Policy Brief Issue 31 WORKING UNDER PRESSURE: IMPROVING LABOR PRODUCTIVITY GENDER INNOVATION LAB THROUGH FINANCIAL INNOVATION The Gender Innovation Lab (GIL) Authors: Eliana Carranza, Aletheia Donald, Florian Grosset, and Supreet Kaur1 conducts impact evaluations of development interventions in Sub-Saharan Africa, seeking KEY MESSAGES to generate evidence on how • In developing countries, financial transfers within social and kin to close the gender gap in networks are ubiquitous and frequent. Though these transfers have social earnings, productivity, assets and agency. The GIL team benefits, pressure to redistribute income can introduce a disincentive to work is currently working on over by reducing the payoff of exerting effort. This comes at a potential cost for the 50 impact evaluations in 21 overall efficiency of the economy. countries with the aim of • We developed a financial innovation to study the impact of this building an evidence base with lessons for the region. redistributive pressure on workers’ labor supply and productivity. This innovation, a direct-deposit commitment savings account, enabled workers to The impact objective of GIL is convert productivity increases into private savings which cannot be accessed increasing take-up of effective by others. policies by governments, development organizations • In the first phase of our project, workers offered the direct-deposit and the private sector in order commitment savings account increased their labor productivity and to address the underlying earnings by 10 percent, which translates into an 18 percent increase causes of gender inequality for workers who opened an account. The effect appears to be driven by in Africa, particularly in terms of women’s economic and workers increasing effort while on the job. social empowerment. The lab • Preliminary results show that the visibility of an account to one’s social aims to do this by producing network and the degree of redistributive pressure a worker faces are and delivering a new body of evidence and developing a strong determinants of account take-up. This suggests that tackling the compelling narrative, geared underlying cause for redistributive norms—the lack of consumption smoothing towards policymakers, on mechanisms—could improve output and growth in developing countries by what works and what does addressing the root cause of the high demand for commitment savings products. not work in promoting gender equality. Aletheia Donald (corresponding author: adonald@worldbank.org), Eliana Carranza (Jobs Group, World Bank), 1  Florian Grosset (Columbia University), Supreet Kaur (University of California, Berkeley). http://www.worldbank.org/en/programs/africa-gender-innovation-lab MOTIVATION HERE’S WHAT WE DID Social and family networks play an important role In our study, we build on existing evidence— as informal safety nets in Sub-Saharan Africa and currently overly reliant on observational studies—by other development contexts, with relatives, friends, constructing an explicit experimental test for the effects colleagues, and neighbors sharing financial resources. of redistributive pressure on the labor productivity of While these arrangements have important private and women. Rather than focusing on the evasion strategies social benefits—for example, by enabling risk sharing adopted by individuals, we examine forced redistribution among kin groups—they may potentially hinder as an impediment to increased productivity in the context of formal off-farm employment. economic transformation, labor productivity and output growth. For instance, individuals who take up more stable and higher-wage employment in the modern IMPLEMENTATION CONTEXT sector might face increased pressure to redistribute We partnered with the Caisse Nationale des Caisses income to their social networks. Knowing they will not d’Épargne (CNCE), the largest savings bank in Côte be able to keep their full earnings, those individuals may d’Ivoire, and OLAM, a leading transnational agro- find it more profitable to stay in the traditional sector or processing firm, to test a financial innovation among become less motivated to work hard and increase their workers in cashew-processing plants: a direct-deposit earnings. commitment savings account designed to make it easier for workers to convert productivity increases While previous work provides tentative evidence into long-term savings which cannot be accessed by that redistributive pressures can have large negative others. The NGO Innovations for Poverty Action (IPA) economic impacts on both men and women (Boltz et al. was contracted by the World Bank to coordinate 2016; di Falco and Bulte 2011; Goldberg 2016; Jakiela the implementation and conduct surveys with study and Ozier 2015; Squires 2017), the consequences participants. The account allows workers to freeze might be particularly important for women due to their for nine months a portion of their income above their lower social and political standing—and thus higher risk current average earnings. This design feature leaves of financial expropriation. For example, Anderson and workers with the same level of disposable cash for daily Baland (2002) examine participation in rotating savings expenses and redistribution to others on average, but and credit associations (ROSCAs) in urban Kenya and makes it more likely that any increases in productivity find that women make a strategic decision to tie up their are retained by workers for their own future use. savings in a ROSCA to prevent their family from taking We chose to implement the program with cashew factory the money. In Chile, Kast and Pomeranz (2014) show workers for several reasons. First, cashew processing that individuals who are not heads of household or who is a crucial sector in Côte d’Ivoire. Although the country face demands from their social network to lend money is one of the largest producers of cashew nuts in the were most likely to take up an offered individual savings world, it misses out on income from value addition account. More broadly, the micro-entrepreneurship by processing only a small proportion of cashews literature finds that anticipation that a share of their domestically, due in part to high unit-labor costs income might be captured reduces women’s incentive stemming from low labor productivity. Second, workers to invest and expand their economic activity (Campos in cashew processing are paid piece-rate based on and Gassier 2017). their individual daily productivity, meaning that changes FIGURE 1: PROBLEMS SAVING MONEY “I have difficulty saving for large goals because if I put money aside, someone else will ask for it.” Strongly Agree 55% Somewhat Agree 20% Do Not Agree 23% Not Sure/ Don’t Know 2% 0 10% 20% 30% 40% 50% 60% in their productivity are easy to observe, and changes in INTERVENTION ROLL-OUT effort translate directly into worker earnings and factory Of the 107 eligible workers in the factory with regular output. Lastly, cashew processing is a key industry for attendance, about half were randomly selected for the increasing women’s economic empowerment, with the program. While workers in the control group continued vast majority of workers involved in manual processing to receive their earnings in cash, workers in the treatment tasks being women. group only received wages under a certain cutoff in cash, and all excess wages over the cutoff as a direct deposit Workers in our setting—often some of the few people into their commitment savings accounts. Workers had in their social network to be employed in the formal the opportunity to choose their cutoffs to correspond sector—expressed facing heavy demands to share to their budgeting needs. Because we were interested earnings with others. For example, 83 percent of workers in observing productivity effects on earning increases, reported giving money to someone multiple times a workers could choose an earnings threshold equal to or month, and 75 percent of workers agreed or strongly greater than their average historic fortnightly earnings. agreed with the statement that redistributive pressures Take the example of a woman who earns $40 per fortnight make it difficult for them to accumulate savings (Figure on average and sets her threshold at $45. She proceeds 1). Although 78 percent of workers strongly agreed with to earn $30 in her next paycheck, and $50 in the one wanting to save more than they currently save, none of after. In this case, she would receive the entire $30 in them had an account in the formal banking sector. The cash for the first paycheck (no deposit) and would receive most popular savings vehicle was an informal rotating $45 in cash ($5 deposit) the fortnight after. savings and credit association (ROSCA), followed by mobile money and keeping money at home. Workers Workers discreetly received receipts confirming that the also felt discouraged to increase their effort at work deposit had been made so that the program remained by their inability to save, though 86 percent reported private, meaning that only the worker would know how wanting to be more productive. much was saved in her account. The account pays out deposited savings in a lump sum after 9 months (at HERE’S WHAT WE FOUND a private date). During the blocked period, deposits Nine months post-intervention, workers who were cannot be accessed—making it possible to credibly offered the direct-deposit accounts (treatment group) deny transfer requests. Once the blocked period ends, increased their earnings by 9.9 percent on average workers can use the saved funds or leave them in the relative to workers who were not offered the accounts (now liquid) bank account. (control group). This estimate implies a remarkably high Selected workers were invited to participate in a brief effect on earnings of 18 percent among workers who sensitization session in which a representative from the actually opened an account. bank ensured that workers understood the program and We do not find significant effects on attendance, how the accounts work. Afterwards, a short individual suggesting that treatment effects are driven by workers survey was administered to gauge each worker’s increasing effort while at the factory. In addition, it comprehension of key program features. Interested appears that rates of job quitting do not differ between workers then met individually with team members from treatment and control workers, indicating that our IPA to choose an earnings threshold and to set up their intervention does not operate through constraining account. These individual sessions—and the continual workers to remain at the plant longer that they would availability of project team members to answer any have liked to stay. questions or concerns about the program—contributed to the fact that among workers who were offered the Moreover, we find that take-up of the commitment direct-deposit accounts, 55 percent decided to take accounts is concentrated among those who report them up. This is 9 percentage points higher than the higher redistributive pressure at baseline using two average take-up rate found by Knowles (2018) in a indicators: those who had any uninvited guests show meta-analysis of 13 randomized controlled experiments up to their house for a meal in the past 7 days— with interventions designed to remove the costs of indicating the presence of kin and network members opening and using a savings account. in close physical proximity—and those who report FIGURE 2: FORTNIGHTLY EARNINGS (FCFA) 20,000 18,000 16,000 14,000 12,000 10,000 8,000 6,000 42,000 2,000 0 Absent Intervention With Intervention FIGURE 3: RESPONDENT INTEREST IN PRIVATE VS PUBLIC ACCOUNT (N=44) 50 Yes 40 No 30 20 10 0 Private Public having to pay school fees for others—a common We will complement these interventions with a series of redistribution request. These individuals were 56 and tests to isolate redistributive pressure from one’s social 33 percentage points more likely to take up the direct- network as the driver of our results—versus competing deposit accounts, respectively. explanations such as workers wanting to constrain future temptation to spend their savings on themselves. Lastly, In addition, we asked a separate sample of workers if we plan to examine whether the impact of redistributive they would take up a blocked savings account if (i) its pressure on workers’ productivity is reduced when existence was private to themselves or (ii) its presence more people in their social network have jobs, which was advertised to their kin network. Figure 3 below would imply positive social externalities from increased shows that expressed demand for the direct-deposit employment. We will test this by randomizing job offers accounts plummets if workers expect that others for factory jobs and varying the fraction of individual in their kin networks will be aware of the account. members in the workers’ network that get a job offer. Moreover, when respondents were asked why they Our findings suggest that redistributive pressures might declined take-up of public blocked accounts, the most constitute a barrier to economic growth in sub-Saharan frequent response was wanting to keep the savings Africa, since new economic activities—from adopting for themselves—consistent with our hypothesis of the cash crops in agriculture to gaining employment in importance of redistributive pressure. modern sector jobs—can be undermined if individuals do not capture the returns from their efforts. This also NEXT STEPS indicates that tackling the likely underlying cause for Building on the above evidence, the team will offer the redistributive norms—the lack of consumption smoothing commitment accounts to a larger sample of workers mechanisms—could improve output and growth. For to see how our results hold up. We will also allow example, support provided by government social safety current workers with savings accounts to renew their net programs could be bolstered so that kin networks commitment period for another 9-month round and reduce pressure on employed family members to look at how effects change over the longer term. redistribute earnings. This could not only improve the welfare of the unemployed, but also boost the productivity of individuals with jobs by reducing their responsibility for redistribution. Similarly, improving access to formal health- or livelihood-insurance might have externality benefits by increasing the aggregate output of others in the same kin network. More broadly, our study will offer evidence relevant for a variety of sectors, from how to improve the viability and productivity of agro-processing to how financial instruments can increase labor productivity—and offer new solutions for policymakers seeking to empower women and lift workers out of poverty. The Jobs Group was created to support World Bank Group (WBG) client countries in the design and implementation of integrated, multi-sector, jobs strategies to reduce poverty and ensure inclusive growth. These jobs strategies articulate policies and programs which address three main challenges most countries face to varying degrees: creating jobs in the formal, private sector; improving the quality of informal jobs; and expanding access to (better) jobs for certain population groups (e.g., women, youth, the poor). The Jobs Group develops solutions to the main jobs challenges and measures the impact of these solutions on jobs outcomes. www.worldbank.org/en/topic/jobsanddevelopment FOR MORE INFORMATION, PLEASE CONTACT Markus Goldstein mgoldstein@worldbank.org afrgenderlab@worldbank.org 1818 H St NW This work has also been funded in part by The Umbrella Facility for Gender Equality (UFGE), a multi-donor trust fund administered by the World Bank to advance gender equality and women’s empowerment through experimentation and Washington, DC 20433 USA knowledge creation to help governments and the private sector focus policy and programs on scalable solutions with sustainable outcomes. The UFGE is supported with generous contributions from Australia, Canada, Denmark, Germany, Iceland, Latvia, the Netherlands, Norway, Spain, Sweden, Switzerland, United Kingdom, United States, and the Bill and Melinda Gates Foundation. The first draft of this policy brief was released in December 2018.