POVERTY, AGENCY, AND DEVELOPMENT

Abstract

This essay provides an account of four interrelated ideas. First, a person who is not poor by the standard conception of poverty can still be functionally poor. Second, poverty is a relationship between the poor and their environment (community, local markets, and local institutions). Third, poverty is a determinant of agency and impedes its exercise. Fourth, promoting agency promotes development. I conclude that agency is central to understanding both poverty and development.

I. Introduction

Consider the lives of a man and woman, Ernest and Hope, as once described to me. ¹ Ernest grew up in the only concrete house in his small village. His father was a jurist and his mother a housewife. They owned a bit of farmland and had high social standing in the local community. Ernest’s two older brothers graduated from high school, got married, and left the village to explore and build their lives. Ernest was attending a local college, living with his parents, and the future possibilities for him looked bright. However, unexpectedly, everything turned topsy-turvy. His father suddenly died due to cardiac arrest and his mother followed three months later. During those days, the medium of long-distance communication was limited to postal mails and it took over two months for Ernest to exchange letters with his brothers. Having dropped out of college, he spent the next year or so aimless, ultimately finding comfort at a shrine three miles away, where he befriended several families who became his support system. There, Ernest met the woman he married, despite some intercaste barriers. ² Then it was time for him to figure out his livelihood. The oldest brother, who had naturally assumed the role of the patriarch in the family, visited the family home with his wife and children and decided that Ernest needed to stay in the village to look after home affairs, their two families, and farming. It was a good plan. The youngest brother had the responsibility for the home while the oldest provided outside income. Land ownership, living in a concrete house, having a source of off-farm support, and some college-level education gave Ernest the profile of an honorable and wise man. In those days, a family’s prestige and codes of conduct were more intergenerationally transferable than they are today. Ernest benefited from his deceased father’s standing in the community.

As Ernest described it, everything was going according to plan, and he had even higher aspirations for his sons. In hindsight, though, the plan was faulty. One day Ernest found himself very ill and soon learned that he had a chronic illness, which ruled out any possibility of him contemplating another career. Soon after, the arrangement with his brother fell apart. Their two families could no longer live as one. The timing could not have been worse. Ernest’s children were growing up and had increasing financial needs. Unable to work, Ernest had no source of cash income, which was necessary for farming and household needs. He started relying on his social capital, borrowing money from relatives and friends, but his ability to do that quickly dried up. Ernest could have eased his financial troubles either by lowering the educational aspirations that he had for his sons or selling parcels of his share of farmland. Determined man that he was, Ernest was unwilling to do either. He saw the redemption of his own perceived failures in his sons’ success and he was unwilling to sell any parcels of farmland because that would tarnish the prestige of his late parents and his brothers as well as his own public image. There was another reason for his unwillingness: Ernest did not have well-defined property rights over his share of the family farmland. ³

As his challenges were turning into constraints, they were also becoming obvious to others. Once a source of wisdom for community members, Ernest became a subject of ridicule. Ernest did not say so, but he had lost his own self-esteem and dignity, which was reflected in his avoidance of others. Ernest lived a good part of his life in struggle.

Now consider the life of Hope. Her parents had a bit of land, but they had to marry three daughters and provide for the education of three sons. Hope learned to read and write at home and had no clear aspirations beyond those that her parents had for her. By her own account, at the age of twenty or so, Hope “married up,” which came with an improved housing situation, a caring husband, and a future family to look forward to. By her fifteenth wedding anniversary, Hope was a mother of two daughters and four sons. She had no direct interactions with the local community, markets, or institutions, and rarely stepped outside of her marital home. When she did, it was for a medical reason or to visit relatives. For the latter, her husband’s consent and cooperation were necessary. A day was not long enough for Hope to do her work, starting and ending with cleaning, washing, cooking, and feeding the family in between. By custom, the women of the house ate after the men had their meals and Hope never called her husband by his name.

Everything seemed to be going fine until her husband’s health started deteriorating. He came to suffer frequent physical pain. Meeting the growing needs of a large family became a daily challenge. The scope of doing something about it was limited for Hope. The possibility of an outside job existed in theory, but it was unavailable to Hope because it was a norm that the women of her husband’s family did not work outside the home. Hope thus became entrepreneurial at home. Curtailing basic needs, economizing on cooking ingredients, reinforcing her daily contribution to the joint household chores, and raising kids to sell them as mature goats were Hope’s way of helping her husband. Hope had high aspirations for her sons and she believed in the power of prayer.

Ernest and Hope were a couple. Their older daughter, the first-born child, attended primary school while the younger daughter graduated from high school. Both are housewives with children. Their oldest son graduated from a university, serving as a role model for his younger brothers and guiding them successfully through universities.

There is clearly a lot to be read into and learned from the lives of Ernest and Hope. The lessons I draw from their story are about the nature and dynamics of poverty, agency, and development. ⁴ Ernest did not end up as poor in any standard sense. Rather, Ernest became “functionally poor,” which I define as the state in which a person exhibits the characteristics of the economic and psychological lives of the poor. Ernest became unable to exercise his agency, with important consequences, including failure to meet the basic nutritional requirement to be determined as nonpoor. Hope did not end up as a poor woman either, but she had poor agency and no ability or scope for exercising the agency she had. Each lived life is a unique journey, yet their lives resonate with more than one billion people around the world who live on under $2 a day.

To be more precise about respects in which their lives are impoverished, the following elements are required. First, we need to consider poverty in terms of the relationship between the poor and the environment in which they live and function. For some, all it takes to have a road out of poverty is a modicum of hope, luck, aspiration, and earnest effort. Understanding persistent poverty, though, requires an understanding of a person’s relationship with his environment and what it means to be functionally poor. Second, we need to consider how poverty factors into the determination of agency and its exercise. Ultimately, however, we need to recognize the primacy of agency in the durable escape from poverty and for development.

The remainder of this essay is organized as follows. In Section II, I briefly describe the evolution of thinking on poverty and draw upon recent empirical studies to characterize the economic and psychological lives of the poor. Drawing on Amartya Sen’s entitlement approach to starvation and famines, I argue that a person who is not poor (either absolutely or relatively) can be functionally poor. In Section III, I distinguish between agency’s constituents and its exercise. In Section IV, using insights from empirical literature in economics and psychology, I discuss the role of poverty both as a determinant of agency and as impediment to exercising agency. In Section V, drawing on literature in political science and economics, I provide empirical support for the thesis that increasing individual (collective) agency promotes individual (collective) development, and then offer some conclusions in Section VI.

II. Poverty

In his classic, Poverty and Famines: An Essay on Entitlement and Deprivation, Amartya Sen describes how to conceptualize poverty:

The first requirement of the concept of poverty is of a criterion as to who should be the focus of our concern. The specification of certain ‘consumption norms’, or of a ‘poverty line’, may do part of the job: ‘the poor’ are those people whose consumption standards fall short of the norms, or whose incomes lie below that line. But this leads to a further question: is the concept of poverty to be related to the interests of: (1) only the poor, (2) only the non-poor, or (3) both the poor and the non-poor. ⁵

Sen dismisses the idea that the concept of poverty should be concerned only with the nonpoor. He finds the last alternative somewhat appealing because it is broad-based and unrestrictive, given that the penury of the poor affects the well-being of the nonpoor. Ultimately, however, Sen argues for a conception of poverty based on the first option while emphasizing that its adoption does not deny that the suffering of the poor themselves may depend on the condition of the nonpoor. Sen is right that, at any given point, to determine whether a person is poor or not, one must look at his or her interests. However, I shall argue that to understand persistent poverty and functionally poor requires us to consider what influences the well-being of the poor. Such influences include the local community (which includes the nonpoor), markets, and institutions.

In the context of a concept of poverty that relates only to the interests of the poor, there has been a significant shift in how poverty is defined. ⁶ On the biological approach, poverty is defined in terms of the sufficiency of total earnings to obtain the minimum necessities to maintain physical efficiency. ⁷ However, this approach has been criticized on three interrelated grounds: (1) variations in the physical features of people, the nature of their work, and the conditions under which they work; (2) the conversion of minimum nutritional requirements into minimum food requirements depends on what people consume; and (3) the difficulty of coming up with an acceptable definition of minimum requirements for nonfood consumption. ⁸ While Sen is sympathetic to these criticisms, he finds some merit in the biological approach to poverty. According to Sen, the concept of poverty need not be sharp.

The relative-deprivation approach defines poverty in terms of deprivation, ⁹ which raises two questions. First, is poverty about conditions or feelings of deprivation? While some have argued that it is about the conditions, ¹⁰ others do not see conditions and feelings of deprivation as independent of each other. ¹¹ Second, how does one determine the appropriate reference group, for example, people living in the same region, city, county, or country?

Widely used poverty measures do not reflect the timing and frequency dimensions of poverty. Official measures of poverty are based on household data that is normally collected in a census or nationally representative sample surveys. A household may be estimated to be nonpoor over the time period measured in the census or survey and yet be poor for several months within that time period. Given the tendency of the poor to fall sick frequently or to feel weak, their income over the course of a year is volatile. As a result, they could be poor for the better part of the year but be identified as nonpoor because their poverty status is based on whether their annual income exceeds the poverty line. ¹²

An important dimension of well-being is that a person might be vulnerable to poverty, so various measures of such vulnerability at a household, regional, and national level have been proposed. ¹³ However, the widely used measures of poverty are based on annual income or consumption. They do not factor in how many times a person falls below the poverty line during a given year.

Income- (or consumption-) based measures of poverty, however, are inadequate because poverty is multidimensional; it entails poor health, lack of education, inadequate living standards, poor quality of work, the threat of violence, and living in areas that are environmentally hazardous, among other things. ¹⁴ Reconsider the lives of Ernest and Hope. By their own accounts, they began their married life living in the only concrete house in their village. Ernest had inherited a bit of farmland and he had some college-level education. However, as they recounted to me, having three meals a day for them was difficult and often the women of their household did not have enough to eat. If one were to look at food and nonfood aspects of poverty and their minimum requirements, Ernest and Hope’s household was likely not deprived in any meaningful sense of what ought to be considered the minimum required for nonfood aspects of poverty. Any reasonable aggregation of their food and nonfood consumption (such as housing) would designate them as nonpoor. Were they nonpoor, especially by the relative-deprivation approach to poverty (relative to their village)? Should it matter that they were living in the best house in the village and owned a bit of land? They were, by conventional standards, nonpoor, yet they were effectively economically poor.

A. The economic and psychological lives of the poor

Table 1 summarizes the economic and psychological lives of the poor.¹⁵ The economic lives of the poor fit some of their typical descriptions: they have large families, do not eat enough, fall sick frequently, do not save, and invest little in education. On the other hand, some of the characteristics are contrary to commonly held beliefs about them: they are entrepreneurial, do not save for lack of access to good financial institutions, do not spend every penny on food because they like to enjoy things that others do (such as festivals and weddings), and seek new opportunities such as migration to cities. However, their migration to cities is temporary because they prefer social ties over money and rural communities provide informal insurance that the cities do not.¹⁶

Table 1. The Economic and Psychological Lives of the Poor

Until a decade ago, it was not mainstream in economics to study the psychological lives of the poor, which is not surprising, given the widely held belief that the poor and nonpoor differ only in economic resources and social capital. We now know that the poor are subject to biases and judgment errors not because they are lazy and unthinking, but because they are preoccupied with hunger, managing sporadic income, and juggling expenses. Thinking and fretting about money taxes their mental bandwidth. ¹⁷

B. Functionally poor

I define a person to be functionally poor if his states fit the economic and psychological lives of the poor described in Table 1. Those who are not poor in terms of the standard definition may experience many of the conditions of poverty, including failing to meet the nutritional requirement. ¹⁸ This can happen for at least three reasons. First, the person can suffer from the psychology associated with poverty. Second, he can experience exchange-entitlement failure. ¹⁹ Third, the person may not be able to exercise his agency and use his resources to meet the requirement to be above the poverty line, but this leads to further questions: Does the lack of agency or inability to exercise agency make a nonpoor person poor? Are poverty and agency inextricably linked?

Clearly, the lack of agency or inability to exercise agency does not make a nonpoor person poor. However, such a lack can render him functionally poor. The ability to get what he wants, such as meeting nutritional needs, depends on his ability to translate resources into consumption, which is mediated by features of his context. Consider the predicament of Britney Spears. ²⁰ She is a wealthy woman, but a conservatorship restricted her agency. Did it render her poor? The answer is straightforwardly “No.” However, if the conservatorship, beyond restricting agency, had caused her to exhibit the economic and psychological features associated with poverty, I argue that she would have been rendered functionally poor. In her case, the capacity to translate resources into agency was blocked by a feature of her context, namely, the American legal system.

The link between poverty and agency is nonlinear. They are inextricably linked at the bottom. If all a person can think about is how to pay immediate bills, he is less likely to make good decisions (see Table 1). Thus, at the lowest levels, poverty and agency go together, where poverty impairs a person’s decision-making. Consequently, escaping poverty is difficult for him.

Ernest’s inability to meet the needs of his family, his chronic illness and physical pain, and thinking and fretting about money taxed his mental bandwidth. In addition, the lack of well-defined property rights over his share of land, compounded by regard for his family’s prestige and honor and his perceived loss of dignity, both impaired and impeded his agency, such as an inability to sell his land, which rendered him functionally poor. Ernest was thus functionally poor even though he was not economically poor.

Whether a person is poor in terms of the standard definition or functionally poor has important policy implications. If the former, then multifaceted anti-poverty programs, involving both policy-level and agency-level interventions, are more likely to be effective. ²¹ If it is the latter, then policy-level interventions, such as property rights and access to formal insurance, may suffice.

C. Poverty is a relationship

Poverty is a relationship between the poor and their environment, that is, the local community, markets, and institutions. To understand poverty, we thus need to approach it as the outcome of a process that binds the poor with their environment.

In economics, a strand of literature on “poverty traps” recognizes that some of the drivers of poverty are external to the poor. Incomplete markets, where the supply of credit is insufficient to meet demand, can lead to a poverty trap. ²² Poverty can beget poverty, if the educational investment in children is sufficiently sensitive to parental income. ²³ Coordination problems, such as when agents fail to coordinate their actions to achieve an optimal allocation of resources, can also lead to a poverty trap. ²⁴ Persistent corruption can lead to a poverty trap. ²⁵ The kin system, an informal system of shared rights and obligations among the extended family for the purpose of mutual assistance, can exacerbate moral hazard problems. ²⁶ The failure to aspire to one’s own potential is also a potential explanation of poverty traps, suggesting that inducing the poor to have higher aspirations may break their vicious cycle of poverty. ²⁷ However, it is possible that inducing higher aspirations may backfire: if the poor set their aspirations too high, they may fail to achieve them, which may actually reduce the investments needed to improve their lives. ²⁸

A recent study tests the theory of aspiration-based poverty traps. ²⁹ The authors collaborated with a microfinance institution in the Philippines that offers savings accounts and group-based microfinance loans to female subsistence entrepreneurs who raise livestock, run small retail businesses, rent tricycles and boats, provide hair dressing, and resell scrap metal. The study covers 2,400 clients in 190 microfinance groups, which were randomly assigned to business as usual (the control group), a financial aspiration treatment, and/or a financial knowledge treatment. In the financial aspiration treatment, the training encouraged participants to set ambitious life goals and to choose savings targets in line with those goals. The researchers cross-randomized with the aspiration treatment a financial knowledge treatment, which provided a more standard financial education approach. The findings in the aspiration treatment are striking: the participants increased their saving goals but failed to achieve them. They borrowed less, so their business investment decreased. This supports the theory that inducing the poor to set their aspirations too high to break out of the poverty trap may backfire.

However, in both traditional and aspiration-based explanations of persistent poverty, the focus remains on the poor per se, their access to markets, coordination failures, and low aspirations. Recent studies provide causal evidence in support of the view that the neighborhoods in which people live have powerful effects on their economic outcomes. In a series of seminal studies, Raj Chetty and his coauthors show that the neighborhoods in which children grow up shape their earnings, college attendance rates, and fertility and marriage patterns; moving to a lower-poverty neighborhood improves college attendance rates and earnings for children who were young (below age thirteen) when their families moved. ³⁰

Thus, to understand poverty, both the poor and the environment in which they live are important considerations. To understand the nature and dynamics of poverty, however, we need to understand the relationship between the two. This idea is not new. Sociologists have contemplated the social world in terms of dynamic, unfolding relations. Relational sociologists have emphasized relationalism, that is, conceiving the social world as consisting in processes, over substantivalism, that is, conceiving the social world as consisting in substances, in explanations of social phenomena. ³¹ Relational ethnographers have studied fields instead of places, boundaries instead of bounded groups, processes instead of processed people, and cultural conflict instead of group culture. ³²

Poverty is a state of the economic and social world. Many people are in poverty not because they lack effort, entrepreneurship, or diversification of risks to reduce their vulnerability. Rather, they often are poor because they are unable to specialize enough and lack access to formal insurance and credit that rich people have; they have psychological impediments that come from being in the state of poverty, which in turn limit their ability to exercise the agency they have. ³³ While a person can slide back into poverty, persistence in poverty is about his relationships with the community, markets, and institutions. Loss of self-respect and dignity can disarm the poor, which has adverse consequences for their agency and its exercise.

Consider Figure 1 below for an illustration of the relationships that the poor, the functionally poor, and the nonpoor, who live in that same environment, may have with the environment. The solid dark line connecting the nonpoor with their environment denotes a strong relationship. For instance, the nonpoor have strong social capital in their local community, which provides informal social insurance in times of need. Likewise, they have a strong relationship with the local market, including access to formal financial institutions, and strong relationships with local public institutions, such as education, health care, and governing bodies.

Figure 1. Poverty is a relationship between the poor and their environment.

In contrast, the dotted line connecting the poor with their environment is a weak relationship. The poor have limited social capital in their community. They usually lack access to credit and insurance from formal financial institutions, frequently rely on informal sources of credits that come at high interest rates (often paid back in kind with labor), and they are less connected with local public institutions (see Table 1 above).

The solid light line connecting the functionally poor with their environment denotes a relationship that is stronger than the relationship that the poor have with their environment, but it is weaker than the corresponding relationship that the nonpoor have. The functionally poor may have good social capital in their local community, linkages with formal financial institutions, and strong connections with the local public institutions. However, they are unable to use them because of their regard for social norms, family prestige and honor, and the psychological consequences of financial stress.

Moreover, not only the strength of these relationships matters. The terms of the relationships are equally important, which are more favorable to the nonpoor than to the functionally poor and the poor. For example, nonpoor farmers get priority during the irrigation seasons from irrigation providers (optimal timing of irrigation affects productivity), the farm laborers are more willing to work on the nonpoor’s farms (crop loss is higher if harvesting is delayed during the peak season), access to credit from formal financial institutions is available on relatively more favorable terms (such as lower interest rates or a longer payment schedule), and local public institutions are more responsive to their concerns. Therefore, to break out of persistent poverty, anti-poverty programs must help markets and local institutions work for the poor—for example, by increasing access to credit and insurance—and provide them with meaningful assets to build upon, induce them to higher aspirations, and build their agency.

III. Agency

David Schmidtz’s conceptualization of a person’s agency is about having goals and their pursuit: “While objects are passive, agents are active. Things happen to objects, but agents are beings that sometimes also do things. Agency is a matter of having goals and making choices about how to pursue them.” ³⁴ Sen conceptualizes agency as a person’s ability to act on behalf of what he or she values and has reason to value. In his account, agency is exercised with respect to goals the person values and may or may not advance his well-being. ³⁵

The agency of an individual can include the following elements: the ability to focus, process and understand the environment, retain and recall knowledge when needed, plan and make choices accordingly, and enforce the choices made. ³⁶ Some aspects of agency are natural, while others can be nurtured and acquired. Just like a person’s executive function, his agency can be expanded through education, exposure, and learning-by-doing.

However, to exercise agency, it is not enough to have agency. The person must also be willing or able to exercise his agency. Often, there is confusion between what constitutes agency and the realization of agency. In classical economics, every individual is “empowered,” so there is no problem of agency:

This is a person with a sense of control over his or her life, who perceives choices, and acts on them. There is not a glimmer in this basic model that any individual might ever feel dominated, oppressed, passive, stuck, ill, unsure about his or her abilities, or unaware of alternatives. There is no recognition that many people in many situations will not perceive that any choices are available to them. ³⁷

This basic model, on Rebecca Blank’s commonsense view, is inapplicable to some people with little education in remote communities in which they might not be included in household decision-making (see Table 2 below), are often restricted to a small social circle, and in which, for example, beating a wife who violates gender norms is deemed legitimate. When a hierarchy is perpetually present in people’s minds through restrictions on mobility, rules of etiquette, gender norms, rules of dress, and symbols, the hierarchy may come to seem part of the natural order of things. Intuition suggests that opportunities individuals have will affect their preferences: their aspirations, their self-construal, and their sense of control over their lives.

Table 2. Married Women’s Say in the Household (Uttar Pradesh and Tamil Nadu, India)

^** Indicates significance at 5%. Source: Author’s own calculation.

Empirical studies support this view. Studies surveyed by Ernst Fehr and Karla Hoff find that historical institutions, even after they are abolished, can shape individuals’ preferences over social insurance and their willingness to accept a labor contract, perhaps by shaping their sense of entitlement. ³⁸ A group’s position at the top or bottom of an extreme social hierarchy can shape group members’ willingness to punish norm violations that hurt a member of one’s community. ³⁹ Some analysts of historical changes in fertility argue that expanded opportunities for contraception made certain choices “thinkable” not only in the sense of their technical possibility, but also because discussions of the expanded opportunities made fertility control socially legitimate. Based on their analysis of changes in fertility, Robert Pollak and Susan Watkins conclude that there was a dynamic “interweaving of opportunities and preferences.” ⁴⁰ The idea that a person’s sense of control over his life (or agency) may mainly be an aspect of social and cultural systems rather than of individual traits and preferences is important because an increase in both agency and the exercise of agency expand well-being and promotes economic development. ⁴¹

IV. Poverty Causes Poor Agency and Impedes the Exercise of Agency

Poverty is highly associated with various deprivations that determine a person’s agency and the exercise of agency. ⁴² The United Nations Development Program (UNDP) underscored this point in 1997:

It is in the deprivation of the lives that people can lead that poverty manifests itself. Poverty can involve not only the lack of the necessities of material well-being, but the denial of opportunities for living a tolerable life. Life can be prematurely shortened. It can be made difficult, painful, or hazardous. It can be deprived of knowledge and communication. And it can be robbed of dignity, confidence, and self-respect—as well as the respect of others. All are aspects of poverty that limit and blight the lives of many millions in the world today. ⁴³

The next two subsections, drawing on recent studies, offer empirical support for this argument.

A. Poverty leads to poor agency through human capital

Poverty is a factor in individual health and education. The poor are less healthy. Poverty also hampers learning through poor nutrition, health, and home circumstances. ⁴⁴ School dropout rate for children from poor families is much higher. ⁴⁵ Thus, the poor are less educated than the nonpoor. Poverty also affects the allocation of resources among household members, which has a strong gender dimension. Given that health and education are vital to a person’s agency, poverty leads to his poor agency by affecting health and education.

The poor face substantial uncertainty and income volatility. ⁴⁶ Long-term exposure to such stress threatens their mental health. ⁴⁷ Recent experimental studies find the efficacy of cash transfers in improving mental and emotional health. ⁴⁸ Ashutosh Kumar and Tauhidur Rahman study the impact of a large, community-driven development program, JEEViKA—which promotes women’s livelihoods and agency through a network of women’s self-help groups (SHGs) in Bihar, India—on women’s mental and emotional health. ⁴⁹ Using a sample of 2,360 SHG women from matched pairs of sixty-six high-exposure and low-exposure JEEViKA villages, they find that the program improved older women’s mental and emotional health.

B. Poverty impedes the exercise of agency

Poverty impedes the exercise of agency through its adverse effects on an individual’s aspirations, self-esteem, dignity, and mental bandwidth, among other potential mechanisms.

1. Aspiration failure

While I have highlighted aspiration failure as a potential explanation of poverty traps, Patricio Dalton and coauthors develop a theoretical model to show how poverty can lead to aspiration failure. ⁵⁰ In the model, both the rich and the poor share the same preferences and behavioral bias in setting aspirations. However, poverty imposes additional external constraints on the poor that exacerbate the adverse effects of the behavioral bias in setting their aspirations: they are more likely to choose a low level of aspiration and effort relative to the best outcome they could have achieved, where low aspirations lead to low effort, reinforcing low aspirations. In other words, poverty curtails the poor’s capacity to aspire. ⁵¹ However, there clearly are other possible causes of aspiration failure, such as cultural norms, kin system, and financial stress. While Ernest and Hope’s household had high aspirations for their sons, they had aspiration failure for their daughters, caused by their increasing financial stress, social and cultural norms, and the regard for their own prestige and honor.

2. Self-esteem

Self-esteem is an individual’s evaluation of his worth as a person. It develops at an early age but changes with changing life conditions. Self-esteem rises with perception of success and falls with perception of failure. ⁵² Self-esteem affects a person’s preferences and low self-esteem harms individual health and well-being. ⁵³ In psychology, there is no consensus on the direction of relationship between low self-esteem and depression. Some studies show low self-esteem contributing to depression, while others show depression eroding self-esteem. In a meta-analysis of the available longitudinal data (covering seventy-seven studies on depression and eighteen studies on anxiety), the effect of self-esteem on depression is significantly stronger than the effect of depression on self-esteem. ⁵⁴ In addition, self-esteem predicts anxiety and anxiety predicts self-esteem.

Game theorists have discussed the relationship between self-esteem and self-handicapping. ⁵⁵ Self-handicapping, a term coined by Edward Jones and Steven Berglas, refers to someone self-inflicting performance impediments. ⁵⁶ These, in turn, reduce the quality of the performance but distort the signal a self-handicapper receives from a poor performance. While self-handicapping behavior may temporarily protect a person’s self-esteem, it may negatively impact long-run outcomes. ⁵⁷

Low self-esteem is associated with poverty. ⁵⁸ In a meta-analysis of 446 studies, low socioeconomic status is associated with low self-esteem. ⁵⁹ Thus, poverty can impede the exercise of agency through affecting a person’s self-esteem. This is also reflected in the life of Ernest. On his own account, as his challenges were turning into constraints, Ernest started avoiding others in his community, which clearly limited the exercise of the agency that he had.

3. Mental Bandwidth

Studies on the psychological consequences of poverty find that poverty impedes the cognitive function of the poor, making them more vulnerable to biases and judgment errors in decision-making. ⁶⁰ Since the poor have less money to buy things, they spend more of their mental bandwidth managing money. In other words, the poor are preoccupied with managing sporadic income, juggling expenses, and making difficult trade-offs. Thus, thinking and fretting about money taxes the poor’s mental bandwidth, which in turn impedes their agency.

C. Stereotype, epistemic discrimination, and agency

Stereotypes and discrimination are among other documented factors of agency, especially for women. Women’s agency can take different expressions, including access to and control over resources, autonomy, decision-making over family formation, having a voice in society, and influencing policy.

To exercise agency, a woman must be heard. To be heard, she needs to have a voice. She may not be heard because of silencing, which occurs when she is prevented from expressing her views. Silencing is passive when she chooses not to communicate and it is active when she chooses not to act on the information or knowledge provided. Compared to men with the same expertise, women with expertise in a male-stereotyped domain (such as sports, history, and geography) are less likely to communicate it and are thus silenced passively. ⁶¹ This may be an explanation of persistent, gender-based occupational stereotypes that have been well documented across societies.

However, constraints on a woman’s ability to exercise her agency may also arise from the prejudiced way that people assess her credibility as a source of knowledge. Such a prejudice would hamper her in every direction. Being discredited in her capacity as a giver of knowledge may undermine her sense of self and, hence, her ability to question conventions and persevere in the face of opposition. Hoff, Rahman, and Vijayendra Rao investigate discrimination against women as a source of information in daily life in Bihar, India. ⁶² They develop three behavioral experiments and use them to study such discrimination against women. In their first experiment of providing nutritional information (a choice between two types of flour), they test whether providing information to the wife is as influential as the husband in the decision of the couple. Couples are randomly assigned one of the three treatments. In one treatment, nutritional information is given to the man. In another treatment, the same information is given to the woman. In the third treatment, no one in the household receives the information. They find that providing information influences the couple’s choice over two types of flour, but it matters more if the information is provided to the husband instead of the wife. This finding is important because many anti-poverty programs specifically target women. An implicit assumption in such programs is that providing information to women is efficient. Their finding shows that providing information to women may not be efficient because men may not listen to them. In their second experiment, they investigate the same question as in the first experiment, but in a different context, where information provided is about optical illusions. Unlike the first experiment, in this case, both (man and woman) have the opportunity to demonstrate their knowledge to the other and there is a strong economic incentive for them to transfer information. They find that men give less credibility to their wife’s word relative to the level that women give to their husband’s word. That is, women have less voice than men and they are less-effective informants.

Their third experiment tests whether information provided by women to the respondents is less persuasive than the same information provided by men. They present respondents with a video of a dispute involving two houses (a red house and a blue house). The disputants are always one man and one woman. They use twelve pairs of professional actors as the disputants. In each pair, one of the disputants is male and the other is female. With each pair they have a video. All videos have the same scripts, pictures, and announcer. In half of the videos, the disputant representing the red house is a male and the disputant representing the blue house is a female. In the other half of the videos, it is the reverse. Respondents are randomly assigned to one of the twelve videos. Each disputant makes three arguments. Once both disputants are done with presenting their arguments, the respondents are asked to evaluate the presented arguments and reveal their preference concerning the solutions presented by the two disputants. More specifically, the respondents are first asked to assess the wisdom, articulateness, and credibility of the disputants. Then they are asked to reveal whose arguments they support. The findings are striking. First, the respondents, on average, find the woman in the dispute to be statistically more credible, articulate, and wise than the man. Second, being perceived wiser, more articulate, and more credible enhances the likelihood of winning the respondents’ support. Nevertheless, the female disputant does not win community (both male and female respondents) support. In other words, a woman can win community support in a debate against a man, but to have an equal chance of winning, she needs to be perceived as wiser, more articulate, and more credible than the man.

V. Promoting Agency Promotes Development

Now I turn to my argument that agency is central to understanding both poverty and development. The relationship between individual (collective) agency and individual (collective) development is bidirectional: promoting agency promotes development and vice versa. The idea that development (such as improvements in income, education, and health) can improve agency is straightforward. Consider the problem of classical consumer utility maximization subject to the budget constraint that is taught in introductory economics. An increase in income, given fixed prices of goods and services, relaxes the budget constraint and shifts the budget line to the right, expanding the consumer’s choice set consisting of different combinations of goods and services that he can buy and consume to maximize his utility. Income is a means to an individual’s ends. It can expand a consumer’s opportunity and choice sets. Therefore, ceteris paribus, it is not difficult to see why an increase in income can lead to improvement in some dimensions of agency. Likewise, better educational and health attainments, ceteris paribus, can improve at least some aspects of agency. However, whether agency is a key to escaping poverty and for development has received renewed attention in recent years.

Therefore, I here focus on the first part of the relationship between agency and development, namely, the role of agency in development. I draw upon empirical studies on agency as a determinant of development and anti-poverty and economic empowerment programs that have targeted individual (collective) agency for individual (collective) development. More specifically, in Section V.A, I briefly discuss evidence showing that building agency through a microlevel, bottom-up approach leads to better development outcomes. In Section V.B, I draw upon empirical literature in economics and political science to support my argument that promoting collective agency through macrolevel, policy-level programs leads to development.

A. Individual agency and development

Evidence on the role of agency in promoting development is available from both observational and experimental studies. Jean Dreze and coauthors examine the determinants of widely used indicators of development—such as fertility, child mortality, and gender bias in child mortality—in India. ⁶³ Using district-level data from the 1981 Census, they highlight three findings. First, women’s agency, captured by female literacy and female labor force participation, has powerful effects on mortality and fertility. Second, higher levels of female literacy and labor force participation are associated with significantly lower levels of female disadvantage in child survival. Third, variables relating to the general level of development and modernization have relatively weak effects on demographic outcomes.

Women’s autonomy results in long-term reduction in fertility and higher survival rates of children. ⁶⁴ Women tend to allocate more resources toward the well-being of their children as compared to men. Therefore, women’s income relative to husbands’ income has a strong bearing on the well-being of children. ⁶⁵ Unearned income under the control of the mother has a greater impact on the health of the family as compared to when it is under the control of the father. ⁶⁶

Among experimental studies, that of Erica Field and coauthors is particularly notable. ⁶⁷ They investigate whether increasing control over earnings can incentivize women to work and thereby influence norms around gender roles. They conduct a randomized controlled trial covering 197 village clusters (gram panchayats [GPs]), in the Indian state of Madhya Pradesh, in partnership with the government workfare program, the Mahatma Gandhi National Rural Employment Guarantee Scheme (MGNREGS). For a female worker, the status quo is for her earnings to be deposited in the bank account of a male household head. To increase women’s control over their earnings, the researchers worked with banks to open individual accounts for women in all treatment GPs and, in one-half of the treatment GPs, they coordinated with state authorities to designate these accounts to receive direct deposit of MGNREGS wages. In addition, the researchers cross-randomized a short training on how to use the local bank kiosks that serviced these accounts. This generated five intervention groups: (i) pure control (status quo); (ii) own account (“accounts only”); (iii) own account and training; (iv) own account and direct deposit; and (v) own account, direct deposit, and training. The findings are striking. In comparison to the accounts-only group, women who received direct deposit and training worked more in public- and private-sector jobs. More importantly, after three years, direct deposit and training liberalized women’s own work-related norms and changed perceptions of community norms. A related strand of studies shows that placing female leaders in positions of power changes public attitudes about gender roles and raises both the aspirations parents have for their daughters and the aspirations teenage girls have for themselves. ⁶⁸

More recent anti-poverty and economic empowerment interventions have been multifaceted, consisting of the components of asset transfer and promotion of agency, but do they work? The answer is a resounding “Yes.” The development impacts of such programs are larger and more durable with a component of agency promotion than without it.

Abhijit Banerjee and coauthors and Oriana Bandiera and coauthors provide experimental evidence from seven countries for a “graduation” program consisting of three components: (a) a transfer of productive assets, (b) two years of training and coaching, and (c) access to a savings account. ⁶⁹ The program increased net worth, income, and consumption. Christopher Blattman and coauthors find that the ultra-poor, war-affected women in northern Uganda have high returns to a package of (a) $150 cash, (b) five days of business-skills training, and (c) ongoing supervision. ⁷⁰ Sixteen months after the program, the microenterprise ownership and income of the program participants doubled. In addition, the study finds that while the ultra-poor have little social capital, their group bonds, informal insurance, and cooperative activities can be induced for positive outcomes. Guadalupe Bedoya and coauthors investigate whether the Targeting the Ultra-Poor program (TUP) can lift ultra-poor households out of poverty in the fragile context of Afghanistan. ⁷¹ In eighty villages in Balkh province of Afghanistan, 1,219 ultra-poor households were randomly assigned to a treatment or control group. Women in treatment households received a one-off “big push” package, which included a transfer of livestock assets, cash consumption stipend, skills training, and coaching. One year after the conclusion of the program, there were large impacts on consumption, assets, psychological well-being, financial inclusion, and women’s empowerment. More specifically, in the treatment group, per capita consumption increased by 30 percent as compared to the control group and the share of households below the national poverty line decreased to 62 percent in the treatment group as compared to 82 percent in the control group. Under modest assumptions about consumption impacts, the program’s estimated internal rate of return is 26 percent, excluding improvements in psychological well-being, women’s empowerment, and children’s health and education. Given these results, the authors conclude that multifaceted interventions can reduce poverty in fragile regions. In other words, it is possible to lift even ultra-poor households out of poverty through multifaceted programming.

Clare Balboni and coauthors study a large-scale randomized asset transfer program (provision of cows) in rural Bangladesh. ⁷² Using eleven years of panel data on 6,000 households that began as ultra-poor, they find that large enough transfers to the ultra-poor, which create better jobs for them, are an effective means of getting them out of poverty traps.

Is a multifaceted program necessary, though, to fight extreme poverty? The preceding studies suggest that it is necessary. Single-faceted programs work if the binding constraint on the ultra-poor is their lack of wealth and access to finances. However, evidence on the psychology of poverty suggests that poverty impedes the cognitive function of the poor and taxes mental bandwidth. ⁷³ Therefore, the poor are more subject to biases and judgment errors. In such contexts, single-faceted programs may not produce desired development outcomes, which is also supported by the success of TUP programs that building the agency of the ultra-poor is required to lift them out of poverty.

B. Collective agency and development

The role of collective agency in development can be understood from at least two points of view. First, one can examine the development impacts of interventions that have attempted to improve group agency, such as with women’s groups and community-driven development programs. Second, one can look at the development impacts of building state capacity, which is a correlate of state agency.

Women’s groups have become a popular approach to promoting women’s empowerment. These groups are utilized to deliver development interventions. Typically, these interventions attempt to change women’s habits—such as how to organize together, save, be accountable, and plan—promote social interaction and group deliberation, provide access to credit and social protection, and target reduction in environmental stress for women. In other words, efforts to create and support women’s groups are also aimed at improving women’s individual agency within their households and their group agency within their local communities. ⁷⁴ A significant body of empirical literature has thus emerged on women’s groups. A recent review of experimental and quasi-experimental studies suggests that group models can achieve positive development impacts by harnessing group interactions and deliberations, which would be difficult to achieve through individual-based interventions. ⁷⁵

To promote development, group monitoring and oversight of public- and private-sector performance by local communities, public information dissemination, public complaint and grievance redress mechanisms, and citizen participation in resource allocation decision-making (such as participatory budgeting) have also been adopted. ⁷⁶ Through these “civic engagements” citizens can exercise their collective agency with the assistance of independent media and civil society organizations. ⁷⁷ Countries with weak civic engagement have conducive environments for corruption and governments are susceptible to being influenced by powerful groups. Consequently, the incentives of governments and citizens are not aligned.

Thus, social accountability of governments, which is greater when the collective agency of citizens is exercised, is vital for development. The collective agency of citizens can be fostered through their direct political participation, increasing access to information, and by supporting civil society organizations. Providing information to citizens also improves their collective agency. ⁷⁸ A well-informed electorate in a democracy can significantly explain the variance in governmental performance and corruption. ⁷⁹ The incentive for government to be responsive to their citizens is stronger if a state has a more informed and politically active electorate. ⁸⁰

The poor in such contexts are also typically less informed and less likely to vote. ⁸¹ By empowering them, though, it is possible to mitigate these effects. ⁸² For example, state governments in India are more responsive to falls in food production and crop flood damage where newspaper circulation is higher and electoral accountability is greater. ⁸³ Civil society organizations can play a powerful role in citizen empowerment by assisting the media in providing unbiased information and through providing citizens with government-monitoring tools. ⁸⁴ For example, Action Aid campaigned against political violence in a dozen villages in Nigeria, held town meetings, and distributed information, reducing the ability of politicians to intimidate voters. As a result, voter turnout increased and the incidence of electoral violence declined. ⁸⁵

Empirical evidence on tangible impacts of these initiatives, however, has been mixed. To make sense of the available evidence, Jonathan Fox conducted a meta-analysis of prominent studies investigating the impacts of “civic engagements” and social accountability measures and documents their positive development outcomes, ranging from education, teacher effort, and infant mortality rates to health and corruption. ⁸⁶ More specifically, the study concludes that enabling environments for collective action combined with strengthened state capacity to respond to citizen voice are promising for development.

To examine the impact of accountability at the country level, Aleksandra Petkovic and Rahman investigate the effects of government accountability on development. ⁸⁷ Using a panel of sixty-four young democracies from 1974 to 2010, they show that countries with stronger institutions of accountability have lower infant mortality rates, a widely used indicator of development. Moreover, they show that while vertical, horizontal, and social accountability are associated with lower infant mortality rates, the latter is a relatively more powerful and robust determinant. These findings suggest that policies that promote social accountability may lead to improvements in development outcomes.

VI. Conclusions

This essay is inspired by the story of Ernest and Hope. The lessons I draw from their story are about the nature and dynamics of poverty, agency, and development. While Ernest and Hope lived in India, the characteristics of their lives and the environment in which they lived resonate with more than one billion people who live on under $2 dollar a day.

I have argued that to have a better understanding of such lives, the following elements are required. First, we need to recognize that a person who is not poor by the standard conception and characterization of poverty can be or be rendered functionally poor. Second, poverty is a relationship between the poor with their environment (community, local markets, and local institutions). Third, poverty has primary roles in understanding agency and the exercise of agency. Fourth, increasing agency promotes development.

Timing and frequency are dimensions of poverty that widely used poverty measures do not reflect. A household may be estimated to be nonpoor over the time period measured in the census or survey, but it may actually be poor for several months during that time period. That is, a household could be poor for the better part of the year but identified as nonpoor because their poverty status is assessed on whether their annual income exceeds the poverty line. The probability that a household might episodically fall into poverty is another important dimension of well-being. However, currently used measures of poverty (both income- or consumption-based and multidimensional) do not factor in how many times a household falls below the poverty line.

A distinction needs to be made between a person who is poor in terms of the standard definition based on income or consumption and a person who is functionally poor. A person is functionally poor if he exhibits characteristics that define the economic and psychological lives of the poor. The latter is possible because a person can suffer from the psychology associated with poverty and experience exchange-entitlement failure.

In both traditional and aspiration-based explanations of persistent poverty, the focus continues to be on the poor per se, which I have argued is misplaced. Drawing on insights from relational sociologists and ethnographers, I considered poverty as a relationship between the poor and their environment. While a person can slide into poverty, persistence in poverty is about his relationships with the community, markets, and institutions. Moreover, both the strength and the terms of these relationships need to be explored to have a better understanding of persistent poverty.

Poverty is associated with various deprivations that determine a person’s agency and exercise of agency. Poverty causes poor agency through affecting human capital formation and poverty impedes the exercise of agency through channels such as aspiration failure, decreasing self-esteem, and taxing mental bandwidth. Whether a person is really poor or functionally poor has important policy implications. If the former, then multifaceted anti-poverty programs, involving both policy-level and agency-level interventions, are more likely to be effective. If the latter, then policy-level interventions, such as property rights or access to formal insurance, alone may suffice.

To help the poor escape from persistent poverty, anti-poverty programs must help markets and local institutions work for them as the agents that they are and that they can be, provide meaningful assets for them to build upon, induce higher aspirations, and build their agency. In other words, a person cannot be nudged out of a poverty trap.