Pakistan’s score on the 2025 Corruption Perceptions Index (CPI) by Transparency International increased by one point to 28 out of 100, ranking the country 136th among 182 nations. However, a closer review of the data and methodology used to generate this ranking reveals significant questions about the score’s derivation, the extent of independent verification, and how well international perceptions of corruption align with the everyday experiences of Pakistani citizens.
These concerns extend beyond the ranking itself, as corruption and governance indicators influence the broader assessment of country risk by foreign investors, businesses, lenders, and international institutions. Although the CPI is not an investment rating or tied to IMF lending conditions, it contributes to the overall picture of economic and governance risks. The IMF’s 2025 Governance and Corruption Diagnostic for Pakistan highlights that corruption and governance weaknesses impede economic growth and investment, noting that each one-point CPI improvement correlates with a 4% increase in investment rates—though this is an association rather than definitive causation.
Transparency International’s 2025 CPI, published in February 2026, showed Pakistan’s score rising from 27 to 28, while its rank slipped one place from 135th to 136th due to changes in the number and performance of countries assessed. TI emphasizes that scores carry more weight than ranks because rankings can shift based on the comparative performance of other nations. Notably, Pakistan’s CPI score was not derived from surveys of its citizens but from eight external data sources, out of a possible thirteen used globally. The CPI measures perceived public-sector corruption as seen by experts and business executives.
Of the eight sources informing Pakistan’s score, five remained steady, two declined slightly, and one—V-Dem—showed a substantial increase. Specifically, Bertelsmann’s score stayed at 21, the Economist Intelligence Unit at 18, Global Insights at 32, PRS ICRG at 33, and the World Bank’s CPIA at 39. Meanwhile, the World Economic Forum’s score dropped from 33 to 32, and the World Justice Project fell from 26 to 25. V-Dem’s score rose significantly from 14 to 19, representing the only notable positive shift among the inputs.
The total of these published source scores is 219, averaging 27.375, slightly below the overall CPI score of 28. This discrepancy is not an error but a result of TI using unrounded underlying data, making exact replication of the CPI score from published whole numbers impossible. Transparency questions focus on the precise unrounded averages and how close Pakistan’s score was to the threshold for a 28 rating, especially since V-Dem’s rise largely accounts for the improvement.
V-Dem’s methodology relies on expert assessments combined with statistical modeling, and its datasets can be updated as new versions are released. Key questions remain about what caused the increase in Pakistan’s V-Dem score: whether it was changes in expert composition, revisions to indicators, updates to historical data, or methodological adjustments. However, the exact unrounded V-Dem score provided to TI is not publicly available.
With seven of the eight source scores either unchanged or lower, the five-point increase in V-Dem’s score is the primary driver behind Pakistan’s one-point CPI rise. It is important to note that these eight data sources are not uniform surveys; they vary widely in approach. For example, the World Economic Forum uses executive opinion surveys, the World Justice Project combines household and expert data, Bertelsmann relies on expert assessments, V-Dem employs expert coding and modeling, while the Economist Intelligence Unit, Global Insights, and PRS utilize country-risk evaluations. The World Bank CPIA assesses institutional performance. This diversity means that the CPI aggregates data from different respondent types, definitions, time frames, and disclosure levels.
Transparency International standardizes these varied scores and assigns equal weight in calculating the overall country score. While this methodological choice is transparent, it raises a fundamental question: should data derived from such different sources carry the same influence in determining a country’s corruption rating? Another issue is independence; although the eight organizations are institutionally distinct, their assessments may not be statistically independent, as they could rely on overlapping information such as court cases, government documents, media reports, or expert networks. This does not imply bias but suggests that the “eight sources” may not represent eight fully independent measures of corruption.
Furthermore, the 2025 CPI score does not reflect a single calendar year’s data. The underlying assessments cover varying periods, with some using 2024 data, others specific months in 2025, and some spanning longer timeframes. Consequently, reforms implemented late in 2025 may not be captured uniformly across all sources, and earlier events could continue to influence scores. Therefore, a one-point movement should not be directly attributed to any particular government action or reform.
Transparency International publishes its methodology detailing source selection, standardization, and aggregation but acknowledges that some underlying data remain private or proprietary. This distinction between methodological transparency and data transparency means that while the CPI’s design is understandable, replicating each country’s source scores from raw data is not possible. Meaningful scrutiny would require access to precise country-level data, assessment dates, expert numbers and types, response rates, sampling and weighting methods, treatment of missing data, unrounded scores, and documentation of methodological or expert pool changes.
In contrast to international assessments, Pakistan’s own surveys reveal a clearer distinction between perception and lived experience. Transparency International Pakistan (TIP) conducts the National Corruption Perception Survey (NCPS), which does not contribute data to the global CPI score. The 2025 NCPS surveyed 3,989 respondents across 20 districts in all four provinces. Results identified the police as the most corrupt sector (24%), followed by tendering and procurement (16%) and the judiciary (14%). Notably, 66% of respondents reported not paying a bribe for public services in the past year, while 77% expressed dissatisfaction with government anti-corruption efforts.
TIP shifted in 2025 from non-probability convenience sampling to a multistage stratified cluster design, allocating roughly 1,000 respondents per province. Participants were selected at public venues such as markets, parks, government offices, and community gathering spots. However, the methodology lacks details on individual respondent selection within venues, handling of refusals, replacements, or selection probabilities. Since provinces differ in population size, equal provincial allocation requires weighting to accurately represent national figures, but TIP does not provide sufficient information to replicate these weightings or selection probabilities.
Methodological rigor is crucial, as demonstrated by a judicial ruling in April 2024 when the Peshawar High Court ordered TIP to withdraw and republish its 2023 NCPS after scrutinizing findings related to the Khyber Pakhtunkhwa judiciary. The court found deficiencies in ethical and methodological standards, deeming the KP judiciary-related findings unreliable.
Leave an opinion