Lably
Skip to content

Corruption Perceptions Index

Country ranking by public sector corruption

  • 11 min read
  • 6 chapters

Related on Lably

Through Public sector, Singapore and World Bank Group

From 1995 to 2011, the index was scored on a scale of 10 to 0. Since 2012, the Corruption Perceptions Index has been ranked on a scale from 100 (very clean) to 0 (highly corrupt).

Of the 182 countries ranked in the 2025 CPI, published in February 2026, the top scorers included Denmark (89), Finland (88), and Singapore (84), while those perceived as the most corrupt included South Sudan (9), Somalia (9), and Venezuela (10).

Although widely used as a key indicator of corruption, the CPI does not capture all forms of corruption. Perceptions about corruption may differ from actual levels of corruption and the index focuses solely on the public sector. For a more comprehensive picture, the CPI should be used alongside other assessments.

Methods

The CPI methodology, revised in 2012, enables consistent comparison of corruption perceptions across countries and years. It involves four stages: selecting data sources, rescaling them to a uniform 0–100 scale, aggregating the results, and estimating uncertainty.

Selection of source data

The goal of the data selection is to capture expert and business leader assessments of various public sector corruption practices. This includes bribery, misuse of public funds, abuse of public office for personal gain, nepotism in civil service, and state capture. Since 2012 CPI has taken into account 13 different surveys and assessments from 12 different institutions. The institutions are:

Countries need to be evaluated by at least three sources to appear in the CPI. The CPI measures the perception of corruption due to the difficulty of measuring absolute levels of corruption. Transparency International commissioned the University of Passau's Johann Graf Lambsdorff to produce the CPI. Early CPIs used public opinion surveys.

Rescaling source data

In order for all data to be aggregated into the CPI, it is first necessary to carry out standardization during which all data points are converted to a scale of 0–100. Here, 0 represents the most corruption and 100 signifies the least. Indices originally measuring corruption inversely (higher values for higher corruption) are multiplied by -1 to align with the 0–100 scale.

In the next step, the mean and standard deviation for each data source based on data from the baseline year are calculated (the "impute" command of the STATA statistical software package is used to replace missing values). Subsequently, a standardized z score is calculated with an average centered around 0 and a standard deviation of 1 for each source from each country. Finally, these scores are converted back to a 0–100 scale with a mean of approximately 45 and a standard deviation of 20. Scores below 0 are set to 0, and scores exceeding 100 are capped at 100. This ensures consistent comparability across years since 2012.

Aggregating the rescaled data

The resulting CPI score for each country is calculated as a simple average of all its rescaled scores that are available for the given country, while at least three data sources must be available in order to calculate the index. The imputed data is used only for standardization and is not used as a score to calculate the index.

Reporting a measure for uncertainty

The CPI score is accompanied by a standard error and confidence interval. This reflects the variation present within the data sources used for a particular country or territory.

Sign in to keep reading

5 more chapters are free with a Lably account.

More in Economics