior in the social sciences where prediction itself is part of the game and forecasters are themselves players”
Adam Smith observed that man wants to be loved and to be lovely—to be esteemed and to deserve that esteem—but this creates a tendency to self-deceive into believing one has earned esteem when one has not.
“man wants to be loved and to be lovely... we want to have people think highly of us and we want to earn that respect... so we have a tendency to delude ourselves that we've actually earned it when in fact we haven't”
Economics sensitizes practitioners to unseen effects—the role of incentives and market forces that are difficult to perceive—and game theory specifically sensitizes one to strategic interactions, which is a genuine but understated benefit of the training.
“it sensitizes you economics does generally it sensitizes you to unseen effects it sensitizes you to the effect of incentives it sensitizes you to the role of market forces which are very difficult sometimes to be aware of”
The daycare-center fine result—where introducing a monetary fine for late pickup increased lateness—is explained by common sense, not by paradoxical demand curves: without a charge, parents act considerately, but a small fine converts a social/moral obligation into a priced transaction they can rationally choose to buy.
“if you don't have to pay anything for some... then you may actually be more considerable... to the interest of the guy who gives you the service... if he charges you... some peanuts then I know... so now it was OK for me to be later... and to pay the $10”
Economics lacks a culture of replication because journals will not publish replications, and even when researchers do replicate they have incentives to alter conditions and claim a novel secondary phenomenon rather than simply confirm or refute the original.
“we lack this... replication... the chances that I will be able to publish it... are very small that makes it very rare people replicate and if they replicate... they'll take... incentives to improve the results”
Game theory's primary object is to study the considerations used in decision-making in interactive situations—it identifies patterns of reasoning and investigates their implications—but it has no normative implications and very limited empirical significance; it is a cousin of logic, which does not screen true from false or right from wrong.
“game theory does not have normative implications and its empirical significance is very limited game theory is viewed as a cousin of logic logic does not allow us to screen out true statements from false ones”
Most of the interesting findings in experimental and cognitive economics do not actually come from experiments but from introspection and skilled observation of how people reason; figures like Tversky and Kahneman are like artists who invent experiments to demonstrate insights they already grasped through introspection.
“introspection... which I believe is the main tool of the experimental economic economics or the cognitive psychologist is... the most important tool in this field”
The word 'solution' in game theory misleads people into believing the field offers concrete answers to real dilemmas and conflicts, when 'solution concepts' like equilibrium are merely logical constructs analogous to logic, not actionable resolutions.
“the term solution we talk about solutions game theory we don't just talk about equilibrium we talk about solutions and this word in English... triggers people to believe that he gives something very concrete really solution to the dilemma”
The genuine value of game theorists and economic theorists comes not from game theory itself but from the fact that the field attracts highly intelligent people who also possess a strong intuitive 'touch' for understanding how people behave and what matters in real situations.
“many of the people work in this field have this special ability of analytical ability and some touch to the world which makes them very valuable independently of... game theory itself”
An elegant game-theoretic mechanism (such as a truth-telling scheme for recovering value of mixed-up cargo after a shipwreck) has logical beauty but is foolish to deploy in reality, because traumatized real people will not behave as the equilibrium assumes.
“the idea that that would work the first time there's a solution in the chaos in the aftermath of a typhoon when people are traumatized that they would just all act like the game theorists say they would is stupid”
People who dress political views in scientific rhetoric are typically not being dishonest but are self-deceiving: their desire for the authority of science is so strong that they genuinely persuade themselves their ideological positions are scientifically grounded.
“I would not say that it is dishonest I would say that... desires people to to make claims which are not just the views... and somehow is so strong that people are just persuading themselves”
A surprising, career-launching psychology finding faded over time when replicated—the effect shrank with each replication until it essentially disappeared—which reflects not changing human behavior but the original study's unreliability and publication's bias toward striking results.
“over the years he's realized that that finding has not stood up that when people tried to replicate it over time it got smaller and smaller and now it's basically it disappears”
Experimental economics' rigid required protocols—small monetary incentives and lab settings—are scientifically weak because the small payments don't meaningfully incentivize subjects; hypothetical-choice or web-based experiments with thousands of participants yield equally or more reliable data because people are good at imagining themselves in hypothetical situations.
“the amounts that are given are not really incentivized people... there is... [little] difference between this protocol and protocol when you're telling people you're not going to get any money”
The key difference between a fairy tale and an economic model is formalism: in a model you cannot simply move from beginning to end as the author wishes but must justify each move using abstract solution concepts, whereas a fable's writer moves freely.
“the big difference... between a fairy tale and an economic model that the formal... model... you have some rules how to move forward... you cannot just say go from the beginning to the end in the way that the writer wants... but you are able to justify the move... using some abstract concept some solution concept”
Economics is not useless despite models being just stories: it is a powerful way of thinking that reveals hidden, unseen forces and magnitudes we cannot measure; the error is scientism—using scientific language and technique to give an aura of certitude it doesn't deserve, as when pretending to run the economy or design public policy.
“economics is a powerful way of thinking it's a way of helping you organize your thinking... economics helps you see things that are hidden that are unseen... we're talking about magnitudes we can't always measure but we can see forces”
A 'wisdom of crowds' bean-jar test illustrated that the mean only matched the true count because one wildly inaccurate outlier (8,000 vs. true ~800) offset systematic underestimation; whether the theory is confirmed depends entirely on the arbitrary decision to keep or discard that outlier.
“every single person almost every single person underestimated how many beans were in the jar... most people said 500 600 but of the 30 people I asked one person says like 8000... when you included that 8,000 it came out very close to the mean”
Supporting political or ideological positions with the rhetoric of scientific justification is a pervasive and dishonest practice, in which people load their biases into scientific language to give their views unearned authority.
“we have the same problem here it's a horrible and it's a dishonesty... pretend we're scientists we're not we're selling we're loading our biases and scientific rhetoric”
Once game theory became applied—with theorists becoming paid consultants and advisors—economic self-interest entered the picture, giving practitioners a vested interest in believing and asserting that their work is concretely useful.
“once game theory became applied in the sense that our companies that... becoming consultants and giving advices and also get some nice amounts of money in return then of course the interest also play”
Skepticism is the fuel of academic life, and a central problem in economics is too little skepticism and too much belief that the field is on the right track toward fulfilling its goals.
“to be skeptic in academics that's the fuel of academic academic life”
An academic has no obligation to produce work that is useful or applied; thinking for the sake of thinking is a legitimate end, and the field should recognize a more modest, intellectual sense of usefulness rather than the everyday sense.
“I've never felt that I asked any obligation to be to do something which is useful which is applied”
An economic model of bargaining need not predict anything or be verified in real life—unlike a physics model—because it is one story among many possible stories, and its usefulness is at most very indirect.
“the model does not have to predict anything it does not have to be verified in life it's not a model from like a model in physics is”
The economics profession behaves like a cartel, imposing rigid experimental protocols as barriers to entry to keep the field from becoming too easy to enter, since the basic methods would otherwise be accessible to too many people.
“the economic position... behave like a cartel and... as barriers to entry... the cartel of... economics... put a barrier to entry because in some sense it sounds very easy”
The advice-giving dynamic is co-created: it satisfies the desires of multiple parties—someone wants to give advice and present it, someone wants to receive it, and someone (like a politician) wants to offload responsibility for a decision.
“the desire to get the advice... is coming from somebody wants to get their advice and somebody wants to present it and somebody would like to listen and somebody would like to tell... responsibility... if it's a politician”
Honestly telling students that economics is not concretely useful for making money or advising ministers paradoxically attracts the best students—those with bright minds, openness to new ideas, and genuine eagerness to solve real problems—because the field truly needs such people rather than those drawn by promised payoffs.
“I some colleagues... criticized me on that I did tell them from studying economics because I tell them exactly what I was saying now... but... what I learned from my experience is the opposite”
In an eye-tracking experiment, the recorded data was entirely wrong because the up/down axes were reversed, yet across 20-30 lectures no audience of experts ever detected the error—demonstrating that researchers and audiences uncritically trust that data was recorded correctly.
“the data was wrong because... everything up down were reversed... and people could realize it because... all the movements the eye movement started actually from the bottom”
It is psychologically easier and more acceptable to claim authority through credentials ('I have a Nobel Prize in game theory, listen to me') than to admit one's insights come merely from being intelligent, so experts cloak personal intelligence in disciplinary authority.
“it's very difficult to say okay I am smart my IQ is such... listen to me... but it's very easy to say I have a... therefore I'm an expert on... whatever and now listen to me”
When people are confronted with the idea that their certainty is not as high as claimed, they typically react angrily and aggressively by attacking the challenger's competence rather than reconsidering their position.
“when you confront people with the idea that their certainty is not as high as they claim it to be the reaction is very angry and aggressive... what they respond with is you don't know what you're talking about”
The best way to genuinely criticize and understand a method is to do it yourself, because practicing it makes you both aware of its flaws and more sympathetic to the difficulties researchers face.
“the best way to criticize something and to understand how people do something is to do it yourself... you are also becoming more sympathetic to the difficulties that the other people face”
Rubinstein is motivated in life by a hatred of pretension and of authority in cases where there is no real basis for that authority.
“one of the things that motivates me a lot in life is my hate to pretension and my hate... of authority in cases there is no... authority to be based on”
The popular literature is full of nonsensical claims about game theory's usefulness.
“the popular literature is full of nonsensical claims to that effect”
Rubinstein trusts and believes in young people more than people his own age, finding positive responses to his skeptical views from students 'more pure'.
“it's nice to get very positive responses from young people from students... I believe in young people more than I believe in people in my age”
Roberts has no answers to economic crises and offers only 'an insight or two,' but people don't want insights—they want the truth, the answer, the solution.
“people come up to me and say in the middle of the crisis... people must be coming to you for answers and I always say not the smart ones because I don't have any answers I may have an insight or two”
It is tempting to claim privileged knowledge over politicians and the public by invoking professorial or Nobel-level credentials, and Rubinstein admits even he has wrestled with this temptation but tries to speak publicly only as a citizen.
“it's tempting to say I know something that you don't know I'm a professor... I know some people who won the Nobel Prize... so listen to me... except I don't know”