Latest research findings
Mass media and public services: more accountability or larger private spending?
Using unique data from Benin, Phil Keefer and Stuti Khemani  find that literacy rates among school children are higher in villages exposed to signals from a larger number of community radio stations. The effect is identified based on a "natural experiment" in the northern communes of Benin where within-commune variation in village access to radio stations is exogenous to observed and unobserved village characteristics. In contrast to prior research, the authors find that this media effect does not operate through government accountability: government inputs into village schools and household knowledge of government education policies are no different in villages with greater access to community radio. Instead, households with greater access are more likely to make financial investments in the education of their children.
Good countries or good projects? Macro and micro correlates of World Bank project performance project performance
Cevdet Denizer, Daniel Kaufmann and Aart Kraay  use data from more than 6,000 World Bank projects evaluated between 1983 and 2009 to investigate macro and micro correlates of project outcomes. They find that country-level "macro" measures of the quality of policies and institutions are very strongly correlated with project outcomes, but that the success of individual development projects varies much more within countries than it does between countries. The authors find that measures of project size, the extent of project supervision, and evaluation lags are all significantly correlated with project outcomes, as are early-warning indicators that flag problematic projects during the implementation stage. They also find that measures of World Bank project task manager quality matter significantly for the ultimate outcome of projects. They discuss the implications of these findings for donor policies aimed at aid effectiveness.
Collective action, political parties and pro-development public policy
Broad consensus exists that the ability of political actors to make credible commitments is key to development. An important and little-explored determinant of the credibility of political commitments is the existence of organizations that facilitate citizen collective action to sanction political actors who renege. In this paper, Phil Keefer  focuses on one essential organization: the political party. Three measures of political parties are used to assess cross-country differences in the degree to which politicians facilitate the ability of citizens to act in their collective interest. Each of these measures is associated with superior development outcomes, above and beyond the effects of competitive elections. These results have implications for understanding the extraordinary economic success of some East Asian countries and notable lags among others: East Asian non-democracies exhibit more institutionalized ruling parties than other non-democracies, while East Asian democracies exhibit equally or less institutionalized parties. The evidence suggests that greater research and policy emphasis be placed on the organizational characteristics of countries that allow citizens to hold leaders accountable.
School inputs, household substitution, and test scores
Studies of the relationship between school inputs and test scores typically don’t account for the fact that households will likely respond to changes in school inputs. Jishnu Das et al.  develop a dynamic household optimization model relating test scores to school and household inputs, and test its predictions in India and Zambia. Consistent with the model, they find in both settings that households offset anticipated grants more than unanticipated grants. They also find that unanticipated school grants lead to significant improvements in student test scores but anticipated grants have no impact on test scores. The results suggest that naïve estimates of public education spending on learning outcomes that do not account for optimal household responses are likely to be considerably biased if used to estimate parameters of an education production function.
Students today, teachers tomorrow? Relaxing the constraints on the provision of educations
With an estimated 115 million children not attending primary school in the developing world, increasing access to education is critical. Resource constraints limit the effectiveness of demand-based subsidies, raising the question of how effective supply-side channels might be. Tahir Andrabi et al.  find that the presence of a government girls' secondary school substantially increases the local supply of skilled women, lowers teacher wages locally, and increases by a factor of three the likelihood that an affordable private school will emerge. These findings highlight the prominent role of women as teachers in facilitating educational access.
Does the association between education and moral behavior affect economic growth?
David Balan and Steve Knack  incorporate morality—defined as lower utility from consuming goods obtained through appropriative rather than productive activities—into a simple static general equilibrium model in which agents choose whether to be producers or appropriators. The authors analyze the relationship between the correlation between morality and human capital on the one hand, and aggregate economic performance on the other. They show that there is a main effect that tends to cause this relationship to be positive, and secondary effects that may reinforce or oppose the main effect. Empirical tests find evidence that higher within-country correlation between morality and ability increases per-capita income levels. Results are robust to correcting for endogeneity and to changes in sample and specification.
Redressing grievances and complaints regarding basic service delivery
Redress procedures are important for basic fairness, but as Varun Gauri  argues, they can also help address principal-agent problems in the implementation of social policies, as well as providing information to policy makers regarding policy design. To function effectively, a system of redress requires a well-designed and inter-linked supply of redress procedures as well as, especially if rights consciousness is not well-developed in a society, a set of organizations that stimulate and aggregate demand for redress. On the supply side, Gauri identifies three kinds of redress procedures: administrative venues within government agencies, independent institutions outside government departments, and courts. On the demand side, the key institutions are nongovernmental organizations/civil society organizations and the news media, both of which require a receptive political and economic climate to function effectively. Overall, procedures for redressing grievances and complaints regarding basic service delivery are under-developed in many countries, and deserve further analysis, piloting, and support.
‘Intersubjective’ meaning and collective action in ‘fragile’ societies
The capacity to act collectively is not just a matter of groups sharing interests, incentives and values (or being sufficiently small), as standard economic theory predicts, but a prior and shared understanding of the constituent elements of problem(s) and possible solutions. From this standpoint, the failure to act collectively can stem at least in part from relevant groups failing to ascribe a common “intersubjective” meaning to situations, processes and events. Though this is a general phenomenon, it is particularly salient in countries characterized by societal fragility and endemic conflict. Michael Woolcock et al.  develop a conceptual account of intersubjective meanings, explain its relevance to development practice and research, and examine its implications for development work related to building the rule of law and managing common pool resources.
New articles and books
Making schools work: New evidence on accountability reforms
A new World Bank book by Barbara Bruns, Deon Filmer and Harry Patrinos  explores threats to education quality in the developing world that cannot be explained by lack of resources. It reviews the observed phenomenon of service delivery failures in public education: cases where programs and policies increase the inputs to education but do not produce effective services where it counts—in schools and classrooms. It documents what we know about the extent and costs of such failures across low and middle-income countries. And it further develops the conceptual model posited in the World Development Report 2004: that a root cause of low-quality and inequitable public services—not only in education—is the weak accountability of providers to both their supervisors and clients.
Investment without Democracy: Ruling-Party Institutionalization and Credible Commitment in Autocracies
What explains private investment in autocracies, where institutions that discourage expropriation in democracies are absent? Scott Gehlbach and Phil Keefer  argue that institutionalized ruling parties allow autocrats to make credible commitments to investors. Such parties promote investment by solving collective-action problems among a designated group, who invest with the expectation that the autocrat will not attempt their expropriation. Gehlbach and Keefer derive conditions under which autocrats want to create such parties, and predict that private investment and governance will be stronger in their presence. They illustrate the model by examining the institutionalization of the Chinese Communist Party.
The cost of complying with human rights treaties: the convention on the rights of the child and basic immunization
The determinants of compliance with human rights treaties likely vary according to the right in question, yet heterogeneity in the pathways through which ratification affects various human rights outcomes has received limited attention. Varun Gauri  first develops an account of treaty compliance that incorporates the intrinsic benefits to the state of compliance, regime costs associated with certain rights, the political costs that NGOs, judges, and others are able to impose for non-compliance, and the fiscal and economic costs of compliance. He argues that for child survival rights, fiscal and economic costs are likely to be dispositive, and that as a result richer countries are more likely to comply. He then uses an instrumental variable approach to investigate whether ratification of the Convention of the Rights of the Child (CRC) was associated with stronger effort at the country level on child survival rights. He finds that ratification of the CRC was correlated with a subsequent increase in immunization rates, but only in upper middle and high income countries.
Fungibility and the impact of development assistance: Evidence from Vietnam's health
The apparent fungibility of aid is a challenge to the evaluation of donor-funded development projects, requiring a comparison of the observed outcomes with the outcomes that would have occurred if the project had not gone ahead. Where projects are targeted on specific geographic areas, counterfactual outcomes in each can differ from observed outcomes because the amount of government spending (gross of aid) differs, the productivity of government spending differs, or both. Adam Wagstaff  estimates the benefits of two concurrent World Bank health projects in Vietnam targeted on specific provinces. He derives estimates from a model linking outcomes (under-five mortality) to government spending before and after the project and in project and nonproject provinces, and are presented for different assumptions regarding fungibility of funds (zero and full fungibility) and the impacts of the project on the productivity of government spending (the project modifies productivity in both sectors equally and in neither sector). The estimated mortality reductions are highly insensitive to the assumed degree of fungibility, but highly sensitive to the assumed productivity effects (the estimates range from 1 to 25%). The wide range reflects the uncertainty due to the lack of a genuine control group of provinces.
The quality of medical care in low-income countries: from providers to markets
It is widely believed that people in low and middle-income countries (LMICs) are in poor health because they cannot reach medical services on time. Yet, recent evidence shows high utilization rates, even among the poor. Jishnu Das  argues that the real constraint is not access but rather quality. The evidence points to a gap between what doctors know and what they do, with the size of the gap varying across provider types: in Delhi, low effort by public-sector doctors reduces the quality of care there to the level of care delivered by untrained private-sector providers. Das notes that while the “know-do” gap is larger with salaried doctors, it also exists in the private sector. This is odd because patients apparently make quite accurate assessments of quality and are more satisfied with better quality care; providers offering better quality ought therefore to be able to attract patients and increase revenues.
Research in the news
In an article in Tehelka (India) suggesting the Supreme Court move out of New Delhi to raise appeal rates among poorer Indians, Paari Vendhan cites Varun Gauri’s  work on public interest litigation in India.
Steve Knack’s work on the impact of aid on the quality of countries’ tax systems  is picked up by Amitai Etzioni writing in the Middle East Quarterly about mission creep in the Afghan conflict.
And on the blogs
In a post on the Bank’s “Governance for Development” blog, Stuti Khemani asks: Do informed citizens hold governments accountable? She concludes that it depends on the political and economic context within which media operates. In Benin, where she found it had no effect via accountability, the context neither encourages media to provide "accountability" information nor facilitates citizen mobilization in response to that information.
Also on the Let’s Talk Development blog, Adam Wagstaff offers four cheers for the “results agenda". It gets us to work back from outcomes to think about multiple ways of achieving better outcomes, including fixing the demand side. It invites us to focus on ways to incentivize results, whether we’re talking about rewarding clinics with better pay or rewarding countries with larger loans. It forces us to get serious about evidence, using impact evaluation but other methods too. And it invites us to think more seriously about the role of knowledge in the Bank’s work.
Steve Knack’s work on aid and the quality of tax administration is picked up by Amitai Etzioni in a piece entitled “Corruption reduction” in the online Harvard International Review. Varun Gauri’s work is picked up in a post “For the sake of fairness: Justice in development” on the Bank’s Governance for Development blog. Tahir Andrabi and Jishnu Das’s work is picked up in a post on the Rug Pundits blog entitled “Doubt and development”. The new book by Barbara Bruns, Deon Filmer and Harry Patrinos is picked up by Ariel Fiszbein in a post on the Education for Global Development blog.
In his maiden blog post, Deon Filmer asks whether disseminating information on school quality can lead to better learning outcomes. He concludes that it can but won’t necessarily do so. He argues: “When the constraint on quality is relatively easily addressed—through squeezing out more instructional time as was the case in Pakistan, or reducing absenteeism as was the case in the Indian three-state study—this is relatively easily done. When constraints are harder to overcome—for example finding the most appropriate pedagogical approach to dealing with a heterogeneous classroom—it is unlikely that information on its own will be enough.”
In a post on Let’s Talk Development, Adam Wagstaff asks whether a consensus on health reform is emerging in Asia. In the six Asian countries he discusses, over one billion people (around 40% of the population) have been covered by new programs that are predominantly tax-financed. Most involve services being ‘purchased’ by an organization that often operates at arm’s length from the health ministry using a payment mechanism that incentivizes the delivery of services by contracted providers, often private ones.
This half-year saw the launch by four Bank researchers—Berk Ozler, David McKenzie, Jed Friedman and Markus Goldstein—of the highly successful Development Impact blog. Posts on public sector and governance topics this half-year have included: Jed Friedman on evaluating drug distribution system reforms, verifying performance in pay-for-performance, what makes health workers get up in the morning and getting people to adopt new health technologies; and David McKenzie on whether teachers matter and fighting malaria with microfinance.
Linking research & operations: anti-corruption and regulatory reform in ECA
Researchers often assist regional staff in Economic and Sector Work, including not only country-specific products such as Poverty Assessments, but also regional “flagship” reports. Below, Steve Knack reflects on his productive collaboration over a 5-year period with ECA staff on anti-corruption issues.
In 1990, legal frameworks and other institutions necessary for market economies to function efficiently and equitably were missing or underdeveloped in the “transition” countries of ECA region, most of which experienced large declines in output in the 1990s. Privatization and adoption of new regulatory systems provided enormous opportunities for bribe-seeking and improper use of connections and influence. By the late 1990s, corruption had emerged as one of the most serious obstacles to growth in the region. Beginning in 2000, a series of reports on Anticorruption in Transition (ACT) has monitored progress on this problem, based on the Business Environment and Enterprise Performance Surveys (BEEPS) conducted every three years by the EBRD and ECA.
The most notable finding in the ACT series is a trend toward lower administrative corruption in the region overall. This headline message has often been obscured by seemingly conflicting evidence, however, and one of the main challenges in recent reports has been to sort through the proliferation of data sources and indicators to provide a clear and consistent story to policymakers. I played a supportive role on the third ACT report, based on the 2005 BEEPS, and was a lead author on the most recent report, based on the 2008 BEEPS.
Anticorruption in Transition 3
The first report (ACT-1)  was based only on the 1999 BEEPS, and set baselines. A favorable trend on administrative corruption was first noted in ACT-2 , based on comparisons from the 1999 and 2002 BEEPS. Some observers pointed out that this trend was not apparent in other well-known data sources on corruption, including the aggregate indexes produced by Transparency International (TI) and the Bank’s own Worldwide Governance Indicators (WGI) project. When ECA staff later discovered another favorable trend in their BEEPS 2002 and 2005 comparisons, they decided to pre-empt the skeptics by requesting me to write a background paper  on corruption indicators that might clear up any apparent inconsistencies. Results from the background paper, later published in the Journal of Public Policy , were summarized in a Box and cited elsewhere in the ACT-3 report . It generally validated the favorable trend in administrative corruption found in the BEEPS, in making the following points:
- Different corruption indicators often measure different aspects of corruption, so less-than-perfect agreement among them is to be expected.
- This heterogeneity is reflected in the BEEPS itself, and its favorable trend in administrative corruption is not exhibited for “state capture”; moreover, administrative corruption in some areas (e.g. tax) improved more than in others (e.g. judicial systems).
- The TI and WGI indicators are aggregated from numerous sources of “expert” perceptions, and some sources do not provide an explicit definition of corruption provided. It is therefore unclear to what extent they “should” behave as measures of administrative corruption, “state capture,” or other types.
- Changes in experts’ perceptions over time are more subject to measurement error than changes in firms’ reported experiences. Often, changes in perceptions indicators represent corrections in response to newly-available information, rather than a belief that corruption has actually changed.
- Evidence from the World Economic Forum’s “Executive Opinion Surveys” was remarkably consistent with evidence from the BEEPS, in showing improvement in administrative corruption but not in state capture, and in showing more improvement in certain areas (licenses and permits, tax and customs, utilities) than in others (public procurement, judicial systems).
Trends in corruption and “regulatory burden,” 2005-2008
The sample design and questionnaire for the 2008 BEEPS were modified for conformity with the Bank’s other Enterprise Surveys. This change facilitated comparisons with other regions, but made it more difficult to track progress over time within ECA leading up to 2008.
With some input from other Research Department colleagues, I provided advice to ECA on several products based on the 2008 BEEPS, including:
- A dataset for 2005 and 2008 that maximizes comparability over time, mostly by dropping firms with certain characteristics (based on size or sectors) from one year or the other.
- A BEEPS at a Glance report and country profiles based on this dataset, posted on the ECA web site. The country profiles are particularly useful in showing which countries deviate from the overall regional trends.
More substantially, I was a lead author on a report  similar to those in the ACT series, assessing progress on corruption and regulatory burden, and providing some evidence on what reforms are more likely to be effective. A major goal in the report was to prevent misleading interpretations of data caused by changes in the way questions were asked. Two illustrations follow:
- The BEEPS appeared at first glance to indicate a doubling – from 6% in 2005 to 12% in 2008 - in the average time spent by managers in the region in dealing with public officials and government regulations in the region. However, this increase turned out to be an artifact of a change in question wording. The “time tax” question in 2008, but not in 2005, included a follow-up statement prompting respondents with examples that, in effect, invited them to interpret the question more broadly. A similar question in the WEF shows no adverse trend over time.
- The three ACT reports made extensive use of a question on the “bribe tax,” i.e. the share of firm revenues paid as bribes to public officials. In the 2005 and earlier BEEPS, used in those reports, respondents were all asked to state their answer as a percentage. In the 2008 survey, respondents were given a choice of (1) stating their answer as a percentage, or instead (2) providing an estimate of total bribes paid to officials in local currency units. For firms responding in this way, the bribe tax is computed using a separate survey question on firm sales in local currency units. The average bribe tax for firms using Method 1 is roughly 50 times as large as for firms using Method 2. Allowing the Method 2 option beginning in 2008 therefore likely creates a bias toward finding reductions in the bribe tax between 2005 and 2008. To correct for this bias, by estimating the average effect on responses of the change in question wording, it would be necessary to conduct a controlled experiment, e.g. in which half of the firms were randomly assigned to use Method 1 and the other half Method 2. In the 2008 survey, however, firms could choose for themselves which method to use, and their decisions were not random: firms reporting more frequent bribe paying (on a separate question) were more likely than other firms to answer the bribe tax question in terms of percentages – the method generating higher values. This selection bias – whereby firms facing more severe corruption problems chose Method 1 - implies that one could not merely drop the firms reporting by Method 2, and compute country aggregates from respondents reporting by Method 1, for comparisons with 2005.
For the next round of BEEPS, I have supported ECA staff in their attempt to add some questions on “state capture” back into the survey. These questions were all dropped for comparability with the other regions’ surveys, but the 2005 BEEPS and more recent WEF data indicate it is a continuing problem in ECA. I have also recommended conducting randomized experiments in countries with larger samples (e.g. Russia), to better understand the implications of asking questions (such as in the two examples above) in different ways.