JUDICIAL CHECKS ON CORRUPTION

Judicial oversight is widely regarded as an important check and balance on the abuse of governmental power. The literature identifies two important components of this oversight function: judicial independence and constitutional review. However, recent work using country-level data indicates that the effectiveness of constitutional review is largely determined by the rigidity of the constitution. My dissertation builds on this work by investigating whether judicial independence and constitutional review are deterrents to a specific type of abuse of power by government officials: corruption in office. Since the appropriate way to measure corruption, which is defined as the abuse of public office for private gain, remains an unsettled issue, I use both measures of corruption derived from criminal convictions data and measures of corruption derived from survey data. In the case of the former, I propose a new empirical strategy that explicitly recognizes the distinction between the unobserved number of corrupt officials and the observed number of convictions for corruption. I begin by examining whether differences in judicial independence and constitutional rigidity at the state level in the United States explain the observed variation in the number of corruption convictions across states. I investigate these issues empirically using annual data for the years 1996-2005. Judicial independence is assessed by looking at judicial remuneration, method of selection, and term length. To assess constitutional rigidity, I use the legislative majority needed in each state to propose constitutional amendments, the provision (or lack thereof) for constitutional conventions, and the provision (or lack thereof) for popular initiatives to amend the constitution. I find that both judicial independence and constitutional rigidity are significant predictors of corruption. The coefficient estimates indicate that in states where judges have a higher degree of independence (higher remuneration, merit plan selection, or longer terms), there is less corruption. Moreover, states that have more rigid constitutions (higher legislative majorities required to propose constitutional amendments) experience less corruption. These findings, which are new to the literature, suggest that policy reforms which promote judicial independence and make it more difficult to alter constitutions have the potential to mitigate the harmful effects of corruption that have been documented in other studies. To develop additional insights, I conduct a similar empirical investigation in a cross-country setting. Since I do not have data on corruption convictions at the country level, I use several survey-based measures, from Transparency International, the World Bank (Kaufmann et al., 2007), and the Political Risk Survey Group. My measures of judicial independence and constitutional review are drawn from a variety of sources. I obtain indices of judicial independence from the Economic Freedom of the World Report, published by the Fraser Institute, from the Political Constraint Index Dataset, and from the La Porta et al. (2004) study cited earlier. The constitutional review index is taken from La Porta et al. (2004) as well. Like most of the literature, I analyze these data using linear regression methods: ordinary least squares, weighted least squares, and panel data techniques. Since data availability varies across indices, I conduct my analysis on cross-sections of between 37 to 165 countries for various years between 1995 and 2005. I then extend my analysis to panels of data with cross-sectional and time-series observations for 1998 to 2005. The evidence on judicial independence from the cross-country analysis is generally consistent with that from the United States data. I find strong support in favor of my hypothesis that higher judicial independence is associated with lower corruption levels, ceteris paribus. This finding is robust to the way corruption and judicial independence levels are measured, to the estimation technique, and, in general, to the inclusion of various control variables. However, I do not find the expected relation between corruption and constitutional rigidity. It appears that countries with more rigid constitutions experience more corruption using the same controls as for the judicial independence models. In their study, La Porta et al. (2004) show that constitutional review is a guarantee of political freedom measured by indices of democracy, human rights and political rights, but not of economic freedom. This suggests that it may be possible to reconcile the findings from the cross-country analysis with the evidence for the United States by gaining a better understanding of the relation between political freedom and economic freedom