Corruption is closely linked to problems associated with money in politics. Indeed, some have argued that an excessive/inappropriate influence of money on elections is corruption (even if it’s not necessarily illegal or currently viewed as unethical). Even for those who (like me) prefer a more restrictive definition of “corruption,” it is widely believed that these issues are related. Many hypothesize that countries with weak or ineffective systems of political finance regulation may experience higher levels of corruption—though at the same time excessively onerous, unrealistic regulations on political spending may also induce corruption in order to circumvent the official rules. Perhaps surprisingly, though, we do not have (or at least I have not yet seen) very much quantitative, comparative research on the relationship between the quality of countries’ laws on the regulation of political finance, on the one hand, and the extent of their corruption problems, on the other.
This may be starting to change, thanks in part to initiatives like the Money, Politics and Transparency (MPT) forum (a collaborative venture of the Sunlight Foundation, Global Integrity, and the Electoral Integrity Project). A few weeks back Rick posted a highly critical assessment of MPT’s volume Checkbook Elections, a collection of qualitative case studies. I haven’t yet read that report, but here I wanted to focus on another aspect of MPT’s work: a quantitative index that purports to measure how well 54 different democratic countries regulate political finance, based on responses to 50 survey questions in five different categories (public funding of elections, contribution and expenditure restrictions, reporting and disclosure, regulation of third-party actors, and monitoring/enforcement). The surveys include questions about both law and practice in all five categories; moreover, in addition to a composite index score, MPT also provides separate scores for the quality of electoral regulation both “in law” and “in practice.” (A detailed description of the methodology is available here.) All the usual caveats and concerns regarding these sorts of composite indexes of course apply here, but at first pass this seems like a useful resource, and potentially helpful in teasing out the relationships between political finance regulation and corruption more generally.
Real progress on this will front require careful research design, more extensive data, and the application of rigorous empirical methods—an enterprise for which I lack both the time and the talent. But just for fun, I played around a bit to see how the MPT index (and each sub-index) correlates with the 2014 Transparency International Corruption Perceptions Index (CPI). Are countries with better regulation of political finance (in law, in practice, or overall) perceived as more corrupt? Less corrupt? I’ll tell you what I found after the break, but just for fun take a guess now, before you know the answer!
OK, here’s what I found:
For the 53 countries that are included in both indexes (the MPT Index has scores for the Solomon Islands, which doesn’t appear in the CPI), the correlation coefficient between the aggregate MPT score and the CPI score (where higher scores indicate better regulation and lower corruption, respectively) is approximately 0.16. That’s a positive correlation, but it’s very small, and not statistically significant at conventional levels. In other words, at least when looking at the aggregate score, we can’t confidently reject the hypothesis that there is no correlation between the quality of political finance regulation (as measured by MPT) and perceived corruption (as measured by the CPI). That is a striking and perhaps counter-intuitive finding. And if one looks at the individual countries on the list, it does indeed seem that although there are some countries that score well on both indexes (Chile, the UK, the US) and some countries that do poorly on both (Kenya, Lebanon, Nigeria, Paraguay, Venezuela), there are also quite a few countries that have high-quality regulation of political finance but high perceived corruption (such as Albania, Argentina, Ecuador, and Mexico), as well as several countries that do well on the CPI but poorly on the MPT (such as Belgium, Botswana, and Sweden).
Things get considerably more interesting when we consider separately the MPT’s “in law” and “in practice” scores. The correlation between the MPT “in law” score and the CPI is actually slightly negative (approximately -0.07) and not statistically significant. But the correlation between the MPT “in practice” score and the CPI is approximately 0.49—a large and statistically significant positive relationship. In other words, although the quality of a country’s laws on political finance has essentially no correlation with that country’s perceived corruption, countries that get better rankings with respect to their actual enforcement and implementation of political finance regulation are generally perceived as less corrupt.
(By the way, you might be wondering about the correlation between the MPT “in law” and “in practice” scores. If better laws lead to improvements in practical enforcement, we would expect this relationship to be positive. If good laws and effective implementation are in some sense substitutes, we’d expect a negative correlation. And if there’s essentially no relationship between the law in the books and the law in action, we’d expect the correlation to be statistically indistinguishable from zero. The correlation between the MPT “in law” and “in practice” scores turns out to be around 0.62—positive, very large, and statistically significant. So generally better laws and better implementation tend to go together in the MPT data.)
One other question I looked into was whether the results were any different for more or less corrupt countries. (Perhaps, I reasoned, some of the results might be driven by a relatively small subset of countries with dysfunctional institutions.) So, as a quick-and-dirty first cut at exploring that issue, I divided my sample into those countries scoring in the top and bottom halves of the 2014 CPI. (Of my 53 countries, 30 fell into the top half and 23 into the bottom half.) Of course, with smaller samples, the statistical power of the tests is much lower, but the results are still intriguing.
For my subsample of 30 “less corrupt” countries, the correlation between the CPI and the aggregate MPT score is approximately -0.06 (very close to zero, and not remotely near statistical significance). But the correlation between the CPI and the MPT “in law” score was large and negative (approximately -0.31), while the correlation between the CPI and the MPT “in practice” score was large, positive, and almost identical in magnitude (0.32). These latter two results are both statistically significant at the 90% level, though not at the 95% level (which may be due partly to the smaller sample size). In other words, among countries perceived as less corrupt (those that rank in the top-80 of the 159 countries in the 2014 CPI), higher-quality political finance laws are correlated with higher perceived corruption, while better political finance practices are correlated with lower perceived corruption.
Things look a bit different for my subset of 23 “more corrupt” countries. Here, the correlations between the CPI score and both the aggregate MPT score and the MPT “in law” score are small and statistically insignificant (0.17 and 0.06, respectively). But the correlation between the CPI and the MPT “in practice” score is 0.38—large, positive, similar in magnitude to the equivalent coefficient for the “less corrupt” subset, and statistically significant at the 90% level despite having only 23 countries in the sample.
None of this should be taken too seriously—it’s just mucking about with the data, really—but still, these simple patterns are intriguing, and perhaps suggest the need for more systematic evaluation. With all the appropriate caveats and grains of salt and so forth, perhaps one could summarize the pattern that emerges as something like this:
- Counties perceived as less corrupt do not appear (on average) to have higher-quality political finance laws. For relatively more corrupt countries, there seems to be no discernable correlation between the quality of political finance laws and perceived corruption levels. Perhaps even more strikingly, in the set of less-corrupt countries, those countries with very low levels of perceived corruption seem to have systematically worse political finance laws. If one goes beyond the numbers, one finds that although the distribution of MPT “in law” scores seems mostly random as one moves up the CPI rank from most- to least-corrupt, the countries at the very high end of the CPI range seem to have a high concentration of countries with quite poor MPT “in law” scores. Of the top six countries in the data on CPI rank, five of them—Australia, Belgium, Germany, Sweden, and the United Kingdom—score in the bottom third of the MPT “in law” index, and the sixth (Japan) gets a middling score.
- Things look different with respect to implementation of political finance rules in practice. Here, there is a strong and systematic positive correlation between the MPT’s “in practice” assessment and the CPI index. And indeed, four of the CPI top-six noted above (Sweden, Australia, Germany, and the UK) score in the top third of the MPT “in practice” index but in the bottom-third of the MPT “in law” index. (Japan is also in the top third for MPT “in practice”; of the top CPI performers in the sample, only Belgium does badly on the MPT across the board.) The association is even more striking at the bottom end of the CPI scale: of the eight countries with the lowest CPI scores in the sample, six score in the bottom third on the MPT “in practice” index, while the other two have middling MPT in practice scores.
What’s going on? Obviously we can’t infer causation from correlations, and it’s dangerous even to make too much of patterns in simple correlations with no controls. But still, here are some off-the-cuff thoughts:
- Probably the most surprising result is the absence of a positive correlation between the MPT “in law” index and the CPI, and indeed a negative correlation at the higher end of the CPI scale. (This is particularly surprising since the reported correlations do not in any way “control for” in-practice implementation. Indeed, as noted above, the MPT “in law” and MPT “in practice” scores are themselves strongly correlated.) What might be going on? One possibility is that more corrupt countries try to adopt stronger political finance laws as anticorruption measures, while countries that have less of a political corruption problem don’t feel as much need to adopt these more stringent laws. Another possibility is that at least for some (more developed?) countries, political corruption is more often the result of overregulation, rather than underregulation, of political finance. Or perhaps the correlation is spurious, and some other factors (economics, history, political structure, etc.) both reduce corruption and make the adoption of robust political finance laws unlikely.
- The fact that there is a consistent and statistically significant negative correlation between in-practice implementation of political finance regulation and perceived corruption is less surprising, but it’s still not entirely clear how we should interpret it. One possibility, perhaps the most intuitive, is that effective enforcement of political finance regulation (for example, on matters like public disclosure of expenditures and regulation of third-party activities) reduces the opportunities for political corruption. Another possibility is that the causal arrow runs the other way: perhaps corruption undermines the effective implementation of the political finance regulations. (These possibilities, of course, are not mutually exclusive—indeed they might be mutually reinforcing.) Yet a third possibility is that the correlation is due to some other factor. For example, perhaps poor countries, or countries with low institutional capacity, have difficulty enforcing both their political finance rules and their anticorruption laws. Again, this possibility is not inconsistent with the hypothesis that there is also a causal connection between the quality of in-practice political finance regulation and corruption, but it does mean we must be cautious about jumping to the conclusion that such a causal relationship in fact exists.
So, there’s still a lot more work to do along these lines (though I probably won’t be the one to do it). The relationship between the regulation of political finance and the phenomenon of corruption more generally is obviously of central importance, so it’s actually a bit surprising how little we know about it so far.