Relating Staff Incentives to Achieving Development Results. The 2006 Annual Evaluation Review was the first systematic attempt at ADB to examine staff accountability for project success. It confirmed that ADB’s current formal and informal incentives reward loan processing and approval, with insufficient focus on project administration, project success, or achieving medium-term results. Adjusting ADB’s incentive systems to encourage more focus on project results and success is consistent with commitments made in both the 2004 human resources strategy and the second medium-term strategy. The question is not whether to change the systems, but how to do it.
The present staff performance evaluation system needs to be complemented by a chain of accountability for achieving development results that begins at the top with ADB’s Management and senior staff and cascades down to all staff. Incentives need to be fully aligned with accountability throughout the chain. The 2006 Annual Evaluation Review recommended that ADB explore the feasibility of providing stronger incentives for staff to focus on project quality at entry and on project administration and supervision. Initial work could include an assessment of best practices in comparable institutions and developing indicators for measuring the achievement of development results. ADB’s Management agreed to examine the feasibility of this approach.
Factors Contributing to Project Quality, including Governance. Good governance and the control of corruption have become important parts of the international development agenda. However, the 2006 Annual Evaluation Review concluded that the relationship between governance and development effectiveness is complex and not straightforward. There are many definitions of governance, and there are many measurement problems that make data less than fully reliable. Analysis of the relationship between measures of governance and macroeconomic indicators suggests the following:
• The Asia and Pacific region is the fastest-growing economic region in the world, but governance is weak in many countries according to standard governance indicators.
• Although there is a strong, positive relationship between the level of economic development and governance variables when a worldwide database is used, the relationships are much weaker if only countries in which the per capita gross domestic product is lower than $5,000 are analyzed.
• No significant relationship can be found between good governance and the rate of economic growth from 1996 to 2004. Some countries in the region with relatively good governance had lower growth rates than countries with lower governance ratings.
The analysis of projects approved in the 1990s suggests evidence of a relationship between some of the dimensions of governance and good project outcomes; however, it is not clear how strong or robust the relationships are. Other factors are also important, perhaps more so than governance, in determining project success (e.g., sector, country characteristics, macroeconomic climate, country ownership, capacity of the executing agency). The relationships between good governance at the macro level and project success may be somewhat weaker than is conventionally assumed. It may be that governance issues at the sector level have a more direct bearing on project success.
Subsequent to ninth replenishment of ADF, ADB harmonized its performance-based allocation procedures with those of the World Bank. ADB’s 2004 performance-based allocation policy for ADF resources increased the weight given to governance in measuring country performance from 30% to 50%, although this is still lower than the 66% assigned in the World Bank’s formula. There was, however, no analysis undertaken by ADB that demonstrated that good governance is the key binding constraint for development in all DMCs and that it merits such a high weight in the performance-based allocation formula. ADB has not undertaken a rigorous study that links good governance, as a causal factor, to economic growth, poverty reduction, development results, portfolio performance, or project success.
The 2006 Annual Evaluation Review recommended that ADB undertake such a study. Depending on the results, consideration may have to be given to including sectoral governance variables in the formula or to reducing the weight for governance in the performance-based allocation formula. Benchmarking a country’s governance performance and then rewarding progress in improving governance would be more consistent with managing for development results than making major decisions on ADF allocations based on the governance score in any one year. Steps also need to be taken to improve the consistency of what is meant by governance in ADB’s various governance-related policies and country governance assessments. Reducing the weight might be perceived as sending the wrong message, given that ADB has recently adopted the second medium-term strategy, which places priority on good governance and controlling corruption. However, this must be balanced against the potential harm that may be caused to DMCs whose ADF allocation is reduced because of a score on a variable that is difficult to define, measure, and compare over time and across countries.
Clearly, good governance and efforts to control corruption are important. The issues identified by the review relate to a lack of clarity of definition, difficulties of measurement, a lack of rigorous analysis of the subject at ADB, and questions about whether governance is the most important binding constraint to development in all countries. These issues are important, because in the performance-based allocation formula, ADB purports to be able to accurately define and measure governance on a numeric scale and uses the results to allocate more or less ADF funding to DMCs.
Technical Assistance. OED maintains a database of evaluation findings from TA performance evaluation reports and project/program performance evaluation reports for TA associated with projects/programs. The ratings included in Table 4, dated 2007, are significantly below the 70% success rate that reflects satisfactory outcomes. Overall, 63% of TA operations evaluated in TA performance evaluation reports were rated as being at least successful, and 48% of activities evaluated in project/program performance evaluation reports were rated as successful. The percentage of activities rated as successful in TA and project/program performance evaluation reports has been fairly static over time. For TA activities approved in the 1980s, the portion rated as successful was 44%; it was 58% from 1990 to 1994, 59% from 1995 to 1999, and 60% from 2000 to 2007. There is thus no trend toward more successful TA outcomes over the past 15 years. ADB’s Management needs to develop specific actions to address strategic and management issues concerning TA activities.20
Table 4: Assessment of Technical Assistance Performance
| Source | Number of Technical Assistance Operations | % Successful |
| Technical Assistance Special Evaluation Studies | 110 | 72 |
| Country/Sector Assistance Program Evaluations | 113 | 71 |
| Technical Assistance Performance Evaluation Reports | 185 | 63 |
| Project/Program Performance Evaluation Reports | 143 | 48 |
| Total | 551 | 63 |
Source: OED.
Directions in Evaluation
The new modus operandi for OED that took effect in 2004 has implications for its performance. The structural changes place more emphasis on OED’s developmental role and make development impact a more explicit consideration for operational decisions. In a changing context, OED has focused its 3-year rolling work program on priority areas, has moved to capture synergies between OED and ADB’s operations departments, and has begun to select evaluation topics in a way that should ensure higher effectiveness and impact.
Assigning Resources. Fewer project or program performance evaluation reports are in OED’s forward work program. Efforts are made to select evaluation topics that are of strategic relevance to ADB and DMCs.21 There are changes, too, in the way that OED assigns staff to evaluation studies. In the past, OED specialists worked individually on