Cognitive ability and corruption: rule of law (still) matters

In this study, we use the terms “cognitive ability” and “intelligence” interchangeably.

In a recent study, Potrafke (2019) shows that there is a positive correlation between risk aversion and intelligence. He argues that intelligent people tend to be patient because they have long time horizons.

This means that high-IQ societies have a forward planning mechanism that helps internalize the negative future effects of corruption.

Other studies such as Kalonda-Kanyama and Kodila-Tedika (2012), Rindermann et al. (2015) and Kodila-Tedika (2014) show that overall quality of governance is positively related to the cognitive human capital of the society as a whole.

In “Appendix”, we also provide some empirical evidence using instrumental variable (IV) approach.

The results are robust if we use transparency international index of corruption. In our sample, the correlation between these two sources of corruption is more than 0.95.

We also found similar results on the negative interaction between the CCA and rule of law using WGI index for the rule of law. There is a high correlation (0.80) between the ICRG and WGI indicators for law and order.

If a country’s Polity score is ≥ 6 (out of 10), we count it as a democracy. For more information on this classification, see http://​www.​systemicpeace.​org/​polityproject.​html.

Also, note that all Models 1–9 in Table 1 use the similar number of observations which were included in the most general specification (Model 8).

It is likely that the quality of the education system and cognitive abilities of students are positively correlated. Farzanegan and Thum (2018) explain the economic and institutional factors which may influence the cognitive skills of students.

Note that in Potrafke (2012), the intelligence variable is assumed to be exogenous in regressions.

It is estimated that 47% of all cases of low vision and blindness worldwide are due to excessive UV radiation (see Ang et al. 2018 for further references).

This is mainly because of the reduction of the expected payoff for individuals as a result of a shorter expected period with sufficiently good eyesight. Such an environment can also reduce the incentives of individuals to invest in their human capital and the cognitive abilities of their children, reducing the ranking of average CCA index in affected countries.

The continental dummies of Africa and America are added to the list of excluded instruments (besides other three excluded instruments) for the purpose of over-identification tests (Hansen test which requires more exogenous excluded instruments than instrumented variables which in our case are three variables: CCA, rule of law and their interaction). If we exclude the continental dummies of Africa and America (for example, in the Model 4 in Table 8), nothing will change on the sign and significance of main results. The Kleibergen–Paap rk Wald F statistic on relevance of excluded instruments is on its acceptable level of 10.3. However, by excluding these dummies from list of instruments, we could not test for the over-identifying restrictions of all instruments (Hansen J statistic) because the equation is then exactly identified.