Isabelle Salcher
PhD Economist and Data Enthusiast
My research is in development economics, with a particular interest in labor markets and gender inequality. I am also interested in how poverty interferes with attention and income generation.
RESEARCH
Joint with Gabriel Brown, Morgan Hardy, Isaac Mbiti, and Jamie McCasland
We use a field experiment to test whether financial incentives can improve the quality of apprenticeship training. Trainers (firm owners) in the treatment group participated in a tournament incentive scheme where they received a payment based on their apprentices' rank-order performance on a skills assessment. Trainers in the control group received a fixed payment based on their apprentices' participation in the assessment. Performance on the assessment was higher in the treatment group. Two years later, apprentices scored 0.15 SD higher on a low-stakes oral skills test and earned 24% more in total earnings, driven by higher self-employment profits.
What We Can Learn When Impacts Fail to Replicate: Lessons from an Empowerment Program for Young Women in Tanzania
Joint with Niklas Buehren, Markus Goldstein, Selim Gulesci, Munshi Sulaiman, and Venus Yam
The Journal of Development Studies, 2024
We evaluate the Empowerment and Livelihood for Adolescents program that combines life and livelihood skills trainings with a safe space where young women can network and learn. The same multifaceted intervention was highly effective in raising adolescent girls' economic and social empowerment when implemented in Uganda. In contrast, this randomized evaluation, using panel data on more than 3,100 adolescents, finds no program impact on any economic, health, or social outcomes in Tanzania. We propose implementation quality as one plausible explanation of why the program's effectiveness could not be successfully replicated in Tanzania.
Working Paper
This paper examines the impact of an aggregate shock to human capital on the marriage market. In Zimbabwe, a large nationwide education reform, together with a large age difference between spouses, allows us to isolate the impact an increase in female education has on matching patterns. Empirically, I exploit the age-specific exposure to the reform in a regression discontinuity design. I find that females who were still young enough to benefit from the reform discontinuously complete more schooling. Husbands of these ”young enough” females are more educated, and the age difference with most of these educated men decreases. Females’ own schooling, however, increases by more than their husbands’, so that the schooling difference between spouses declines. I then introduce a theoretical framework to examine whether the observed changes in spousal characteristics were associated with a change in complementarities and marital sorting patterns. The estimation of a so-called affinity matrix reveals the reform has led to trade-offs in the marriage market, but not reinforced sorting on education.
Working Paper
Joint with Morgan Hardy, Isaac Mbiti, and Jamie McCasland
This paper examines the effects of a government-sponsored apprenticeship training program designed to address high levels of youth unemployment in Ghana. The national-scale program placed unemployed young people with small firms that regularly train apprentices, approximating the institutional framework in which the vast majority of training occurs in West Africa. Our evaluation design features two sequential randomizations. First, we randomized which eligible youth applicants received an apprenticeship offer, allowing us to causally estimate the labor market returns to apprenticeship training. We find that short-run total earnings effects are negative. Young people who received an apprenticeship offer shift out of wage employment, and earnings losses from wage employment are not offset by other sources of income. Second, we randomized the matching between apprentices and training firms, allowing us to causally study the importance of trainer characteristics for apprentice outcomes. We estimate that training with larger and more profitable firms is equivalent to a doubling of total earnings. Taken together, our findings are less than encouraging for the promise of apprenticeships to address youth unemployment. They do, however, suggest apprenticeship programs can be made more effective through targeted recruitment of training firms.
Gender Norms and Female-owned Enterprises in Ethiopia
Work in Progress
Joint with Eyoual Demeke
This paper studies the extent to which the norm that women should do most housework and childcare constrains female-owned enterprises. In Ethiopia, gender norms continue to prescribe the burden of domestic duties to women, even as many women run their own business. First, we make weekly work offers with randomized wages to female entrepreneurs and their husbands. This step allows us to quantify the premium that female entrepreneurs require to accept additional market work relative to their husbands. Second, we experimentally elicit participants’ (un)willingness to perform gender-coded tasks and combine it with survey-based belief measures over the division of home production. This step allows us to examine how much of the reservation-wage premium could be explained by gender norms. Finally, we correlate our uncovered norm measure with business profits, time use, and household income. Overcoming measurement challenges is crucial for understanding the restrictions imposed by gender norms and for designing interventions to ease them.
Hello, Goodbye: What Predicts Early Job Separation Among Female Factory Workers
Work in Progress
Joint with Kehinde Ajayi, Niklas Buehren, and Rachel Cassidy
This paper explores potential explanations for the high turnover of female factory workers in Ethiopia. The Government of Ethiopia has embarked on an ambitious effort to plan, construct, and operationally support a series of industrial parks. Existing research on the impacts of factory work inside these industrial parks has found low take-up rates, high turnover, and biased expectations about job attributes. We use survey measures and incentivized games to measure job applicants’ traits, preferences, and expectations about work inside and outside industrial parks. We then study whether these measures predict women’s retention in factory jobs and other employment outcomes. Among the applicants, only 59% of women who expressed interest in a factory job and passed an initial screening actually started a factory job. Of those, less than a third were still working in the same factory three months later. Inaccurate beliefs about factory base wages as well as about the repetitiveness and healthiness of the factory work predict whether workers are retained in the factory job. Workers with more grit and planning capacity are also more likely to still be working in the same factory job.
How Much Does Attention Cost? How to Calculate the Bill When Asked to Pay Attention
Joint with Esther Owelle and Chalmers Mulwa
Attention is an important input in many work activities and, as such, matters for generating income. A variety of poverty hardships can interfere with attention and thereby, diminish worker productivity and earnings. Sufficiently strong feedback effects could even result in self-reinforcing cycles of poverty. In order to study the role of attention more carefully, one needs to measure attention first. However, since attention is inherently invisible and intangible, this is not entirely straightforward. Estimating attention cost curves provides us with a systematic and robust, yet intuitive and simple, framework to quantify attention empirically. Intuitively, while paying more attention does not have a monetary price tag, it comes at an internal utility cost and can lead to higher payoffs through improved output quality. Observing mistakes at varying stakes helps trace out the cost-benefit trade-off individuals face when choosing how much attention to pay. We design three tasks to recover attention cost curves in practice, ranging from a real-world production to a computer-based lab task.
Are Workers Aware of their Productivity? The Case of Attentional Fatigue
Work in Progress
This paper examines the extent to which individuals are unaware of how their productivity evolves over time, and assesses the earnings implications of this unawareness. In an online experiment, I elicit incentivized beliefs over the quantity and quality of output produced over time on a representative real-work task. Uncertainty in belief estimates is captured in addition to average estimates. For a randomized subset, I also elicit beliefs prior to working on the task. After completing the work, workers can choose between two compensation schemes. The scheme they choose will be applied to their past performance to determine their earnings. The optimal compensation scheme thus depends on the quantity and quality of workers’ output. First, I document how workers’ productivity varies over time. Then, I assess the extent to which individuals hold incorrect beliefs about how their productivity evolves over time and update their beliefs by working on the task. Finally, I assess the stakes of workers’ unawareness of their performance in terms of foregone earnings.