سیاست های ارفاق و معاملات غیر قانونی

We study the consequences of leniency—reduced legal sanctions for wrongdoers who spontaneously self-report to law enforcers—on sequential, bilateral, illegal transactions, such as corruption, manager–auditor collusion, or drug deals. It is known that leniency helps deterring illegal relationships sustained by repeated interaction. Here we find that—when not properly designed—leniency may simultaneously provide an effective governance mechanism for occasional sequential illegal transactions that would not be feasible in its absence.

The economic and social costs of organized illegal transactions are enormous. Financial frauds, earning management schemes, firm/auditor and investment bank/analyst collusion have recently undermined the smooth operation of financial markets in more than one country.1 The direct and indirect costs of drugs, arms, toxic waste, and people trafficking are obvious and huge. Terrorism can also be seen as a form of organized crime. And even economists' traditional benevolence towards corruption, once seen as a way to overcome excessive regulation, has been heavily questioned by recent studies showing that corruption may reduce investment, financial development, and growth.2 Since illegal transactions involve at least two parties and require trust among them—their potential opportunism cannot be limited by court-enforced contracts—one way law enforcement agencies traditionally fight them is undermining trust by shaping incentives to play one party against the other(s): ensuring that they find themselves in a situation as close as possible to a Prisoner's Dilemma.3 Law enforcers do this by awarding leniency—typically a reduction or cancellation of legal sanctions accompanied by protection from retaliation and related benefits—to wrongdoers that self-report helping to convict “the rest of the gang”. Formal or informal exchanges of leniency against information are common in most world countries, and have been extensively and successfully used to fight Mafia, drug dealing and terrorism. It is by now well known that on multi-agent criminal relationships sustained by repeated interaction leniency for self-reporting wrongdoers can have a particularly strong deterrence effect: Spagnolo, 2000 and Spagnolo, 2004 showed that leniency for undetected wrongdoers that spontaneously self-report could in principle costlessly deter all forms of organized crime enforced by reputational considerations, making the investigation activity redundant. To have such pervasive effects, however, leniency should not only reduce sanctions, it should also reward wrongdoers that spontaneously report and turn in their partners. Instead, in reality leniency is often “moderate”: it only reduces or at most cancels legal sanctions against a self-reporting party (obvious examples are the US and EU Leniency Programs for cartels). The main reason why such policies often do not offer rewards to applicants appears ethical and political, as many people consider immoral to reward someone who acted illegally, and fear the lack of trust of an “informants' society”, although rewards to whistleblowers are and have been successfully used in many situations.4 Moderate leniency may still produce significant benefits as it reduces law enforcement cost for individual crimes (Kaplow and Shavell, 1994), and may have deterrence effects on long term illegal relationships as it may facilitate prosecution (Motta and Polo, 2003), and increase both the incentives to defect and report and the “strategic risk” of entering an illegal arrangement (Spagnolo, 2004). However, as it turns out, moderate leniency may also have puzzling side effects on occasional illegal transactions, which add a potential cost besides their just mentioned benefits. Occasional sequential illegal transactions are extremely exposed to “governance” or “hold up” problems, because to constrain opportunistic behavior from the party delivering last, the party delivering first cannot normally rely upon explicit contracts enforced by an external system.5 We show that poorly designed leniency policies may solve this governance problem by inducing transacting parties to collect hard evidence on the illegal agreement and use it as a “hostage”, threatening to report it to law enforcers in case of “hold up”. A poorly designed/managed leniency policy may ensure that this threat is credible, so that all parties comply with the otherwise unfeasible occasional illegal exchange. Williamson (1983) first discussed this “third way” by which parties may govern economic transactions and hold-up problems: transferring “hostages” between parties.6 An occasional sequential illegal exchange is a transaction in need of governance because the party delivering first risks being “held-up” by the one delivering last. The mechanism we uncover here is an example of this third form of governance, with the hard evidence on the illegal transaction playing the role of the hostage: the information is valueless in itself—as hostages typically are for the party that holds them—but a party can credibly use it in response to hold-up to seek amnesty and punish the partner that did not deliver. We model asymmetric bilateral sequential illegal transactions, like a corrupt deal between an entrepreneur or manager and a bureaucrat or auditor, where the existence/production of hard evidence on the illegal agreement, the sequence of the exchange, and the distribution of illegal gains are all endogenously chosen by the parties to ensure that the transaction can be implemented, given the legal framework. We characterize the effects of leniency under all parameter configurations, and find that the moderate forms of leniency frequently implemented in reality may indeed have a paradoxical side effect: they may provide an effective enforcement mechanism for occasional, sequential illegal transactions that would not be enforceable in their absence—nor in the absence of law enforcement altogether. We mentioned that Kaplow and Shavell (1994) highlighted how the policy of offering a lenient treatment to wrongdoers that spontaneously self-report reduces the pool of individuals to monitor without reducing deterrence, thereby directly reducing law enforcement costs (see also Malik, 1993); and that when agents are risk averse, this policy also increases welfare by reducing the overall risk agents bear. Innes, 1999a and Innes, 1999b stressed the value of the early remediation of damages from a crime that encouraging self-reporting entails.7 These analyses focus on the case of single wrongdoers committing isolated crimes, in the tradition of Becker (1968). In a novel strand of literature, Motta and Polo (2003), Spagnolo, 2000 and Spagnolo, 2004, Rey (2003), Aubert et al. (2004), Harrington (2005), and others analyse groups of organized wrongdoers who must govern/sustain internal cooperation (e.g. collusion) through self-enforcing illegal agreements, as illegal explicit contracts cannot be enforced by courts by definition. They show that the problem becomes then fundamentally different, dynamic, because repeated interaction or reputational concerns must be relied upon to govern internal opportunism; and that leniency deters organized crime—tightening “self-enforcing constraints”—by facilitating prosecution, increasing incentives to cheat on partners and report, protecting those who cheat and report from former partners' revenge, and increasing the “riskiness” of the criminal agreements.8 The present paper extends the analysis to occasional, asymmetric, sequential opportunities for illegal cooperation that cannot be governed by long term considerations, and highlights a potential perverse effect that awarding leniency to wrongdoers that self-report may have on these. Potential perverse effects of law enforcement instruments have been highlighted before, among others, by Mookherjee and Png (1995), Livernois and McKenna (1999), and Boadway et al. (2002), but these effects are very different from the one we point out here, and concern sanctions rather than leniency. Section 2 presents the simplest model we could think of to identify the alterations of individual incentives induced by a moderate leniency program that may implement sequential illegal transactions. Section 3 analyzes the model and presents the main results. Section 4 discusses extensions, tests the robustness of our results, and relates them to some existing literature, traditional arguments in favor of leniency, and reality. Section 5 briefly concludes.

Within a stylized model we characterized the effects on the viability of occasional, sequential illegal transactions of all conceivable parametrizations of leniency policies for wrongdoers that spontaneously self-report to the law enforcing agency, turning in their partners. The results highlight a risk involved in using moderate forms of leniency in law enforcement, forms that only cancel or reduce self-reporting parties' sanctions below the expected fine they would be subject to without reporting. Economists are aware that poorly designed incentive schemes may have counterproductive side effects, and leniency programs—as the law enforcement system in general—are incentive schemes. The moderate form of leniency often implemented around the world, while plausibly deterring some long-term criminal arrangements sustained by repeated interaction, may have a counterproductive side effect: it could be exploited by sophisticated wrongdoers to implement occasional illegal transactions that would not be feasible otherwise. More generous forms of leniency that offer a reward to at least one party if he or she self-reports, if appropriately designed and managed, are not subject to this potential drawback and may further increase deterrence of long-term criminal relationships. Enriching the model by introducing asymmetric information, uncertainty, mistakes in law enforcement, and other important features of reality we kept out of our model will surely modify the boundaries of our characterization, and it will be very interesting to understand in which direction. However, we are fairly confident that our qualitative conclusions will not change with future work. Leniency for wrongdoers that spontaneously self-report is a very powerful but also a potentially dangerous tool, that must therefore be properly designed and managed with care.