HD Update Quarterly Q3 2010 September 2010

Human Development Quarterly Update Q3 2010 Latest research findings

New articles and books

Research in the news

And on the blogs

Linking research and operations

References

Previous issues

Latest research findings The World Bank is a “knowledge bank”

he World Bank has produced a huge volume of books and papers on development—20,000 publications spanning decades, but growing appreciably since 1990. Martin Ravallion and Adam Wagstaff [1] find evidence

that many of these publications have influenced development thinking, as indicated by the citations found using Google Scholar and in other bibliometric databases. However, they also find that a non-negligible share of the Bank’s publications have received no citations, suggesting that they have had little influence on development thinking, although they may well have helped connect ideas among development thinkers to a broader audience. Individually authored journal articles have been the main channel for scholarly influence: they are the most common type of publication and the type with the highest average citations. The volume of the Bank’s research output on development is greater than that of any of the comparator institutions identified, including other international agencies and the top universities in economics. The bibliometric indicators of the influence of the Bank’s portfolio of journal articles are found to be on a par with, or better than, those of most top universities. How to measure the influence of a publication portfolio with minimal assumptions Bibliometric measures based on citations are widely used in assessing the scientific publication records of authors, institutions and journals. Yet currently favored measures lack a clear conceptual foundation and are known to have

counter-intuitive properties. Martin Ravallion and Adam Wagstaff [2] propose a new approach grounded on a theoretical "influence function" that embodies explicit prior beliefs about how citations reflect influence. They provide conditions for robust qualitative comparisons of influence—conditions that can be implemented using readily available data. They provide an example using the economics publication records of selected universities and the World Bank. The "brain drain" among the best and brightest yields large economic returns John Gibson and David McKenzie [3] present the results of innovative surveys that tracked academic high-achievers from five countries to wherever they moved in the world. These surveys allow direct measurement at the micro level of the channels through which high-skilled emigration affects the sending country. The results show that there are very high levels of emigration and of return migration among the very highly skilled; the income gains to the best and brightest from migrating are very large, and an order of magnitude greater than any other effect; there are large benefits from migration in terms of postgraduate education; most high-skilled migrants from poorer countries send remittances; but involvement by migrants in trade with and foreign direct investment in the sending country is rare. There is considerable knowledge flow from both current and return migrants about job and study opportunities abroad, but little net knowledge sharing from current migrants to home country governments or businesses. The authors find that the fiscal costs vary considerably across countries, and depend on the extent to which governments rely on progressive income taxation. Experimental approaches in migration studies help identify a counterfactual The decision of whether or not to migrate has far-reaching consequences for the lives of individuals and their families. But the very nature of this choice makes identifying the impacts of migration difficult, since it is hard to measure a credible counterfactual of what the person and their household would have been doing had migration not occurred. Migration experiments provide a clear and credible

T

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

2

HD Update Quarterly Q3 2010 September 2010

way for identifying this counterfactual, and thereby allowing causal estimation of the impacts of migration. David McKenzie and Dean Yang [4] provide an overview and critical review of the three strands of this approach: policy experiments, natural experiments, and researcher-led field experiments. Economic freedom raises the returns to human capital According to T.W. Schultz, the returns to human capital are highest in economic environments experiencing unexpected price, productivity, and technology shocks that create "disequilibria." Elizabeth M. King, Claudio E. Montenegro and Peter F. Orazem [5] test the hypothesis that the returns to skills are highest in countries that allow individuals to respond to shocks. Using estimated returns to schooling and work experience from 122 household surveys in 86 developing countries, this paper demonstrates a strong positive correlation between the returns to human capital and economic freedom, an effect that is observed throughout the wage distribution. Economic freedom benefits those workers who have attained the most schooling as well as those who have accumulated the most work experience. Despite greater use of contract teachers, India’s courts are increasingly unsympathetic to their demands Since the early 1990s, the Indian government has increasingly employed contract workers to perform state functions, including teaching. Yet little research has been done on how courts have reacted to this policy shift. Nick Robinson and Varun Gauri [6] look at all relevant cases in the Indian Supreme Court and four High Courts over the last thirty years, and finds that the judiciary has increasingly become less sympathetic to contract teachers’ demands. The paper argues that the Court could use judicial review to engage the government and guide the creation of a labor policy that achieves a better school system for students and teachers. Questionnaire design does affect child labor statistics Andrew Dillon, Elena Bardasi, Kathleen Beegle and Pieter Serneels [7] present the results from a randomized survey experiment in Tanzania focusing on two aspects of household surveys aimed at eliciting information on child labor: different questionnaire design to classify child work, and proxy response versus self-reporting. Use of a short module compared with a more detailed questionnaire has a statistically significant effect, especially on child labor force participation rates, and, to a lesser extent, on working hours. By contrast, proxy reports do not differ significantly from a child’s self-report.

School enrollment rates are lower among children with higher-ability siblings Using data they collected in rural Burkina Faso, Richard Akresh, Emilie Bagby, Damien de Walque and Harounan Kazianga [8] examine how children's cognitive abilities influence resource-constrained households' decisions to invest in their education. This paper uses a direct measure of child ability for all primary school-aged children (regardless of enrollment), and it also uses measures of siblings’ ability, to test how sibling rivalry affects parents’ decisions on investing in education of their child. The findings indicate that children with one standard deviation higher ability are 16 percent more likely to be currently enrolled, but that having a higher-ability sibling reduces the likelihood of enrollment by a similar amount. How to control for unobserved heterogeneity in a dynamic probit model–and how failing to do so leads to an underestimate of the likelihood of a poor household staying poor John Giles and Irina Murtazashvili [9] propose a parametric approach to estimating a dynamic binary-response panel-data model that allows for endogenous contemporaneous regressors. Their approach is of particular value in settings where one wants to estimate the effects of an endogenous treatment on a binary outcome. An application of the method shows that migration within China is important for reducing the likelihood that poor households remain in poverty and that non-poor households fall into poverty; the authors find that failure to control for unobserved heterogeneity leads to an underestimation of these effects. Why poor voters may demand infrastructure spending yet not get it Stuti Khemani [10] explores a puzzle in the political economy of infrastructure in India—the coexistence of relatively low shares of capital spending in public budgets alongside evidence of large demand for village infrastructure from poor voters. It argues that this pattern is due to infrastructure projects being used at the margin for political rent-seeking, while spending on employment and welfare transfers are the preferred vehicles to win votes for re-election. Khemani presents new evidence on the variation of public spending composition across states and within states over time that is consistent with this argument.

The high willingness-to-pay to reduce the risk of death in China Hua Wang and Jie He [11] estimate individuals’ willingness to pay for cancer risk prevention in three provinces of China.

3

HD Update Quarterly Q3 2010 September 2010

Their results imply that the mean value of willingness to pay for a cancer vaccine that is effective for one year is 759 yuan, with a much lower median value of 171 yuan. The estimated income-elasticity of willingness to pay is 0.42. Using data on the incidence of cancer illness and death in the population, these willingness-to-pay figures imply that the marginal value of reducing the anticipated incidence of cancer mortality by one in the population (i.e. the value of a statistical life) is 73,000 yuan, with an average value of 795,000 yuan; these figures are about six and 60 times average household annual income, respectively.

New articles and books The mixed results of hospital payment reform in ECA

hile there is broad agreement that the way that health care providers are paid affects their performance, the empirical literature on the impacts of provider payment reforms is

surprisingly thin. During the 1990s and early 2000s, many European and Central Asian (ECA) countries shifted from paying hospitals through historical budgets to fee-for-service (FFS) or patient-based payment (PBP) methods (mostly variants of diagnosis-related groups, or DRGs). Using panel data on 28 countries over the period 1990-2004, Rodrigo Moreno-Serra and Adam Wagstaff [12] exploit the phased shift from historical budgets to explore aggregate impacts on hospital throughput, national health spending, and mortality from causes amenable to medical care. They find that FFS and PBP both increased national health spending, including private (i.e. out-of-pocket) spending. However, the two payment methods had different effects on inpatient admissions (FFS increased them; PBP had no effect), and on average length of stay (FFS had no effect; PBP reduced it). Of the two methods, only PBP appears to have had any beneficial effect on "amenable mortality", but the authors found significant impacts for only a couple of causes of death, and not in all model specifications. The better educated in the US responded faster to information on the dangers of smoking, confirming the beneficial effect of education on health Damien de Walque [13] tests the hypothesis that education improves health and increases life expectancy. The analysis of smoking histories shows that after 1950, when information about the dangers of tobacco started to diffuse, the prevalence of smoking declined earlier and most dramatically for college graduates. He constructs panels based on smoking histories in an attempt to isolate the causal effect of smoking from the influence of time-invariant unobservable characteristics. The results suggest that, at least among

women, college education has a negative effect on smoking prevalence and that more educated individuals responded faster to the diffusion of information on the dangers of smoking.

Research in the news

xtensive media coverage followed Damien de Walque and Berk Ozler’s presentations at the global AIDS summit in Vienna. Both outlined innovative experiments that appear to have reduced HIV/AIDS in

Africa. De Walque gave cash to people who stayed free of sexually transmitted illnesses, while Ozler simply gave cash to girls to stay in school; as it turned out, those who did stay on also didn’t marry older men with high rates of infection, which is what those who left school often did. Their work was featured in, among other places, The New York Times, The Australian, the Financial Times, the Guardian, the Thai News Service, and the Wall Street Journal, as well as on the BBC. Writing in The Internationalist, David Woodward argues against the Bank’s dollar-a-day poverty line, marshalling as evidence Adam Wagstaff’s work showing the variation across countries in child survival prospects and child malnutrition rates among families living on at a dollar a day. Associated Press (via Google) reported on a yet-to-be-published study by Jishnu Das and Tahir Andrabi on trust of Westerners among victims of the 2005 Pakistan earthquake. The study found higher trust of Westerners among those living close to fault line, attributable to the foreign emergency relief effort at the time according to the authors. 40 percent of quake survivors said they had been helped by foreigners or the UN; less than 5 percent received assistance from Islamist charities. Writing in the Australian magazine The Monthly (billed by the publishers as “An intelligent Australian magazine”), Anne Manne mentions Monica Das Gupta’s work on gender and son preference.

And on the blogs

alsey Rogers’ paper on “The Global Financial Crisis and Development Thinking” [14] gets coverage in the blogs. Amol Agrawal calls it a “nice overview of literature on impact of crisis on development

economics”. Daniel Ben-Ami, writing in his Ferraris for All

W

E

H

4

HD Update Quarterly Q3 2010 September 2010

blog, is also complimentary, as is Ryan Hahn in a posting on the Bank’s PSD blog called “A road network for private firms”. Several blogs picked up the experiments of Damien de Walque and Berk Ozler, including Alex Ogle’s post on NGO K'ARALE’s blog, Mead Over’s post on the Center for Global Development’s blog, and the US Kaiser Daily Global Health Policy Report. Jishnu Das’s study of trust and foreign assistance following Pakistan’s 2005 earthquake gets picked up in a Hudson Institute blog. Finally, Damien de Walque turns his hand to blogging with a blog in French on the Africa Can End Poverty blog about his study of school enrollment among children with higher-ability siblings.

Linking research and operations

arrying out impact evaluations in the context of a World Bank project is often touted as a real comparative advantage of the Bank. Rather than working with small-scale NGO projects, the Bank

works with governments, and programs implemented by governments. Learning from these projects goes beyond showing that an intervention can work, but allows us to find out whether it works under “real world” conditions. Moreover, the close link between the project and evaluation allows real-time feedback and adjustments. But how can such a program be set up? And how can it work in practice? Deon Filmer relays below how his engagement in the research and operations in the education sector in Cambodia led to a virtuous cycle of learning and adjustments to implementation. Designing the CESSP Scholarship Program It started with an invitation to join a Project Preparation mission in late 2004. The Bank was designing a new education lending operation in Cambodia and was keen to balance the supply-side assistance—building and improving school buildings, investing in quality—with a demand-side program. For a number of years, Cambodia had had programs that offer “scholarships” to poor children making the transition from primary to lower secondary school. One of these efforts was a program funded by the Japan Fund for Poverty Reduction (JFPR) and implemented in 100 schools. It targeted girls at the beginning of secondary school and offered selected girls a scholarship of $45 per year if they attended school regularly and maintained adequate performance. The Bank team and government counterparts

shared the view that this was a good model to build on, but on what basis— and how could the program be improved? Thanks to careful monitoring data that had been collected by government, and with assistance from the Asian Development Bank, it was possible to compile data on all program applicants. These data were complemented with data from surprise visits to schools to collect information on the attendance of applicants. While this impact evaluation design was not ideal, the results were encouraging: the scholarship increased enrollment and attendance by an estimated 20 to 30 percentage points, with larger increases among the poorest girls [15]. So the feedback to the project team was positive—this was certainly a program to build on. But the analysis, combined with a review of the government’s Education Management Information System data as well as a Poverty Map that had been carried out by the World Food Programme, suggested several changes—and raised several questions. The changes included adopting a more systematic approach to selecting scholarship recipients, to ensure that all primary-school completers could apply and that the program successfully identified the poorest applicants. Two important unanswered questions related to the size of the scholarship—should it be increased, or was $45 adequate?—and to whether the program should be limited to girls only—or should it include boys? The analysis and subsequent discussions between the research team, the Bank’s operational team, and the government counterparts, led to the CESSP Scholarship Program (CSP), launched in 2005. This program was more consistently targeted to schools located in the poorest communes; encouraged all students finishing primary school to apply for scholarships; and used a more consistent and transparent system for determining eligibility. It also allowed boys to become recipients, and introduced two levels of scholarships—$60 for the poorest recipients, and $45 for the slightly less poor. Importantly, it included a much more rigorous evaluation strategy to ensure that the outstanding questions—as well as others—could be answered. The CSP Evaluation Like the JFPR program, the CSP evaluation collected information on all the applicants to the program. But it also carried out multiple rounds of school visits to collect data on applicant attendance, as well as a large-scale household survey that covered a much wider set of outcomes. The evaluation design built on the way students were selected to receive scholarships. Based on responses to a simple survey, 6

th-grade applicants were ranked by a

C

5

HD Update Quarterly Q3 2010 September 2010

composite “dropout-risk score.” Within each participating lower secondary school, a fixed number of scholarships were awarded to children with the highest dropout risk. Comparing applicants “just below” to “just above” the cutoff effectively compares extremely similar applicants—who differ only in that the former were offered a scholarship. The scholarships were found to raise the school participation of recipients by about 20 percentage points. Enrollment among non-recipients was about 60 percent—so the impact of the program was to increase recipients’ enrollment rate from 60 to 80 percent. By global standards, this is a large impact. A recent review of conditional cash transfer programs around the world found impacts ranging from zero to 13 percentage points in other countries—depending on the size of the transfer, the age/level of school targeted, and the baseline enrollment rate [16]. So even though its transfers are relatively small by international standards, the CSP has the largest impact of any program that the review could document. The evaluation also addressed the two main outstanding questions: the results showed virtually no difference in the impact on girls and boys, and showed virtually no additional impact of the $60 transfer over and above the impact of $45 [17]. Other issues were analyzed too— such as how school expenditures, child labor, and intrahousehold resource allocation responded to the scholarships [18]. Another surprising result was that despite their increased school attendance of recipients performed no better on vocabulary and mathematics tests. These results were discussed with government as soon as they were produced—well in advance of the formal write-ups of the econometric analyses. This meant they could feed into program revisions. For example, when additional resources became available for the scholarships, the government decided to increase coverage at $45 rather than increase the amount of the transfer, since that would have the maximal impact. More important in the long run, as the CESSP phases out, these lessons are being applied to the government-financed National Secondary Scholarship Program. Together with other evaluations and stakeholder consultations, the evaluation enabled a high degree of confidence in the soundness of the approach and provided answers to important design questions. Beyond the CSP As often happens, the evaluation of the CSP itself raised many new questions, which is trying to answer. One question is why the impacts on learning were so small despite large impacts on enrollment and attendance. A second round of data collection and analysis, now underway,

asks whether this is just because the program hadn’t yet had enough time to operate. The second-round analysis will track outcomes over four years after the students originally applied for their scholarships—allowing analysis of impacts not just on learning, but also on labor-market activity and family formation. Another lingering question from the CSP evaluation concerned the issue of targeting. Careful analysis of data on program applicants revealed that they were not necessarily the poorest children in the targeted communes—because the poorest of the poor had already dropped out of school before the end of the primary cycle. These findings fed back into the policy dialogue—and in the context of the Royal Government of Cambodia’s 2009 grant from the Fast Track Initiative (FTI), a scholarship program for the upper primary grades is being piloted in three poor remote provinces. That pilot is being carefully evaluated, with initial results expected in 2011. A third question is how to improve low levels of school readiness, which are related to both primary-school dropout and the disappointing results on learning at the secondary level. So in the context of the FTI grant, three alternative models for delivering services to promote early child development are being analyzed through rigorous impact evaluations. Results are expected in 2011.

References 1. Ravallion, M. and A. Wagstaff, The World Bank's publication record.

2010, Washington, DC: World Bank, Policy Research Working Paper #5374.

2. Ravallion, M. and A. Wagstaff, On measuring scientific influence. 2010, Washington, DC: World Bank, Policy Research Working Paper #5375.

3. Gibson, J. and D. McKenzie, The economic consequences of "brain drain" of the best and brightest: microeconomic evidence from five countries. 2010, Washington, DC: The World Bank, Policy Research Working Paper Series #5394.

4. McKenzie, D. and D. Yang, Experimental approaches in migration studies. 2010, Washington, DC: The World Bank, Policy Research Working Paper Series #5395.

5. King, E.M., C.E. Montenegro, and P.F. Orazem, Economic freedom, human rights, and the returns to human capital: an evaluation of the Schultz hypothesis. 2010, Washington, DC: The World Bank, Policy Research Working Paper #5405.

6. Robinson, N. and V. Gauri, Education, labor rights, and incentives : contract teacher cases in the indian courts 2010, Washington, DC: World Bank, Policy Research Working Paper #5365.

7. Dillon, A., E. Bardasi, K. Beegle, and P. Serneels, Explaining variation in child labor statistics. 2010, Washington, DC: The World Bank, Policy Research Working Paper Series #5414.

8. Akresh, R., E. Bagby, D. de Walque, and H. Kazianga, Child ability and household human capital investment decisions in Burkina Faso. 2010, Washington, DC: The World Bank, Policy Research Working Paper #5370.

9. Giles, J. and I. Murtazashvili, A control function approach to estimating dynamic probit models with endogenous regressors, with an application

6

HD Update Quarterly Q3 2010 September 2010

to the study of poverty persistence in China. 2010, Washington, DC: World Bank, Policy Research Working Paper Series #5400.

10. Khemani, S., Political economy of infrastructure spending in India. 2010, Washington, DC: The World Bank, Policy Research Working Paper #5423.

11. Wang, H. and J. He, The value of statistical life: a contingent investigation in China. 2010, Washington, DC: The World Bank, Policy Research Working Paper Series #5421.

12. Moreno-Serra, R. and A. Wagstaff, System-wide impacts of hospital payment reforms: Evidence from Central and Eastern Europe and Central Asia. Journal of Health Economics, 2010. 29(4): p. 585-602.

13. de Walque, D., Education, Information, and Smoking Decisions: Evidence from Smoking Histories in the United States, 1940–2000. J. Human Resources 2010. 45(3): p. 682-717.

14. Rogers, F.H., The global financial crisis and development thinking. 2010, Washington, DC: World Bank, Policy Research Working Paper #5353.

15. Filmer, D. and N. Schady, Getting girls into school: Evidence from a scholarship program in Cambodia. Economic Development and Cultural Change, 2008. 56(3): p. 581-617.

16. Fiszbein, A., N. Schady, F.H.G. Ferreira, M. Grosh, N. Kelleher, P. Olinto, and E. Skoufias, Conditional Cash Transfers: Reducing Present and Future Poverty. 2009, Washington, D.C.: World Bank.

17. Filmer, D. and N. Schady, Does more cash in conditional cash transfer programs always lead to larger impacts on school attendance? Journal of Development Economics, 2010.

18. Ferreira, F.H.G., D. Filmer, and N. Schady, Own and sibling effects of conditional cash transfer programs: theory and evidence from Cambodia. 2009, Washington, DC: World Bank, Policy Research Working Paper #5001.

Previous issues: HD Quarterly Previous Issues