Progress in corruption research and anticorruption policies has been hampered by the lack of adequate measurement in at least two ways. On the one hand, shortcomings of corruption measurement limits our capacity to understand what works and hence develop effective policies. On the other hand, in the absence of sufficiently sensitive measures of corruption frequency and distribution, we simply cannot know whether we are making progress or not. To address these interrelated pressing problems, this article reviews a wide array of corruption measurement instruments across diverse disciplines and offers a balanced assessment of their quality and outlines where to deploy which instrument. It also compiles a comprehensive state-of-the-art repository of data on corruption measurement largely missing from the literature.
Methods reviewed fall broadly in the following categories:
• expert scores such as the V-Dem corruption indices.
• population surveys, perception as well as self-reported experience-based such as Transparency International’s Bribe Payers Index;
• national context indicators such as the Index of Public Integrity,
• micro-level enforcement-based indicators such as criminal statistics in the US),
• micro-level transactions-based indicators in
o companies such as indicators of ownership networks,
o public procurement such as measures of limited competition,
o public sector employment such as surveys of civil service hiring, and
o personal connections such as measures of revolving door.
The review will put a particularly strong emphasis on more recent innovations in corruption measurement, most of which fall in the micro-level transaction-based indices category. These indicators constitute a diverse group of corruption indices with the common trait that they all measure corruption on the level of economic transactions where corruption actually takes place. Furthermore, these indicators invariably resort to proxying corruption rather than measuring it directly which means that they have to rely on theory and statistical evidence to underpin their validity. As corruption can manifest itself in a range of economic and political transactions these indicators are quite different from each other looking at, for example, welfare payments, government contracts, payments and donations, public officials’ asset declarations, or law making.
While transaction-level measurements hold the promise of greater detail and sensitivity to change, they are often plagued by false positives, that is signalling corruption where there is none. Hence a careful approach is needed when using proxy measures. For example, indicators of revolving door or corporate political connections are widely used and all too often uncritically. However, contrary to the dominant corruption risk interpretation of public officials working in the private sector and having strong ties between public and private entities also carry a host of benefits such as improved information flows between sectors and the spread of entrepreneurial values to the public sector to name a few. Hence, it is only the misuse of connections which poses risks which requires the measurement of the impacts not only the existence of a particularistic link. For example, in public procurement, most studies look at the amount of contracts won due to connections with only some trying to also link to corrupt means of obtaining contracts enabled by connections.
In spite of the widely held views that there is a paucity of indicators overall, this broad review finds that the field is diverse, advanced in many respects, and highly specialised. Unfortunately, different literatures and results are largely disconnected from each other, hampering cumulative knowledge generation. Given this surprising abundance of measurement instruments, this review offers a carefully crafted guide as to which measurement instrument is best suited to which research or policy problem. Such a guide is greatly needed also because many of the different indicators actually measure a different corrupt phenomenon and they are often uncorrelated across and within countries. Crucially, researchers and policy makers have to precisely define the kind of corruption they are interested in and the level of precision they need for deriving useful and actionable insights (e.g. targeting audits requires a lot higher degree of precision than guiding high-level policy interventions).
The review article concludes with an agenda for future research which both enriches the measurement landscape by addressing notable gaps, while also outlining ways of using different measurement instruments in tandem. In particular, it is proposed that a successful future agenda shall develop new corruption indicators which
• rest on a theoretically sound understanding of the corruption process,
• derive from objective data describing actor behaviour,
• are defined on the micro level such as individual transactions, and
• allow for consistent comparisons across countries, organisations, and time.