Mihály Fazekas

Despite the considerable interest, there is little evidence on the suitability of beneficial ownership data for systematic corruption risk assessment. This paper aims to validate common beneficial ownership risk indicators for proxying public procurement corruption. By implication, it offers practical insights for research, policy, and investigations. It also generates hypotheses regarding the impact of beneficial ownership registers on the organisation of financial crime. We match administrative data of 8 million government contracts with 11 million companies’ beneficial ownership records in Denmark, Estonia, Latvia, Slovakia, Ukraine, and the UK. We estimate fixed effects regressions tailored to capture non-linear relationships between company risk indicators of beneficial ownership and corruption risk indicators of public procurement. Correlations among two sets of differently constructed, yet conceptually related risk factors are interpreted as evidence for measurement validity. We find that BO-based risk indicators capturing unusual and outlier BO features - high company frequency of BO, frequent information change, outlier BO age, and no BO data - all perform in line with expected results. However, BO-based risk indicators relating to BO countries, such as sanctioned jurisdictions, largely fail to relate to public procurement corruption risks in line with expectations. Finally, BO-based risk indicators, which have already been widely validated in the literature using different data sources - company age and political connections - also turn out to be valid. Our findings lend support to the systematic use of beneficial ownership-based risk indicators in research, policy, and investigations. Our new risk assessment tools enable investigators to generate new investigative leads and policymakers to track the scale of likely corrupt transactions in public procurement.

It has long been argued that paying politicians higher salaries should help decrease corruption. However, the empirical evidence is mixed, partly due to the large variation in contexts, research designs, conceptual definitions and measures of corruption, and the predominance of case studies with potentially limited generalizability. To alleviate these challenges, we evaluate uniformly defined and validated corruption risk indicators from an original dataset of more than 2.4 million government contracts in eleven EU countries, covering more than half of the European Union population and gross domestic product. To aid causal identification, we exploit sizable changes in salaries of local politicians tied to population size across close to 100 discrete salary thresholds. Applying fixed effects estimators, regression discontinuity, and difference-in-discontinuities designs, we consistently find that better-paid local politicians (by about 15 per cent on average) oversee less risky procurement contracts, by a third to one standard deviation on our measure of corruption risk.

Theory and case studies suggest that emergencies and disasters increase corruption, especially in public procurement, hampering relief and reconstruction efforts. Despite a growing interest in the topic, including in research, there is still little systematic evidence about these effects, their structure and trajectories. We set out to investigate the medium‐term impact of disasters on corruption risks, using large‐scale administrative data on public tenders in Italy from 2007 to 2020, combined with data on 5 natural disasters. We employ logistic regression, coarsened exact matching and difference‐in‐differences estimators. We find that disasters increase corruption risks in the medium‐term (3 or more years after the disaster), even more than on the short term (1 year after the disaster). In the matched and diff‐in‐diff analyses, we find 3%–10% points more non‐open procedures used, 19%–21% points fewer call for tenders published, 19%–29% points more tenders with short advertisement period and 14%–17% points more single bidding tenders. Our findings highlight the importance of ring‐fencing corruption risks associated with disaster response, especially in the medium to long term.

Background: Increased costs in the health sector have put considerable strain on the public budgets allocated to pharmaceutical purchases. Faced with such pressures amplified by financial crises and pandemics, national purchasing authorities are presented with a puzzle: how to procure pharmaceuticals of the highest quality for the lowest price. The literature explored a range of impactful factors using data on producer and reference prices, but largely foregone the use of data on individual purchases by diverse public buyers. Methods: Leveraging the availability of open data in public procurement from official government portals, the article examines the relationship between unit prices and a host of predictors that account for policies that can be amended nationally or locally. The study uses traditional linear regression (OLS) and a machine learning model, random forest, to identify the best models for predicting pharmaceutical unit prices. To explore the association between a wide variety of predictors and unit prices, the study relies on more than 200,000 purchases in more than 800 standardized pharmaceutical product categories from 10 countries and territories. Results: The results show significant price variation of standardized products between and within countries. Although both models present substantial potential for predicting unit prices, the random forest model, which can incorporate non-linear relationships, leads to higher explained variance (R² = 0.85) and lower prediction error (RMSE = 0.81). Conclusions: The results demonstrate the potential of i) tapping into large quantities of purchase-level data in the health care sector and ii) using machine learning models for explaining and predicting pharmaceutical prices. The explanatory models identify data-driven policy interventions for decision-makers seeking to improve value for money.

By now, most political systems around the world hold regular multiparty elections of different quality and type. However, we know relatively little about the effect of elections on corruption, especially in high-discretion, public procurement contracts implementing development aid. To address this gap in the literature, we employ unmatched comparisons and matching estimators to analyze a global government contracting dataset that provides an objective proxy for corruption: the incidence of single bidding in competitive markets. We find that, all things being equal, corruption risks increase in the immediate pre-election period: single bidding is higher by 1.3–6.1% points. We demonstrate that the corruption-enhancing effect of elections is stronger under conditions of (i) high electoral competitiveness, (ii) medium-level party institutionalization, and (iii) “localized collective goods” clientelism.

Previous research has shown that corruption risks may distort market incentives in high-risk contexts. However, there is a dearth of evidence on the potential impact of corruption in settings characterized by low corruption and high-quality institutions. Against that background, this paper delves deeper into the alleged consequences of corruption by examining the link between corruption risks in public procurement and the profitability of firms in the Swedish construction industry. We introduce a novel measure of corruption risk based on the share of single bidder contracts that a firm has won. Validity analysis confirms that our measure is correlated with an alternative corruption measure and local tender winners. Our results reveal that firms that win many single bidder contracts have higher profitability than other firms in the sector: 10 percentage points higher single bidding rate firms have a 0.2–0.6-percentage-point higher sales margin. The findings underscore that public procurement corruption risks distort markets and economic incentives, and that this risk is present even in low-corruption contexts such as in Sweden.

The impacts of money in US politics have long been debated. Building on principal-agent models, we test whether and to what degree companies’ political donations lead to their favored treatment in federal procurement. We expect the impact of donations on favoritism to vary by the strength of control by political principals over their bureaucratic agents. We compile a comprehensive dataset of published federal contracts and registered campaign contributions for 2004–15. We develop risk indices capturing tendering practices and outcomes likely characterized by favoritism. Using fixed effects regressions, matching, and regression discontinuity analyses, we find confirming evidence for our theory. A large increase in donations from $10,000 to $5m (in USD) increases favoritism risks by about 1/4th standard deviation (SD). These effects are largely partisan, with firms donating to the party that holds the presidency showing higher risk. Donations influence favoritism risks most in less independent agencies: the same donation increases the risk of favoritism by an additional 1/3rd SD in agencies least insulated from politics. Exploiting sign-off thresholds, we demonstrate that donating contractors are subject to less scrutiny by political appointees.