Economists often describe households as small units of cooperation — spaces where members, bound by affection and mutual interest, pool resources and make decisions to maximise the family’s well-being. Classic economic models assume that household members know one another’s preferences, act benevolently, and make choices that are fair and efficient, that is, they use their resources in ways that maximize the overall wellbeing of the family.
But what if that is not how real households work? What if even people who live together, eat together, and care for each other routinely misjudge what others want, or insist on deciding for themselves even when there is nothing material to gain? If households are not always benevolent or perfectly coordinated, then the way the resources are distributed within them can produce very different outcomes. Who receives money, credit, or other types of support within the household begins to matter, because decision-making power determine whose preferences guide how those resources are used.
These questions motivated a series of experiments I conducted with Uzma Afzal, Giovanna d’Adda and Marcel Fafchamps in Pakistan.1 We wanted to test how people make consumption decisions for themselves and for others within the same household, and what those choices reveal about the demand for agency, control, and recognition
INSIDE THE EXPERIMENT
Our first two ‘lab-in-the-field’ studies in Punjab were conducted with nearly 500 pairs of household members. Most were spouses, but some were other close relations. The setting was deliberately simple: a small hall with tables, chairs, and trays of food and drinks. Respondents were provided a modest amount of money for participating in the session. Each participant tasted different food and drink items, ranked them by preference, and guessed what their household partner would prefer. They then chose food items – a snack or lunch menu - for that partner, and for themselves. Before the selected items were served, each participant faced one more decision: they could accept what their partner had chosen for them, or they could give up a small amount of money to guarantee that they received their own preferred items. In effect, they could pay for control. If both partners truly understood and trusted each other’s preferences, as most economic theory would predict, there should be little reason to do so.
WHAT WE FOUND
The results offered a nuanced insight into the dynamics of household decision-making. First, household members often know surprisingly little about one another’s preferences.
THIS PATTERN ECHOES BROADER GENDER DYNAMICS: WHEN WOMEN HAVE LIMITED CONTROL IN MAJOR HOUSEHOLD DOMAINS, EVEN SMALL ACTS OF CHOICE CAN CARRY DEEP SYMBOLIC MEANING.
When asked to guess their partner’s favourite flavor of drink, or their most preferred lunch item, most participants performed no better than if they had been guessing for a complete stranger. This finding challenges the idea of perfect information within families — a cornerstone of standard household models. Even in close relationships, people act on assumptions rather than accurate knowledge.
Second, participants valued decision-making power for different reasons.
When men learned that their guesses about their partner’s preferences were wrong, many revised their choices: they tended to respect the partner’s stated preference once informed. Women, however, were much less likely to change their selection, even after learning what the other preferred.
In other words, men’s behaviour was more outcome-oriented, focused on getting the decision right; while women’s was more process-oriented, reflecting a desire to exercise agency itself. This pattern echoes broader gender dynamics: when women have limited control in major household domains, even small acts of choice can carry deep symbolic meaning.
Third, about one in ten participants paid for agency even when there was nothing to gain. Approximately 11 percent were willing to sacrifice part of their participation fee to ensure that their own choice prevailed, even when they already knew that their partner had selected their favourite drink. This reflects a demand for pure agency, which cannot be explained by efficiency or mistrust alone Women were more likely than men to defer to another’s choice only when that partner was a stranger, suggesting that expression of agency within the household reflects a wish to decide for themselves, rather than to achieve any material gain.
FAMILIES ARE NOT PERFECTLY BENEVOLENT COLLECTIVES, BUT SITES OF NEGOTIATION SHAPED BY NORMS, HABIT, AND HIERARCHY.
WHY THIS MATTERS
From a traditional economic perspective, what these behaviours represent may seem ‘inefficient’ – people giving up money for choices that don’t change outcomes. But viewed through the lens of gender and power, they reveal how control itself can be valuable. For women in Pakistan, the act of deciding, even over something as small as what to eat or drink, can signal recognition and self-worth.
To test whether this demand for agency could shift with genuine empowerment, we extended the work into a field experiment embedded in a microfinance programme. In this study, nearly 2000 women were randomly offered access to a small financial product in their own name, a tangible resource over which they could exercise real decision-making power.
When these women later participated in a similar choice experiment, something striking emerged: those who had received a microfinance loan were less likely to spend their own earnings to ensure that their personal choice prevailed, and their male household partners, in turn, were more willing to let them decide on their own. In essence, when women had substantive control in one domain, both partners became less preoccupied with her asserting control in trivial ones, suggesting that empowerment can ease the quiet tensions around everyday decision-making.
RETHINKING EFFICIENCY AND THE QUIET POWER OF CHOICE
Our findings challenge the assumption that inefficiency in households stems only from poor information or misaligned incentives. Even when people know what others want, they may still insist on deciding for themselves, or for others, because the act of choosing carries meaning. In our experiments, some participants selected for their household partners items they knew the other did not prefer, suggesting that decision-making can also be used to signal authority or care, depending on context.
This has important implications for how we think about household welfare. Families are not perfectly benevolent collectives, but sites of negotiation shaped by norms, habit, and hierarchy. True cooperation requires not only shared goals but shared voice.
Dr. Farah Said is Assistant Professor of Economics and Executive Director of the Mahbub ul Haq Research Centre at LUMS. Her research focuses on gender, household decision-making, and behavioral economics.
- Afzal, U., d’Adda, G., Fafchamps and Said, F. (2022). Intrahousehold Consumption Allocation and Demand for Agency: A triple experimental investigation, American Economic Journal: Applied Economics, 14 (3), 400-444.