When justice answers to the president: Reexamining the effect of cabinet partisanship on human rights in presidential democracies

ABSTRACT

Recent research suggests that the administrations of presidents with diverse cabinets are less likely (and less able) to violate human rights. I posit that a more clear image of how (and where) repression occurs can be obtained by looking at the cabinet portfolios that have the greatest capacity to repress, such as the ministry of justice (which typically oversees a state’s justice system and prison apparatus). In this research note, I reexamine the relationship between cabinet partisanship and government respect for human rights by empirically testing the effect that presidential control over the ministry of justice has on human rights violations in presidential democracies. I find that states are more likely to experience greater government respect for human rights when the justice minister is in a different party than the president’s party. I conclude by suggesting that given the expansive power justice ministers have with regards to the administration of justice, perhaps this position is best occupied by an individual whose first loyalties are the law and the rights of citizens, rather than political party and presidential piety.

KEYWORDS:

• Presidentialism
• Human rights
• Repression
• Democracies

Acknowledgements

An earlier version of this research note was presented at the 2018 Annual Meeting of the International Studies Association-Midwest in Saint Louis, Missouri. I would like the thank several anonymous reviewers whose comments have greatly improved this manuscript. Additionally, I would like to thank Sera Chiuchiarelli for all of her steadfast support. Any remaining errors are my own.

Notes

¹ Human rights issues are complex and can often impinge on other priorities of a president. Even well-meaning presidents may come to see greater value in protecting human rights as a result of a justice minister that is able to offer a contrasting perspective by playing ‘devil’s advocate.’

² In some states this ‘ministry’ is called a ‘department.’

³ In some states the minister of justice is called the minister for justice, the secretary of justice, or the attorney general.

⁴ Note that the CIRI scale is only available up until 2011, which provides a temporal limitation of this study.

⁵ Available at: http://https://www.cia.gov/library/publications/world-leaders-1/.

⁶ The year 2001 is the first year for which this directory has data, which provides another temporal limitation of this study.

⁷ Note that he only examined years 2001–2011 because, as mentioned, 2001 is the first year that the CIA directory has data, while 2011 is the last year that the CIRI scale has data.

⁸ Note that in order for Holzer to calculate his percent of the cabinet not from the president’s party variable, he needed to ascertain the political affiliation of every member in any given cabinet for every given year. Being unable to determine the political affiliation of even a single member of the cabinet results in that country-year having no percent of the cabinet not from the president’s party value, thereby resulting in missing data as a result of listwise deletion. In contrast, it is relatively more manageable to ascertain the political affiliation of only two individuals (i.e. the president and his or her justice minister). As a result, my is the justice minister’s party different than the president’s party? variable has 430 observations from 46 countries in comparison to Holzer’s percent of the cabinet not from the president’s party variable, which only has 241 observations from 35 countries.

⁹ Emphasis in original.

¹⁰ Note that I have reversed their original ‘0’ to ‘1’ scale such that higher values now indicate greater presidential constraints (as opposed to greater presidential power).

¹¹ As mentioned above (in footnote 8), this 78% increase is a function of the relative ease to ascertain the political affiliation of only two individuals (i.e. the president and his or her justice minister) versus an entire cabinet (which is required in order to calculate the percent of the cabinet not from the president’s party variable). Whereas available data was only able to produce 241 observations from 35 countries for Holzer’s percent of the cabinet not from the president’s party variable, that same raw data was able to produce 430 observations from 46 countries for my is the justice minister’s party different than the president’s party? variable.

¹² To further investigate this disagreement within the literature, I have performed a robustness check where I have constrained the coefficients of my is the justice minister’s party different than the president’s party? and is the justice minister nonpartisan? variables to equality and have used a nested F-test to discern any reduction in fit (Baum, 2006, p. 92). The associated p-values, however, were well above 0.1, which suggests that the coefficients for these two variables (reported in Table 1) do not statistically differ. In other words, nonpartisan justice ministers and justice ministers in a party different than the president’s party do not seem to have a statistically different effect on government respect for human rights, at least at the 90% (or better) level, suggesting that Holzer (2018a) is justified when aggregating these two groups for his percent of the cabinet not from the president’s party variable.

¹³ Note that country-years where both of these variables are equal to ‘1’ implies that for that country-year the justice minister is in the same party as the president (i.e. they are a copartisan).

¹⁴ In Table 3 I also report the change in baseline probability when the justice minister is nonpartisan (versus in a different party than the president’s party). However, note that the 95% confidence for all of these changes (across all three scenarios) overlap with zero. This suggest that these positive/negative changes are not statistically different than no change at all (i.e. zero). As mentioned above (in footnote 12), this suggests that although Lee (2018a, 2018b) argues that nonpartisan ministers are typically more loyal than ministers within the president’s own party (i.e. copartisans), nonpartisans are perhaps less loyal than copartisans with regards to policies related to government respect for human rights.

¹⁵ Alternatively, rather than a president using his or her copartisan justice minister to repress citizens’ rights, repressive justice ministers may find themselves less constrained by a copartisan president and/or more constrained by a president of a different party. Either way, this research note warns of the human rights repercussions when one party controls both the presidency and the ministry of justice.