Doug Hadden, VP Products
Alex Cobham, a Research Fellow at The Center for Global Development followed his twitter rant in early July about the Corruption Perception Index (CPI) from Transparency International in Foreign Policy. Aptly entitled Corrupting Perceptions, Cobham describes the limitations of perception versus reality. He also points out to limitations of the “more appealing” Global Corruption Barometer (GCB).
Supporting Smug Perceptions of Corruption
I’ve been somewhat vocal about how these two landmark surveys are socialized. It seems to me that they help persist inaccurate narratives about developing countries. That’s not to say that these surveys are not helpful in the anticorruption fight. It’s that observers can extrapolate some fairly fantastic narratives from the tool. Cobham points out that “in general the media emphasis on corruption in lower-income countries is, if anything disproportionate.”
I took exception to an article yesterday that suggested that because of higher trust of NGOs compared to governments, donors should reconsider the use of country systems. I collected a set of odd conclusions from the GCB last week. Alternative explanations to perception changes over time do not seem to be considered. For example, increased transparency and accountability can expose more corruption leading to the perception that there is an increase. Or, the increase in corruption reporting may have the same perception as crime statistics in many countries where the incidence of violent crime is reducing but the perception is the opposite. And, the GCB focus on bribery, although useful, measures but one aspect of the corruption problem.
Cobham minces no words in his assessment that “CPI embeds a powerful and misleading elite bias in popular perceptions of corruption, potentially contributing to a vicious cycle and at the same time incentivizing inappropriate policy responses”
My colleague Carlos Lipari has pointed out some of the problems in measuring corruption. Cobham describes the “sample size” problem of the CPI that looks at perceptions of a small sample size of elites and GCB that has a broader sample. Cobham points out that in CPI “the same kinds of people are being asked for their perceptions.”
Where does that leave us then? How can we find more effective measurement for corruption?
Policy responses and legal reform such as the introduction of an anti-corruption commission is often window dressing, because, in the words of Matt Andrews, often “what you see is not what you get.”
Nexus of Corruption Measurement
We are in a technology age that analysts the Gartner Group calls the “nexus of forces”. Before I get accused of technology solutionism a la Evgeny Morozov, it’s important to note that there are some interesting tools that we can leverage for anti-corruption.
- Open data standards IATI (aid), EITI (resource transparency), CoST (construction) and the emerging OpenContracting (procurement) provides “big data” of structured and unstructured data. Many countries have transparency portals that provide additional open information. Much of this information can be processed on inexpensive cloud computing platforms using open source tools.
- Social media and the use of mobile devices provide more tools in the hands of citizens. We’ve seen the impact in the Arab Spring and recent anti-corruption protests in Brazil, Bulgaria and Turkey. Crowdsourcing open-source tools like Ushahidi provides platforms to increase the sample size for corruption perception while collecting real data on corruption incidents. Big data tools used for marketing like “sentiment analysis” could also be leveraged.
- Civil society reports on corruption incidents. This information can be captured via RSS feeds and other mechanisms to build up a statistical picture that could be compared to other mechanisms. This means that “top down” and “bottom up” measurement methods could be used. This will expose potential gaps that will help to improve the measurement process.
My sense is that breakthroughs in anti-corruption enablement will occur when a critical mass of open information can be analyzed through the use of semantic web techniques will expose patterns. This will show us the weak corruption signals in countries that we do not have significant amounts of data. And, it will put pressure by the international community on what kinds of information is useful for combating corruption.