S alt H pe h e ci an al d ResearchDigest Fo N World Bank cu ut s riti on o n VOLUME 13 NUMBER 1 FALL 2018 The Critical Role of Health and Nutrition in Supporting Development H ealth matters for development: development, and jobs, this issue of IN THIS ISSUE Healthier children learn better the Research Digest features recent and are more likely to stay in World Bank studies that highlight the The Critical Role of Health and Nutrition school. Healthier adults live longer role of health and nutrition in building in Supporting Development … page 1 and are more productive. human capital and alleviating poverty. Recent decades have seen advances One study shows that childhood stunt- The Income Losses from Childhood in health care and in the availability ing caused by malnutrition lowers per Stunting—and the Returns to Nutrition of food. Yet poor health and nutrition capita income in developing countries Programs Aimed at Reducing It … page 2 persist as major hurdles to develop- by an average of 7 percent. Another Measuring Physical Activity to ment. Half the world’s population lacks study shows that malaria testing and Investigate the Link between Health access to basic health services, and treatment can increase the productiv- and Productivity … page 3 every year nearly 100 million people in ity of adult agricultural workers by as developing countries are pushed into much as 25 percent. Yet another finds … page 4 Bed Nets and Politics in Tanzania extreme poverty because of health ex- that the distribution of bed nets in penses. According to the World Health Tanzania improved public trust in pol- Investigating the Joint Effects of Organization’s World Health Statistics icy makers—spotlighting a crucial fac- Parenting and Nutrition on Child 2018, nearly 15,000 children under tor in the successful take-up of health Development … page 5 age five die every day. Malnutrition re- and other development programs. mains a chronic problem even in coun- Other studies featured in this issue How Do Targeted Cash Transfers Affect tries where food is abundant. highlight innovative approaches to Nutrition among Nonbeneficiary Health and nutrition are at the improving children’s nutrition status Children? … page 6 center of the World Bank Group’s long- as well as diagnosing and treating standing strategic support for human preventable threats to maternal and Cities, Slums, and Child Nutrition capital development. This includes the child health. And the last two offer in Bangladesh … page 7 World Bank’s recently launched Human insights into overcoming financial and The Dynamics of Child Development in a Capital Project, designed to help infrastructure challenges in improving Very Low-Income Country … page 8 developing countries increase their access to health care. investments in people. And over the Continued progress on health and Can Offering People Gifts Increase Their period 2000–16 the World Bank invest- nutrition in developing countries will Use of Health Services? … page 9 ed $35 billion in health and nutrition require concerted and collective ef- initiatives while doubling its average forts that are inclusive of the poorest Looking into the Black Box of Performance- annual lending to the sector from $1.3 populations and adequately supported Based Financing … page 10 billion to $2.6 billion. by the stakeholders—donors, policy The importance of such efforts is makers, and the public. It also calls for The Importance of Good Roads in Access underscored by World Bank research. more research to support programs to Health Care … page 11 Mounting evidence shows that without focusing on such issues as universal investments in health countries cannot health coverage, early childhood de- Recent World Bank Research Publications sustain economic growth, prepare their velopment, and the management of on Health and Nutrition … page 12 workforce for the more highly skilled pandemics. For its part, World Bank re- jobs of the future, or compete effec- search will continue to provide rigorous tively in the global economy. analysis and insights aimed at helping Building on previous issues cov- countries improve access to quality ering such topics as education, skill health care and eliminate malnutrition. 2 World Bank ResearchDigest The Income Losses from Childhood Stunting—and the Returns to Nutrition Programs Aimed at Reducing It New estimates show that childhood The authors estimate the median age The results show an internal rate stunting, through its effects on of today’s workers using the distri- of return of doing so that averages 17 today’s workforce, lowers per capita bution of the current adult popula- percent, and a benefit-to-cost ratio tion by country. Drawing on the Joint averaging 15 to 1. The internal rate of income in developing countries by Malnutrition Estimate database pre- return varies significantly across re- an average of 7 percent pared by UNICEF, the World Health gions. It is highest in East Asia and the A Organization, and the World Bank, they Pacific (24 percent), reflecting the low n estimated 150 million chil- take as the average rate of under-five per capita program cost, the high rate dren under age five—one in five stunting the rate in the year in which of return to education, the high initial worldwide—are stunted (with a the median-age worker was two years GDP per capita, and the high GDP height-for-age more than 2 standard old. They then apply their development growth rate. And it is lowest in Sub- deviations below the median for a accounting methodology to compute Saharan Africa (15 percent), reflecting healthy reference population). As has the country-specific income penalty the high per capita program cost, the been extensively documented, stunt- associated with childhood stunting relatively low initial GDP per capita, ing in childhood is associated with among today’s workers. and the relatively low GDP growth rate, adverse outcomes throughout the life The results show that the income offset in part by the relatively high rate cycle. The undernourishment and dis- penalty from some of today’s work- of return to education. The estimates ease that cause stunting may impair ers having been stunted in childhood are robust to sensitivity analysis that brain development, leading to lower averages 7 percent of GDP per capita. doubles the program costs or that cognitive and socioemotional skills, The estimates vary significantly across halves the estimated benefits from the lower levels of educational attain- countries and regions, with the larg- reduction in stunting. ment, and thus lower incomes. est average penalties in Sub-Saharan Overall, the exercise is likely to rep- How large is the aggregate cost Africa and South Asia—around 10 per- resent a lower bound on the returns: at a societal level? And how big are cent of GDP per capita. a more complete cost-benefit analysis the potential rates of return to health The prevalence of stunting declines would capture the intrinsic value as- and nutrition interventions to reduce with economic growth, at an estimated sociated with the lower child mortality stunting? A recent study by Galasso rate of 1.5 percent a year, though not that would result from the nutrition and Wagstaff tackles these questions. fast enough to reach the Sustainable and health programs, as well as the The authors use a development ac- Development Goals target of a 40 per- externalities and potential other chan- counting framework, common in stud- cent reduction by 2025. Programs that nels of social returns that arise from ies that account for some or all of the tackle the most proximate determi- human capital. sources of differences across countries nants of stunting—that help provide in per capita income, to model how adequate nutrient intake (in utero and childhood stunting affects income in after birth) and that reduce exposure adulthood. They rely on micro-econo- to disease—have the potential to sig- metric estimates and factor in the ef- nificantly accelerate this decline. fects of childhood stunting on adult The second, forward-looking ex- wages through their effects on years of ercise looks at the potential rate of schooling, cognitive skills, and height return from such programs, aimed in adulthood, parsing out the relative at reducing stunting among today’s contribution of each set of returns to children. The authors rely on the esti- avoid double counting. They use this mated costs and estimated impact of a setup to perform two calculations, the package of 10 nutrition and health in- first for all developing countries, the terventions (including salt iodization, second for the 34 developing countries different forms of supplementation that together account for 90 percent of during pregnancy and in childhood, the world’s stunted children. complementary nutrition education, The first exercise looks at how and management of severe acute mal- much lower a country’s per capita nutrition). They use their framework to income is today as a result of some estimate the net present values of the Emanuela Galasso and Adam Wagstaff. 2018. of its workers having been stunted costs and benefits of scaling up this “The Aggregate Income Losses from Childhood in childhood. This is a backward- package of interventions to 90 percent Stunting and the Returns to a Nutrition Inter- looking exercise asking, in effect, coverage over a 10-year horizon in the vention Aimed at Reducing Stunting.” Policy what the costs are today of not hav- 34 developing countries accounting for Research Working Paper 8536, World Bank, ing eliminated stunting in the past. 90 percent of stunted children. Washington, DC. World Bank Research Digest 3 Measuring Physical Activity to Investigate the Link between Health and Productivity Treating Nigerian sugarcane cutters physical movement across the three For those tested and found to be for malaria leads to a shift from spatial dimensions and, in report- malaria positive, the shift out of sed- sedentary to physically active time ing, typically aggregate the numeric entary and light activity (again con- counts into minute-level aggregates ditional on working) is larger—about of about 90 minutes per workday of four levels of activity: sedentary and twice the estimated effect of simply P lightly, fairly, and very active. In addi- being offered testing and treatment. roductivity is a key determi- tion to recording physical activity, the Among malaria-positive workers, as ex- nant of economic growth, but study tested each worker for malaria pected, those treated for malaria show measuring it at the level of in- and prescribed effective treatment for higher productivity than those not yet dividual workers remains a challenge. those found to be malaria positive. treated. But an effect is also identi- Economists consider worker effort to The testing was rolled out over time, fied for workers who test negative be unobservable, and for good reason. with the order of testing randomized for malaria. Conditional on working, For physically demanding occupa- across workers. This enabled the study these heathy workers see a significant tions, however, direct measures of to explore not only the link between increase in active hours, largely from physical activity may serve as a valid physical activity and work output but a decline in sedentary time. A healthy proxy for productivity. If so, new wear- also the influence of malaria on physi- diagnosis appears to motivate workers able technologies that record physical cal activity and productivity. to work harder and be more active, at movement—accelerometers—might As would be expected, the level of least over the short run of the study provide individual measures of pro- physical activity in a day is strongly as- period (a surprising finding first identi- ductivity in settings where it would sociated with the worker’s decision on fied by the earlier study cited above). otherwise be unknown. whether to work that day. An addition- One challenge confronting the This possibility is probed in a al hour of light activity increases the study involves worker attrition and recent study by Akogun, Dillon, probability of a day of work by 10 per- compliance with its protocols: of the Friedman, Prasann, and Serneels. centage points, moderate activity by 83 workers initially assigned devices to In addition, the study builds on an 7.5 percentage points, and heavy activ- wear, 25 had lost them by the end of earlier one that estimates the pro- ity by 6 percentage points. Conditional the study. (This attrition did not seem ductivity costs of adult malaria infec- on working, moderate physical activity to lead to bias, as no observable vari- tion (Dillon, Friedman, and Serneels, is the level most associated with earn- ables predict the loss.) Such challeng- “Health Information, Treatment, and ings. Because the algorithm used by es with incorporating wearable devices Worker Productivity: Experimental the accelerometer to calibrate physi- into a survey are not uncommon; other Evidence from Malaria Testing and cal activity as heavy is akin to intense studies have found up to 40 percent Treatment among Nigerian Sugarcane physical exercise, the association of of individual-day data to be unusable. Cutters,” Policy Research Working cane cutting with light and moderate Clearly, further work on this issue is Paper 7120, World Bank, 2014). Both activity appears sensible. All told, the needed. But combining wearable tech- focus on a large sugarcane plantation study calculates that about 60 percent nologies with multitopic surveys offers in Nigeria. of the activity association with earn- promising potential. This setting is particularly suited to ings is driven by the extensive margin investigating the relationship between (that is, by supplying labor or showing labor productivity and physical activity up for work) and 40 percent by more because the output of the workforce is intense activity. directly measurable. Workers cut stalks Following the earlier study, the au- of cane, measured to a standardized thors then estimate the impact of sim- size, and are paid on a piece-rate ba- ply being offered malaria testing and sis. Both the plantation and the worker treatment; the impact of being tested, keep careful track of the quantity of found to be positive for malaria, and cane cut, so the productivity of the treated for the disease; and the impact workday is clearly observed. Finally, of being tested but found not to be workers are transported in and out of infected. Overall, testing leads to a the plantation, so all are on-site for shift from light to moderate activity. the same number of hours each day Conditional on working, the effects Oladele Akogun, Andrew Dillon, Jed Friedman, of work. are stronger, with a similar shift out Ashesh Prasann, and Pieter Serneels. 2017. A random subset of workers was of light activity to moderate or heavy “Productivity and Health: Alternative Productiv- assigned belt-mounted accelerometers activity on the order of 0.8 hours per ity Estimates Using Physical Activity.” Policy to wear for a six-week span over the workday. This shift in turn is linked to a Research Working Paper 8228, World Bank, study period. Accelerometers record 25 percent increase in output. Washington, DC. 4 World Bank ResearchDigest Bed Nets and Politics in Tanzania Tanzania’s distribution of bed nets, redeemable at the distribution point. These results are robust to alterna- though designed to fight malaria, More than 17 million nets were distrib- tive treatment bandwidths of 45, 60, also led to a short-term boost in the uted, at a total cost of $96.4 million, fi- 120, or 150 days as well as to the incor- nanced primarily by the Global Fund to poration of statistical controls (for age, popularity of local politicians Fight AIDS, Tuberculosis and Malaria. education, and household consump- M Estimating the effects of universal tion) and zonal fixed effects. In addi- any African countries have programs like this one is typically chal- tion, the analysis shows no effect from implemented public health lenging—because there is no natural hypothetical “placebo” campaigns 100 programs in the past decade comparison group with which the pro- or 200 days before or after the actual targeting major diseases such as HIV/ gram recipients can be compared. But campaign, further suggesting that the AIDS and malaria. Numerous studies the analysis in this case benefits from findings are not spurious. have examined the effect of these pro- an unusual coincidence: Tanzania was The study provides what is thought grams on health outcomes. But little fielding an unrelated national panel to be the first quasi-experimental evi- is known about their effect on politics. survey at the same time as the bed net dence of the relationship between a Indeed, how public programs af- campaign. This allows causal inference malaria prevention program and pub- fect politics has been a long-standing through a regression discontinuity de- lic opinion in a developing country. question in political economy. A criti- sign, in which respondents interviewed Moreover, it shows a large effect on cal assumption of democratic theory just before and just after bed net dis- public opinion despite a relatively low is that voters reward politicians who tribution are compared. program cost ($7.07 per delivered net choose good policies and deliver bene- The survey team interviewed 3,200 in 2010). The results suggest that for ficial programs. There is a large empiri- households in 2010  –11, with the some types of health programs there cal literature, focused chiefly on devel- survey visits randomized into an ear- may be clear compatibility between oped countries, that examines whether lier cycle and a later one. Household the interests of political leaders and voters reward politicians who deliver members were asked their opinions those of their constituents. In quantify- economic growth. In developing coun- about local public officials, and these ing the political benefits of one such tries, however, many political economy opinions are used as the outcome vari- program in Tanzania, the study may models suggest that these account- ables for the study. Meanwhile, as the point the way to a broader research ability relationships are weaker. Causal survey was taking place, the bed net agenda on the political returns to de- evidence on the relationship between distribution campaign was occurring livering lifesaving interventions. More public programs and public opinion in phases by zone, starting in southern work is needed to estimate the politi- and voting is sparse. While an emerg- Tanzania and reaching new zones every cal benefits of other critical interven- ing literature shows that incumbent one to two months. Thus the study tions, such as providing basic primary politicians sometimes benefit from can exploit the variation in possession health care services or antiretroviral the introduction of conditional cash of a free bed net driven by whether treatment for HIV/AIDS—interven- transfer programs in Latin America and the distribution campaign happened tions that are high on the global health Southeast Asia, there has been little before or after an individual was sur- agenda but too often underprovided in evidence on the political effects of veyed. Households interviewed after developing countries. health programs in Africa. the campaign were 30–  40 percentage To fill this gap in the literature, a points more likely to own a treated bed recent paper by Croke looks at a ma- net than those interviewed before the jor health campaign in Tanzania, the campaign. 2010–11 Universal Coverage Campaign. Receipt of the free bed nets led to This program distributed long-lasting significant changes in recipients’ polit- insecticide-treated bed nets to every ical opinions. Approval ratings for the household in the country, for each village chairman, the ward councilor, sleeping space not already covered the ward executive, and the local mem- with a net. The bed net manufacturer ber of parliament increased by 7–13 and NGOs handled distribution and percentage points, from base approval logistics. Ward and village executives ratings of 70–80 percent. The effect was identified and supervised four com- larger for households interviewed with- munity volunteers in each village, who in 30 days of the bed net distribution, over a five-day period visited each with the effect gradually fading over a Kevin Croke. 2017. “The Impact of Mass Bed Net home and registered each sleeping period of several months. And it was Distribution Programs on Politics: Evidence from space. The volunteers issued coupons strongest in districts with the highest Tanzania.” Policy Research Working Paper 7963, for each space that warranted a net, prevalence of malaria. World Bank, Washington, DC. World Bank Research Digest 5 Investigating the Joint Effects of Parenting and Nutrition on Child Development A study in rural Cambodia looks at families, collected in rural Cambodia one of the important explanatory fac- the extent to which parenting can in 2016 and 2017. The child survey tors of cognitive differences associated offset the adverse effects of poverty comprises a set of cognitive tests (lan- with family wealth. The measures of guage, early numeracy, and executive parenting behaviors are strongly cor- on children’s development function) as well as anthropometric related with children’s performance C measures (height and weight). The (even after controlling for other fac- hildren in poor families are battery of child tests was developed in tors), accounting for about 8  –14 per- more likely to experience early cooperation with researchers from the cent of the gap in cognitive competen- stunting, poor health, or lack of Measuring Early Learning and Quality cies observed among three-to-five- educational opportunities—and, as Outcomes project and the local survey year-olds. Yet the size of the parenting a result, less likely to achieve their firm. The caregiver survey includes effect is relatively small: a large (1 developmental potential. Indeed, fam- basic socioeconomic characteristics, standard deviation) improvement in ily poverty has been shown to have a questions about the child’s socio- the main parental competencies—an wide range of negative effects on child emotional development (Strengths improvement beyond the scope of development in many countries, with and Difficulties Questionnaire), and most parenting programs—would gradients in child development due to 25 questions about parental practices compensate for only about half the ef- family wealth evident as early as four (measuring cognitive, emotional, and fect of stunting. months of age and growing through- “negative” parenting). This survey thus Fourth, the positive effect of parent- out childhood. enables the study to distinguish the ing is generally stronger for children Developing effective interventions role of socioeconomic characteristics who are not stunted. This finding is to reduce such inequalities requires and that of specific parental practices important for two reasons: It shows the understanding the mechanisms that in explaining children’s development. importance of designing and testing account for the correlation between The analysis reveals four main find- integrated programs for child devel- family wealth and child development. ings. First, there is a substantial cogni- opment that anticipate the complex Yet these mechanisms remain imper- tive gap between children from differ- interactions that will arise between fectly described. They likely include ent backgrounds in rural Cambodia child characteristics and environmen- a range of factors, including health (confirming earlier findings). In the tal stimuli. And it contributes to the and nutrition as well as the quality of language test the children in the growing body of research documenting home environments. Perinatal health poorest quartile of families score 0.60 the variable effects of environments on and nutrition conditions are known standard deviations lower than those child development, by adding under- to affect later cognition. And parent- in the wealthiest quartile. The gap nutrition as a condition that may mod- ing quality has been shown to have is larger for cognitive competencies erate the effects of later environmental strong and long-term effects on child (language, early numeracy, executive stimuli on child development. development. But less is known about function) than for noncognitive or Encouraging stimulating and re- how parenting may help compensate socioemotional outcomes. sponsive parenting and improving for disparities in development associ- Second, the gap in child develop- children’s educational environments ated with poverty, particularly in low- ment widens with age (again consistent are key to reducing inequalities in cog- income countries. with earlier findings from Cambodia nitive outcomes and improving chil- Parenting quality may be lower and elsewhere). Between ages three dren’s performance during that critical among families living in poverty—as and five the gap widens especially in time when the cognitive gap widens. a result of the stress of living without language (from 0.31 to 1.12 standard Yet to fully compensate for initial adequate economic resources and the deviations) and early numeracy, but inequality, improvements in the mea- concomitant characteristics of poverty less so in executive function (from sured dimensions of parenting are not (such as low levels of parental educa- 0.37 to 0.54 standard deviations). enough; factors such as nutrition and tion)—which in turn contributes to the It also increases substantively for poverty are also important. These find- adverse effects of poverty on children. socioemotional outcomes. That the gap ings highlight the need for integrated Thus parenting quality has been posit- in executive function is already pres- interventions that address both par- ed to mediate the association between ent at age three and increases by only enting and early nutrition, especially family poverty and child development. 46 percent between ages three and five for the poorest families. This key hypothesis is investigated in a suggests that before age three children new study by Berkes, Raikes, Bouguen, are already experiencing substantial Jan Berkes, Abbie Raikes, Adrien Bouguen, and and Filmer. growth in executive function that is af- Deon Filmer. 2018. “Joint Effects of Parenting and The study is based on a unique fected by family socioeconomic status. Nutrition Status on Child Development: Evidence panel data set of almost 7,000 children Third, cognitively stimulating and from Rural Cambodia.” Policy Research Working (ages two to four in 2016) and their emotionally responsive parenting is Paper 8529, World Bank, Washington, DC. 6 World Bank ResearchDigest How Do Targeted Cash Transfers Affect Nutrition among Nonbeneficiary Children? By raising local food prices, targeted and suppliers. If the relevant market may contribute to the worsening of cash transfer programs can worsen is principally local and not fully inte- child growth is that of utilization spill- nutrition outcomes among children grated with the wider economy, the overs in the formal health care sys- presence of oligopolistic producers tem. Access to key maternal and child in nonbeneficiary households or, if the local market is competitive, a health services significantly increased A rising marginal cost curve of local pro- among beneficiary households, while id programs such as cash duction will translate the increase in the utilization of a few health services transfers—a type of assistance demand into higher prices. In addition, declined among nonbeneficiary moth- reaching 380 million people in any general equilibrium effect of a cash ers and children. Whether this decline developing countries—often introduce transfer should be magnified where a was due to an increase in financial or large amounts of money into small larger share of local households are convenience costs, or perhaps due to village economies. Such cash infu- beneficiaries. a decline in the perceived quality of sions can bring substantial benefits Using administrative data on pro- care, cannot be determined through to a village. But they may also raise gram rollout and food prices, the au- available data. The authors investigate local prices, which can have negative thors show that the Pantawid program and rule out a variety of other poten- consequences for household welfare, significantly raised the prices of perish- tial channels, including behavioral re- particularly among those not receiving able protein-rich foods while leaving sponses to the program by adults and the transfers—the nonbeneficiaries. the prices of other foods unaffected. older children that could result in a re- While an extensive economic lit- The price changes were largest in areas duced availability of caregivers, as well erature has considered both direct with the highest program saturation, as the possible influence of subgroup and indirect effects of cash transfer where the shock to village income was imbalance in baseline characteristics. programs, the study of possible local on the order of 15 percent, and they The results show that failure to market effects has focused on posi- persisted more than two and a half consider such local general equilib- tive channels that operate through the years after the program’s introduc- rium effects can lead to overstatement labor market or through informal in- tion. While beneficiary households are of the net benefit of targeted cash surance and credit markets. However, (likely more than) compensated for the transfers. But they also bring to the emerging evidence suggests that such price increases, this is not the case for fore the issue of program targeting. programs can generate significant nonbeneficiary households. Indeed, Like most cash transfer programs, the negative general equilibrium effects consistent with a negative real income Pantawid program is targeted to indi- that can detract from their antipov- shock, consumption of the now more- vidual households on the basis of a erty goals. In a recent paper Filmer, expensive protein-rich foods such as proxy-means test score. The authors Friedman, Kandpal, and Onishi test eggs and dairy declined among non- suggest that an alternative could be to for such local general equilibrium beneficiaries, including young children. offer the program on a universal basis effects on food prices and a range Concomitant with these price and in the subset of villages that are par- of outcomes for beneficiaries and consumption changes, the Pantawid ticularly poor or remote, an approach nonbeneficiaries. Their analysis uses program increased the prevalence of that would compensate all households the randomized evaluation of a large child stunting among nonbeneficiaries for any rise in local prices and thereby conditional cash transfer program by 11 percentage points while reducing avert increases in child stunting. in the Philippines, the Pantawid it among beneficiaries. The effects are Pamilya Pilipino Program. This pro- observed only for children who were gram provides cash transfers to poor in the vulnerable first 1,000 days of life households conditional on house- when the program was introduced and hold investments in child education not for older children. Moreover, where and health as well as use of maternal program saturation was higher, the health services. detrimental effects on nonbeneficiaries The additional income from cash were larger. These are not short-run ef- transfers can increase demand for nor- fects: the transfer program had been in mal goods by beneficiaries. Demand place for 31 months at the time of the for nutritious foods may increase even follow-up survey. The observed decline further because antipoverty programs in egg consumption associated with typically broadcast messaging on rec- higher prices alone explains 40 percent Deon Filmer, Jed Friedman, Eeshani Kandpal, and ommended child feeding practices. of the negative effect on the stature of Junko Onishi. 2018. “General Equilibrium Effects The price response to such an increase nonbeneficiary children. of Targeted Cash Transfers: Nutrition Impacts on in demand will depend on the mar- Besides the food price channel, Non-Beneficiary Children.” Policy Research Work- ket structure of producers another, complementary channel that ing Paper 8377, World Bank, Washington, DC. World Bank Research Digest 7 Cities, Slums, and Child Nutrition in Bangladesh Children’s nutrition status is differences between slum and non- are representative for urban and rural particularly poor in Bangladesh’s slum areas. areas indicate that mean child nutri- urban slums. Household wealth and Theory and evidence suggest that tion status is poorer in slum areas the study of slum health and nutri- than in rural areas. But it is the same the mother’s educational attainment tion should be treated as distinct in nonslum areas of large cities as it is are both important factors from the study of urban health or of in all urban areas, pointing to a cause U poverty and health. One argument for concern for children in these cities, rbanization offers countries for this is that the physical and social particularly for those in slum areas. opportunities to broaden and environments of slum settlements Child’s age, mother’s age at child’s boost welfare gains. But it also may amplify health risks for residents, birth, mother’s educational attain- poses development risks arising from particularly for young children, whose ment, and household wealth all have inadequate infrastructure and lack of immune systems are still developing. significant effects on height-for-age basic services. The tension between Moreover, there is scant literature in z-scores. Living in a nonslum neigh- the potential opportunities and risks this area, and research is needed to borhood area has a positive effect that is arguably greater for developing guide interventions related to develop- remains significant even after control- countries. One reason is that their ment, and particularly health, in slum ling for a range of factors. Access to already large urban populations are settlements. improved toilets that are shared with a expected to grow substantially in the To measure children’s nutrition large number of other households has coming decades. Another is that their status, the study uses height-for-age a negative effect in slum areas, while a governments typically have poor re- z-scores, which show how much a handwashing site with soap and water cords in designing and implementing child’s height at a particular age differs at the dwelling has a positive effect in policies and regulations and in provid- from the international reference medi- nonslum areas. ing services. an. Height is regarded as the most rel- The difference in mean height- Leveraging the potential socio- evant measure of child nutrition, and for-age z-scores between slum and economic benefits of urbanization, and child stunting (a height-for-age that is nonslum areas in 2013 is driven by mitigating the potential risks, have more than 2 standard deviations below differences in the mean levels of fac- been recognized as a global imperative the international reference median) as tors, particularly mother’s educational by the United Nations. This is reflected the key indicator for tracking progress attainment and household wealth. in, for example, the 2030 Agenda for in addressing child undernutrition. Similarly, the increase in mean height- Sustainable Development (adopted The study first investigates the ef- for-age z-scores between 2006 and in 2015) and the New Urban Agenda fects of child, maternal, household, 2013 in both slum and nonslum areas (2016), a key element of which is pro- and neighborhood-area factors on is driven by increases in the mean tecting and promoting the health of height-for-age z-scores, using the 2013 levels of factors, again particularly urban residents. survey data. It then extends its analysis mother’s educational attainment and Yet urban health and nutrition in in four potentially policy-relevant di- household wealth. developing countries remain under- rections: Using the 2013 data, it exam- Because the study is descriptive, studied. Rigorous empirical research ines the effects of the local availability its policy conclusions are necessarily has been constrained by a lack of of maternal and child health services circumspect. Policies, programs, and large-scale data that are representa- and of the use of these services; the partnerships aimed at boosting house- tive across and within urban areas in a effects of access to, or use of, potential hold wealth and women’s educational country and are extensive in capturing health-protective household amenities attainment are consistent with its potential determinants. One country related to cooking fuel, dwelling floor, results. So are initiatives to enhance for which such data are available is drinking water, sanitation, and hygiene; maternal and child health services and Bangladesh, thanks to a large-scale and the contributions of different fac- the physical environment in homes household survey, administered in tors to the slum-nonslum difference and neighborhoods, particularly in 2006 and 2013, that provides data on in mean height-for-age z-scores. And slum areas. adult and child health and nutrition using both the 2006 and 2013 data, it outcomes and a wide range of poten- examines the contribution of differ- tially relevant factors. Using data from ent factors to the increase in mean this survey, a recent study by Raju, height-for-age z-scores in slum and Dhushyanth Raju, Kyoung Yang Kim, Quynh Kim, Nguyen, and Govindaraj exam- nonslum areas. Thu Nguyen, and Ramesh Govindaraj. 2017. ines the nutrition status of children The study finds that mean child “Cities, Slums, and Early Child Growth: Empiri- under age five in the country’s major nutrition status is substantially cal Evidence from Bangladesh.” Policy Research cities, investigating patterns, trends, poorer in slum than in nonslum areas. Working Paper 8094, World Bank, Washington, and determinants and focusing on Comparisons with 2014 statistics that DC. 8 World Bank ResearchDigest The Dynamics of Child Development in a Very Low-Income Country Among a cohort of children in a cohort of children from the age of to influence wealth disparities in child Madagascar, wealth gaps in 0–3 years with follow-up assessments development in the short term, this cognitive outcomes translate at preschool age (3–6 years) and research suggests that programs that school age (7–10 years). improve home environments and early into similar gaps in later learning In addition, while the literature on child stimulation have the potential to outcomes child development in resource-poor close 10 –20 percent of the wealth gap. Y settings has focused on height, weight, A full assessment of specific policies ears before entering school, chil- or morbidity outcomes or on a single would require more in-depth knowl- dren living in poor families differ domain of child development (such as edge about the costs of such policies in cognitive and noncognitive receptive language), this study exam- as well as knowledge of how mothers abilities from those living in richer ines multiple domains of child devel- and families make decisions about the families. There is substantial evidence opment, including measures of cogni- inputs they provide for their children. that wealth gradients in child devel- tion, language, and executive function. Finally, the study shows that wealth opment start early, change over time, It also evaluates the cohort through gradients observed in child develop- and accumulate with age. early math and literacy measures at ment outcomes map closely onto But much of this evidence comes school age. The paper relies on both socioeconomic differences in learning from studies in high-income countries, the cross-sectional and longitudinal outcomes (math and literacy), with where the availability of longitudinal aspects of the data to document the differences between the richest and panel data enables researchers to critical timing of the emergence and poorest quintiles of about 1.3 standard follow the same cohort of children evolution of wealth differences in child deviations, even after controlling for over time while assessing repeated development and learning outcomes. lagged outcomes and initial endow- measures of child development out- The study documents substantial ments. The close mapping between comes at critical periods. While wealth gradients in school readiness out- school readiness and learning out- gradients in child development out- comes by household wealth even comes suggests that school readiness comes have also been documented in a very low-income setting. These measures are a reasonable way to as- in low- and middle-income countries, gradients, already evident when chil- sess children’s potential, even in set- the evidence in these countries has dren were 3–  4 years old, widen with tings with low school attendance. been based largely on cross-sectional increasing child age and flatten out data, covering different age groups at a by ages 9–10 for all domains of child single point in time. development. This has left important questions The magnitude and evolution of the largely unanswered: What are the tim- wealth gradient vary by domain, widen- ing and evolution of the wealth gaps ing earlier (between ages 3 and 6) for in child development outcomes in a receptive vocabulary and sustained at- very low-income setting? How do these tention tasks and later (between ages gaps change over time and accumulate 7 and 8) for cognitive composite and with age? Do the gaps and their evolu- memory of phrases tasks. The largest tion differ across different domains of wealth gradients are observed for vo- child development? And how do early cabulary and sustained attention and gaps in school readiness map into are robust to controlling for lagged learning outcomes? outcomes, maternal endowments, and A recent paper by Galasso, Weber, home environment. and Fernald takes up these questions. The study also assesses how much The paper is among the few examples of the wealth gap in child develop- of a longitudinal cohort study set in ment outcomes can be accounted for a very low-income country in Sub- by household inputs such as home Saharan Africa that has assessed stimulation. The results show that children through direct measures of home stimulation, as measured at cognitive development at two criti- the time of school age, accounts for cal stages of child development. The 12 –18 percent of the wealth gap (in Emanuela Galasso, Ann Weber, and Lia C. H. analysis draws on a nationally rep- z-scores) across the different domains. Fernald. 2017. “Dynamics of Child Development: resentative survey of nutrition and The largest effects are observed for Analysis of a Longitudinal Cohort in a Very Low child development in Madagascar receptive vocabulary and for memory Income Country.” Policy Research Working Paper (Enquête Anthropométrique et de of phrases, which is a verbal measure 7973, World Bank, Washington, DC. Also forth- Développement de l’Enfant), following of working memory. While it is difficult coming in World Bank Economic Review. World Bank Research Digest 9 Can Offering People Gifts Increase Their Use of Health Services? A study in Rwanda shows that health centers in randomly selected was introduced alongside the ongoing women are more likely to seek subdistricts. Facilities received this Performance-Based Financing Program timely pre- and postnatal care if they funding for a period of about two and for health facilities. Because all health a half years. They were given sugges- centers received performance-based expect to receive gifts for doing so tions on the content of the incentive payments, the effectiveness of provid- W packages, and ceilings were set on the ing these incentives for health provid- ith the aim of increasing the monetary value of the gifts for each ers cannot be compared with the effec- coverage of health care, a type of care. tiveness of providing incentives only growing number of low- and The results show that offering gifts for users of the health care system. It is middle-income countries have intro- as demand-side incentives can ef- possible that there were synergies be- duced programs of conditional re- fectively increase the use of targeted tween the incentives for facilities and wards for either providers or users of health services. The intervention led to those for users. While the incentives health services. Several Sub-Saharan an increase of 7.7 percentage points in for users improved indicators that were African countries have implemented the share of women who initiated pre- not affected by those for health provid- pay-for-performance programs at natal care within the first four months ers, health centers could independent- a national scale, while other coun- of their pregnancy and an increase ly provide women with gifts with the tries have piloted such programs or of 8.6 percentage points in the share aim of improving their performance on implemented them at a regional level. who received postnatal care in the 10 the targeted indicators. These programs typically give health days following delivery. The evalua- As a growing number of countries facilities financial rewards conditional tion finds no significant effect on the in Sub-Saharan Africa introduce or on the number of targeted services share of women who gave birth in a scale up pay-for-performance programs provided as well as the quality of care. health facility attended by a skilled for health providers, this study shows But if the policy goal is to increase the health provider. But the share of deliv- that conditional rewards for users can use of health services, independent eries taking place in such conditions enhance the effect of these programs of the quality of care, rewarding us- increased sharply nationwide, reaching on the use of health services. ers rather than providers can be more 95 percent at the time of the follow- effective. While providers can exert up survey. greater effort in expanding outreach The results are consistent with and improving services, it is users who earlier studies of programs using con- ultimately decide whether to go to a ditional transfers (cash or in-kind) to health facility. encourage the use of health services. A recent paper by Shapira, Kalisa, But what is remarkable is that most Condo, Humuza, Mugeni, and Walldorf eligible women reported not receiv- evaluates the effects of an intervention ing the gifts; health centers reported in Rwanda that offered gifts (in-kind frequently running out of the gifts be- conditional transfers) as an incen- cause of a lack of funds. What explains tive for the use of particular health the results? Women might have gone services. Specifically, the intervention to the health centers with the expecta- rewarded women with gifts for receiv- tion of receiving the gifts and learned ing timely pre- and postnatal care about the lack of stock only after re- as well as for giving birth in a health ceiving the services. In addition, the facility. The aim was to increase the intervention might have changed wom- use of these targeted services so as to en’s behavior by changing their percep- improve the diagnosis and treatment tions. For example, women might have of preventable threats to maternal and assigned greater importance to the neonatal health. timing of care, whether offered gifts or This demand-side incentive not, after observing that resources had strategy was introduced as part of been invested in rewarding early pre- the Community Performance-Based and postnatal care. And for women Gil Shapira, Ina Kalisa, Jeanine Condo, James Financing Program, itself an extension who did receive the gifts, this might Humuza, Cathy Mugeni, and Jeanette Walldorf. to the national Performance-Based have improved their general attitude 2017. “The Effects of In-Kind Demand-Side Con- Financing Program for health facilities. toward the health centers. ditional Transfers for Improving Uptake of Mater- The evaluation relies on an experi- One important aspect to keep nal and Child Health Services in Rwanda.” Policy mental design in which funding for the in mind is that the Community Research Working Paper 8060, World Bank, demand-side incentives was given to Performance-Based Financing Program Washington, DC. 10World Bank ResearchDigest Looking into the Black Box of Performance-Based Financing Performance-based financing led to centers)—were randomly assigned to and the additional financing group. positive results in Cameroon’s health four study groups for comparison. The Indeed, many (though not all) of the system. But was linking payments to first group received the standard PBF improvements measured for the first package (the PBF group); the second group were also observed in the sec- performance the critical factor? received the same level of financing, ond one—though few improvements C though not linked to performance, were observed in the group given en- onfronted with slow progress as well as the same levels of supervi- hanced supervision without additional toward health-related targets of sion, monitoring, and autonomy (the financing or financial incentives. Yet the Sustainable Development additional financing group); the third the comparison between the PBF and Goals, some countries have experi- received no additional resources or au- additional financing groups is delicate mented with results-based financing tonomy but the same levels of supervi- because the two share many similari- approaches in the health sector. These sion and monitoring; and the fourth ties: the same supervision and moni- aim to improve health systems and received no PBF components. toring mechanisms, the same level health outcomes through the use of The evaluation relied on two main of managerial autonomy, the same financial incentives, paid only after sources of data: a household survey level of increased financing. The only targeted results have been attained carried out before PBF was implement- difference is that in the PBF group and verified. ed (at baseline) and after it had been the additional financing was linked One such approach is performance- in place for two years (at endline) and to the performance of the individual based financing (PBF), a specific a facility-based survey also conducted facility while in the additional financ- supply-side intervention comprising at baseline and endline. ing group it was linked to the average a set of health system reforms meant Overall, the results indicate that performance of the PBF group facilities to increase the coverage and quality PBF in Cameroon is an efficient mech- in the same district. Among the man- of essential health services as well anism for channeling payments and agement and staff of health facilities, as equity and efficiency, often with a funding to providers. It led to signifi- this distinction might not have been special focus on maternal and child cant increases in utilization in the PBF salient enough to be reflected in how health. While models differ, all PBF group for several services (child and they modified their practices. programs involve the purchasing of maternal vaccinations, use of modern The weaker effects in the group of- health services using a predefined family planning), though not for oth- fered enhanced supervision but no list of services and prices. They also ers (such as prenatal care visits and additional financing or financial incen- include a strong verification system facility-based deliveries). It also im- tives suggest that reinforced supervi- that relies on systematic and detailed proved structural quality as measured sion is not enough to change behav- review of health facilities’ records as by staff presence, staff satisfaction, iors and improve outcomes. Additional well as community-level client tracking and the availability of equipment. financing appears to be required, and in which reported patients are asked a But despite an increase in provid- its impact seems in some instances to series of questions to confirm their re- ers and supplies available at health be stronger when linked to results. ceipt of health care. And many involve facilities, PBF did not increase the greater autonomy for health facilities. completeness of service provision dur- Is it possible to isolate the influ- ing prenatal care and child health con- ence on health outcomes of different sultations. Importantly, out-of-pocket components of the PBF approach— health spending (including unofficial such as explicit financial incentives payments) fell for households in the linked to results, additional resources PBF group, and this fall in revenue did available at the point of service de- not come at the cost of process qual- livery (not linked to performance), ity: there were no negative effects on and enhanced supervision, coaching, completeness of services and advice and monitoring? A recent paper by de provided during prenatal visits and Walque, Robyn, Saidou, Sorgho, and consultations for children under age Steenland presents the results of an five. Perhaps not surprisingly, given impact evaluation in Cameroon that the lower out-of-pocket costs and im- Damien de Walque, Paul Jacob Robyn, Hamadou sought to do just that. provements in structural and process Saidou, Gaston Sorgho, and Maria Steenland. For purposes of the evaluation, quality, client satisfaction increased for 2017. “Looking into the Performance-Based two types of primary health centers— consultations for children under five. Financing Black Box: Evidence from an Impact those with a medical doctor on staff Nevertheless, for many of these Evaluation in the Health Sector in Cameroon.” (medicalized health centers) and those outcomes there was no significant Policy Research Working Paper 8162, World without a doctor (integrated health difference between the PBF group Bank, Washington, DC. World Bank Research Digest11 The Importance of Good Roads in Access to Health Care Better roads would improve access disrupted during the rainy season, also of institutional issues relating to to health care in rural Madagascar— when rural roads frequently become health services. For example, they show by helping both patients and impassable. And about 25 percent of that poverty is among the most signifi- all health facilities are disconnected cant factors affecting people’s demand medicines reach health facilities (located more than 5 kilometers away) for health services. This suggests the A from the official road network. importance of ensuring financial afford- ccess to health care is an impor- A new paper by Iimi and Rajoela ability in the health sector, which might tant challenge in the rural areas analyzes this nexus of transport call for targeted subsidies, an expan- of many developing countries, connectivity, availability of medical sion of insurance mechanisms, free particularly in Africa. As the literature supplies, and people’s demand for health care programs in rural areas, or shows, the reasons for this generally health services in rural Madagascar some combination of these. And policy include a variety of constraints on by estimating a system of equations. coordination among different sector both the demand for and the supply of These are based on a hypothesis that ministries is essential. While health health services. people’s demand for health services sector reforms need to be advanced to On the demand side, people in ru- may depend on road connectivity, their ensure that health care is affordable ral areas are often much poorer than income level, and the availability of and the supply of medicines reliable, those in urban areas. They are also medicines, while supply conditions, developing and maintaining transport less likely to have health insurance. notably the availability of medicines, connectivity to priority health facilities And many rural people do not have may depend on road connectivity. By is also important. good access to the road network, let shedding light on the critical question alone to health facilities. Therefore, of what determines people’s access to all else being equal, the demand for health care, the analysis is designed to health services in rural areas tends to help inform policy makers about ways be relatively low. to address the issues through a multi- On the supply side, the availability sectoral approach. The analysis com- of doctors and nurses is often biased bines micro and spatial data from dif- toward urban areas. As a result, ru- ferent sources, such as poverty maps, ral doctors often have to see more road network inventory data, and the patients than urban doctors do. And country’s health management informa- because of poor transport connectivity, tion system. medical supplies may be inadequate in The results show that transport rural and remote areas. Both these fac- connectivity is essential on both the tors can often compromise the level of demand and the supply side. The dis- health services provided in rural areas, tance from villages to health facilities and this may further weaken rural de- is found to be an important deter- mand for health care. minant of people’s access to health In Madagascar both demand- and care. More people could benefit from supply-side constraints have long health services if they lived closer to hampered people’s access to health a health facility. Similarly, transport care in rural areas. Madagascar is one connectivity between health facilities of the poorest countries in the world, and the nearest district capital affects with a GDP per capita of about $400. the availability of medical supplies. More than three-quarters of its rural Larger stocks of medicines are avail- population lives under the national able in health facilities that are better poverty line. In addition, most of the connected to district capitals, which rural road network is in poor condition. play an important part in the country’s According to the Rural Access Index, medical supply network. The better a measure of rural transport acces- availability of supplies can in turn sibility, only 11.4 percent of the rural stimulate people’s demand for health population has access to the official services: more people are willing to road network in good condition. About visit health facilities when those facili- Atsushi Iimi and Voahirana Hanitriniala Rajoela. half of villages (fokontany) are located ties are equipped with sufficient medi- 2018. “Transport Connectivity, Medical Supplies, more than 10 kilometers away from the cal supplies. and People’s Health Care Access: Evidence from nearest basic health center. Moreover, The results highlight the importance Madagascar.” Policy Research Working Paper the medical supply chain is often not only of physical infrastructure but 8487, World Bank, Washington, DC. 12World Bank ResearchDigest Recent World Bank Research Publications on Health and Nutrition Journal Articles Neelsen, Sven, and Owen O’Donnell. 2017. “Progressive Based Motivations in Health Care.” Policy Research Universalism? The Impact of Targeted Coverage on Working Paper 8338, World Bank, Washington, DC. Booysen, F. L. R., E. Wouters, D. de Walque, and M. Over. Health Care Access and Expenditures in Peru.” Health 2017. “Mutual HIV Status Disclosure Is Associated with Bossavie, Laurent, Harold Alderman, John Giles, and Cem Economics. First published online February 16, 2017. Consistent Condom Use in Public Sector ART Clients in Mete. 2017. “The Effect of Height on Earnings: Is Stat- Free State Province, South Africa: A Short Report.” AIDS Nguyen, H. T. H., S. Bales, A. Wagstaff, and H. Dao. 2017. ure Just a Proxy for Cognitive and Non-Cognitive Skills?” Care 29 (11): 1386–90. “Getting Incentives Right? The Impact of Hospital Capi- Policy Research Working Paper 8254, World Bank, Wash- tation Payment in Vietnam.” Health Economics 26 (2): ington, DC. Briceño, Bertha, Aidan Coville, Paul Gertler, and Sebastian 263–72. Martinez. 2017. “Are There Synergies from Combining Cawley, John, Damien de Walque, and Daniel Grossman. Hygiene and Sanitation Promotion Campaigns? Evidence Ozier, Owen. 2018. “Exploiting Externalities to Estimate 2017. “Effect of Stress on Later-Life Health: Evidence from a Large-Scale Cluster-Randomized Trial in Rural the Long-Term Effects of Early Childhood Deworming.” from the Vietnam War Draft.” Policy Research Working Tanzania.” PloS One 12 (11): e0186228. American Economic Journal: Applied Economics 10 (3): Paper 8063, World Bank, Washington, DC. 235–62. Bundy, D. A. P., L. J. Appleby, M. Bradley, K. Croke, T. D. Dizon, Felipe, and Anna Herforth. 2018. “The Cost of Nutri- Hollingsworth, R. Pullan, H. C. Turner, and N. de Silva. Reichert, Arndt, and Harald Tauchmann. 2017. “Work- tious Food in South Asia.” Policy Research Working 2018. “100 Years of Mass Deworming Programmes: force Reduction, Subjective Job Insecurity, and Mental Paper 8557, World Bank, Washington, DC. A Policy Perspective from the World Bank’s Disease Health.” Journal of Economic Behavior and Organization Fitzpatrick, Anne, and Rebecca Thornton. 2017. “The Effects Control Priorities Analyses.” Advances in Parasitology 133: 187–212. of Health Insurance within Families: Experimental 100: 127–54. Shen, Gordon, Ha Thi Hong Nguyen, Ashis Das, Jumana Evidence from Nicaragua.” Policy Research Working Coarasa, Jorge, Jishnu Das, Elizabeth Gummerson, and Qamruddin, and Jed Friedman. 2017. “Incentives to Paper 8115, World Bank, Washington, DC. Asaf Bitton. 2017. “A Systematic Tale of Two Differing Change: Effects of Performance-Based Financing on Haque, Sabrina S., George Joseph, and Nazia Moqueet. Reviews: Evaluating the Evidence on Public and Private Health Workers in Zambia.” Human Resources for Health 2017. “Does Arsenic-Contaminated Drinking Water Limit Sector Quality of Primary Care in Low and Middle Income 15 (1): 20–27. Early Childhood Development in Bangladesh?” Policy Countries.” Global Health 13 (1): 24. Published online Sylvia, Sean, Hao Xue, Chengchao Zhou, Yaojiang Shi, Hong- Research Working Paper 8172, World Bank, Washing- April 12, 2017. mei Yi, Huan Zhou, Scott Rozelle, Madhukar Pai, and ton, DC. Daniels, B., A. Dolinger, G. Bedoya, K. Rogo, A. Goicoechea, Jishnu Das. 2017. “Tuberculosis Detection and the Martinez, Sebastian, Sophie Naudeau, and Vitor Pereira. J. Coarasa, F. Wafula, N. Mwaura, R. Kimeu, and J. Das. Challenges of Integrated Care in Rural China: A Cross- 2017. “Preschool and Child Development under Extreme 2017. “Use of Standardised Patients to Assess Quality Sectional Standardized Patient Study.” PLoS Med 14 Poverty: Evidence from a Randomized Experiment in of Healthcare in Nairobi, Kenya: A Pilot, Cross-Sectional (10): 1–20. Rural Mozambique.” Policy Research Working Paper Study with International Comparisons.” BMJ Global Tiwari, S., H. G. Jacoby, and E. Skoufias. 2017. “Monsoon 8290, World Bank, Washington, DC. Health 2 (2): 1–11. Babies: Rainfall Shocks and Child Nutrition in Nepal.” Mete, Cem, Laurent Bossavie, John Giles, and Harold Alder- Das, Ashis Kumar, Jed Friedman, and Eeshani Kandpal. Economic Development and Cultural Change 65 (2): man. 2017. “Is Consanguinity an Impediment to Improv- 2018. “Does Involvement of Local NGOs Enhance Public 167–88. ing Human Development Outcomes?” Policy Research Service Delivery? Cautionary Evidence from a Malaria- Wafula, Francis, Amy Dolinger, Benjamin Daniels, Njeri Working Paper 8074, World Bank, Washington, DC. Prevention Program in India.” Health Economics 27 (1): Mwaura, Guadalupe Bedoya, Khama Rogo, Ana Goicoe- 172–88. Milazzo, Annamaria, and Dominique van de Walle. 2018. chea, Jishnu Das, and Bernard Olayo. 2017. “Examining “Nutrition, Religion, and Widowhood in Nigeria.” Policy Das, Jishnu, Liana Woskie, Ruma Rajbhandari, Kamran the Quality of Medicines at Kenyan Healthcare Facilities: Research Working Paper 8549, World Bank, Washing- Abbasi, and Ashish Jha. 2018. “Rethinking Assumptions A Validation of an Alternative Post-Market Surveillance ton, DC. about Delivery of Healthcare: Implications for Univer- Model That Uses Standardized Patients.” Drugs: Real sal Health Coverage.” The BMJ. Published online May World Outcomes 4 (1): 53–63. Raju, Dhushyanth, and Ritika D’Souza. 2017. “Child Undernu- 21, 2018. trition in Pakistan: What Do We Know?” Policy Research Wagner, Zachary, Erick Gong, Damien de Walque, and Working Paper 8049, World Bank, Washington, DC. Do, Quy-Toan, Shareen Joshi, and Samuel Stolper. 2018. William H. Dow. 2017. “The Impact of Positive Income “Can Environmental Policy Reduce Infant Mortality? Shocks on Risky Sexual Behavior: Experimental Evidence Sato, Ryoko, and Yoshito Takasaki. 2018. “Psychic vs. Evidence from the Ganga Pollution Cases.” Journal of from Tanzania.” AIDS and Behavior 21 (3): 650–54. Economic Barriers to Vaccine Take-Up: Evidence from a Development Economics 133: 306–25. Field Experiment in Nigeria.” Policy Research Working Woolcock, Michael. 2018. “Enhancing Public Health Paper 8347, World Bank, Washington, DC. Kim, Young Eun, Wilma A. Stolk, Marcel Tanner, and Fabrizio Outcomes in Developing Countries: From Good Policies Tediosi. 2017. “Modelling the Health and Economic and Best Practices to Better Implementation.” Scandina- Shapira, Gil, Ina Kalisa, Jeanine Condo, James Humuza, Impacts of the Elimination of River Blindness (Onchocer- vian Journal of Public Health 46 (S22): 10–18. Cathy Mugeni, Denis Nkunda, and Jeanette Walldorf. ciasis) in Africa.” BMJ Global Health 2 (2). 2017. “Effects of Performance Incentives for Commu- Miller, Rosalind, Jishnu Das, and Madhukar Pai. 2018. “Qual- Working Papers nity Health Worker Cooperatives in Rwanda.” Policy Research Working Paper 8059, World Bank, Washing- ity of Tuberculosis Care by Indian Pharmacies: Mystery Banuri, Sheheryar, Philip Keefer, and Damien de Walque. ton, DC. Clients Offer New Insights.” Journal of Clinical Tubercu- 2018. “Love the Job…or the Patient? Task vs. Mission- losis and Other Mycobacterial Diseases 10: 6–8. Papers can be downloaded at www.worldbank.org/dec. The World Bank Research Digest is a quarterly publica- The Research Digest is financed by the World Bank’s Editorial Committee: Shanta Devarajan, David tion disseminating findings of World Bank research. Research Committee and managed by DECSO, the Rosenblatt, Aslı Demirgüç-Kunt, Roberta Gatti, The views and interpretations in the articles are those Strategy and Operations unit of the Development Hiau Looi Kee, Chorching Goh, and Shiva S. Makki of the authors and do not necessarily represent the Economics Senior Vice Presidency (DEC). The Research (managing editor). Editor: Alison Strong; production: views of the World Bank, its Executive Directors, or the Digest is not copyrighted and may be reproduced with Roula Yazigi. For information or free subscriptions, countries they represent. appropriate source attribution. send email to research@worldbank.org or visit http://www.worldbank.org/researchdigest. The World Bank 1818 H Street, NW Washington, DC 20433, USA Printed on Recycled Paper