The Role of the Fund in Governance Issues – Review of the Guidance Note – Preliminary Considerations – Background Notes

The Role of the Fund in Governance Issues - Review of the Guidance Note - Preliminary Considerations - Background Notes


The Role of the Fund in Governance Issues - Review of the Guidance Note - Preliminary Considerations - Background Notes

Note III. Fund use of Third-Party Corruption Indicators1

This background paper discusses the Fund’s use of third-party governance indicators with a focus on measures of corruption. It opens with a discussion of the characteristics of available indicators; discusses Fund practices in using third-party indicators; explores tentative findings with regard to evenhandedness; and concludes with areas for further work.

Preliminary findings are that third-party governance indicators are used infrequently and inconsistently in surveillance and use of Fund resources. In some cases, where indicators pointed to high levels of corruption coverage of corruption issues in Fund documents was limited. Although there is generally a close correlation between the different third-party corruption indicators considered in this study, the relationship between these measures and the Fund’s coverage of governance/corruption in country Board papers is much weaker. This suggests that greater use of third-party indicators could contribute to more evenhanded treatment of corruption by the Fund.

Annex I discusses the different approaches third-party indicators take to measuring corruption. Annex II explores how a few of the third-party indicators are constructed and examines how methodological approaches are constantly evolving to address the difficulty of measuring corruption. Annex III provides the average historical global percentile rankings of countries under three third-party indicators.

A. Introduction

1. A range of third-party indicators are typically available to provide perspectives on corruption in any given country. As noted in Annex I to this background note, the available governance/corruption indicators have different characteristics; they measure a variety of aspects of corruption, have different scope and depth of coverage, and have changed methodologies over time.2 While the Executive Board endorsed the greater use of indicators in general, and although staff guidance has further highlighted the usefulness of indicators produced by the World Bank, Transparency International (TI), the International Country Risk Guide (ICRG), and others, there is no definitive list of indicators approved for Fund use and little guidance on when their use might be appropriate.3 This Note explores the characteristics of a range of available indicators (see table below).

Note III, Table 1.

Description of Selected Third-Party Indicators

article image

Based on a long-standing research program of the World Bank, the WGI do not reflect the official views of the Natural Resource Governance Institute, the Brookings Institution, the World Bank, its Executive Directors, or the countries they represent. The WGI are not used by the World Bank Group to allocate resources.

B. Characteristics of Third-Party Indicators

2. There is no single state-of-the-art corruption indicator. Operational considerations for using corruption indicators were explored in the 2015 UNDP’s “User’s Guide to Measuring Corruption and Anti-Corruption” (UNDP Guide) 4. The Guide reviewed a range of tools and indicators and concluded that there is no single formula or one-size-fits-all approach given the complex nature of corruption. It recommends that a variety of methods and tools be used to capture the progress and evaluate the effectiveness and impact of anti-corruption programs and projects. In considering which indicators to use, several factors should be taken into consideration.

3. A key consideration is the methodology used in compiling a particular indicator. A key distinction is between three broad types of indicators: (a) those that measure corruption perception; (b) those that measure the actual experience of citizens or businesses with corruption; and (c) measures based on assessments of the characteristics of anti-corruption laws, policies, and institutions (see Annex I for further details). A further distinction relates to the extent of country coverage of different measures. Some measures are based on fixed methodologies and data sources, while these elements have evolved over time for other indicators, so that comparisons over time are not fully reliable. In some cases, methodological approaches are constantly evolving to address the difficulty of measuring corruption. Most measures are also compiled by aggregating different data sources, and procedures are needed to handle missing data sources. The aggregation process can also conceal individual areas of better- or worse-than-average performance in a particular country. Corruption perception measures are perhaps most straightforward to compile, and tend to have the widest country coverage. But even where perceptions are based on expert opinions, the latter can be influenced by scandals that may not reflect the underlying conditions (Olken and Panda (2012)).5 Similarly, corruption perception indices do not directly measure the volume of bribes, the incidence of corruption, or its actual impact.

4. Country corruption indicators tend to evolve only slowly, over time. While a premium is usually assigned to the most recent macroeconomic data, this is less critical for corruption indicators, which tend to move more slowly over time. This reflects, several factors. The political economy institutions that influence the extent of corruption tend to change only slowly (See Figure 1). This is reflected in a tendency for perceptions of corruption to change only slowly. Last, where there are gaps in data (perhaps relating to the availability of component perception surveys), the aggregate indicators are often compiled by substituting the preceding year’s data, which gives the measure a “moving average” quality.

Note III, Figure 1.
Note III, Figure 1.

WGI – Control of Corruption Index by Region, 1998 vs 2015

Citation: Policy Papers 2017, 005; 10.5089/9781498346481.007.A003

5. Corruption measures are often closely correlated. While different measures bring different perspectives, they often produce similar messages. The correlation is highest between measures that have the same methodological approach: for example, perceptions-based measures tend to show similar cross-country patterns. Table 2 reports correlation coefficients for four measures (TI, WGI, ICRG, and Maple). The three perceptions-based measures (TI, WGI, and ICRG) are most highly correlated with each other (correlation coefficients around 0.95) while the institutions-based measure (Maple) has a lower correlation coefficient with the perceptions-based measures (around 0.70–0.75). This suggests that corruption perceptions do not always move in line with measures of the adequacy of anti-corruption institutions, and Fund staff may want to consider country ratings according to both concepts. There are also challenges in comparing country rankings over time and across countries (see Annex I of this Note).6

Note III, Table 2.

Correlation Matrix for Third-Party Corruption Indicators (2012–15)

article image
Note: All correlations significant at conventional levels.

6. Other external measures of governance or corruption also result in high correlations. Indicators such as BTI, WJP, EIU and GI which are also inputs to TI’s measure of corruption, are strongly correlated with each other (0.75-.0.88) and with indicators of WGI, ICRG and Maple.

C. Use of Governance Indicators by Fund Staff

7. Use of existing governance indicators in the Fund’s work has been encouraged by the Executive Board and in staff guidance. The Fund itself does not produce measures of governance and corruption and thus relies on external governance indicators for comparisons over time and across members.7 In the surveillance context, Directors have called for wider use of appropriate outside sources of information, including governance indicators.8 Use of such indicators is viewed as facilitating the selective but evenhanded coverage of governance/corruption issues. The use of indicators has also been encouraged in the area of Jobs and Growth to identify constraints on growth such as poor institutional quality.9 In the context of use of Fund resources, indicators provide a useful basis on which to assess the strength of members’ institutional policy framework, have helped assess capacity to manage public resources, and assist in identifying structural constraints to growth.10 See Annex I of this Note for an overview of the role, benefits, and shortcomings of external governance and corruption indicators and a discussion of how methodological approaches are constantly evolving to address the difficulty of measuring corruption.

8. However, the stocktaking exercise found that external governance indicators do not appear to be used regularly in staff’s analysis in the context of surveillance and the use of Fund resources. The stocktaking exercise shows that only 20 percent of country reports reviewed for 2005–16 used third-party indicators.

9. The stocktaking exercise also found that, where staff cite third-party indicators, most use World Bank or Transparency International corruption perceptions measures. However, a wide variety of other measures have also been used. The World Bank’s Doing Business Indicator, as well as three of the four indicators discussed above (TI, WGI and ICRG) have most often been used in Fund documents; the Maplecroft measure is a relatively new index (since 2009) with similar country coverage but has not been widely cited. It appears that the choice of these indicators is primarily driven by their extensive country coverage along the lines of the Fund’s membership, stability of the criteria/definition used to construct the indicators and the availability of time series. According to the results of the stocktaking exercise, staff selected a mix of perception-based indicators of corruption and indirect corruption related indicators to analyze corruption and governance in a country. The stocktaking exercise shows that of the reports that relied on third-party indicators, perception based indicators such as the TI’s Corruption Index, and World Bank’s WGI were most often cited. However, in many instances, staff used indicators that have a less direct relationship with corruption and governance issues (e.g., World Bank’s Doing Business, Heritage Foundation’s Index, and the World Economic Forum’s global competitiveness report).11 See Annex II of this Note for a summary of how some of these indicators are constructed. Based on a survey of Fund mission chiefs, about half rely on third-party corruption indicators and other external information but noted familiarity with shortcomings in some indicators (see Note VI).

Note III, Table 3.

Third-Party Indicators of Corruption Used in Fund Documents (2005–2016)

article image

10. There also appears to be an uneven relationship between a country’s rank on external governance indicators and staff’s treatment of corruption in Fund documents. In accordance with the 1997 Guidance Note and the Fund’s mandate, governance should be addressed when it has been judged to have a significant macroeconomic impact. However, staff’s judgment of macroeconomic impact has been informed by a wide variety of sources; it does not appear to have been made solely based on third-party indicator rankings. In general, staff made use of the third-party indicators largely in order to compare a country’s indicator rank to its peers in the region. In other reports, staff’s limited usage of third-party indicators also briefly described whether the country had moved up positions in the third-party indicator rank or simply mentioned whether the country’s rank was too low compared to other countries.

D. Correlation Between Fund Engagement and Third-Party Indicator Rankings

11. As a supplement to the qualitative review of staff reports, a word search was conducted to gain insights into the correlation of Fund engagement on governance and corruption issues with third-party indicator rankings. Specifically, the number of references to governance-related terms in Fund documents was used as a proxy for the extent of engagement These data were compiled for two country samples:

  • a sample of all country documents in the Institutional Repository for the period 2012–2015; and

  • a narrower sample of country documents for 1976–2016 covering surveillance, UFR, press releases, TA reports, and assessment letters.12

12. For each sample, countries were ranked according to the number of references to “corruption,” “governance,” or “transparency” (or all three concepts together) (for methodology see Box 1). This ranking of countries was then compared to external corruption/governance indicators. For the latter, in addition to the four third-party indicators discussed above (TI, WGI, ICRG, and Maple), staff also looked at corruption indicators which were readily available for 2012–15. These are the World Economic Forum corruption index (WEF), the transformation index of the Bertelsmann Foundation, the rule of law index of the World Justice Project, the country risk assessment of EIU, and the country risk rating of Global Insight13

13. It is recognized that drawing conclusions and implications from a word search of key terms can be highly speculative. In particular, many policy measures relevant to governance and corruption can be discussed without using key terms such as “governance,” “transparency,” or “corruption,” and terms such as “governance” and “transparency” can have multiple meanings depending on the context. Also, the qualitative review of staff reports showed that even when these key terms are used, this does not necessarily mean that the Fund engages on the issue. Nevertheless, when combined with the detailed qualitative review, the word search may reveal certain broad trends that may be of interest.


14. Fund references to corruption, governance, and transparency for a given country are correlated but not to a high degree. Countries for which there are a high number of references to corruption in Board papers also tend to rank high in number of references to governance or transparency. The relationship is stronger over the longer 1997–2016 period than over the more recent few years (2012–15) (Table 4). The number of references in a country to governance and transparency are more similar than the number of references to either governance or transparency relative to corruption.

Note III, Table 4.

Correlation Matrix of Country Ranking Based on References in Fund Documents

article image

15. Rankings of countries based on the number of references to corruption or governance are also correlated with third-party corruption indicators. Results are presented below using the World Bank WGI data for Control of Corruption. However, similar results are obtained using other measures. The significant correlation between the number of references to corruption in Fund documents (See Box 1 on methodology) and the World Bank WGI metric suggests that the Fund’s engagement has some basis in independent data. The fact that Fund engagement as measured by references to governance or transparency is less closely correlated may reflect the different meanings of these terms, and their less direct relationship to the WGI Control of Corruption measure. The fact that correlations are much weaker over a short period may not be a cause for concern, given that the focus on corruption and governance issues can be expected to be irregular and addressed occasionally in annexes or special issues papers rather than on a regular basis.

Note III, Table 5.

Correlation of Country Ranking Based on References in Fund Documents with Third-Party Indicators

article image

16. Outliers on the evenhandedness comparisons. Countries such as Somalia, Equatorial Guinea, Libya, and Syria that rank high on corruption according to the WB WGI measure score low in terms of frequency of Fund references to corruption, primarily due to the scarce Fund interaction/engagement with these countries. Conversely, in Georgia, Indonesia, Kenya, and Ukraine the references to “corrupt” in Fund documents were more frequent than expected based on their ranking on third-party indicators of corruption. This is attributable to the deeper Fund engagement on governance/corruption issues with these countries.

17. Different external metrics. It is unclear whether more broad third-party indicators of “governance” might better explain the Fund’s references to governance issues than corruption-specific measures such as the WGI’s control of corruption.

Methodology for Measuring the Degree of Fund Attention to Corruption Issues

As a supplement to the qualitative review of staff reports, the number of references to corruption-related terms in Fund documents was generated in order to estimate the degree of the Fund’s attention to governance/corruption. Staff used Python word search across all documents in the Institutional Repository (IR) for 2012–15 to generate word count data related to corruption/governance and ranked the countries accordingly. Similarly, the word search was conducted on selected IR documents for 1997–2016 (TA reports, surveillance and UFR staff reports, press releases), controlled for duplication of reports.1 Reports that focus on more than one country were also dropped. The word-search counted the number of times the keywords “corrupt,” “governance,” and “transparency” occurred. The search did not analyze the context such words were included in the documents.

The data generated using Fund IR documents for 2012–15 were aggregated by country and year so as to obtain the total number of times these three keywords were mentioned in Fund documents in any given year.

For data generated using the three keywords on selected documents for 1997–2016 the result was aggregated by country only to obtain the total number of times the keywords “corruption” and “governance” were mentioned during a 19-year period.

1/ Documents included: TA reports classified as TAR, BUFFs, excl. ED statements, EVC docs related to country FO/Dis (such as assessment letters), MD/SP (related to a country); NB, PDP, EBS, PIN, PR, SM, and SUR.
Note III, Table 6.

Correlation Matrix: Third Party Indicators of Corruption and Word Count in Fund Documents (2012–15)

article image

E. Parallel Fund Work on the Use of Third-Party Indicators

18. In parallel with this stocktaking exercise, the Fund is also reviewing the use of third-party indicators in surveillance. This review is expected to result in a guidance note to staff on a principles-based approach to using third-party indicators in surveillance to ensure minimized reputational risks to the Fund while preserving flexibility for staff to make progress on issues identified in the Integrated Surveillance Decision14 and the Triennial Surveillance Review. The guidance note will be accompanied by an “Indicators Digest,” which is being developed to inform staff’s judgement on the use of potentially controversial third-party indicators.

Annex I of Note III. External Sources of Measurements on Corruption

1. Measuring corruption and its costs, consequences, and impact is critical for a wide range of stakeholders at a global, regional, and national level. Measurement of corruption itself is important in order to facilitate the analysis of corruption and its impact and to monitor and assess the results of measures to address corruption. Recognition of this fact has renewed interest by international organizations—and among aid donors, aid recipients, investors, and civil society—in developing measures of corruption, both in aid-financed projects as well as more broadly in both advanced and developing countries.

2. However, there are challenges in the measurement of corruption, and often a variety of indicators are used. The UNDP Guide has reviewed a selection of these indicators and acknowledged that trying to quantify the exact costs of corruption by using a single indicator is difficult given corruption’s hidden nature and the variety of forms it can take.1 2

3. Given the difficulty of capturing precise data on corruption, there has been a reliance on perceptions and experiences of corruptions. The UNDP Guide notes that there is a dearth of knowledge about the effectiveness of anti-corruption interventions due to weak reporting and evaluation standards. Currently, indicators tell us nothing about how corruption operates, nor do they differentiate between corruption that represents a transfer of funds and corruption that distorts the allocation of resources. Finally, these indicators do not directly measure the volume of bribes, the incidence of corruption, or its impact. Rather, they are better proxies of transparency, accountability, and integrity.

4. Organizations that produce governance indicators have used differing approaches to measuring corruption. This means that different organizations may use different quantitative and qualitative data of perceptions, experience, external assessments, administrative data, and a hybrid thereof as well as surveys, monitoring and evaluation systems, crowdsourcing, compliance reviews, and indicator/scorecard case studies to obtain the data. Understanding the type of input into an indicator and its limitations helps prevent misuse of the indicator. For example, an indicator measuring perceptions of corruption should not be assumed to reflect actual corruption; perceptions can rise and fall independently of levels or extent of corruption, and only a political economy analysis could clarify why perceptions are changing.

5. In general, there are two different types of corruption indicators: i) composite indicators; and ii) policy-relevant indicators. A snapshot of different indicators and the type of data used in each in indicator is reflected in Figure 1.

  • Composite indicators gather many different data points with the goal of broad topical coverage, global country coverage, or both. These indicators use statistical means of summarizing, combining, and organizing different types of data (e.g., perceptions, experiences, administrative information). In principle, the use of many data points can potentially result in greater accuracy, provided that the same concepts are measured consistently over time. Composite indicators, however, generally fall short since the methodologies and sources of data of many composite indicators change from year to year. Further, given the many different data inputs, the rankings of individual countries may take years to change, particularly if similarly-ranked countries are also undergoing reforms. The aggregation of data may also obscure low scores on some underlying indicators. Finally, given that the composite indicators give varying weight to underlying input, it is important that the user has a clear understanding of the relationship between the input used and the final composite measurement, which is often not possible.

Figure 1.
Figure 1.

Subjective vs. Objective

Citation: Policy Papers 2017, 005; 10.5089/9781498346481.007.A003

Source: UNDP 2015

6. Examples of composite indicators are Transparency International’s Corruption Perceptions Index (CPI), and World Bank’s Control of Corruption Index (CCI). The CPI has been published annually since 1995.3 The CPI is a compilation of data from other sources that are merged to generate a single number for a country. CPI’s scores tend to persist over time, with only few countries showing marked improvement or deterioration. The other major cross-country index is the CCI, which is part of the Worldwide Governance Indicators. The CCI is also a compilation, including most of the same sources and countries as the CPI.4 The methodology is somewhat different, but the two indices are highly correlated, and scores generally fall within the margin of errors of each.

7. Policy-relevant indicators are useful, necessary, and complement composite indicators. They can serve as the underlying data aggregated in composite indicators or can stand alone as individual data points. Such indicators can demonstrate variation in outcomes of anti-corruption policies within a country. They also facilitate benchmarking across provinces and within national boundaries, given their narrower focus, and provide more robust information about the local drivers of change. However, these datasets necessarily build local context into their frameworks, potentially preventing meaningful cross-country analysis. Finally, this kind of data can be very difficult to collect, especially since corruption is a covert phenomenon.

8. In some external composite indicators (e.g., WGI, Transparency International), the underlying data may range from subjective perceptions to specific laws and statistics.

9. Perceptions-based input includes opinions by ordinary citizens, business owners, or experts on specific topics, often expressed through public opinion surveys and hypothetical vignettes. These types of inputs are helpful for capturing information about topics that are difficult to conceptualize for objective data collection, such as public trust, civic space, grand or political corruption, and client preferences. They also fill a gap where administrative data are unavailable, such as for the quality of public administration or governments, and can include anecdotal information on the frequency, location, and cost of bribes, or the incidence and severity of crimes, as well as the extent of knowledge about specific laws, policies, or practices.

10. Corruption perceptions are an important part of citizens’ attitude towards political systems and leaders and affect the level of political trust in a society. It is well known that, in turn, this trust can be an important determinant of investment decisions, political participation, and other behaviors with economic consequences. It should be highlighted, however, that corruption perception data do not measure actual levels of corruption, point to what policies are most effective at changing perceptions of a country’s level of corruption, or show what the impact of perceptions on real variables is. Examples of perceptions-based indicators include Gallup public opinion polls, and Transparency International’s Bribe Payers Index Experiences.

11. Experiences-based input provides specific citizen experiences (or knowledge). These data are commonly used in service delivery fields, such as health, education, law enforcement, and transportation. Experiences data are often collected through surveys, but face-to-face survey-based interviews are also common. As with perceptions data, surveys of experiences may result in higher data-collection costs in order to ensure a sample size that reduces the margin of error. They are useful when administrative data are unavailable and can be used to measure the extent and nature of petty corruption, such as bribes, in particular sectors. They are also very helpful in supplementing performance data collected by government agencies and can identify bottlenecks and problems at the government-citizen interface. Examples of experiences-based indicators include the World Economic Forum Competitiveness Report, the Latin American Public Opinion Survey (LAPOP), crime victimization surveys, the Kenya Urban Bribery Index, and Ushahidi platforms (crowdsourcing).

12. External assessments capture data through scoring, rating, or ranking and are often some of the most popular global datasets. This type of data may constitute either expert assessments or citizen assessments. Often, the former tend to focus on country or institutional performance (e.g., grand corruption in extractive industries, transparency in public finance), whereas the latter often concentrate on micro-level impacts (e.g., petty corruption in education ministries, fraud in the provincial health system). External assessments can provide much more specific data on petty corruption and service delivery at the country or community level than other types of indicators. Because of the lower cost involved in data collection and quality control (e.g., online surveys, no travel, no interviews), it is easier to cover a large number of countries. These assessments are based on administrative data (further described below) or third-party reports—such as case studies, audit reports, or agency statistics—and in this way can be understood as “evidence-based” assessments of corruption and governance. Examples of external assessment-based indicators include the Open Budget Index, the Financial Secrecy Index, Benchmarking Public Procurement, and the Campaign Finance Indicator.

13. Administrative data consist of agency statistics or performance data generated by governments about their own activities, as well as audit reports or project reports, compliance or field tests, and citizen feedback or observations. (e.g., data from the anti-corruption agency). These data are useful for assessing the quality of government resources, processes and performance. These are also the easiest data to translate into action, since the data already closely adhere to existing public sector functions. However, there are questions about the reliability of self-reported data in government monitoring and evaluation systems.

14. In general, methodological approaches are constantly evolving to address the difficulty of measuring corruption. While some progress has been made in the areas of data collection, analysis and dissemination, challenges remain in the areas of coverage (global vs. local), subjectivity (perception-based not being fully based on facts), comparisons over time (given changing methodologies and data sources), data quality and reliability, and cross-country comparisons.

Annex II of Note III. Methodology: External Governance/Corruption Measures

This Annex explores some of the key external governance/corruption measures. An analysis of some the benefits and shortcomings of each of the measurements are set out in the UNDP Guide and summarized in Annex I.

Transparency International -Corruption Perceptions Index (CPI)1

1. The CPI ranks countries and territories based on how corrupt their public sector is perceived to be since 1995 (latest 2016). It is a composite index—a combination of polls-drawing on corruption-related data collected by a variety of reputable institutions. It is calculated using 12 different data sources from 11 different institutions that capture perceptions of corruption within the past two years. The index reflects the views of observers from around the world, including experts living and working in the 180 countries and territories evaluated.

2. Select data sources: Each data source that is used to construct the CPI must fulfil the following criteria to qualify as a valid source:

  • Quantifies perceptions of corruption in the public sector

  • Be based on a reliable and valid methodology, which scores and ranks multiple countries on the same scale

  • Performed by a credible institution and expected to be repeated regularly

  • Allow for sufficient variation of scores to distinguish between countries.

3. Standardize data sources: Data scores are standardized to a scale of 0–100 where a 0 equals the highest level of perceived corruption and 100 equals the lowest level of perceived corruption. This is done by subtracting the mean of the data set and dividing by the standard deviation and results in z-scores, which are then adjusted to have a mean of approximately 45 and a standard deviation of approximately 20 so that the data set fits the CPI’s 0–100 scale. The mean and standard deviation are taken from the 2012 scores, so that the rescaled scores can be compared over time against the baseline year. Because of the update in the methodology CPI scores before 2012 are not comparable over time.

4. Calculate the average: For a country or territory to be included in the CPI, a minimum of three sources must assess that country. A country’s CPI score is then calculated as the average of all standardized scores available for that country. Scores are rounded to whole numbers.

5. Report a measure of uncertainty: The CPI is accompanied by a standard error and confidence interval associated with the score, which capture the variation in scores of the data sources available for that country/territory.

Transparency International -Global Corruption Barometer (GCB)2

6. The GCB draws on a survey of 100,000+ respondents in 100+ countries. The surveys are conducted through face-to-face or telephone interviews. It addresses people’s direct experiences with bribery and details their views on the overall prevalence of corruption in the main institutions in their countries. It also captures their perception of the government’s anti-corruption performances, amongst others. Since the survey reflects on experiences of the population at large, it only captures street-level corruption experiences and not grand corruption by high-level officials. Corruption incidences reported in the GCB measures the user-based incidences by tabulating experiences based on survey of respondents’ use of various services and if they paid a bribe.

7. Beginning in 2015, the GCB questionnaire was fielded in collaboration with a number of regional surveys networks and has been published as regional GCB reports for Sub-Saharan Africa, Middle East & North Africa, Europe and Central Asia, and Asia-Pacific. The Americas regional report as well as a summary Global Corruption Barometer report will be published by mid-2017. The most recent complete GCB survey is the 2013 Barometer which reflects the responses of more than 114,000 people in 107 countries• The survey sample in each country has been weighted to be nationally representative where possible. Global results are based on the entire sample, i.e., one response is counted as one vote. For most countries, the sample size is approximately 1000.

Worldwide Governance Indicators (WGI)3

8. The WGI reports aggregate and individual governance indicators for over 200 countries and territories over the period 1996–2015, for six dimensions of governance:

  • Voice and Accountability

  • Political Stability and Absence of Violence

  • Government Effectiveness

  • Regulatory Quality

  • Rule of Law

  • Control of Corruption.

9. These aggregate indicators combine the views of a large number of enterprise, citizen and expert survey respondents in industrial and developing countries. They are based on over 30 individual data sources produced by a variety of survey institutes, think tanks, non-governmental organizations, international organizations, and private sector firms.

10. The last dimension in the WGI, the Control of Corruption (CCI) uses a broad concept of corruption, covering not only perception of corruption but also victimization and anticorruption institutions such as electoral integrity and freedom of the press. The CCI is reported as a normalized distribution, with zero mean and a standard deviation equal to one, hence avoiding arbitrary cutoff points at the top and bottom of the scale. One disadvantage of such scaling is that it cannot be used to measure global trends but only how countries fare relative to each other.

International Country Risk Guide (ICRG) Corruption Risk Index4

11. The PRS Group Inc.’s ICRG has been published monthly since 1984. ICRG rating and forecasting methodology comprises 22 metrics affecting three categories of country risk: political, financial, and economic. Multiple tables are available for each of these subcategories.

12. ICRG’s political risk rating includes 12 weighted variables (and 15 subcomponents) covering an extensive menu of risks, including those that affect the overall stability of government, socio-economic conditions, investment risks, and other risks related to a variety of conditions internal and external to the country in question. Assessments are made using both qualitative and quantitative data with the former ‘harnessed’ through the application of proprietary risk bands.

13. PRS has a roster of analysts located globally, and the firm’s universe of country coverage includes 140 developed, emerging and frontier markets. However, the firm’s coverage does not include small islands/states, except for four states (Bahamas, Guyana, Surinam, and Trinidad &Tobago), but provides its clientele, on request, with specialized risk analyses on countries not covered on a regular basis.

14. One sub-component of the political risk index is the level of corruption within the political system. There is a 6-point corruption risk assessment that is utilized when measuring corruption. The low score (0) represents a very high risk, whereas a high score (6) means the risk posed by corruption is minimal. Scores are rounded to the nearest 0.5.

15. One of the main indicators used in assessing country graft is government’s longevity in office, which may reflect corrupt practices. While the measure takes financial corruption into account, through demands for special payments and bribes connected with import and export licenses, exchange controls, tax assessments, police protection, or loans, the index is more concerned with actual or potential corruption in the form of excessive patronage, nepotism, job reservations, ‘favor-for-favors’, secret party funding, and suspiciously close ties between politics and business.

16. ICRG’s corruption component measures the risk posed by corruption to the private sector and factors in incidents of graft. However, since corruption is typically hidden, perception indices are taken as a proxy for corruption and, in this context, PRS relies on a network of individuals and institutions (“experts”) to uniformly assess corruption across countries.

Maplecroft Corruption Risk Index (CRI)5

17. A sub-component of Maplecroft’s governance index is the Corruption Risk Index (CRI). It has been developed to enable companies to identify the countries where the risk of association with corruption is highest. The CRI evaluates 198 countries on the reported prevalence and persistence of corruption in the public and private sectors, as well as the efficiency of governments in tackling the issue.

18. CRI is a qualitative survey index. Analysts assess ten key indicators relating to the legislative framework (structure); anti-corruption implementation bodies and the practical applications of the law (process), and the frequency with which various kinds of corruption occur (outcomes). It assesses risk by modelling the strength of anti-corruption legislation, the efficacy and independence of anti-corruption bodies and the prevalence of corruption from a business perspective, including distribution, petty and grand corruption.

19. The index is constructed by averaging scores for Structure and Process questions (see below) and combining that with the Outcome score. The outcome score is weighted twice as much as the average Structure and Process score. Thus, the outcome pillar is the most important in determining a country’s overall score in the index.

20. Structure and Process questions used for the index are related to whether: legislation is in place to tackle corruption in the public sector; political appointments are subject to patronage or nepotism; the public budget process is transparent; the recipients of government contracts are published; there are anti-corruption investigation or prosecution bodies; and corruption is a particular concern for business from an operational perspective, for example regarding business permits and planning.

21. The scores for grand and petty corruption are multiplied and assigned a score based on an S-curve model where the highest frequencies of grand and petty corruption are assigned higher-risk scores.

22. The CRI methodology has been updated to provide a more granular assessment of the issue. While highly correlated, and capturing the same issue, the methodologies are not directly comparable around this break, and time series analysis should be handled with care.

World Bank’s Country Policy and Institutional Assessment (CPIA)6

23. The CPIA is a diagnostic tool compiled by the World Bank. It is designed to assess the quality of a country’s policies and institutional arrangements according to 16 criteria currently covering economic management, structural policies, policies for social inclusion and equity, as well as public sector management and institutions.

Selection/Performance Criteria:

24. The 16 CPIA criteria are grouped into four clusters:

  • Economic Management: Macroeconomic Management; Fiscal Policy; Debt Policy

  • Structural Policies: Trade; Financial Sector; Business Regulatory Environment

  • Policies for Social Inclusion/Equity: Gender Equality; Equity of Public Resource Use; Building Human Resources; Social Protection and Labor; Policies and Institutions for Environmental Sustainability

  • Public Sector Management and Institutions: Property Rights and Rule Based Governance; Quality of Budgetary and Financial Management; Efficiency of Revenue Mobilization; Quality of Public Administration; Transparency, Accountability, and Corruption in Public Sector.

25. The country teams prepare rating proposals that are based on their informed professional judgment. To ensure that scores are consistent across countries and regions, the country team proposals undergo a series of checks and balances in which they are reviewed first within each operational Bank Region by the respective Chief Economist, and then by sector specialists in Global Practices and Cross-Cutting Solutions Areas and by staff from central departments. For each criterion, countries are rated on a scale of 1 (low) to 6 (high). Each of the four clusters gets equal weight in overall rating. Also within each cluster, each criterion receives equal weights. The CPIA data is produced for all World Bank clients and are publicly disclosed only for IDA-eligible low-income countries.7

Annex III of Note III. Average Historical Global Percentile Rankings of Economies by Indicator

Note III, Annex III, Table 1.

Perceptions Index (World Bank WGI CCI), Average of 2005–151/

article image
article image

These comparative indicators are developed by external organizations. While each offers insights on the extent of corruption, they also have inherent limitations and do not by themselves, or collectively. represent the Fund’s assessment of corruption in any country

Note III, Annex III, Table 2.

Experience Index (Transparency International GCB), Average of 2005–131/

article image

These comparative indicators are developed by external organizations. While each offers insights on the extent of corruption, they also have inherent limitations and do not by themselves, or collectively. represent the Fund’s assessment of corruption in any country

Note III, Annex III, Table 3.

Institutions Index (Maplecroft CRI), Average of 2013–161/

article image
article image

These comparative indicators are developed by external organizations. While each offers insights on the extent of corruption, they also have inherent limitations and do not by themselves, or collectively. represent the Fund’s assessment of corruption in any country


Prepared by Julianne Ams, Cecilia Marian, Concha Verdugo Yepes (all LEG), and Haimanot Teferra (SPR).


A detailed analysis of the different measures of corruption and their pros and cons can be found in User’s Guide to Measuring Corruption and Anti-Corruption, United Nations Development Programme, 2015.


For the Executive Board’s position, see The Chairman’s Summing Up—Biennial Review of the Implementation of the Fund’s Surveillance and of the 1977 Surveillance Decision, 7/23/2004 (encouraging greater use of existing indicators generally as well as ROSCs and other available materials to refine coverage of governance issues); see also the underlying staff paper Biennial Review of the Implementation of the Fund’s Surveillance and of the 1977 Surveillance Decision—Content of Surveillance, 7/2/2004, ¶139 (concluding that the use of indicators such as those developed by the World Bank and Transparency International should be encouraged). For examples of staff guidance, see Guidance Note on Jobs and Growth Issues in Surveillance and Program Work,9/30/2013, Annex I ¶11 (noting usefulness of governance indicators generally to examine economic constraints); Flexible Credit Line—Operational Guidance Note, 6/2/2015 (relevant indicators to inform judgment on strength of member’s institutional policy framework are the government effectiveness and control of corruption indicators from the World Bank); Precautionary and Liquidity Line—Operational Guidance Note, 6/2/2015 (same); Review of the Flexible Credit Line, the Precautionary and Liquidity Line, and the Rapid Financing Instrument—Specific Proposals, 5/1/2014 (noting that “no single indicator currently exists that can sufficiently summarize information about the quality of a country’s institutional policy frameworks,” but suggesting use of the ICRG and WGI); Debt Limits in Fund-Supported Programs—Proposed New Guidelines, 8/7/2009, ¶¶ 9, 10 and Table 2 (for assessing debt capacity, highlighting the use of, inter alia, the Public Expenditure and Financial Accountability Framework; the World Bank’s CPIA, DeMAP, and WGI; and Project Performance Assessments); Staff Guidance Note on the Application of the Joint Fund-Bank Debt Sustainability Framework for Low-Income Countries, 10/8/2008, Box 2 (listing indicators that can help establish link between public expenditure and growth, including CPIA, Public Expenditure and Financial Accountability Framework, public expenditure management analyses, and “public governance indicators” more generally). See also Review of Some Aspects of the Low-Income Country Debt Sustainability Framework, 8/7/2009 (discussing how CPIA scores may be used in determining debt distress thresholds).


UNDPs Guide to Measuring Corruption and Anti-Corruption (2015).


“Corruption in Developing Countries,” Annual Review of Economics, 2012, vol. 4, issue 1, pages 479–509.


For example, the World Bank provides a disclaimer on the use of the WGI to compare perceptions across countries and over time, noting that “a key feature of the WGI is that all country scores are accompanied by standard errors. These standard errors reflect the number of sources available for a country and the extent to which these sources agree with each other (with more sources and more agreement leading to smaller standard errors). These standard errors reflect the reality that governance is difficult to measure using any kind of data.” For more details, see Annex II and


While the Fund does produce indicators on fiscal, budget, and investment issues, these are not directly focused on governance. Fiscal Transparency Evaluations (FTEs) are the Fund’s fiscal transparency diagnostic. FTEs provide countries with: a comprehensive assessment of their fiscal transparency practices against the standards set by the Fiscal Transparency Code; rigorous analysis of the scale and sources of fiscal vulnerability based on a set of fiscal transparency indicators; a visual account of their fiscal transparency strengths and reform priorities through summary heat maps; a sequenced fiscal transparency action plan to help them address those reform priorities; and the option of undertaking a modular assessment focused on just one pillar of the Code. A number of FTEs have been conducted to date in countries across a wide range of regions and income level.


The Chairman’s Summing Up—Biennial Review of the Implementation of the Fund’s Surveillance and of the 1977 Surveillance Decision, 7/23/2004; Biennial Review of the Implementation of the Fund’s Surveillance and of the 1977 Surveillance Decision—Content of Surveillance, 7/2/2004, ¶139 (providing as examples indicators developed by the World Bank and Transparency International).


Guidance Note on Jobs and Growth Issues in Surveillance and Program Work, 9/30/2013, Annex 1 ¶11.


Flexible Credit Line—Operational Guidance Note,, 6/2/2015 (listing the government effectiveness and control of corruption indicators from the World Bank); Precautionary and Liquidity Line—Operational Guidance Note, 6/2/2015 (same); Review of the Flexible Credit Line, the Precautionary and Liquidity Line, and the Rapid Financing Instrument— Specific Proposals, 5/1/2014 (suggesting use of the ICRG and World Bank Governance Indicators Database); Debt Limits in Fund-Supported Programs—Proposed New Guidelines, 8/7/2009, ¶10 (highlighting, inter alia, the Public Expenditure and Financial Accountability Framework, the World Bank’s DeMAP and WGI, Project Performance Assessments); Staff Guidance Note on the Application of the Joint Fund-Bank Debt Sustainability Framework for Low-Income Countries, 10/8/2008, Box 2. See also Review of Some Aspects of the Low-Income Country Debt Sustainability Framework, 8/7/2009.


See Annex II on how these indicators are constructed and a discussion on how methodological approaches are constantly evolving to address the difficulty of measuring corruption.


In comparison to the first sample of reports, this excludes Board grays, duplication of country documents (ISCR), Decisions, Executive Board documents such as PRSP and related government documents, and Working Papers.


These indicators are also inputs for the index of Transparency International.


See “Modernizing the Legal Framework for Surveillance—An Integrated Surveillance Decision.”


A user’s guide to measuring corruption (UNDP 2015)


See also Rose-Ackerman & Palifka, Corruption and Government (2016). This annex also includes input from the IADB Seminar on the use of Corruption Indicators in Latin American countries.


The CPI is available on Transparency International’s website.


The CCI is reported as normalized distribution, with a zero mean and a standard deviation equal to one. This form has the advantage of not imposing cutoff points at the top and the bottom of the scale, but it is centered at zero each year. Hence, it cannot measure global trends, but can only show how countries fare relative to each other. See discussions on World Bank and Transparency International indicators on corruption in Donchev and Ujhelvi (2014), What Do Corruption Indices Measure?


For more information, see Transparency International’s Corruption Perceptions Index website.


For more information, see Transparency International’s Global Corruption Barometer website.


For more information, see the World Bank’s Worldwide Governance Indicators website.


For details see PRS/ICRG Methodology document


For details, see


For details See World Bank Group on CPIA methodology


CPIA data is not compiled for OECD and high-income countries.

The Role of the Fund in Governance Issues - Review of the Guidance Note - Preliminary Considerations - Background Notes
Author: International Monetary Fund