This is delightful -- from Anthony Watts:
Could this study on honesty and government service explain the EPA climateer fraud and �Climategate�?Anthony spends a lot of time talking about Climategate and citing links to various damning emails. He then concludes talking about the paper itself and offers this selection from the paper:
A new paper published the National Bureau of Economic Research has given an insight that may explain some of the personal decisions that led to the recent EPA corruption fiasco Massive fraud at the EPA from agency�s top paid climate official (where a top climate specialist defrauded the taxpayers out of millions of dollars and made wild claims about being on CIA missions) and to Climategate, since I see some significant parallels between the two and this study. Links to a story about the paper and the paper itself follow.
CONCLUSIONS:You can read the full paper here (PDF).
In this paper, we offer evidence that the college students who cheat on a simple task are more likely to prefer to enter government service after graduation. This relationship does not appear to vary by ability, suggesting that screening on ability does not change the level of honesty of those chosen for government service among the pool of applicants. Importantly, we show that cheating on this task is also predictive of fraudulent behaviors by real government officials, which implies that the measure captures a meaningful propensity towards corruption. Given that the existing methods of measuring corruption only apply for those who are already entrenched in the bureaucracy, our validation of a measure of cheating against real-world corruption outcomes offers an important tool for future research on selection and corruption.
These findings are important because they demonstrate that the variation in the levels of observed corruption may, in part, be driven by who selects into government service. In addition, they offer two key policy insights. First, the recruitment and screening process for bureaucrats may be improved by increasing the emphasis on characteristics other than ability. It is important to note that individuals may not want to reveal their characteristics, especially their propensity for dishonesty, so the method of measurement matters. The simple, experimental measure we employed predicted the corrupt behaviors of the government employees, but the game in which corruption was explicitly framed and the fairly standard attitudinal questions had little predictive value. Second, while recent empirical papers have shown that reducing the returns to corrupt behavior decreases the probability that bureaucrats engage in corruption, our work suggests that these interventions may have had even broader effects by changing the composition of who might apply.
