Corruption, like all contested concepts, is hard to measure. That hasn’t stopped ever more organizations from trying. The most well-known of these is Transparency International (TI), a Berlin-based NGO.
Its most famous product, the Corruption Perceptions Index, has been a regular feature of the anti-corruption landscape since the mid-1990s. It pulls together a variety of surveys on corruption and related concepts in a glorified poll of polls, giving (in 2016) 176 countries a score between 1 and 100; the most recent version was just published on Jan. 25.
The pace-setters in the 2016 version were, as they usually are, the Nordic countries, with Denmark (90), Finland (89), Sweden (88) and Norway (85) all in the top six. New Zealand (90) and Switzerland (86) were the only others who could keep pace. At the bottom, Somalia recorded a dismal 10 points while South Sudan (11) and North Korea (12) were little better. The United States came in 18th with a score of 74.
The problem of perceptions
The CPI has its critics. Boiling a country’s corruption problems down to one number is a heroic attempt to simplify a very complex phenomenon.
Further, the CPI doesn’t actually claim to measure corruption at all; it measures perceptions of corruption. The two are not synonymous.
The CPI also looks solely at perceptions of public sector corruption. But private sector actors also have (at times very significant) roles.
The CPI has its limitations. But it has been important in prompting more nuanced attempts at quantifying corrupt practice. That’s where the future of corruption measurement lies. As I document in a new book, there are now a plethora of competing measurement tools. Some offer interesting insights into various parts of the corruption mosaic.
From perceptions to experiences
We now know, for example, a lot more about the corruption that people actually experience. At the level of ordinary citizens, the World Values Survey asks a number of questions on corruption while the Global Corruption Barometer also asks tens of thousands of people worldwide whether they themselves have experienced corruption over the past 12 months. Furthermore, the Business Environment and Enterprise Survey (BEEPS) turns the spotlight onto the business community, surveying over 15,000 business managers across 30 countries to map the scope of the bribes they were required to pay to win deals.
Proxy indicators
The challenge of measuring corruption has also led observers to develop potentially useful proxy indicators. These indicators can reveal corruption that is either very hard to see or very hard to measure (or both). They range from surveys of where and how money from the public purse gets allocated (and misallocated) to analyzing which companies get public sector contracts and on what terms.
Public expenditure tracking surveys (PETS) are arguably the most well-known. They highlight cases where public money has either not ended up where it should have or can’t actually be accounted for at all. That money could, of course, have been accidentally misallocated or lost by incompetence. But when we find systematic patterns of misallocation, that pattern of behavior may well point to corrupt practices.
Exploring patterns of what’s often known as “leakage” is therefore an attractive way of illustrating where problems may indeed lie. A great example is Jonathan Stromseth, Edmund Malesky and Dimitar Gueorguiev’s work in China. They use data from the China National Auditing Office to put together a data set illustrating how much money was misused as a proportion of each of China’s 33 provincial budgets. They wanted to know, in other words, how much money each Chinese province was unable to account for. They used that figure as a proxy for what they term “macro-corruption”.
Some leakage will happen everywhere. But in 2011, nearly 20 percent of the budget of Heilongjiang Province (in China’s North-East) was misused. It is highly unlikely that that was all down to mistakes and errors.
It is not hard to see that proxy measures, such as these, offer policymakers useful tools. Unaccounted-for money is rarely dumped in a lake or burned on a bonfire. It ends up somewhere. Frequently that somewhere is an illegitimate home.
Big data can help spot corruption in public procurement
A set of proxies has also been developed specifically to identify corruption risks in another area that’s traditionally been seen as corruption-prone: public procurement. These take advantage of the fact that large contract data sets — or “big data” — are now published by many national and local governments. Thus they can be analyzed for irregularities that might indicate corruption.
The approach relies on identifying “red flags” often associated with corrupt practices. These include, for example, a very short time between the announcement of a tender and the deadline for submitting bids, which often reflects hidden practices to favor a certain bidder. Frequent use of noncompetitive or “negotiated” procedures, perhaps citing “emergency” conditions, can be another red flag.
By looking for systematic occurrences of these red flags in certain contracting authorities or relating to particular bidders, researchers, such as Mihaly Fazekas, have created tools such as the Corruption Risk Index.
The future of corruption measurement
Corruption is complex, multifaceted and riddled with nuance, and this makes aggregate indicators, such as the CPI, problematic. But reactions to the CPI’s inadequacies have led others to develop more focused and sophisticated tools. The very best of these can be used to help spot trends and to illustrate the scale and scope of particular types of corruption. They can also make a contribution to helping us develop explanations of what might be causing corruption. They can’t provide definite answers.
But if used carefully, the data produced by these proxy indicators can help us think just a little more about how to battle corruption.
Dan Hough is a professor of politics at the University of Sussex and director of the Sussex Centre for the Study of Corruption (SCSC).