THE WORLD
3. Reform proposals
Theoretical studies discuss a large array of reform measures. Only a few of them, however, have been subject to empirical research. For example, although penalties are assumed to deter corrupt behavior and play a major role in theory, related empirical studies are still lacking. The difficulty of finding cross-country comparable data on penalties is one reason for the absence of such studies. Research on legal standards is also rare.
Nevertheless, the research reported above on the causes of corruption does suggest how reformers might focus their energies to limit the damage caused by corruption. Cross-country research suggests that it is not the overall size of government that matters but the nature of government pro- grams and regulations. Thus regulations that create corrupt incentives should be redesigned, not necessarily abandoned, and policies to promote competition and openness to foreign trade and investment could limit cor- ruption. Corruption may be limited by policies that engage more women in politics. Devolving government responsibilities to lower levels of govern- ment does not seem to be a generally desirable solution. Some evidence sug- gests that the form of government affects the level of corruption, but at this point the results are too contested to provide clear guidance.
Some cross-country research has an explicit policy focus. In this section I review a few such studies that consider the impact of civil service wage
policy, press freedom and the judiciary. Nevertheless, cross-country research provides only rough guidance on where to set priorities with little detail on which reform programs to implement.
Official wages
High salaries provide office holders with prospects of a future income premium that would be lost if they lose their job. Also, the intrinsic motiv- ation of public servants may increase with salary. Thus, some propose that civil service salaries be the focus of reform. However, the empirical evidence in support of such policies is poor.
Evans and Rauch (2000) investigate the impact of merit-based recruit- ment on corruption in 35 developing countries. Higher values in the merit- based recruitment index are associated with a greater proportion of higher-level officials in the core economic agencies who are either in pos- session of a university degree or who enter the civil service through a formal examination system. Controlling for income, this index is negatively asso- ciated with corruption.
Van Rijckeghem and Weder (2001) examine the extent to which the level of public sector salaries is linked to the amount of corruption. They argue that low salaries force public servants to supplement their incomes illicitly.
At the same time, high salaries are a premium that is lost if a public servant is caught and fired. In a small sample of 31 developing countries, they find a significant negative correlation between higher civil service wages (rela- tive to manufacturing wages) and corruption levels. Doubling the civil service wage will improve the corruption index on the order of 1 point of the TI index.19 The authors also point out that the association may be driven by reverse causality: corrupt countries tend to have a poor budgetary performance or may subscribe to the view that civil servants earn sufficient income from corruption, prompting them to reduce civil service pay. Such endogeneity problems diminish the prospects of fighting against corrup- tion by increasing wages. Even disregarding these issues, pay increases are a costly approach to fighting corruption.
Other studies provide poor results for the impact of wages on corruption.
Treisman (2000), Swamy et al. (2001) and Manow (2005) investigate the ratio of average government wages to per capita GDP, controlling for a variety of other influences. The results are ambiguous and mostly insignificant, depending on the indicator for corruption employed and the inclusion of control variables.
Press freedom
Another more promising candidate for reform relates to the freedom of the press. More than the fear of being prosecuted and sentenced, politicians
may be deterred by the danger of tarnishing their reputation. Although political reporting may not necessarily be fair and impartial, it nonetheless induces public office holders to refrain from corruption. One basic precon- dition for a press to contain corruption is its freedom and independence.
Indeed, demonstrated by the empirical studies, substantial evidence sup- ports a negative correlation between corruption and press freedom. Thus, reform aimed at improving the quality, freedom and independence of the media is influential in reducing corruption.
By regressing various indices of corruption on indicators of press freedom, Brunetti and Weder (2003) show that a free press effectively deters corruption. The press freedom variables consist of ‘laws and regulations that influence media content’ ‘political influence over media content’, ‘eco- nomic influence over media content’ and ‘repressive actions’, as compiled by Freedom House. These four separate indices and an aggregate index of press freedom all correlate negatively with the level of corruption in various specifications. This negative association between freedom of the press and corruption is confirmed by Lederman et al. (2001). Sung (2002) also reports this result, albeit without controlling for income per capita. A free press appears to be a solid deterrent to corruption. Brunetti and Weder (2003) corroborate these findings by using alternative indicators of corruption, providing us with additional confidence. Corrupt authoritarian regimes may restrict press freedom, suggesting that part of the causality may run the other way. Nevertheless, Brunetti and Weder (2003) show that their findings survive the use of instruments, claiming that a good share of the causality runs from a free press to less corruption. Adsera et al. (2000) employ data on daily average newspapers per person. These figures vary from 0.7 daily copies per person in Hong Kong to 0 in Mauritania. They show that the number of newspapers per person is negatively associated with corruption, particularly in democracies.
The judiciary
A high quality judiciary acts as a deterrent to corruption. Even in the dismal case where verdicts can be bought and judges bribed, a judiciary might still reduce corruption. As long as the judiciary is independent, courts endanger the corrupt transactions of a country’s elite. In contrast, in some countries politics has a strong influence on the judiciary, making it possible for big fish to escape prosecution. In some cases the judiciary might even operate in favor of corrupt elites by enforcing their corrupt deals. Empirical evidence supports the claim that judicial independence lowers corruption. This result should also hold for the independence of prosecutors. But more than just changing laws is necessary;de factoinde- pendence is the key. Reform proposals should certainly not overlook the
possibility that freeing the judiciary from corruption is also an important contribution to reform.
An approach by the World Bank (1997: 104, 168) focuses on the quality of the judiciary. Controlling for other explanatory variables, an index of the predictability of the judiciary from WB/UB significantly influences the level of corruption in 59 countries.
Ades and Di Tella (1996) propose a correlation between corruption and the independence of the judicial system. Sung (2002) also reports this result, albeit without controlling for income per capita. Damania et al.
(2004) show that corruption diminishes with a composite index of diverse indicators including people’s tendency to abide by the rules of society, per- ceptions of violent and non-violent crimes, predictability and efficiency of the judiciary, and the enforceability of legal contracts. Given the broad definition of this variable and its potential endogeneity, however, the coefficient could be biased upward.
Voigt et al. (2004) investigate the impact of prosecutorial indepen- dence on corruption. They measure independence with the help of a questionnaire sent to supreme court judges, law professors, lawyers and anti-corruption activists. The authors distinguish between de jure independence (for example, life tenure, appointment and promotion by others than politicians, lack of executive power to substitute prosecutors working on a specific case) and de factoindependence (dependence would be assumed, for example, in a case of forced retirement, frequent changes in legal foundations, and decreasing budgets of prosecutors). They find that de factoindependence decreases corruption, and relate this to the disciplining effect on the executive and other influential politicians.
Surprisingly, they find that de jure independence increases corruption.
De jure reform might be futile in some cases, but that it would increase corruption is counterintuitive. One possible explanation relates to endogeneity. Rather than de jure independence affecting corruption, corruption may bring about de jureindependence. The more pervasive corruption is among the executive, the greater the willingness to pay lip service to prosecutorial independence.