Show Summary Details

Page of

Printed from Oxford Research Encyclopedias, Economics and Finance. Under the terms of the licence agreement, an individual user may print out a single article for personal use (for details see Privacy Policy and Legal Notice).

date: 20 March 2023

Behavioral Development Economicsfree

Behavioral Development Economicsfree

  • Karla HoffKarla HoffInternational and Public Affairs, Columbia University
  •  and Allison DemerittAllison DemerittDepartment of Sociology, University of Washington


Economics, like all behavioral sciences, incorporates premises about how people think. Behavioral economics emerged in reaction to the extreme assumption in neoclassical economics that agents have unbounded cognitive capacity and exogenous, fixed preferences. There have been two waves of behavioral economics, and both have enriched development economics. The first wave takes into account that cognitive capacity is bounded and that individuals in many situations act predictably irrationally: there are universal human biases. Behavioral development economics in this first wave has shown that low-cost interventions can be “small miracles” that increase productivity and well-being by making it easier for people to make the rational choice. The second wave of behavioral economics explicitly takes into account that humans are products of culture as well as nature. From their experience and exposure to communities, humans adopt beliefs that shape their perception, construals, and behavior. This second wave helps explain why long-run paths of economic development may diverge across countries with different histories. The second wave also suggests a new kind of intervention: Policies that give individuals new experiences or new role models may change their perceptions and preferences. New perceptions and preferences change behavior. This is a very different perspective than that of neoclassical economics, in which changing behavior requires ongoing interventions.


  • Economic Development

Since the creation of neoclassical economics in the 1870s, theoretical research and much of the practice of economics have used a very special model of the individual—an actor with unbounded rationality and exogenous, ordered, and stable preferences. Such a person is the rational actor, or Homo economicus. Many neoclassical economists believe that this model of the individual, however unrealistic, is the best basis on which to explain economic outcomes and to design policy. Until the late 20th century, economics did not have evidence of systematic and widespread violations of rational choice. Economics also did not have quantitative evidence that history and experiences can durably change preferences. In many circles, invoking preference change as an explanation for changes in behavior was condemned as bad science (Fehr & Hoff, 2011; Stigler & Becker, 1977).

Development economists could explain a lot with neoclassical theory, and powerfully so. Yet the theory also left a lot of important phenomena inadequately explained—perhaps most especially, the nonconvergence of countries onto similar paths of development. As the precision of empirical and experimental work on judgment and behavior grew, most economists recognized that there was a great deal that rational choice theory did not explain. The first wave of behavioral economics is based on the discovery by Daniel Kahneman and Amos Tversky (1979, 1984) that much of human behavior is based on intuitive—not deliberate—thinking, and certain biases—such as loss aversion and present bias—appear to be hardwired in the human brain. (These two biases are also known to be in the brains of the primate ancestors of humans). Present bias means that time preferences are inconsistent—planned future decisions are systematically different from the actual decisions people make when the future comes.

The idea of incorporating experimental and empirical evidence of how the mind works into economics is the crux of all behavioral economics. Drawing on behavioral economics, development economists discovered that interventions could sometimes be designed to improve decision making and avoid making choices that they would later regret. For example, an estimated 2 to 3 million people die each year from diseases that can be prevented with vaccines. In a field experiment in India, giving parents small nonfinancial rewards every time they bring a child to an immunization clinic for a needed vaccination helped to overcome the parents’ present bias and boosted vaccination rates. For some parents, the intervention transformed a loss frame (a parent gives up a day by going to the immunization clinic) into a gain frame (the parent gets a small reward on the spot when he brings his child to the clinic). This behavior change is one of many “small miracles” made possible by interventions that overcome biases, inattention, or limited willpower and thereby increase investment in productive inputs, education, water chlorination, and preventative health products such as bed nets (see Bernard et al., 2019; Duflo et al., 2008; Dupas & Robinson, 2013; Kremer et al., 2011; an overview is Demeritt & Hoff, 2016).

Colin Camerer (2005) and Karla Hoff and Joseph Stiglitz (2016) distinguish two waves of behavioral economics. The first wave, which began in the 20th century, focused on systematic deviations from simple rationality principles and how to correct them. The second wave, which began in the 21st century, is informed by the sociological and anthropological perspective that experience and exposure influence how people see and understand the world: Society, so to speak, exists within persons. This idea is common sense to most parents, who know that the social environments that their children spend time in influence their values and behaviors. New techniques—including psychological tests and quasi-natural experiments—as well as innovative historical analyses have provided rigorous evidence in the 21st century that sociocultural experiences, including experiences of one’s ancestors in the remote past, may influence how people think and behave and the per capita incomes of nations.

Although many people associate only the first wave of behavioral economics with the term behavioral economics, the second wave can be understood as an extension of the first. By emphasizing the effects on behavior of biases (some of which are learned) and of irrelevant contextual features (some of whose meanings are also learned), the first wave of behavioral economics implicitly brings in culture, since it recognizes the power of cultural cues (e.g., Kahneman, 2011, p. 56) and cultural categories. Categories and their meanings differ across cultures (Malt, 1995). To take two obvious examples, most North Americans automatically categorize each other by race and do not see race as a social construct, even though, to maintain the racial divide, interracial sexual relations and marriage were long forbidden by law; in rural North India, most people do not see castes as social constructs but rather as orders of being, even though what maintains caste are choices to marry within caste and social sanctions against inter-caste marriage. Throughout much of history, race and caste defined relations of power.

The second wave of behavioral economics explicitly brings in culture: it studies how experience and exposure of people in a community lead to the formation of shared categories and the emergence of meanings and cause them to evolve. Research in the second wave has shown that some biases that the first wave suggested were universal are actually cultural. An example is the endowment effect, the tendency of people to value owned items more than nonowned items. Experiments in the United States show that people demand substantially more to give up an object than they would be willing to pay to acquire it. But this result does not hold everywhere. Comparing two groups in Tanzania that are genetically and largely culturally homogeneous, Coren Apicella et al. (2014) show that only the group that was exposed to the concepts of ownership, the selling and buying of goods, and the payment for labor has the endowment effect.

Economists have only recently asked, Which biases are hardwired in the brain, and which ones are learned through sociocultural experience? As of the early 21st century, most experiments in psychology and behavioral economics have been conducted in Western, educated, industrialized, rich, and democratic societies, famously named WEIRD by Henrich et al. (2010). Thus, it is too early to answer this question.

The differences in emphasis between the first and second waves of behavioral economics mirror changes that have occurred in the discipline of psychology. Before 1990, psychology emphasized universals of human thinking; after 1990, a new subfield called cultural psychology emerged, drawing on the new subfield of cognitive sociology. Researchers in these two subfields viewed the evidence as overwhelming that “people think and feel and act in . . . ways that are shaped by particular patterns of historically derived meanings, practices, products and institutions” (DiMaggio & Markus, 2010, p. 348). The second wave of behavioral economics investigates the economic consequences of culturally shaped cognition and preferences. The section in this essay on the second wave will give many examples.

A policy implication of this second wave of research is that what worked “there” may not work “here.” Thus, implementation of new interventions should test several interventions, each based on different assumptions about how people think. The process of refinement should continue after the intervention is scaled up (World Bank, 2015).

Behavioral economics does not offer a new grand theory as an alternative to neoclassical theory. Instead, it offers realism-improving middle-range theories. While an expressed need for midlevel generalizations is recent in economics, it has long propelled work in other social sciences (Merton, 1949). The behavioral sciences are converging on a shared view of the value of middle-range theory (e.g., Deaton, 2010; Elster, 2007; Hedstrom & Ylikoski, 2010; Nunn, 2012, 2020). Such theory does not match the scope and power of rational choice theory but is necessary for understanding many historical changes and providing a framework for policy interventions to address the impediments to economic development that rational choice theory does not explain well. The remainder of this essay provides a deeper look at the applications to economic development of each wave of behavioral economics.

The First Wave of Behavioral Economics

The first wave investigates a decision maker who has bounded cognitive capacity, thinks intuitively, and uses rules of thumb (a thorough overview is Kremer et al., 2019). One important theme is the psychology of poverty. Anxiety over an inability to pay for essential needs and repay debts takes up some of the mental “bandwidth” that a person has. Thus, the situation of poverty tends to lower cognitive function and economic productivity.

In the “hungry” season in rural India, when there is little agricultural work to do and many people run up debts to moneylenders, Supreet Kaur and colleagues set up a firm to run a field experiment to assess the impact on workers’ productivity of variation in the timing of cash income, while holding overall income fixed (Kaur et al., 2021). Workers in the firm used a technology that was simple but required constant attention: By hand, the workers sewed together leaves to make plates that would be used for serving food. A random selection of the workers received earlier payment of 2 to 3 weeks’ worth of earnings. They used much of the funds to reduce their debt overhang. After receiving their early pay, the workers made fewer errors and increased their hourly output compared to the control group. As expected, the effects were larger on the poorer workers.

The depressing effect of poverty on the quality of work is supported by studies that the quality of decision-making declines when individuals are put in a situation of scarcity of time or opportunities (e.g., Shah et al., 2012). The psychological consequences of poverty may produce a feedback loop through which poverty perpetuates itself (Haushofer & Fehr, 2014).

As noted, the heart of the first wave of behavioral economics is research that identifies deviations from rationality and designs interventions to induce rational behavior. Three examples from behavioral development economics are illustrative.

Overcoming a Commitment Problem

Esther Duflo et al. (2008, 2011) documented the low take-up of fertilizer in Kenya, which they estimated had an average return of just under 50%. Since fertilizer had long been used in the survey area, how could the low take-up be reconciled with the large increases in yield from fertilizer? Just as biases limit Westerners’ investment in attractive pension plans, biases also limit profitable investment in fertilizer by farmers in Kenya. Take-up was increased by 47% to 70% by giving farmers at the time of harvest, when they have ready cash, the opportunity to commit to fertilizer purchase. The commitment helped farmers overcome procrastination problems, while minimally distorting the investment decisions of farmers who did not suffer from present bias. Small, time-limited reductions in the cost of buying fertilizer at the time of harvest induced increases in fertilizer use comparable to those induced by much larger price reductions later in the season.

Overcoming Present Bias

Abhijit Banerjee et al. (2010) document the underresponsiveness of parents in India to reliable, free vaccination programs for children. The researchers ran an experiment to increase take-up. Some parents of children under age 2 were given small nonfinancial rewards after each of the four visits to the clinic that were needed to complete the vaccination series that immunizes a child against multiple diseases. In the sparsely populated rural area of India where the experiment was conducted, the small rewards raised the percentage of children who completed the series of vaccines from 18% to 39%. Since a large part of the cost of immunization is a fixed cost of setting up the immunization camp, the intervention greatly reduced the average cost of fully immunizing a child: This cost fell from $56 to $28. Because the 39% proportion of children who were immunized still falls far short of what is needed to achieve herd immunity, Jishnu Das (2010) views the study as “a wake-up call for more experimentation and evaluation on the fundamental design of vaccination programmes” (p. 1258).

Increasing Willpower

Pascaline Dupas and Jonathan Robinson (2013) document that many households in Kenya do not have adequate savings to purchase bed nets and cover emergency medical costs. They find that the mere provision of a lockbox, together with a passbook on which a woman writes down what she wants to purchase with her savings, increases the amount she saves for health needs by 66%. The mechanism appears to be a kind of “mental accounting”: Having mentally committed to save a certain amount enables women to resist temptation and requests for help from friends. Thus, they save much more.

The book entitled Nudge (Thaler & Sunstein, 2008) was the spearhead of the first wave of behavioral economics. A nudge is a policy that changes the way options are presented or designed in order to make it easier for individuals to make the rational choice. One kind of nudge is a commitment device. Since commitment devices may help people who are present biased to save, there might be a demand for commitment devices to increase savings. Nava Ashraf et al. (2006) offered a commitment savings account to a random subset of participants through a bank in the Philippines. Under the commitment, the individuals had restricted access to their savings until they net their savings goal or until a target date. Twenty-eight percent of respondents took up this offer. A year later, the individuals who had been offered commitment savings had 82% higher savings balances than those in the control group.

Another mechanism that has been used to “nudge” people to make rational decisions to save is to make saving for retirement be the default option for employees at a firm. (The default option is the outcome that people get if they make no explicit decision over a set of options.) An intervention of this kind in the United States that gained a lot of attention is Save More Tomorrow. It asks employees to commit to save for retirement a small portion of their earnings, and it links planned increases in savings rates to future pay raises. At the first company that implemented Save More Tomorrow, most of the employees joined, and they ended up quadrupling their saving rate in 4 years (Benartzi & Thaler, 2013). As of this writing, over 15 million employees in the United States participate in Save More Tomorrow.

The Second Wave of Behavioral Economics

The second wave of behavioral economics recognizes that culture influences preferences, perception, and cognition (e.g., Henrich et al., 2001). Individuals do not confront a new situation tabula rasa; they bring to it simple theories about how things work and associations that they have learned. Before individuals can tackle a problem, they must conceptualize or visualize it in their minds; cultural categories, cultural meanings, narratives, and identities may influence how they do that.

To promote progress in understanding how culture affects preferences, perception, and cognition, researchers need a unit of analysis for the cognitive structures that culture shapes. A consensus has emerged among many researchers that widely shared mental representations of reality, called “mental models” or “schemas,” are a useful unit of analysis (e.g., Bartlett, 1932; Bruner, 1990; D’Andrade, 1995; DiMaggio, 1997; Fiske & Taylor, 1984; North, 1994; World Bank, 2015). This essay will use the term mental models. Culture works through “the interaction of shared cognitive structures and supra-individual cultural phenomena . . . that activate those structures to varying degrees” (DiMaggio, 1997, p. 264). Michael Tomasello (2014) refers to shared mental models as a kind of “cooperative cognition and thinking” (p. 6). Mental models are simple enough that a person can hold them in their head. They take a wide variety of forms—including stereotypes, causal narratives, and simple theories of the world.

Whereas it used to be believed that all categories were defined by rules, and later it was thought that they were best understood in terms of prototypes or exemplars, a modern view is that categories are best understood as mental models that are “mini-theories” of the world (see, e.g., Hampton, 1997, p. 135). They are epistemological resources (Douglas, 1986, p. 10). People have an immense repertoire of them (e.g., mental models for what is a “real man,” for how to drive a manual transmission car, for whom to trust, and for how organizations work). While some mental models are the result of personal experience, the focus in behavioral economics is on mental models that are widely shared among the members of a social group, that is, cultural mental models. They may be passed down across generations. They are in part held unconsciously, and so an individual may hold inconsistent mental models. Ann Swidler (2001, p. 36) finds that most Americans in speaking about love express both the romantic view (“a true love is forever”) and the prosaic-realist view (love is fragile and requires hard work to maintain it). Which view a person expresses depends on how a question is asked and/or how the respondent frames a given situation.

A common metaphor for shared mental models is “cultural lens,” since mental models mediate individuals’ experience of the world. Individuals do not see things fully or objectively. Mental models do many things—shape what an individual sees, add things in (that are not there), subtract things (that are there), give meaning, and give causality that may or may not be there. Cultural mental models may also become a part of individuals’ identity and therefore valued as ends. Mental models hold information in memory with labels so that it can be retrieved. The human mind thus constructs perceptions of the world from a combination of sensory matter and mental models.

Douglass North (1994) argued that by distorting perception, mental models may foster the underdevelopment of a subculture or an entire society. He lamented economists’ failure to study the shared mental models that hinder economic development.

Consider, as an example, the problem of learning to cooperate—an essential skill to promote well-being, since many economic interactions are not mediated by markets or controlled by law. Experiments with North American subjects show that fixed pairs of players in a repeated assurance game quickly form the efficient convention. It was believed that this pattern would apply universally. To test that, Brooks et al. (2018) implemented a repeated assurance game in 10 villages in North India. In the final period of a fixed pairing (periods 5 and 10), the proportion of pairs at the cooperative equilibrium was more than two times larger for low-caste than for high-caste pairs: 73% compared to 32%. The apparent obstacle to the high-caste men to form the cooperative convention was that after a coordination failure, the high-caste men who had the loss from coordination failure perceived it as an insult to their honor. After obtaining such a loss, 32% of the players in high-caste pairs did not take the cooperative action in the next round of play, whereas 68% of the players in low-caste pairs did.. This is the only history of play after which there was a statistically significant caste difference. It appears that the high-caste men who suffered a loss were motivated to retaliate against their partners. Retaliation hurt their ability to signal that they valued the cooperative convention and so generally forestalled its emergence. In contrast, low-caste men appeared to perceive coordination failures that caused them losses as mere mistakes. By continuing to take the cooperative action after incurring a loss, they signaled their interest in forming the cooperative convention, which made it likely to emerge.

A survey of men in the villages in which the assurance game was implemented provides supportive evidence of the importance to high-caste but not low-caste men of retaliation for even minor slights: Compared to low-caste respondents, high-caste respondents were much more likely to think it appropriate to beat up an individual who, because of adverse or ambiguous circumstances, had caused one to suffer a loss. Additional supportive evidence in anthropology is that in rural North India, for high-caste but not low-caste men, a “defining characteristic of masculinity has been the concept of revenge” (Chowdhry, 2015).

Extended exposure to social patterns—for example, patriarchy, self-government, or interethnic violence—gives individuals guiding metaphors and the sense of a priori rightness of some ideas and behaviors (Douglas, 1986, p. 10). When members of a group share similar cultural experiences, their beliefs and attitudes can become entrenched, even as available information and technologies change. This creates a gap between beliefs and reality. Fictions can be sustained as truths, hampering social and economic progress (Hoff & Stiglitz, 2010). Individuals and social patterns are mutually constituted and evolve in complementary ways (Markus & Kitayama, 2010).

There are no forces to ensure that cultural mental models promote well-being (Hoff et al., 2022). We next consider four examples in which exposure to particular historical social patterns has had a persistent impact on attitudes and behavior. In the first two cases, it reduced national income. One interpretation (but not the only possible interpretation) is that the social patterns at one point in time shaped mental models that became maladaptive once the environment changed.

An Influence on Female Labor Force Participation: The Plough in Preindustrial Agriculture

Preindustrial agriculture relied primarily on either shifting cultivation or plough cultivation. Unlike the hoe or digging stick used to prepare the soil in shifting cultivation, the plough requires significant upper body strength either to pull it or control the animal that pulls it. In areas that were topographically well suited to use of the plough (i.e., well suited to grow wheat, barley, and rye rather than maize, sorghum, and millet), the plough created gendered occupations: Men specialized in agriculture and women specialized in domestic activities.

Alesina et al. (2013) find evidence that this historical social pattern creates in the modern world a very specific aspect of gender inequality: It lowers female participation in the labor market, entrepreneurship, and national political positions. In regions that are topographically well suited to use of the plough, compared to regions that are not, female labor force participation in the year 2000 was more than 10 percentage points lower. There was also lower participation of women in firm ownership and in national politics. The qualitative results extend to second-generation immigrants in the United States and Europe from historically plough-using regions. This is strong evidence that cultural beliefs and values, based on historically gendered occupations, influence attitudes today of the “natural” or “appropriate” role of women. A possible interpretation is that the historical experience of gendered occupations changed the mental models of gender—the “mini-theory” of the kinds of work that it is appropriate for men and women to do.

An Influence on Mistrust: Slave Exports in the Atlantic Slave Trade

For nearly 500 years, there was a slave trade in Africa. To protect themselves from enslavement, Africans bought guns. To purchase the guns, many kidnapped or tricked others and sold them to slave traders. It is an “historical fact that by the end of the slave trade, it was not uncommon for individuals to be sold into slavery by neighbors, friends, and family members” (Nunn & Wantchekon, 2011, p. 3222). Nathan Nunn and Leonard Wantchekon show that heavy targeting of some ethnic groups by the slave trade in the past 500 years adversely affects the modern-day level of trust in these the ethnic groups in their relatives, neighbors, and local government. This is measured by responses to specific questions in the 2005 Afrobarometer survey. The average level of trust is very low. For example, 18% of all respondents report that they trust their relatives only a little, and only 30% report that they trust their elected local government council a lot (Nunn and Wantchekon, online Appendix Table 12). The heavy targeting of some ethnic groups has also lowered per capita incomes in their countries today (Nunn, 2008). The effect of the slave trade on mistrust arises much more from the level of slave exports from an individual’s own ethnic group than from the level of slave exports in the location in which they currently live (p. 3249). Both channels influence trust, but the slave exports from the individual’s own ethnic groups affect their level of trust at least twice as much. This is supportive evidence that individuals are shaped by the beliefs of their cultural groups. The slave exports from each ethnic group damaged the group’s social fabric and shaped its cultural mental models, and the mental models were passed down across generations. Embedded in the mental models are norms of mistrust.

An Influence on Civic Culture: The Experience of Self-Government in the Middle Ages

In 1970, a national law in Italy created from scratch 15 regional governments with identical formal rules. Many of the new governments in north Italy became strong, responsive, and effective. Most in the south did not. There is evidence that the divergence can be explained by regional differences in the strength of civic-spirited behavior that date back to the late Middle Ages (Guiso et al., 2016; Putnam, 1993). Around the turn of the last millennium, southern regions of Italy were brought under the Norman Empire. In contrast, in north Italy, many cities formed self-governing communes when the Holy Roman Empire weakened. People in municipalities with a history as self-governing communes appear to have a different psychology. Luigi Guiso and colleagues found differences between municipalities with and without a history as city-states on the basis of responses to survey questions posed to students in Italy.

Here is one of the questions: “Your performance is so bad that you have to do the assignment over. How do you explain this?” The set of possible answers includes the following: No one helped me; I was unlucky; it was difficult; I lack ability; and I did not put effort on it. The answer that indicates high self-efficacy is the last one. Individuals with stronger self-efficacy expect their own effort to determine the outcome.

Consistent with the higher self-efficacy scores of pupils in municipalities with a history as self-governing communes, such cities have a greater number of nonprofit organizations, a greater probability of having an organ donation program (a direct measure of individuals’ internalization of the common good and ability to act on it), lower frequency of cheating in a national mathematics exam that fifth graders take, more resistance to the Nazis during World War II, and higher participation of citizens in referenda. The authors suggest that the institutional shock of being a medieval city-state influenced the regions’ psyches (p. 1434) in a way that made it possible for the citizens today to form effective regional governments.

An Influence on Discrimination by Judges: Terrorism in the Neighborhood of the Courts

Moses Shayo and Asaf Zussman (2011, 2017) find a persistent effect of a period of intense Arab-Jewish conflict in Israel in 2000–2004 on discrimination in the aftermath of the violence, 2007–2010. Most of the ethnic violence took the form of suicide bombings by Arabs, rather than violence by the local population. Shayo and Zussman created a measure of ethnic bias from decisions made by judges in Israel’s small claims court in cases that involved either an Arab plaintiff suing a Jewish defendant or a Jewish plaintiff suing an Arab defendant.1

Judges in Israel’s small claims court are assigned to cases in an effectively random process. The litigants present their case, which are typically “fender-bender” accidents. The judge alone makes the decision. This is a situation where an effectively randomly assigned judge allocates money between two individuals, one of whom belongs to their ethnic group, while the other one does not.

Shayo and Zussman find little ingroup bias in small claims court before the outbreak of political violence in the neighborhood of the court. But during and after a period of intense conflict in the neighborhood of the court, judges from each ethnic group (Jewish and Arab) were more likely to make judgments in a way that took into account the welfare of members of their ethnic group. Jewish judges were 15 to 20 percentage points less likely than Arab judges to accept claims filed by Arab plaintiffs. The persistence of discrimination was not due to the judges’ personal exposure to violence but rather to violence within the area of the court as many as several years before. Thus, what elicited the discrimination and likely made it salient was the common knowledge of the events and the culture of the court.

Policies to Change Maladaptive Mental Models

As Glenn Loury (2002, p. 37) argues, if a representation of a group is a human product (i.e., a social construct), then should not humans be able to control it? Deviations from reality of a mental model are, in fact, an entry point for policy to change behavior. Recent research suggests that coaching, collaborative intergroup contact, and affirmative action are among the tools through which interventions can make it possible for new mental models to emerge.


In many cases, boys growing up in violent neighborhoods learn to “act tough” when others assert authority in order to develop a reputation that deters potential predators. Having learned this strategy, males may automatically respond with aggression to all assertions of authority. In places where social control is strong, such a response is not adaptive: in particular, it leads to high rates of suspensions from high school, high rates of avoidable violence, and high risk of incarceration.

Interventions with at-risk, disadvantaged young men in high schools and prisons in the Chicago area helped participants reflect on whether aggression was the best response in a wide variety of situations (Heller et al., 2017). Coaches presented participants with many situations and, in each case, asked participants to consider a variety of interpretations of the situation and to practice ways to interact. In one exercise, called “the Fist,” participants were divided into pairs. One member of each pair was given a ball, and the other was told he had 30 seconds to get the ball from his partner. Many boys took the ball by force from their partners. The counselor then asked questions to show that their partners would have been willing to give up the ball if they had been asked calmly to do that. A typical response was, “I would have given it: it’s just a stupid ball” (p. 3). In this and many other exercises, the participants practiced overcoming automatic “fast” thinking with “slow” reflection. They were never told that aggression was wrong but rather that situations varied and aggression was not always an appropriate way to interact. The coaching programs substantially decreased suspensions from high school, increased high school graduation rates, and reduced violent-crime arrests. For the participants in prison, the programs decreased recidivism.

Sara Heller and colleagues have limited information on the mechanisms driving the change in behavior. Their preferred interpretation is that the participants learned to “think slow.” This interpretation falls within the first wave of behavioral economics. But as Kahneman (2011) emphasizes, to think slow is hard to maintain in the uncountably many social interactions people have in their everyday lives. A more plausible interpretation may be that through practice, the participants learned a new mental model, and it was activated in situations where social control was strong. Participants learned to make a distinction between dangerous situations, where aggressive responses were appropriate, and safe situations, where aggression was not appropriate. To use the language that describes mental models, participants gained from the activities that they did in the intervention an epistemological resource, and it mediated their experience of the world. It led them automatically to respond cooperatively in settings that they saw were probably safe. Instead of learning to think “slow,” they learned a new way to think “fast.”

A related intervention, evaluated by Blattman et al. (2017, 2022), coached criminally engaged men in Liberia in ways that led them to reduce reliance on harmful, automatic patterns of thinking linked to their criminal identity. They developed a different representation of themselves (a different self-schema). The intervention reduced the participants’ criminal and violent activities. Based on self-reports, the effect persisted 10 years later.

Collaborative Intergroup Contact

In India, individuals in the lowest castes were historically called “untouchables.” They were represented as innately polluted. At independence, India abolished untouchability, but the “pollution barrier” between former untouchable castes and all others remains strong in North India today. This barrier depresses the performance of the low caste in school (Hoff & Pandey, 2006, 2014). This barrier also causes gains that could be realized from trade between high-caste and low-caste individuals to be foregone: Anderson (2011) finds that low-caste farmers have agricultural yields 45% higher if they reside in a village where the majority of groundwater sellers are of low caste rather than of high caste, since the latter generally refuse to sell water to low-caste farmers. The obstacles to low castes to develop a positive group identity may explain the much lower willingness of low-caste than high-caste individuals to punish norm violators who hurt members of their own subcaste; this makes cooperation within a low compared to a high subcaste more difficult to sustain (Hoff et al., 2011). Social punishment is believed to play a large role in the enforcement of contract and property rights and, thus, in economic development.

Gordon Allport (1954, p. 276, cited in Lowe, 2021) hypothesized that intergroup collaborative contacts, in which people do things together for a common goal, changes attitudes that the groups had about themselves and the other groups. He argued that “the principle is clearly illustrated in the multi-ethnic athletic team . . . It is the cooperative striving for the goal that engenders solidarity.” Matt Lowe (2021) tested the hypothesis in a field experiment in North India; he recruited volunteers from young men in mixed-caste villages to play cricket. He created cricket leagues of teams that were either mixed low and high caste or only high caste or only low caste. The mixed-caste teams were the treatment group (with intergroupcollaboration) and the single-caste-status teams were the control group (without intergroup collaboration). Each team played eight games over the course of the month-long experiment. Lowe shows that collaborative contact (playing on a team with members from both high and low castes) reduces caste prejudice in young men, while competitive contact (playing against others of a different caste status) increases caste prejudice. Having all other-caste teammates instead of none increased by one the number of friendships between high- and low-caste men in the cricket leagues. Having most rather than few opponents in cricket matches be of the other caste status decreases, on average, by 3.5 the number of other-caste-status friends that a participant has in the league.

After the cricket matches ended, Lowe implemented a second field experiment with the same participants. He gave each person two left flip-flops or two right flip-flops and two left-handed gloves or two right-handed gloves. There was a monetary incentive to engage in trade with other participants in order to get correctly matched flip-flops and gloves. Being on a team with both low- and high-caste individuals increased by up to 21% the participation in trade between low- and high-caste individuals. Thus, Lowe’s experiments strongly bear out Alport’s hypothesis that intergroup collaborative contact reduces prejudice.

Supporting evidence are the results of the natural experiment in New Delhi evaluated by Gautam Rao (2019). He shows that the integration of rich and poor children in schools from the age of 4 makes the rich children more inequity averse and less discriminatory toward the poor. The change in rich students’ preferences occurs only if they have regular contact with poor students in a study group, to which students were as-good-as-randomly assigned. No effect on rich students’ prejudice against the poor arises merely from being in a classroom of which 20% of the students are poor.

Affirmative Action and Other Strategies for Changing the Status of Groups

Despite formalequality across men and women in political rights in most countries, women are underrepresented in positions of political leadership. To address this, many countries have adopted affirmative action quotas for women in certain elected positions. In one case—India—reservations in the elected position of village chief for women were randomly assigned to local governments. This makes it possible to causally assess the impact.

A study in the state of West Bengal, India, found that 7 years of experience of living under a woman village chief erased male villagers’ prejudice against women leaders by many measures—an Implicit Association Test, the evaluation of political speeches in the Goldberg paradigm, and the assessment of the quality of actual female village chiefs (Beaman et al., 2009). In villages that had been exposed to women village chiefs for at least 7 years, parents’ aspirations for their teenage daughters and the daughters’ aspirations for themselves were higher, and girls went to school somewhat longer and did somewhat fewer hours of housework (Beaman et al., 2012). Affirmative action for women in the local village councils made female victims of crime more willing to report the crimes to the police and made the police more willing to record the reports, even though the local village council has no formal authority over the police (Iyer et al., 2012).

All these changes in attitudes and behavior suggest that the creation of exemplars of women in high-status political positions changed widely shared mental models of women. Supportive evidence comes from a field experiment where the employment, for the first time, of women in call center jobs (an average of just three new employees per village) increased the care that parents gave to their daughters, as measured by the daughters’ and sons’ body mass index (Jensen, 2012).

But success in using affirmative action to change attitudes towards a disadvantaged group depends on how the dominant group responds. In contrast to the positive effects in India of affirmative action for women in village government, there is no evidence of such effects in the case of affirmative action for the formerly untouchable castes. These reservations did not reduce the practice of untouchability after the period of reservation for the village chief had ended (Girard, 2018). In many districts of India, affirmative action for the low castes induced a backlash. In the public schools, on which low-caste children disproportionately depend, there was increased absenteeism of high-caste teachers. The high caste teachers reduced their teaching effort, and there was greater embezzlement of scholarships intended for low-caste students (Pandey & Pandey, 2022). The process of debiasing perceptions and improving interactions across a social cleavage does not follow automatically from affirmative action. There can be a powerful backlash.


To acquire knowledge about behavior, neoclassical economics gives precedence to theoretical investigation based on a rational actor model of the decision maker. In contrast, behavioral economics gives precedence to experimental and empirical investigation and makes no assumptions about rationality. Both types of investigation are necessary. We have limited knowledge about how the mind works and how people make decisions, and so we need experiments and empirical investigation of real-life economic behavior. But an experiment is about only one situation in one place at one time. Without theory, its results cannot be generalized, and the mechanisms at play cannot be understood (Deaton, 2010).

Many discoveries in development economics in recent decades were made possible by experiments whose outcomes were subsequently analyzed with midlevel theory. The contribution of randomized controlled trials (RCTs) to reducing poverty was acknowledged in the award of the Nobel Prize in Economic Sciences to Abhijit Banerjee, Esther Duflo, and Michael Kremer in 2019. RCTs made possible the testing and scaling up of interventions that increased well-being. Some of these interventions used incentives—the central lever of policy in neoclassical economics. But many others were in the new field of behavioral development economics—the interventions made it easier for people to make the rational decision, or reframed situations in a way that changed people’s preferences.

Work that examines the impact of culture on how people think and what they want extends the first wave of behavioral economics. In the second wave of behavioral economics, researchers have measured shared beliefs, perceptions, and preferences that differ across communities and countries. For example, responses in Afrobarometer surveys make possible measures of trust across Africa, and Italian national surveys make possible measures of self-efficacy across municipalities in Italy. This second wave of behavioral economics shows that experience and exposure shape minds and the effects can be very durable. It also demonstrates new policy levers that can sometimes change maladaptive behavior: in particular, interventions that give individuals new experiences may lead to the emergence of new mental models that change how people see the world and, thus, how they behave. The two waves of behavioral economics have enriched economists’ understanding of why some groups and some regions of the world fall behind and have enlarged the set of policy tools to help lagging regions and groups catch up.

Further Reading

  • Acharya, A., Blackwell, M., & Sen, M. (2020). Deep roots: How slavery still shapes southern politics. Princeton University Press.
  • Akerlof, G. A., & Snower, D. J. (2016). “Bread and bullets.” Journal of Economic Behavior & Organization, 126(Part B), 58–71.
  • Demeritt, A., & Hoff, K. (2018). “The making of behavioral development economics.” History of Political Economy, 50(2018), 303–322.
  • Drexler, A., Fischer, G., & Schoar, A. (2014). “Keeping it simple: Financial literacy and rules of thumb.” American Economic Journal: Applied Economics, 6(2), 1–31.
  • Hoff, K., & Walsh, J. (2018). “The whys of social exclusion: Insights from behavioral economics. World Bank Research Observer, 33(1), 1–33.
  • Hochschild, A. R. (2016). Strangers in their own land: Anger and mourning on the American right. The New Press.
  • Jha, S., & Shayo, M. (2019). Valuing peace: The effects of financial market exposure on votes and political attitudes. Econometrica, 87(5), 1561–1588.


  • Alesina, A., Giuliano, P., & Nunn, N. (2013). On the origins of gender roles: Women and the plough. Quarterly Journal of Economics, 128(2), 469–530.
  • Allport, G. (1954). The nature of prejudice. Addison-Wesley.
  • Anderson, S. (2011). Caste as an impediment to trade. American Economic Journal: Applied Economics, 3(1), 239–263.
  • Apicella, C. L., Azevedo, E. M., Christakis, N. A., & Fowler, J. H. (2014). Evolutionary origins of the endowment effect: Evidence from hunter-gatherers. American Economic Review, 104(6), 1793–1805.
  • Ashraf, N., Karlan, D., & Yin, W. (2006). Tying Odysseus to the mast: Evidence from a commitment savings product in the Philippines. Quarterly Journal of Economics, 121(2), 635–672.
  • Banerjee, A., Duflo, E., Glennerster, R., & Kothari, D. (2010). Improving immunization coverage in rural India. British Medical Journal, 340(7759), 1291.
  • Bartlett, F. C. (1932). Remembering: A study in experimental and social psychology. Cambridge University Press.
  • Beaman, L., Chattopadhyay, R., Duflo, E., Pande, R., & Topalova, P. (2009). Powerful women: Does exposure reduce bias? Quarterly Journal of Economics, 124(4), 1497–1540.
  • Beaman, L., Duflo, E., Pande, R., & Topalova, P. (2012). Female leadership raises aspirations and educational attainment for girls: A policy experiment in India. Science, 335(6068), 582–586.
  • Benartzi, S., & Thaler, R. (2013). Behavioral economics and the retirement savings crisis. Science, 339, 1152–1153.
  • Bernard, T., Dercon, S., Orkin, K., & Taffesse, A. S. (2019). Parental aspirations for children’s education: Is there a ‘girl effect’? Experimental evidence from rural Ethiopia. American Economic Review, 109, 127–132.
  • Blattman, C., Chaskel, S., Jamison, J., & Sheridan, M. (2022). Cognitive behavior therapy reduces crime and violence over 10 years: Experimental evidence [NBER working paper 30049].
  • Blattman, C., Jamison, J. C., & Sheridan, M. (2017). Reducing crime and violence: Experimental evidence from cognitive behavioral therapy in Liberia. American Economic Review, 107(4), 1165–1206.
  • Brooks, B., Hoff, K., & Pandey, P. (2018). High‐caste culture in North India impedes learning to cooperate. Proceedings of the National Academy of Sciences of the United States of America, 115(45), 11385–11392.
  • Bruner, J. (1990). Acts of meaning. Harvard University Press.
  • Camerer, C. (2005). Comment on “Development economics through the lens of psychology” by Sendhil Mullainathan. In Lessons of experience: Annual World Bank conference on development economics 2005 (pp. 71–78). The World Bank
  • Chowdhry, P. (2015). Popular perceptions of masculinity in rural North Indian oral traditions. Asian Ethnology, 74, 5–36.
  • D’Andrade, R. G. (1995). The development of cognitive anthropology. Cambridge University Press.
  • Das, J. (2010). Improving immunization coverage in rural India: Incentives help, but not nearly enough. British Medical Journal, 340(7759), 1257–1258.
  • Deaton, A. (2010). Instruments, randomization, and learning about development. Journal of Economic Literature, 48(2), 424–455.
  • Demeritt, A., & Hoff, K. (2016). Small miracles—Behavioral insights to improve development policy: World development report 2015. In T. Besley (Ed.), Contemporary issues in development economics (pp. 19–43). Palgrave Macmillan.
  • DiMaggio, P. (1997). Culture and cognition. Annual Review of Sociology, 23, 263–287.
  • DiMaggio, P., & Markus, H. (2010). Culture and social psychology: Converging perspectives. Social Psychology Quarterly, 73(4), 347–352.
  • Douglas, M. (1986). How institutions think. Syracuse University Press.
  • Duflo, E., Kremer, M., & Robinson, J. (2008). How high are rates of return to fertilizer? Evidence from field experiments in Kenya. American Economic Review, 98(2), 482–488.
  • Duflo, E., Kremer, M., & Robinson, J. (2011). Nudging farmers to use fertilizer: Theory and experiments evidence from Kenya. American Economic Review, 10(6), 2350–2390.
  • Dupas, P., & Robinson, J. (2013). Why don’t the poor save more? Evidence from health savings experiments. American Economic Review, 103(4), 1138–1171.
  • Elster, J. (2007). Explaining social behavior. Cambridge University Press.
  • Fehr, E., & Hoff, K. (2011). Introduction—Tastes, castes and culture: The influence of society on preferences. Economic Journal, 121(556), F396–F412.
  • Fiske, S. T., & Taylor, S. E. (1984). Social cognition. McGraw-Hill.
  • Guiso, L., Sapienza, P., & Gonzales, L. (2016). Long‐term persistence. Journal of the European Economic Association, 14(6), 1401–1436.
  • Hampton, J. (1997). Conceptual combination. In K. Lamberts & D. Shanks (Eds.), Knowledge, concepts, and categories (pp. 133–159). Psychology Press.
  • Haushofer, J., & Fehr, E. (2014). On the psychology of poverty. Science, 344(862), 862–867.
  • Hedstrom, P., & Ylikoski, P. (2010). Causal mechanisms in the social sciences. Annual Review of Sociology, 36, 49–67.
  • Heller, S. B., Shah, A. K., Guryan, J., Ludwig, J., Mullainathan, S., & Pollack, H. A. (2017). Thinking, fast and slow? Some field experiments to reduce crime and dropout in Chicago. Quarterly Journal of Economics, 132(1), 1–54.
  • Henrich, J., Boyd, R., Bowles, S., Camerer, C., Fehr, E., Gintis, H., & McElreath, R. (2001). In search of homo economicus: Behavioral experiments in 15 small-scale societies. American Economic Review, 91(2), 73–78.
  • Henrich, J., Heine, S. J., & Norenzayan, A. (2010). The weirdest people in the world? Brain and Behavioral Science, 33(2–3), 61–83.
  • Hoff, K., Demeritt, A., & Stiglitz, J. E. (2022). The other invisible hand: The power of culture to either enable or stymie progress. Manuscript in progress for Columbia University Press.
  • Hoff, K., Kshetramade, M., & Fehr, E. (2011). Caste and punishment—The legacy of caste culture in norm enforcement. Economic Journal, 121, F449–F475.
  • Hoff, K. & Pandey, P. (2006). Discrimination, social identity, and durable inequalities. American Economic Review, 96(2), 206–211.
  • Hoff, K., & Pandey, P. (2014). Making up people: The effect of identity on performance in a modernizing society. Journal of Development Economics, 106, 118–131.
  • Hoff, K., & Stiglitz, J. E. (2010). Equilibrium fictions: A cognitive approach to societal rigidity. American Economic Review, 100, 141–146.
  • Hoff, K., & Stiglitz, J. E. (2016). Striving for balance in economics: Towards a theory of the social determination of behavior. Journal of Economic Behavior and Organization, 126(B), 25–57.
  • Iyer, L., Mani, A., Mishra, P., & Topalova, P. (2012). The power of political voice: Women's political representation and crime in India. American Economic Journal: Applied Economics, 4(4), 65–193.
  • Jensen, R. (2012). Do labor market opportunities affect young women's work and family decisions? Experimental evidence from India. Quarterly Journal of Economics, 127(2), 753–792.
  • Kahneman, D. (2011). Thinking, fast and slow. Farrar, Straus, and Giroux.
  • Kahneman, D., & Tversky, A. (1979). Prospect theory: An analysis of decision under risk. Econometrica, 47, 263–291.
  • Kahneman, D., & Tversky, A. (1984). Choices, values, and frames. American Psychologist, 39(4), 341–350.
  • Kaur, S., Mullainathan, S., Oh, S., & Schilbach, F. (2021). “Do financial concerns make workers less productive? Randomized Social Experiments eJournal.
  • Kremer, M., Miguel, E., Mullainathan, S., Null, C., & Zwane, A. P. (2011). Social engineering: Evidence from a suite of take-up experiments in Kenya. Unpublished Working Paper.
  • Kremer, M., Rao, G., & Schilbach, F. (2019). Behavioral development economics. In B. D. Bernheim, S. DellaVigna, & D. Laibson (Eds.), Handbook of behavioral economics, (pp. 345–458). North Holland.
  • Loury, G. (2002). The anatomy of racial inequality. Harvard University Press.
  • Lowe, M. (2021). Types of contact: A field experiment on collaborative and adversarial caste integration. American Economic Review, 111(6), 1807–1844.
  • Malt, B. C. (1995). Category coherence in cross-cultural perspective. Cognitive Psychology, 29, 85–148.
  • Markus, H., & Kitayama, S. (2010). Cultures and selves: A cycle of mutual constitution. Perspectives on Psychological Science, 5(4), 420–430.
  • Merton, R. (1949). Social theory and social structure. Free Press.
  • North, D. C. (1994). Economic performance through time. American Economic Review, 84(3), 359–368.
  • Nunn, N. (2008). The long-term effects of Africa’s slave trades. Quarterly Journal of Economics, 123(1), 139–176.
  • Nunn, N. (2012). Culture and the historical process. Economic History of Developing Regions, 27(Suppl. 1), 108–126.
  • Nunn, N. (2020). The historical roots of economic development. Science, 367(6485), 1441.
  • Nunn, N., & Wantchekon, L. (2011). The slave trade and the origins of mistrust in Africa. American Economic Review, 101, 3221–3252.
  • Pandey, P., & Pandey, S. (2022). Political Reservations and Service Delivery in Village Government. Economic & Political Weekly, vol. LVII, no. 28, (48–56).
  • Putnam, R. (1993). Making democracy work: Civic traditions in modern Italy. Princeton University Press.
  • Rao, G. (2019). Familiarity does not breed contempt: Generosity, discrimination, and diversity in Delhi schools. American Economic Review, 109(3), 774–809.
  • Shah, A., Mullainathan, S., & Shafir, E. (2012). Some consequences of having too little. Science, 338(6107), 682–685.
  • Shayo, M., & Zussman, A. (2011). Judicial ingroup bias in the shadow of terrorism. Quarterly Journal of Economics, 126(3), 1447–1484.
  • Shayo, M., & Zussman, A. (2017). Conflict and the persistence of ethnic bias. American Economic Journal: Applied Economics, 9(4), 137–165.
  • Stigler, G. J., & Becker, G. (1977). De gustibus non Est disputandum. American Economic Review, 67(2), 76–90.
  • Swidler, A. (2001). Talk of love: How culture matters. University of Chicago Press.
  • Thaler, R., & Sunstein, C. (2008). Nudge. Penguin Books.
  • Tomasello, M. (2014). A natural history of human thinking. Harvard University Press.
  • World Bank. (2015). World development report: Mind, society and behavior.


  • 1. The measure of bias is how much Jewish judges are more likely than their Arab colleagues to accept a claim filed by a Jewish plaintiff rather than by an Arab one. It is positive if there is bias by Jewish or Arab judges, or by both.