Research 

The resilience of social protection systems in the Middle East and North Africa (MENA) region faced significant challenges due to COVID-19. Governments and non-governmental organizations responded by expanding their social safety nets (SSNs) and/or initiating new cash transfer programs. We evaluated the impact of SSNs on mitigating food insecurity in MENA during COVID-19. This evaluation is based on data from the COVID-19 MENA Monitor Household (CMMHH) survey, a panel phone survey conducted across four MENA countries (Egypt, Jordan, Tunisia, and Morocco) from November 2020 to June 2021. Using a staggered difference-in-difference (DD) technique, we found that those who received non-usual government support in Tunisia were 15 percentage points (ppts) less likely to be unable to buy their typical amount of food due to price increases than those who did not receive support. Similar associations were not observed in Egypt, Jordan, or Morocco. As for non-governmental organizations, we found that individuals who received non-usual support in Morocco and Jordan were 22 and 15 ppts less likely to report being unable to buy their typical amount of food due to decreased income, respectively. Our estimates also showed that government SSNs helped mitigate the negative effect of food insecurity on resorting to adverse coping strategies during COVID-19, especially selling assets. 

COVID-19 has had various effects on women’s labour supply worldwide. This study investigates how women’s labour market outcomes in the MENA region have been affected by the stringency of governments’ COVID-19 responses and school closures. We examine whether women, particularly those with children at young age, reduced their labour supply to take care of their families during the pandemic. To investigate whether having a family results in an extra penalty to women’s labour market outcomes, we compare single women to married women and mothers. Using the ERF COVID-19 MENA Monitor Household Surveys, we analyse the key conditions underlying women’s labour market outcomes: (1) wage earnings and labour market status including remaining formally employed, informally, unpaid or self-employed, unemployed or out of the labour force and (2) becoming permanently terminated, being suspended, seeing a reduction in the hours worked or wages, or seeing a delay in one’s wage payments because of COVID-19. Ordered probit and multinomial logit are employed in the case of categorical outcomes, and linear models for wage earnings. Findings – Women, regardless of whether they have children or not, appear to join the labour market out of necessity to help their families in the times of crisis. Child-caring women who are economically inactive are also more likely to enter the labour market. There is little difference between the negative experiences of women with children and child-free women in regard to their monthly pay reduction or delay, or contract termination, but women with children were more likely to experience reduction in hours worked throughout the pandemic. These findings may not have causal interpretation facilitating accurate inference. This is because of potential omitted variables such as endogenous motivation of women in different circumstances, latent changes in the division of domestic work between care-giving and other household members, or selective sample attrition. 

Economists often use balance tests to demonstrate that the treatment and control groups are comparable prior to an intervention. We show that typical implementations of balance tests have poor statistical properties. Pairwise t-tests leave it unclear how many rejections indicate overall imbalance. Omnibus tests of joint orthogonality, in which the treatment is regressed on all the baseline covariates, address this ambiguity but substantially over-reject the null hypothesis using the sampling-based p-values that are typical in the literature. This problem is exacerbated when the number of covariates is high compared to the number of observations. We examine the performance of alternative tests, and show that omnibus F-tests of joint orthogonality with randomization inference p-values have the correct size and reasonable power. We apply these tests to data from two prominent recent articles, where standard F-tests indicate imbalance, and show that the study arms are actually balanced when appropriate tests are used.  

This study examines the impact of male international migration on the labor market outcomes and empowerment of women left behind in Egypt (WLB). Data is obtained from 2006, 2012, and 2018 Egyptian Labor Market Panel Survey (ELMPS). Two batteries of outcomes variables are modeled including the labor supply of WLB and three pillars of women empowerment. The findings show that the migration of a male member in the household is associated with a lower level of labor market participation for women. However, these effects are not driven by lower empowerment, but mostly because WLB replaces the role played by the male migrant in subsistence and non-paid family work, particularly in rural areas. Both WLB living in rural and urban areas are less likely to ask for permission their husband or another family members when going out and are more likely to have personal savings than women living in a non-migrant household. Our results show that socio-economic changes happening in the destination countries may have contributed to the progress we observe in terms of WLB empowerment. 

There is substantial inequality in educational achievement in Jordan, but few studies examine its extent and the way forward to reduce inequality and achieve social mobility. This paper attempts to examine the sources and trends of inequality in educational outcomes in Jordan using Household Income and Expenditure Surveys (HIES) from 2008 to 2017. In addition, using the D-index and shapely decomposition, we model the sequential nature of attaining education to give an overall view of the influence of circumstances on child progression through schooling years. We find that inequality is high, especially with respect to transition from basic completion to secondary education and above. Parents’ education, in particular fathers’ education, and family financial resources are considered the main drivers of the inequality in the transition from basic to secondary and tertiary education levels. In the 2017 HIES wave, shapely decomposition shows an increasing gender gap over time where girls have lower completion rates compared to boys. 

This paper investigates the coupled dynamics of climate and growth in the Middle East and North Africa (MENA) region, considering exogenous shocks such as the COVID-19, with no presumptions on the direction of causality. We utilize temperature and precipitation departures from their historical trends to quantify climate change. We estimate country-specific and panel vector autoregressive models with exogenous variables to supplement existing attempts that integrate climate change in economic modelling. Our findings lend evidence to the interrelation between climate change and economic growth without ruling out bi-directional causality, a reality that many climate studies tend to overlook. 

A large majority of forcibly displaced people are hosted in low- and middle-income countries, where access to labour markets is severely constrained. Many of these people have skills that are demanded in high-income countries, and they aspire to migrate; yet they are often locked out of skilled migration systems due to not having travel documents and other administrative barriers, not being visible to employers, or facing difficulties identifying relevant job opportunities. Talent Beyond Boundaries (TBB), a global non-profit, is the first organization in the world to focus on refugee labour mobility as a complementary solution to traditional humanitarian resettlement. The aim of our research project is to study the impact of TBB’s labour mobility programs on the outcomes of the displaced talents and their families using a staggered difference-in-difference approach. The project is funded by the Innovation Challenge on Jobs and Migration World Bank Initiative. 

The livelihood of farmers in Egypt is highly threatened by climate change. Yet, many farmers may hold incorrect beliefs about the impacts of climate change on agriculture and ignore how to utilize adaptation strategies. We propose running an experiment in which 1) households will be provided with information about climate change and on-the-farm adaptation strategies, and 2) a sub-set of households in the treatment and control groups will be offered off-the-farm adaptation opportunities. This design will allow us to identify the effect of relaxing information frictions about climate change on agricultural practices and on employment in non-agricultural sectors. It will also allow us to examine whether the returns to off-the-farm opportunities are higher for households facing less information frictions. The RCT rolls out over three agriculture seasons starting April 2025. The last final follow up is expected in September 2026. The impact evaluation is funded by the HAPIE initiative from Egypt Impact Lab and J-PAL MENA. 

In the course of implementing a BDM mechanism to allocate job-matching services in a field experiment in Egypt on young farmers’ adaptation to climate change, we identified a potential concern that has received little attention in the literature: collusion between enumerators and participants. Such collusion could arise for several reasons. For instance, participants who do not win the auction in the BDM may regret bidding too low and request that enumerators increase their reported willingness to pay in order to buy the service. Conversely, participants who are unable to pay for the service after winning the auction may ask enumerators to adjust their reported willingness to pay in order not to buy the service. Enumerators may have incentives to accommodate such requests, for example, to reduce the likelihood of reporting payment failures or disputes to supervisors, or simply to ease interactions with participants. Such collusion can bias estimated willingness to pay and compromise the identification strategy in selective trials. This project investigates the extent and implications of this issue. First, we conduct a systematic review of the studies using a BDM mechanism to document whether and how the possibility of collusion is discussed. Second, we propose a statistical test to detect collusion. Third, we implement the test across all studies using BDM with publicly available data. Fourth, we introduce a BDM protocol that is robust to collusion between enumerators and participants, and implement it in our field experiment in Egypt. Finally, we show how in our field experiment estimates on the impacts of the job-matching services would have differed had we not implemented this robust protocol.

Smallholder corn farmers in Upper Egypt face growing exposure to climate change, such as heat waves during critical growth stages, and shifting pest and disease pressure, threatening yields and livelihoods. This project builds on a prior RCT showing that half-day climate-adaptation awareness sessions, provided in collaboration with the Ministry of Agriculture, foster farmers’ adaptation efforts. This project evaluates whether adding a WhatsApp-based AI advisory chatbot can help provide the more frequent, individualized support requested by farmers. In collaboration with the LemonTree Lab, we developed a WhatsApp-based chatbot grounded in content validated by local agronomists and extension agents. Specifically, we test whether providing access to this chatbot on top of the existing half-day session fosters the adoption of climate-smart agricultural practices, yields, and farmer resilience. To study selection patterns, we randomize farmers into the status quo sessions, or into sessions topped up with an opportunity to get the chatbot through a Becker-DeGroot-Marschak mechanism. Results are expected to inform the Ministry of Agriculture’s extension strategy and the design of low-cost, scalable AI-based climate adaptation tools for smallholder farmers.

The consequences of climate change for agricultural livelihoods in developing countries are well-documented. Yet far less is known about how individual farming households perceivethese long-run risks, and how such perceptions shape how much to invest in the human capital of the next generation, and through which pathway. Egypt represents a compelling setting where climate change is characterized by slow-onset processes, gradual shifts in temperature, that will progressively and substantially increase the riskiness and reduce the profitability of agriculture, particularly for smallholders who lack the resources to adapt. Using the experiment of (Gazeaud & Rostom, forthcoming) in which farmers are provided with information about how climate will change in the next 15 years and we introduce ways to adapt now, and in the future. We aim to identify the effects of relaxing information friction about climate change on (i): Changing perceptions returns from agriculture in the future and the occupational risks their children may face (iii) Perception of children’s value of education in formal school enrollment, but also through on the job-vocational training and (iii) changes their plans for future expenditures and their children’s time allocation between schooling, farm labor and farm or non-farm labor.