Identity and beliefs: How biaised are our cognitive abilities and decision-making? – Interview with Roland Bénabou

How do we process information? How selective is our memory? What information does our mind easily retain? Why do we need beliefs to shape our identity?
These and other questions underlie the issues we address in this article, thanks to the inspiring expertise of Princeton Professor Roland Bénabou.

On January 15th and 16th, the Institute for Advance Studies in Toulouse (IAST) organised a series of seminars on Cognitive Control and Social Decision Making. On this occasion, we met Professor Roland Bénabou from Princeton University, who has lately been working in Economics and Psychology, and Behavioural Economics.

He joined the forum to present some of his latest work summarised in a presentation entitled “The Economics of Motivated Cognition”. The latter refers to the unconscious mental processes that makes us assimilate only those information that go “in our favour” and best suit our goals and needs. This is also called biased information-search and assimilation. Sometimes we are motivated to deny or ignore information to protect a positive self-image, to reassure ourselves, to avoid displeasing feelings and unfavorable conclusions. Besides psychological motives, there may also be “instrumental” ones: decision-makers can also seek (or avoid) information that is more (less) useful to favour bargaining and to preserve others’ motivation, groupthink, and identity. Professor Bénabou investigates , often in collaboration with Jean Tirole, a new class of models analyzing cognitive attitudes – like information seeking or avoidance, realism or denial – their economic consequences, and when the resulting equilibrium achieved through such motivated cognition is socially beneficial or detrimental.

Neurosciences are increasingly intersecting with the less and less closed sphere of economics, and luckily many economists are facilitating this interdisciplinary process. Provided with an incredible competence and experience in these (and many other) fields, Professor Bénabou belongs to this group. Therefore, we felt extremely eager to interview him during the IAST forum in January, and we report here his very insightful answers.

1. We can imagine an individual as an open entity operating within society surrounded by an incredible amount of information, inputs and feedback. How do individuals deal with all these external inputs given their constrained capacity to process and internalize them?

Attention is limited and so is memory, thus, necessarily, we are going to focus on certain things and omit the rest. For instance, many researchers are working on Rational Inattention, which means precisely to have such constraints and, simply put, be subject to these constraints, which inevitably imply making certain decision errors. There are also models of bounded memory, still relatively simple at this stage, where the individual can hold only a limited number of “pieces” of information in memory. In our economic models it is very useful to have, as a benchmark, the rational agent without cognitive limitations, or just with some costly information acquisition, but it is also important to realize that cognitive resources are limited and therefore mistakes will arise due to the fact that simple, “quick” heuristics and, very importantly, emotions compete with the rational part of the brain. Most of these bounded-rationality models, while important are still missing the key influence of emotions on cognition, however.

So mainly selective memory?

It depends on what you mean by selective memory. Memory is necessarily selective because we cannot remember everything, so we tend to remember those things that are more likely to be useful to us, in particular to predict what we should do in the future, based on past experience. We also tend to remember or misremember things depending on how pleasant or unpleasant they are to have in mind. Even a computer, which has a finite memory, optimizes its memory management for the tasks it is facing or likely to face. Emotional humans pick very different pieces of information from what a computer, or an ideal statistician, would have selected.

2. Up to which point do you think individuals’ identity and self-cognition are influenced by the external events of the world? And why does external information have different effects on different people in terms of behaviour and identity?

One may live in an information-rich or information-poor environment, one where everybody is subject to very similar stimuli all the time and everyone thinks alike. Alternatively, one may be born into, or choose, an environment where one will meet many different people and be confronted to different opinion, like in a university setting. It depends on both your circumstances and your preferences. What psychologists tell us is that there are personality types who are novelty seeking and others who prefer to be in familiar situations, who either fear or look forward to changes, etc. The “type” you become has largely to do with your childhood experiences, the type of education you receive, and perhaps genetics as well.

It is interesting to see if and why beliefs are irrational. What do you think?

Yes, they can be irrational indeed. This happens for instance when there is a strong need at stake, such as the need to feel reassured about your opinions and ideas, or feel good about yourself. This need will then compete with standard value for accurate beliefs, and it may distort the latter severely.

Do you think that this may be modeled by including, in optimization problems, other components besides individual utility?

We model it as a maximization process with, typically, a nonstandard utility function – for instance one in which beliefs enter directly or indirectly. For example, if I care about pride and self-image, this means I “consume” the beliefs I have about myself, as the thought of myself being a good person provides me with utility. Thus, I want to think I am a good person, as opposed to “rationally” (in the standard sense) wanting to know exactly how good or bad person I am. People thus derive utility from self-reputation or self-esteem, concerning for instance their IQ, attractiveness or morality. Once we have that type of utility function, then we want to achieve and maintain beliefs that are in part serving the usual role of making good decisions, in part serving the new role of satisfying those needs.

Sometimes we are affected by external inputs involuntarily as well, for example marketing can induce people liking things they didn’t before, or it affects the willingness to pay…

Sure, there is recent literature on persuasion and advertising as well. Of course we are not in full control of our beliefs. We may have some desire and some ability to shape them, but a lot of other things are affecting our beliefs: one is reality, hopefully, and another one you are alluding to is other people trying to distort our beliefs for their own purposes. It could be marketing, ideology, politicians and so on.

La Reproduction Interdite. Renè Magritte, 1937

3. In our globalised era, identification can be challenging, given the disappearance of national communities and identities and the continuous questioning of one’s personal values. As a consequence we can imagine two possible extreme scenarios: one dominated by excessive individualism, and the other based on the search for identification in a relatively small closed group.
Do you see these two extreme scenarios in reality? Is there an in-between situation, which may be a realistic outcome nowadays?

This is a question for the young people like you, since you will be the ones who forge the outcome! Each of us has some need to identify with a group, searching for an identity, which could be professional, cultural or related to some other group. This need is highly variable, however, both individually and across societies. There are people who are relatively happy considering themselves not closely identified with any particular group – except for humanity in general, which is a form of identification but not linked to a specific religion, culture or place. Others need strong group identification. What formal analysis allows us to do is to study (and make testable predictions about) how these “types” reflect not just exogenous preferences but also the different kinds and amounts of assets that people have (whether by circumstance or purposeful accumulation), how likely each of these is to be a source of future utility or disutility and to what extent they are complements or substitutes. For instance, if you have ample social capital in term of friends, family, etc., or a solid professional capital, you have less need for a strong form of national, cultural or ideological identity, and vice versa.

I am not that pessimistic about our world becoming increasingly composed of completely atomistic and selfish individuals. Many people feel a need to “find a meaning of life” other than maximizing own material consumption and will thus find other causes or groups to identify with: environmentalism, defence of human rights, etc. In general, I am more worried about going excessively into a collectivist dimension – which involves many forms of identification as ideology, nationalism or religion – rather than into an individualistic direction, but that’s just my own personal view.

4. Looking at the current period of economic and political crisis, we might think that the crisis is perceived as a big shock that could shift individual and collective preferences. So, how can our beliefs be affected by political events, in particular referring to France?

If for some reason whether external or internal, the economy is not functioning well, a significant number of people lose much of the assets (professional, financial, social) they used to have, and so can no longer look forward to a strong or desirable identity as a source of future well-being. They are then going to look for one somewhere else, redefining their values and molding their beliefs in accordance with them. For example, if citizens have been disillusioned by mainstream politics, they look the extreme as a source of hope even if a rational assessment of the evidence does not support it. People have a need to believe in something that will bring them good stuff, in other word, utility. Traditionally, these assets were represented by one’s village, neighbours, family, church, etcetera. If these are no longer as relevant, then it will be one’s profession and human capital, and more broadly the economy and political system in which I have to function. If I get disappointed by these as well due to bad outcomes, or someone convinces me that they do not hold much of a promising future for me, I will find yet other things to believe in – maybe in fact more far-fetched – as sources of future utility.

Does the crisis create incentives for politicians to be more efficient? Due to resource constraints we may think that people would demand more, so that politicians may face a change in the “demand of politics”, towards better performance.

When the stakes rise, competition will intensify. A priori competition can be good, but there is also the risk, given all we have talked about concerning motivated belief distortions, that politicians may just play on voters’ psychological needs, exploiting their desire for hope and something to believe in at almost all costs. So indeed there will be more political competition, however, just as in a product market, it may not be about improving quality, it can also take the form of excessive differentiation, and thus too many varieties of, let’s say, ideologies.

5. About the recent tragedy of Charlie Hebdo (on January 7th). People reacted by identifying themselves with Charlie and with what it represents. The sentence “Je suis Charlie” may be emblematic. People’s reaction culminated in the mass demonstration in Paris. Now, starting from this evidence, what is the reason for this strong need of identification?

I think there was a need for those who participated in that movement – either in person or by expressing support – to send a signal to the world: to demonstrate that even if individually we are all very different from the other, we have core common values, and in particular we strongly believe in freedom of speech.

In terms of identity, there was definitely also a feel-good aspect going on: those who demonstrated together – even though they came from different social strata and held very different opinions on daily issues – felt good about thinking that “we” are not just selfish individuals, each maximizing their own “utility”. They wanted to show that they shared common values and thus felt good seeing others do so at the same time. The latter comes from the fact that normally we tend to like when people around us agree upon our preferences, and conversely it is not so nice when they don’t. Demonstrating together and identifying (literally) with Charlie, not only made the people involved (others disagreed, of course) feel stronger in front of the threat, but also gave a sense of belonging to a larger community. To conclude, few more questions about your and, generally, economists’ work.

6. You devoted a large part of your career to macroeconomics, how did you end up researching in behavioural economics and cognitive sciences? What are your research plans in the short-run?

Well, I think it’s easier to start from your last question…From what seems more logical to me [laughs]. In the short term, I am interested in topics related to religion. This turns out to be a natural evolution of my research agenda as I devoted much work on studying beliefs, and religion is indeed the most important belief to many people in the world. Religion is that kind of belief people really want to pass on to the next generation, are ready to sacrifice a lot for, and even to die or kill for. In this and many other ways, it is one of the best examples of motivated cognition. Moreover, differently from self-confidence or anxiety, religious belief is also something on which there is a lot of data, and on which we can therefore carry out some empirical work.

Now, going back to your first question, I think that switching from macroeconomics was a more complex decision, it was a gradual process that I would also like to fully understand. I first got interested in the growth implications of income distribution, education systems and social interactions. In a sense, my concerns started with inequality, and developed to sociology and ultimately psychology. These transitions are hard to explain, I think they are the result of observing the world under different lenses and responding to it. Being French, and arriving to the US (first to Cambridge) I was puzzled by things that looked totally different from here [France]. For example, discovering the US educational system at the primary and secondary levels especially, which is completely dissimilar to ours. Its structure is decentralized, with huge disparities in the allocation of even the most basic inputs, which depends critically on how each city or metropolitan area is organised into different jurisdictions, and geographically stratified more generally. Thus, I certainly don’t mean to say that I got bored doing research in macroeconomics, but, at some point – especially if you are a theorist – you have said what you have to say on a given topic, and you feel that literature needs to move on to investigating empirical questions and to come back later on with new puzzles in need of analysis. And in the meantime, there are just so many other tempting subjects to get interested in…

7. About the job as a researcher in economics or in any other science: why do we need to model reality, trying to make visible the invisible? What drives you to research?

The main reason is that the world is complex and we want to understand it either because we want to do something for it, like policy advising, or just because we are curious. In order to understand it, we need to simplify and organise it, and that’s what models in science do. This is the first purpose. The second step is to use models to make predictions. Last but definitely not least, perhaps even first, the researcher is searching for a certain beauty in models, an ideal mixture of simplicity and explanatory power.


Among his contributions to the field, Bénabou’s curriculum enumerates various collaborations with Professor Jean Tirole. In “Identity, Morals and Taboos: Beliefs as Assets” [1] they try to explain pro-social behaviour by modelling it as an investment choice, which gives benefit to the individual if she is pro-social, i.e. altruistic. The choice is non-trivial because if in a first period the individual receives a signal about her type (altruistic or not) and bases her first investment on that, in a second period she might forget about her characteristic, and will infer it from her past decisions. In another paper titled “Belief in a Just World and Redistributive Politics” [2] they develop a theory of “collective beliefs and motivated cognition” to study people’s need to trust others in a “just” world, and its implications. The model studies the voting preferences and the effort choice with respect to a future individual outcome, and a redistribution policy, which will be applied in a successive period. The agent’s choice, which is based on a signal of the long-run return to effort, may be based on a “distorted awareness/recall of past signals”. In “Laws and Norms” [3] , Bénabou and Tirole develop a framework to analyze how both private decisions and public policies are shaped by personal and societal preferences (“values”), material or other explicit incentives (“laws”), and social sanctions or rewards (“norms”). In the model, agents’ preferences include intrinsic motivation, extrinsic incentives, and (social or self ) esteem-related concerns. Further, in a recent paper called “Groupthink: Collective Delusions in Organizations and Markets” [4], Professor Bénabou studies group behavior, developing a model of (individually rational) collective denial and willful blindness. Here, agents engage in a joint enterprise where their final payoff will be determined by their own action and those of others, all affected by a common productivity shock. These works contribute to the growing literature on identity and beliefs. Among other important contributions to the topic, we here list, as a reference, “Economics and Identity” [5] and “Identity and the Economics of Organizations” [6] by Akerlof and Kranton.

1 R. Bénabou, J. Tirole, (2011) Identity, Morals and Taboos: Beliefs as Assets, Quarterly Journal of Economics, 126, 805-855.
2 R. Bénabou, J. Tirole, (2005) Belief in a Just World and Redistributive Politics, NBER Working Paper Series
3 R. Bénabou, J. Tirole, (2012) Laws and Norms, Discussion Paper series, Forschungsinstitut zur Zukunft der Arbeit, No. 6290.
4 R. Bénabou, (2013) Groupthink: Collective Delusions in Organizations and Markets Review of Economic Studies, 80, 426-462.
4 G. A. Akerlof, R E. Kranton, (2000). Economics and Identity, The Quarterly Journal of Economics, 3, 715 – 753.
5 G. A. Akerlof, R. E. Kranton, (2005) Identity and the Economics of Organizations, The Joural of Economics Perspectives, 1, 9-32.