Across the world, men are more likely to hold jobs than women. This is consistent with a widespread belief that “when jobs are scarce, men should have more right to a job than women.” What explains these patterns? On the one hand, perhaps households prefer men to work because labor markets tend to offer men better opportunities. Alternatively, these patterns may reflect underlying preferences for men’s employment. For example, men working may align with traditional gender norms that households value, while women’s employment may raise concerns about reduced time for childcare.
In this paper, the authors study these questions by running a randomized controlled trial (RCT): A scientific study in which participants are randomly assigned to either a treatment group receiving the intervention being tested or a control group receiving a placebo or standard treatment. This method is used to objectively evaluate the effectiveness of a new treatment or intervention. with married couples living in Rohingya refugee camps in Bangladesh, the largest refugee settlement in the world, where unemployment is widespread and future job prospects are scarce. Their analysis proceeds in four steps:
1: Documenting Baseline Preferences
The authors begin by asking participants: “Imagine we could offer six months of employment to you or your partner. Would you prefer to take the opportunity yourself or have your partner take it?” The majority of men, at 71%, and the majority of women, at 59%, prefer that the man take the job.
2: Understanding What Drives These Preferences
The authors next explore what is driving these preferences by randomly assigning the same work task to either the husband or the wife, then comparing psychosocial wellbeing for both the worker and their spouse.
While employment significantly increases well-being for the individual who is employed—whether a man or a woman—the authors document a clear asymmetry in its effects on the partner. The wives of employed men experience large and statistically significant psychosocial gains from their husband being employed, reflected in less severe depression and an improved sense of purpose and self-worth. In contrast, the husbands of employed women exhibit no detectable change in their wellbeing from their wife being employed.
3: Testing Whether Preferences Are Malleable
Having established that partnerships appear to fare better when men work, which could explain the preferences that households express for men’s work, the authors next investigate whether these preferences are malleable. They revisit households fifteen months after the initial experiment and offer each spouse a one-week paid job, with only one job available between the two partners. The authors privately elicit whether each partner prefers to keep the opportunity for themselves or pass it to their spouse, progressively increasing the proposed wage for the non-preferred partner until the participant states indifference, enabling them to capture the strength of each partner’s gendered preference for work.
The experience of a woman working for a mere six weeks appears sufficient to meaningfully shift both the woman’s and her husband’s revealed preferences toward the woman working more than one year later.
Among female respondents, women who were formerly employed are 19.1 percentage points (30%) more likely than the wives of formerly employed men to prefer taking the job themselves. Their premium required to give the work opportunity to their husbands also increases significantly, reaching 70% of the base wage rate offer of 200 Tk.
Among men, husbands of formerly employed women are 25.8 percentage points (103%) more likely than formerly employed men to prefer giving the job opportunity to their wives. The average man whose wife had previously worked prefers that his wife take the job, but will choose to take it himself for a negligible wage premium. In contrast, the average man who was previously employed will not give the work opportunity to his wife unless her wage is 63% higher than his own.
4: What Drives These Shifts in Preferences
It appears that the experience of working enables women to revise their beliefs about the benefits and costs of women’s employment. Indeed, while women were working, their psychosocial wellbeing improved, and their husbands’ psychosocial wellbeing did not decline. Fifteen months later, formerly employed women are also seven percentage points less likely (a 70% reduction) to view female employment as socially inappropriate relative to their stay-at-home female counterparts.
Exposure to women’s employment also enables men to update their beliefs. While women were working, men did not assume greater household responsibilities, nor did they experience declines in self-worth or sense of purpose or lose bargaining power to their working wives. Consistent with this absence of negative impacts, one year later these men are also significantly less likely to express concerns about women’s employment:
- Nine percentage points (38%) less likely to believe that employed women will have less time for household tasks
- Seven percentage points (30%) less likely to say employment changes women’s attitudes
- Seven percentage points (44%) less likely to say it creates household tension
- Ten percentage points (23%) more likely to report that there are no costs associated with women working, relative to their stay-at-home female counterparts
The results present a complex trade-off for policymakers: households appear to feel happier when the man works, but they are more likely to prefer that the woman work if they have prior experience with her doing so. To reconcile these findings, policymakers might facilitate learning opportunities that enable both members to experiment with employment, and then enable beneficiaries themselves to choose who in the household should be targeted with an employment opportunity – a decision that necessarily considers the full set of impacts they experience, many of which the policymaker herself may not be able to observe.