رشوه خواری، بی کفایتی، و تاخیر بوروکراتیک

We examine bribery in a dynamic setting with a partially honest bureaucracy. The presence of honest officials lowers the bribe that a corrupt official can extract from efficient applicants. Consequently, efficient applicants may face delays in receiving licenses, or may never receive licenses, as inefficient applicants are serviced ahead of them. An increase in the proportion of honest officials, while directly increasing the incidence of socially optimal decisions, can also induce greater inefficiencies (misallocation and delay) from corrupt officials. As a result, social welfare may be non-monotonic in the proportion of honest officials, and small differences in monitoring costs may lead to very different levels of corruption. Competition between officials can hurt efficiency; its desirability depends on the prevalence of applicants' ability to pay bribes.

In many developing economies, business ventures require official sanction, in the form of licenses or permits, to operate.1 If bureaucrats maximize personal gain and are not sufficiently monitored, as is often the case, their discretionary power in awarding licenses gives them the opportunity to extract bribes. This raises important questions. What is the impact on efficiency? Who ends up with licenses? How is timeliness of bureaucratic service affected? How can licensing be set up in order to maximize efficiency? To what extent should officials be monitored or trained to be honest? One stand of the literature has defended bribery as a mechanism that ensures both allocative efficiency and punctuality in official decisions. For example, it is argued that the most productive applicant will be served since she is willing to pay the highest bribe.2 Similarly, it is argued that bribes can provide bureaucrats with the incentive to “speed up”, and that the one with the highest value on time will be willing to pay most to move ahead in the queue.3 In short, allowing bribery may both improve and speed up the allocation of goods and services. A counterargument is that the obstacles to business–red tape, required permits, and slow service–are not exogenous barriers, but bureaucratic actions designed to raise revenue. If the existence of these barriers is endogenized, bribery may no longer be efficient.4 While this argument is undoubtedly important, we take a different approach here. We employ a standard bribery model in which the licensing requirements–the barriers, so to speak–are perfectly in line with efficiency. Further, the socially efficient candidates are the ones to whom the license is worth most.5 We augment this standard model in two basic ways. First, we extend it to be dynamic. Licenses are distributed over two periods. Applicants can re-apply to a new official in the second period if denied a license in the first period. Second, some officials are honest and act in the interest of efficiency. This assumption may be less standard, but can be motivated by assuming these agents are either intrinsically motivated or externally monitored. Intrinsically motivated agents, via a sense of identification with the organization or its goals, have recently received greater theoretical attention.6 They are arguably as relevant in the study of corruption as anywhere else. These two additional features overturn both types of efficiency results. Rather than speeding up licensing, bribery can give corrupt officials incentive to delay disbursal. It also may result in the inefficient producers being licensed ahead of, or instead of, the efficient ones. Bribery may, therefore, result in the inefficiencies of misallocation and delays that are frequently, in popular perception, regarded as the twin vices of a corrupt bureaucracy. Inefficiencies arise for the following reasons. Productive applicants are rationally searching for conscientious service that will recognize the merits of their petition. In contrast, less deserving applicants do not anticipate gains from searching for honest officials and can be more ready to bribe. Necessary for this is sufficient prevalence of honest officials, paradoxically. Past some threshold of bureaucratic honesty, efficient applicants' value of waiting is elevated to the extent that they are no longer willing to pay the most for licenses. Consequently, inefficient applicants can be serviced while efficient applicants are delayed service or denied it altogether. Though we are unaware of empirical evidence verifying the efficiency or inefficiency of bribery-mediated allocations, it is not hard to imagine well-qualified applicants finding it profitable to pursue a principled approach longer than less-qualified ones. The results point to the following complexity in optimal organizational design: efficiency can be non-monotonic in the prevalence of honest officials. An increase in the number of honest officials has two opposing effects. The positive effect is that there are more officials who act in the public interest. But there is an indirect and negative effect: a greater prevalence of honest officials lowers the willingness to pay of efficient applicants, making the remaining corrupt officials' behavior more inefficient. When the indirect effect dominates, social welfare is locally decreasing in the proportion of honest officials. Thus, replacing some subset of corrupt, bribe-maximizing bureaucrats with perfectly honest ones, or spending more resources on monitoring or on fostering organizational identity, can actually reduce efficiency. A partially honest bureaucracy can be worse than either extreme.7 and 8 These results are different from the arguments of Leff (1964) and Huntington (1968), among others, who have argued that honest officials enforcing irrational regulations can lead to inefficiency. In our model, the regulation is justified on efficiency grounds. Relatedly, here the perverse effects of honesty are local; efficiency is always maximized by a perfectly honest bureaucracy. Nonetheless, if the degree of honesty is interpreted as the result of some expenditure (to monitor, train, or foster identity, for example), the optimal degree may vary widely with small cost differences, due to the non-monotonicity. This would predict a bimodality in corruption levels across sectors or countries. A final result is that efficiency can be reduced by competition, that is, the ability to leave one official and search for another. It is precisely the option of re-applying that lowers productive applicants' willingness to pay and delays or rules out their receiving licenses. Re-application also allows corrupt officials to price discriminate dynamically and license more inefficient candidates. This result provides a caveat to the literature's call for competition between bureaucrats as a key means for reducing corruption.9 and 10 Of course, our setting is not perfect competition, in the sense of costlessly observable prices. Here, applicants must search and can only observe one price per period; reputation-building is also ruled out by assumption. However, it seems likely to us that many settings of bureaucratic corruption involve publicly unobservable prices and minimal dissemination of information across applicants. Realistically, the choice is often between monopolistic licensing and some form of imperfectly competitive licensing. The key condition in our model for competitive licensing to lower efficiency is that the ability to pay the bribes be sufficiently widespread. If instead many lack wealth and access to credit, competitive licensing does represent an improvement over monopolistic licensing. Consequently, bureaucratic competition seems especially justified in the context of a less developed economy. Our paper is closest to Banerjee (1997) and the two-period models utilized by Choi and Thum, 2003 and Choi and Thum, 2004 to analyze bureaucratic corruption. In the former, delay induced by red tape can be a useful mechanism for separating out the socially desirable recipients, when ability to pay is imperfect.11 The latter papers analyze issues that arise when an official can demand payments from an entrepreneur repeatedly or when the timing of application is endogenous—effects on the level of investment, the entry of new firms, and the efficacy of job rotation. The distinguishing feature of our model is that we allow for applicants' ability to re-apply within a pool of heterogeneous officials. This assumption gives rise to very different results. The baseline model is outlined in Section 2. Our main results on misallocation and delay, the non-monotonicity of social welfare with respect to honest officials, and the benefits of competition, are derived under the assumption of unlimited licenses in Section 3. Section 4 extends our analysis to the case of limited licenses. We discuss robustness of the results in Section 5 and conclude in Section 6.

In summary, we find that bribery is inefficient in a simple dynamic setting, and in fact creates a non-monotonic relationship between the degree of honesty of the bureaucracy and efficiency. The key mechanism is that the option of re-applying in hopes of finding an honest official is worth most to productive candidates. This lowers their initial willingness to pay a bribe and thus can lead to misallocation of licenses toward less productive candidates and bureaucratic delay in awarding licenses to productive candidates. Nonetheless, allowing applicants this option to search (competition) is optimal provided that ability to pay bribes is not too widespread. Thus bureaucratic competition may be most beneficial in developing countries.