Worldwide Governance Indicators

Governance, the traditions and institutions by which authority in a country is exercised, is widely recognized as a critical driver of development outcomes. Theoretical and empirical research shows that countries with inclusive and accountable institutions tend to achieve higher levels of economic development, better public service delivery, and more robust job creation.

The Worldwide Governance Indicators (WGI) are one of the leading global initiatives for measuring governance. The WGI draw on 35 cross-country data sources—including household surveys, firm surveys, and expert assessments—and are annual composite indicators based on perceptions of governance quality across more than 200 economies from 1996 to 2024.

Across these 35 data sources, more than 400 underlying indicators are used to capture the six dimensions of governance included in the WGI:

The WGI draws on 35 data sources that capture the perceptions of citizens, entrepreneurs, and experts on the six dimensions of governance in their countries.

All data sources must meet the following six criteria:

• Relevance: Sources must capture subjective perceptions relating to at least one of the six WGI governance dimensions.
• Originality: Sources must be based on primary data collection from household/firm surveys or expert assessments. Compilations or secondary products derived from other governance indicators are excluded.
• Cross-country coverage: Sources must cover multiple countries to permit meaningful international comparison, typically at regional, global, or major economic-grouping scales.
• Regular updates: Sources must be updated on a recurring basis. Annual frequency is preferred, but lower-frequency updates (e.g., every three years) may be considered for high-quality sources with broad coverage, particularly large-scale surveys. 
• Transparency: Sources must provide well-documented information on their methodology, including respondent selection, data collection processes, scoring methods, and quality assurance procedures
• Independence: Sources must be produced by recognized multilateral organizations, non-governmental organizations, academic institutions, or commercial data providers. Indicators produced solely by national governments are excluded.

The 35 data sources included in the revised WGI fall into two broad categories: surveys of firms and households, and expert assessments. Those expert assessments are produced by multilateral organizations, nongovernmental organizations, or commercial data providers:

• Surveys of households and firms, including the Gallup World Poll, the regional Barometers, the World Bank's Enterprise Surveys, and the World Economic Forum's Executive Opinion Survey.
• International organizations (IO), including the Country Policy and Institutional Assessments (CPIA) assessments of the World Bank and regional development banks.
• Non-governmental organizations (NGO), including Freedom House, Reporters Without Borders, and World Justice Project.
• Commercial data providers (CDP), including the Economist Intelligence Unit, the PRS Group, and S&P Global.

Please find the full list of WGI data sources and the individual questions and variables from each source used in the WGI in the attachments below.

This release of the WGI supersedes previous releases.  To access previous versions of the WGI dataset, please click here.

Suggestions for additional data sources that meet these criteria are welcome-–please email suggestions to wgi@worldbank.org.

A country’s governance scores for each of the six WGI governance dimensions are estimated using the following four steps:

STEP 1:  Assigning indicators from the underlying sources to the six governance dimensions.  Individual questions or variables from the underlying data sources are mapped to up to two of the six governance dimensions. For example, a firm-survey question on the regulatory environment would be assigned to Regulatory Quality, or a CPIA indicator on transparency, accountability, and corruption would be assigned to both Voice and Accountability and Control of Corruption. The list of variables used in the WGI (for the recent year for each underlying data source), and how they are mapped to the six dimensions, can be found by clicking on the dimension names listed here. Note that not all data sources cover all countries, hence the governance scores for each country are based on different sets of underlying data.

STEP 2:  Rescaling the individual source data to range from 0 to 1.  Each question from the underlying data sources is rescaled to range from 0 to 1, with higher values corresponding to better governance outcomes. For example, if a survey question uses a scale from 1 to 4, a response of 2 would be rescaled as (2 – 1) / (4 – 1) = 0.33. When a data source provides more than one question or variable relating to a given governance dimension, the rescaled values are averaged to form source-by-dimension value. This ensures that each source contributes only one signal per dimension and prevents data sources with more questions or variables from exerting undue influence.

A combined dataset containing all source data in a single file is available in both Excel and Stata formats. Although the variables are nominally rescaled to the same 0–1 range, this underlying data are not necessarily comparable across sources. For example, one data source may use a 0–10 scale but have most scores clustered between 6 and 10, while another uses the same 0–10 scale but spreads responses across the full range. The max–min rescaling described above does not correct for these source differences, but the Unobserved Components Model (UCM) procedure used to construct the governance scores does (see below).

STEP 3:  Using an Unobserved Components Model to construct a governance estimate for each dimension by taking a weighted average of the source-by-dimension data. To aggregate data across multiple sources, the WGI uses a statistical technique known as an Unobserved Components Model (UCM). The UCM first makes the rescaled 0–1 data comparable across sources and then constructs a weighted average of the data for each country and each dimension. The model assumes that the observed data from each source are a linear function of the unobserved level of governance plus an error term. Because this linear relationship differs across sources, the UCM adjusts for the remaining non-comparability in units noted above.

The resulting governance estimates are a weighted average of the available source data, with weights determined by the pattern of correlations among sources. Summary information for the weights applied to the component indicators, as well as the parameter estimates from the UCM, can be downloaded in Excel and Stata format. The UCM assigns greater weights to sources that are more strongly correlated with the others. The resulting governance estimates are expressed in units of a standard normal distribution, ranging from approximately –2.5 to 2.5, with higher values indicating better governance.

STEP 4:  Transforming the UCM-generated governance estimates to a 0–100 absolute governance score. The WGI transform the governance estimates for each country, year, and dimension—typically ranging from approximately –2.5 to 2.5 —into absolute scores on a 0–100 scale, with 100 representing the best absolute governance performance. This transformation requires establishing fixed upper and lower performance benchmarks for each dimension. Two hypothetical benchmark countries are therefore constructed: a “best-case performer’’ and a “worst-case performer.’’ The best-case performer is assigned a value of 1 on every representative source indicator, and the worst-case performer is assigned a value of 0. Their governance estimates define the upper and lower bounds of the governance scale for that dimension and correspond, by definition, to absolute scores of 100 and 0 on the 0–100 scale.

The benchmark estimates are calculated separately for each dimension and each year. All UCM-generated country–year estimates are mapped linearly between these two benchmarks to obtain absolute governance scores: absolute score for a country = 100 × (governance estimate – worst-case performer estimate) / (best-case performer estimate – worst-case performer estimate). 

Details on the aggregation procedure can be found in the WGI methodology paper. A full reproducibility package is available here.

The six estimated WGI dimension measures, along with the underlying source data, are available interactively here. The full dataset can also be downloaded as an Excel file here. The six dimension measures are reported in two ways: (1) a governance estimate, which is normally distributed with a mean near zero and typically ranges between –2.5 and 2.5, and (2) as an absolute score, ranging from 0 to 100.

A key feature of the WGI is that all country scores are accompanied by standard errors. These reflect both the number of sources available for a country and the degree of agreement among them—more sources and stronger agreement result in smaller standard errors. This acknowledges the reality that governance is inherently difficult to measure with precision. While most governance or investment climate measures leave this uncertainty implicit, the WGI methodology makes it transparent.

When comparing governance scores between two countries, it is important to take margins of error into account. A useful rule of thumb is that if the two confidence intervals overlap, the difference is not statistically, or likely practically, significant. In other words, small differences in aggregate governance scores should not be over-interpreted when they fall within the reported margins of error.

As with comparisons across countries, comparisons of a country’s governance scores over time must also take margins of error into account. If the confidence intervals for two periods overlap, the WGI should not be interpreted as signaling a meaningful change in governance. Most year-to-year movements in the WGI are small relative to their margins of error and therefore are neither statistically nor practically significant. For this reason, they should not be over-interpreted as indicating real short-run shifts in governance performance. However, over longer periods—such as a decade—the WGI do show statistically significant trends for a number of countries.

Changes in a country’s WGI score over time reflect a combination of two factors: (1) changes in the underlying source data (the survey responses or expert assessments themselves), and (2) the addition of new data sources that became available only in later years or the removal of sources that stop updating. For large, statistically significant changes over longer periods, movements in the underlying source data are usually the primary driver. For smaller, often insignificant changes over shorter periods, both factors typically paly a role. Users can consult the underlying data sources to better understand the reasons for changes in a country’s governance scores.

The papers listed in this page describe previous rounds of the WGI. Many of them also contain analytical work on specific methodological issues relating to the measurement of governance using perception data.

WGI Working Papers

• The Worldwide Governance Indicators: 2025 Methodology Revision

• The Worldwide Governance Indicators: 2024 Methodology Update

• The Worldwide Governance Indicators: 2010 Methodology and Analytical issues

• Governance Matters VIII: Governance Indicators for 1996-2008

• Governance Matters VII: Governance Indicators for 1996-2007

• Governance Matters VI: Governance Indicators for 1996-2006

• Governance Matters V: Governance Indicators for 1996-2005

• Governance Matters IV: Governance Indicators for 1996-2004

• Governance Matters III: Governance Indicators for 1996-2002

• Governance Matters II: Governance Indicators for 2000-2001

• Governance Matters

• Aggregating Governance Indicators

Critiques and Responses

• The Worldwide Governance Indicators project: Answering the Critics

• Governance and Growth: Reply/Rejoinder

External Panel Review

• External Review of the Worldwide Governance Indicators