Encyclopedia of Law and Economics
Show Less

Encyclopedia of Law and Economics

Edited by Gerrit De Geest

The second, expanded edition of the acclaimed Encyclopedia represents a major update of the most authoritative reference work in the field of law and economics and the nine print volumes are now released online as a single integrated product.The Encyclopedia provides balanced and comprehensive coverage of the major domain in law and economics, including: criminal law, regulation, property law, contract law, tort law, labor and employment law, antitrust law, procedural law, and the production of legal rules. Each theme or volume is overseen by a leading scholar and each of the 166 entries is prepared by an expert in the field, providing an in-depth and authoritative overview of the individual topic, combined with an exhaustive bibliography, allowing users to access and filter the entire corpus of literature in law and economics.As with the print edition, the Encyclopedia is unique in serving both as an entry point and a platform for advanced research. The online edition is enhanced with Elgaronline’s powerful search tools, facilitating the search for key terms across the entire Encyclopedia, whilst the browse function allows users to move seamlessly between the volumes. These elements combine to create a powerful research tool for any researcher or scholar in the field of law and economics.
Show Summary Details
This content is available to you

Chapter 11: Corruption

Roger Bowles

[In: Volume 3, Nuno Garoupa (ed) Criminal Law and Economics]

1. Introduction

Corruption takes many forms and is encountered in a bewilderingly wide range of settings. It can be an elusive notion ranging as it does across a wide spectrum of activities from complex dealings involving politicians, government officials and business executives to simple bribery to activities which fade imperceptibly into ‘parochial corruption’ (Kingston, 2007) or ‘custom and practice’. It can be confined to a small number of specific individuals or pervade a whole organisation and can be in the form of a one-off casual event or embedded in a longer-term relationship. That it is widespread in many countries is not in any doubt. The World Bank estimates that more than $1 trillion is lost each year to pay bribes, draining the world economy of investment and developing countries of funds to reduce poverty and address development objectives.1 The United Nations (UN) has set up a Convention against Corruption,2 the Organisation for Economic Co-operation and Development (OECD) has a convention against bribery and many countries have set up anti-corruption commissions (Doig et al. 2005).

Definitions of corruption tend to rely on the notion of dishonesty, and particularly on the idea of a betrayal of public trust. They often refer to the improper use of influence or power for purposes of private gain. Public officials and politicians, for example, are particularly susceptible because they are in a position to ‘misuse’ their power or position to grant favours in exchange for bribes, campaign funding, votes or whatever. They are by no means alone in having the capacity to subvert a position of trust for private financial gain, however: corporate managers, for example, can p. 272collude with others to enrich themselves at the expense of shareholders. For the purposes of this chapter we are going to exclude this kind of fraud. We confine ourselves, as is customary (Ogus, 2004), to a relatively narrow definition of corruption that has a focus on misuse of public office and we exclude corporate corruption.

It is a characteristic of the economic approach, which we follow, to treat corruption as an issue of incentives and organisation rather than as a matter of public morals and norms (Bardhan, 2006). We treat corruption as a mutual activity, as a transaction with willing traders as both buyer and seller. Both parties are assumed to be driven by their own private interests as well as, or instead of, the ‘public interest’. This does not necessarily imply moral unworthiness, but it does mean that the design of government organisations needs to reflect private incentives if the bodies are to be proof against opportunism and exploitation.

Many definitions of corruption, implicitly at least, identify the initial breach of trust as the key point at which damage is done to a public institution. From a moral perspective this may be right: the breach of public trust is a serious matter and may be very costly to remedy. But such a breach is not of itself the only costly consequence. From a social cost perspective, the harm from corruption can be split into at least two components. One is the ‘institutional damage’ which occurs at the point when the possibility of a corrupt transaction is aired: the breach of trust, say, by an official asking for a bribe or indicating willingness to accept one. This single action may damage irrevocably the integrity and reputation of the agency. The second component is the substantive social cost of the corrupt transaction itself: for example, the cost to taxpayers of having goods supplied to government at a price above the market rate and any further adjustments that might be triggered on the part of agents anticipating a repetition of the corruption.

Our purpose in this review is to focus on the costs of corruption. We do this by bringing together contributions on the rent-seeking facet of corruption with the literature on the costs of other types of crime. We suggest that such a fusion offers a useful framework for the analysis of anti-corruption policy.

2. The literature on corruption

The breadth of the topic of corruption is reflected in the wide ambit of its coverage in the economics literature. It is a matter that has been treated in many sub-branches of the discipline, as evidenced by the very large number of Journal of Economic Literature (JEL) categories to which the papers in the bibliography to this chapter refer. Mainstream micro and macro, economic organisation, law and economics, development economics, applied micro and applied macro economics, public economics, public choice are p. 273just some of the subject areas within which substantial numbers of papers on corruption have been published: useful surveys include Andvig (1991) and the large collection in Fiorentini and Zamagni (1999). It goes without saying that there has been a great deal written about corruption outside the discipline of economics. The disciplines of law, philosophy and political science are a few examples and of course corruption is a mainstream matter of interest to non-academic audiences.

Even from a ‘law and economics’ perspective, corruption has been studied from many different angles. It can be treated variously as a crime (Kugler et al., 2005; Bowles and Garoupa, 1997; Chang et al., 2000; Polinsky and Shavell, 2001), as a transaction cost (Murphy et al., 1993), as an issue of asymmetric information (Laffont and Martimort, 1997), as an agency cost (Bac, 1996) or as a violation of an allocation of rights. It crops up sometimes in an abstract way and at other times in an empirical form. In some settings, it takes a one-shot, casual form, while in others it is more deeply embedded within organisations (Laffont, 1990; Kofman and Lawarree, 1993) and/or takes a repeated form (Mishra, 2006).

In order to give some shape to this vast area we exploit some of these distinctions. But we start by trying to capture some of the common themes underpinning most analysis of corruption.

3. Illustration of the three-party structure of a corruption model

The basic structure of most economic analysis of corruption can be illustrated in a simple three-party model. The notion of a betrayal of trust requires some form of collusion between an employee of an organisation with responsibilities to the public and someone outside the organisation seeking a good or service from it. The benefits of the collusion are split in some way between the two parties at the expense of the organisation itself and ultimately society at large, the third player or interest.

To give concrete expression to the setting, suppose an official in a passport office demands payment of a bribe (over and above the cost of the visa) to issue a visa without further delay. Provided that the official sets the bribe at a level below the amount the citizen is prepared to pay then the visa is issued and both parties have ‘gained’. If the maximum amount the citizen is prepared to pay for the visa is p (in addition to the cost of the official fee) then a well-informed official could charge a bribe b(<p), leaving both parties better off. The citizen enjoys a surplus of (pb) while the official keeps the bribe, b.

This is, of course, a simplification for purposes of modelling. For both sides, there may be reasons why this overstates the gains. The citizen has an entitlement to obtain a visa without paying a bribe, and is thus worse off than she would be had the entitlement been respected. The official may p. 274incur some moral qualms or ‘psychic costs’ in taking a bribe. This reduces their net gain from taking a bribe, but may not be sufficient to prevent them taking the bribe.

The gain (b) may appear not to affect anyone else but usually in reality it will. The official may still honestly pass the revenue from the visa fee to his office, in which case there may seem just to be a transfer payment from the citizen to the official: the transaction seems ‘victimless’. But in fact these sorts of transactions take longer when bribes are involved than when they are not and this gives rise to costs to third parties, such as those waiting in line for a visa: Murphy et al. (1993). Irrespective of how the negotiations are conducted, the net result will be that the official ties up time in the transaction and this will reduce the number of other visa applications the official can handle. This reduces productivity and causes greater delay for other visa applicants.

In addition to the transactions costs there will likely be further rounds of costs. We look more closely at the composition of these costs in the following section. For the immediate purpose of setting out the structure of our three-player model we will summarise the transactions costs (or ‘rent-seeking costs’ as they might be termed) as RSC and the other downstream costs of corruption as D. The transactions costs (RSC) will be shared by the parties to the corruption while the downstream costs (D) fall on third parties.3 For the citizen involved in the corruption, RSC will depend on the opportunity cost of time, previous experience of bribery transactions and so on. In any event, these costs will reduce the net rent (p) being shared by the conspirators. The scale and distribution of the downstream costs will be highly context-specific, as will the relationship between the size of the rent and the third-party costs.

From a social perspective, there is, in essence, a principal-agent problem. The corruption choices of the agent (the official) are not observable. Typically, the official will be paid a salary that is not (and cannot be) linked to her choices about taking bribes. The incentive to avoid corruption may thus be rather weak. There are several strategies that might be considered as a means of mitigating this moral hazard.

One is to re-design the employment contract of the official. This might entail a model based on the ‘tax farming’ approach used historically in many societies (Azabou and Nugent, 1988; Besley and Maclaren, 1993). The tax collector (or in our case, the visa issuer) pays a lump sum in exchange for the right to collect taxes (issue visas) according to the laws of p. 275the land. A good deal of public sector reform in recent decades has been motivated by this sort of approach, although it is important to keep in mind that such changes may be motivated by their incentive effects as well as their anti-collusion properties. In practice, there may be substantial cultural and other barriers to implementing radical solutions of this kind.

Another approach is to use enforcement powers under the criminal law to tackle corrupt practices directly. This treats corruption like any other offence and involves investigation of officials, with the prospect of prosecuting those who are found to be taking bribes. Models based on the approach of Becker (Becker, 1968; Becker and Stigler, 1974) can be readily adapted for a corruption setting by treating the corruption decision for an official as a simple gamble (Cadot, 1987; Bowles and Garoupa, 1997; Garoupa, 2000; Garoupa and Jellal, 2007; Polinsky and Shavell, 2001). The official’s share of the rent (b) net of any transaction costs represents the potential gain while the expected cost is the product of the probability of the corruption being detected (π) and the sanctions imposed in the event of conviction. If we assume for simplicity that transactions costs are zero and that the sanction is a fine of size f, then the corruption choice for a risk-neutral official with no psychic costs of taking a bribe is:

Accept bride iff:(1π).b>π.f

This basic structure can of course be modified in many different ways. For example:

  • the size of the bribe can be endogenised (since it will quite plausibly be a function of the probability of conviction and the size of the sanction) or fixed at some reference level that reflects the surplus gain to each side of the corrupt transaction such as the Shapley value;

  • moral restraints can be introduced that limit the net gain to the official of taking a bribe;

  • the sanction can be expanded to allow for loss of job and pension rights and other elements.

The decision rule in equation (11.1) can be used to estimate the incidence of corruption on various assumptions about the effectiveness of anti-corruption technology, as captured through the relationship of the probability of detection to the resources going into anti-corruption work. But in any event, an informed judgement about the net return to investment in corruption-reduction will require evidence about the degree of harm being caused by corruption. Particularly if, as is likely, it is very costly to p. 276prevent corruption, it is essential to have a good sense of the savings that will result.

4. The costs of corruption

4.1 Typology of costs

The approach we take here to measuring the costs of corruption makes use of a blending of the ‘rent-seeking’ analysis of corruption referred to above and the empirically based research that has been done on the economic and social costs of crime. This latter has focused primarily on ‘street crime’ such as burglary, theft and assault. Its literature normally identifies three principal components of the economic and social costs of crime (Cohen, 2000, 2005; Brand and Price, 2000; Dubourg et al., 2005; Mayhew, 2003 and others), namely: costs to victims, costs in anticipation of crime and costs of the criminal justice system. In effect, then, we are arguing for a four-way classification of the costs of corruption that incorporates rent-seeking costs plus the three components just listed. We consider each category in turn.

4.1.1 Rent-seeking costs

In the model of the previous section, corruption centred on the possibility of the official and the citizen sharing a rent (p) created by the citizen being prepared to pay a sum in excess of the visa fee. The rent-seeking costs are the costs the two sides invest in securing, sharing and protecting their interest in this rent. Corrupt transactions are illegal, with the result that the normal protection of obligations under contract law does not apply. The parties have to spend time agreeing first that a bribe is to be paid, then how much it is to be, how the agreement is to be made self-enforcing and possibly also how it is to be kept secret (Rose-Ackerman, 1999).

For a one-off transaction this rent-seeking activity can take considerable time, since the official has to probe the citizen to make an estimate of their reservation price (p). There may be haggling over the size of the bribe, particularly if the citizen either feels able to complain about treatment at the hands of the official or can credibly threaten to go to another official who will provide the service for a smaller bribe.

The scale of the potential benefits from a continuing corrupt relationship may be such as to promise substantial returns from economising on transactions costs. By entering informal agreements to govern these relationships, the parties may be able to reduce the proportion of the rents going in transactions (or rent-seeking) costs. This kind of trust strategy underlies a lot of corruption, particularly in fields such as tax administration, procurement contracting and policing. Such conspiracies are, of course, illegal p. 277and unenforceable. But because they tend to be deeply rooted in social or family networks, they can be very difficult to uncover or counter effectively (Lambsdorff, 2002). Another development that high transactions costs may prompt is the emergence of a market for agents or intermediaries with specialist knowledge of the markets in question (Hasker and Okten, 2008). Closely related arguments underpin the emergence of multi-tier corruption of whole organisations (Tirole, 1986; Laffont, 1990; Kofman and Lawarree 1993).

To digress briefly from the field of corruption, we argue that there might be merit in including rent-seeking costs when looking at the costs of all or most offence types. In the context of burglary, for example, we could argue that real resource costs are entailed in the committing of the crime itself. Burglars spend time searching for appropriate targets, acquiring tools to speed entry, acquiring burglary-specific skills and so on. Clearly in the case of offences such as fraud and embezzlement, substantial resources may be invested by offenders in finding a modus operandi and in concealment and other activities that quite closely resemble the kinds of rent-seeking costs identified by Rose-Ackerman in her analysis of corruption. These costs are all real resource costs because the time and effort could otherwise be going into productive activity.

4.1.2 Victim costs

The victim costs of street crime are usually fairly obvious: an identifiable person is injured, robbed or burgled or whatever. The victim costs of corruption are generally substantially more diffuse. In our visa example, they are shared by all the other citizens waiting to get a visa. This diffuseness impedes estimation of the extent of the victim costs. But the obvious importance of such costs for purposes of promoting economic development as well as evaluating anti-corruption policy has spurred the development of methods for inferring something about their scale.

The notion of ‘victim costs’ in the costs of crime literature is a fairly direct one: it refers to the impact of a particular offence and excludes the costs incurred though subsequent adjustments. So it could refer, in a corruption application, to the price paid over and above the lowest tender for a project, given that the two bids were similar in terms of the quality they offered. Or it could measure the loss of surplus to a group of citizens as a result of their paying a bribe for access to a service to which they were legally entitled to access at zero user cost. Or it could refer to the additional delay caused in getting a visa because an official had spent too much time on a prior corrupt transaction. It could even include losses from corrupt officials or politicians moving the proceeds off-shore to protect them, since this reduces the supply of funds available to domestic financial institutions for recycling into investment in the country.

p. 278Considerable care is needed at this point because much of the work on corruption does not distinguish very carefully between the harm to victims caused by corruption once it has occurred (‘victim costs’) and harm caused because of the effect on actions to mitigate losses from corruption that might occur in the future (costs in anticipation of crime). In some of the empirical literature, particularly cross-section analysis of the relationship between the extent of corruption and the rate of a country’s economic growth, these two elements are clubbed together.

As in other areas of crime where victim costs have been estimated, the first task is to explore the extent of victimisation. This can generally be done either by looking at the incidence of the crime as recorded in official data or it can be estimated by using surveys. In the case of corruption it is this latter methodology that is most widely used for the simple reason that there are strong grounds for believing that officially recorded corruption will identify no more than a tiny proportion of the corruption going on. A number of international organisations, including Transparency International4 and the World Bank,5 publish findings from surveys of firms, citizens, experts and others asked about their experience of being asked for bribes. For example the World Bank survey asks about three aspects of corruption, namely: unofficial payments for a typical firm to get things done (percentage of sales); firms expected to give gifts in meetings with tax inspectors (per cent) and the value of gift expected to secure a government contract (percentage of contract). The survey findings do not generally give direct estimates of the costs of the corruption involved, but it is possible to imagine the sorts of questions that would need to be asked as a prelude to making such estimates.

As we will show below in Section 5, assessing victim costs will be substantially more complex than this when we move away from petty bureaucratic corruption to more elaborate corrupt edifices where the polity and bureaucracy have become entwined.

4.1.3 Criminal Justice System costs

The costs of corruption to the criminal justice system (CJS) vary widely across countries, depending on how widespread corruption has been and how vigorously it has been pursued. Normally the three principal organisations whose costs contribute to CJS costs are the police, the courts and offender management services (prison, p. 279probation and so on). Substantial amounts of time are required for serious investigation of corruption by police (or commissions of enquiry) and prosecution. Terms of imprisonment may be lengthy, extending even to life imprisonment, and thus costly.6 In practice the investigation work is often given to independent commissions, particularly if there is a suggestion that a country’s police force is corrupt.

There are some data on the number of prosecutions being brought for corruption. In the US, the Transactional Records Access Clearinghouse (TRAC) publishes a monthly time series, taken from Justice Department sources, on new official corruption prosecutions. The monthly data, which run from October 2000, show the number of new prosecutions fluctuating between about 20 and 100 per month and suggest a slightly declining trend.7 Although useful for certain purposes, such as exploring time trends within a country (Del Monte and Papagni, 2007), such data are clearly not a reliable guide as to how widespread corruption might be, particularly when making international comparisons. Variations in legal definitions of corruption, in the volume of law enforcement resources devoted to anti-corruption work and in the political will to explore corruption, may all play a significant role.

The possibility that prosecuting bodies, police or judicial authorities are themselves corrupt also influences the propensity to bring corruption charges. Further, there are countries where it is alleged that corruption charges are motivated purely by a wish to discredit political opponents.

It is probably fair to say that there are few countries where a substantial proportion of CJS resources is going into corruption. A few costly, high-profile cases can be identified, but the CJS costs are unlikely to represent more than a small proportion of the total costs of corruption.

4.1.4 Costs in anticipation of corruption

The costs of burglary include the costs to householders of installing locks and bolts and taking other avoidance steps that are welfare-reducing. In the case of corruption, the corresponding avoidance measure may include the diversion of investment resources over the longer term away from sectors in which corruption is high. If the corruption becomes deeply embedded across many sectors in the economy, then there is a danger that investment resources will flow abroad or that consumption will rise at the expense of saving. Likewise, the prospect p. 280of high corruption may deter inward investment. The principal purpose of many indices of corruption is precisely to enable international investors to take an informed view about the countries and sectors to avoid.

Corruption prevention actions may also take the form of developing procedures and practices within organisations. Greater transparency brings its own potential costs to organisations but may also be used to strengthen a reputation for straight, honest dealing that is profitable over the longer term. Many large corporations have anti-corruption compliance programmes explicitly aimed at developing a good name on which they can trade. Likewise, accounting bodies may find themselves in a position to insist on strong enforcement of anti-corruption policy by their members as a means of improving their international standing.

Studies of international investment flows can be expected to give some indication of the extent to which countries with high corruption are losing out in the marketplace.

However, some care is needed with this argument lest greater corruption become an attraction to unscrupulous investors. If officials are prepared to take bribes in exchange for turning a blind eye to infringements of regulatory requirements, then this may reduce operating costs relative to those in competing locations. However, such advantages may be only skin deep because this same corruption can undermine significantly the quality of government services and infrastructure (Ogus, 2004). In any event, depressed investment inflows into a corrupt economy will help drive up interest rates and impoverish the economy and is thus a real cost of corruption to residents in the country.

An influential strand of the corruption literature seeks to relate national growth rates (and investment rates) to explanatory variables, including the extent of corruption in a country (Mauro, 1995; Keefer and Knack, 1997; Gyimah-Brempong, 2002) and the rate of growth of firms or the efficiency of firms to the amount of corruption (Fisman and Svensson, 2007; Boa and Rossi, 2007). These contributions typically find that corruption plays a significant role in retarding growth. Empirical estimates have been made of the strength of such effects but we do not document that here.

In extreme cases where political and government institutions have fallen prey to corruption, a widespread loss of trust can threaten social stability and even a government’s own legitimacy. Citizens may become increasingly disinclined to pay taxes or to cooperate with state institutions, resulting eventually in revolution, a coup or military takeover.

One important conclusion we draw from this review of categories of the costs of corruption is that it is the costs in anticipation of corruption that may have the most serious economic consequences. In some ways, this reflects the position in relation to certain other kinds of crime types. p. 281For example, terrorism derives a lot of its political leverage not from its immediate consequences for victims, devastating though these can be, but from the fear it creates and from the high costs that result as agents (and organisations) adjust their behaviour in line with events.

5. Political corruption

The political arena has historically been central to a great deal of what has been written about corruption, although it figures less prominently in a lot of the economics literature on the subject. Control of legislation and the granting of public office give politicians a great deal of patronage. This power can be used benignly and in the public interest or it can be misused for selfish gain. The temptations are obvious. Control of large public budgets makes a politician vulnerable to many sorts of pressure, ranging from a straight bribe to more subtle forms of influence such as offers of lavish hospitality from business leaders or lobbyists in exchange for opportunities to press a particular interest.

But there can be more subtle and sinister interpretations of corruption within the political sphere. Politicians have powers they may exercise to control key features of the landscape within which bribery occurs. In some countries, for example, they have influence over the appointment and transfer of civil servants, including tax inspectors, procurement officers and budget-controlling officials. In other countries, ministers may be in a position to approve special legislative provisions in the form of special tax exemptions, or special treatment for particular sectors or firms. Likewise, they may be able to make appointments to state-operated enterprises, such as financial institutions with discretionary powers to approve loans at special interest rates or to write off bad debts. All of these manifestations of weakness in governance arrangements have their dangers. The obvious danger is that the political will to remedy such weaknesses may itself be limited by the likelihood that it is political leaders themselves who have most to lose from reform.

There are echoes of these concerns in some of the models of corruption that explore the links between layers of hierarchical organisations (Bac, 1996). Managers of front-line bureaucrats taking bribes can usually ensure a cut for themselves. If politicians have control of the upper echelons of the bureaucratic triangle, then they too can join in. And if this is widely known, then it is only a short step to finding that the greediest will fight hardest for positions of political influence. The implication is that corruption can run very deep and result in contamination of the most central institutions. The most corrupt individuals will have an incentive to find their way to the most vital positions and exploit them for their own advantage.

p. 282Targeting enforcement efforts on front-line junior bureaucrats can thus rather miss the point. A deliberate blurring of distinctions between the branches of government, and gaining influence over critical points in the tax, procurement or budgeting systems of a country are the kinds of strategies that serious players may use. These ploys can be difficult to track. An anti-corruption prosecutor may yield, for whatever reason, to the temptation to go after the more amateurish front-line targets.

6. Anti-corruption policy

The breadth of the discussion of the design of policy to prevent or reduce corruption reflects the wide variation in the underlying analysis of corruption. The policy prescriptions produced by more abstract models are sometimes ambiguous (because optimal policy may be sensitive to parameter choice) and may have practical limitations.

One of the key lessons from observation of corruption is that great ingenuity may go not only into creating corrupt practices but also into defending the rents they generate. It is clearly essential, therefore, that the incentive structures within organisations are well understood before efforts to prevent corruption are implemented. But it is also clear that the corruption may spill over organisational boundaries, and that there may be scope for argument about the real source of corruption. For example, it is commonplace for government officials in low income countries to be very poorly paid. The responses by officials to this vary: they may spend afternoons away from the office tending market gardens, they may take bribes or whatever. The anti-corruption policy design problem is that just raising the salary levels of the officials to a reasonable level where they can survive without bribes is unlikely to be a sure-fire way of eliminating corruption (or moonlighting for that matter). The practices may have become deeply engrained in wider social relationships and networks and not susceptible to change overnight. The important point is that the ‘engraining effect’ is not accidental. It is of the essence of corrupt behaviour to want to compromise co-conspirators and to manoeuvre them into a position where they will hurt themselves more by confessing their corrupt involvement than by keeping it secret. The government of a country with a corrupt bureaucracy may, in any event, have insufficient resources to raise civil service income levels. By turning a blind eye to the corruption of department officials, a ruling party may be able to maintain civil service employment levels. Morally at least, the party has become involved in the corruption even if legally it has played no active part in its promotion. Of course it is a short step in this kind of setting for a politician to become directly implicated by taking a share of the bribe income from the top level of the civil service in exchange for protecting the status quo.

In many economic models of corruption the wage level of bureaucrats p. 283plays an important role in determining the propensity to engage in corruption. Particularly in developing countries suffering bouts of rapid inflation, public sector wages can fall to negligible levels, making jobholders highly vulnerable to bribe offers. Many models demonstrate, however, that a policy of raising wages in such circumstances is unlikely to be an effective form of corruption control (Mookherjee and Png, 1995; Besley and MacLaren, 1993; Van Rijckeghem and Weder, 2001).

Staff rotation is widely canvassed as a precautionary measure against corruption in public administration (Abbink, 2004; Rose-Ackerman, 1999), and is used by a number of governments, including the German federal government. To study the effect of this policy, an experiment was conducted where pairs of potential bribers and public officials were randomly re-matched in every round. A comparison of the outcome with the case where the identity of the pairs interacting remained fixed indicated that rotation of interacting pairs significantly reduced both the levels of bribes and the frequency of inefficient decisions due to bribery.

In a model of bribes paid for speeding up document processing, Bose (2004) shows that a reward/penalty scheme based on the average time taken to clear applications has superior properties to an enforcement-based approach.

From a business perspective, the attitude to corruption is generally rather pragmatic: if you have to pay a bribe to get a foreign contract, then that’s what you do. This of course is a non-cooperative solution. It would be better for firms as a whole if none of them ever agreed to pay bribes. But if you are in a setting where you suspect other firms are paying bribes, then the profitable policy will often to be to join in with the corruption. Of course a superior solution may be to implement international agreements to outlaw bribery, but there are clearly instances where corruption goes on despite the existence of such agreements (Rose-Ackerman, 2002).

The effectiveness of measures against corruption can be difficult to gauge. Despite the existence of some tough national legislation, such as in the US, and despite there being a number of international agreements and conventions in place, it is suspected that a great deal of corruption continues to flourish. There are many possible reasons for this. Some are simple. For example, enforcers using powers under criminal law confront a high burden of proof. A corrupt individual with a good lawyer and the foresight to have maintained a degree of ambiguity in their corrupt dealings may often be able to escape conviction. Others are more complex.

Players in these kinds of positions take care to remove themselves from positions where there is a significant probability of being detected and prosecuted. There are many defensive moves they can make to avoid detection and skilful manipulators can find ways of weakening the capacity or p. 284will of anti-corruption enforcers to convict them. Standard devices include the use of off-shore bank accounts and complex corporate structures to make it difficult to track funds, operating through intermediaries who are compensated in a roundabout way, the mixing of legitimate and illegitimate sources of income, making counter-claims of victimisation, calling in political protection, threatening to reveal harmful details of an organisation’s workings or connections and implicating co-conspirators in other illegal activities.

For the most part this multi-layering of corruption and subterfuge can successfully cover up corruption. But, particularly where corporate interests are concerned, there may be very powerful financial incentives for private or media investigation to uncover secrets that government-sponsored anti-corruption agencies cannot reach. Stock price movement analysis can sometimes be used to reveal the best informed market opinion about events behind the scenes.

As a first example, Fisman (2001) looked at Indonesian firms during the Suharto era to see whether, on days when scares about his health were made public, firms with close ties to Suharto would see their stock prices drop. The inference from the analysis of stock price movements drawn in the paper is that political connections accounted for a quarter of the value of well-connected Indonesian firms during the Suharto era.

Another intriguing example of the interplay between business and politics is the Al-Yamamah agreement signed between the governments of Great Britain and Saudi Arabia in 1985 for the provision of Tornado jets. Allegations had been made that corrupt payments, including commissions of around £1 billion, were made to intermediaries close to the Saudi government as part of the £40 billion deal. The British company at the centre of the deal, BAE Systems, claims it did nothing wrong and was working with the full knowledge of the British and US governments. An investigation of the deal by the UK’s Serious Fraud Office (SFO) was discontinued at the end of 2006 by the British government’s Attorney General, on the grounds that it represented a threat to national security and manufacturing jobs. This decision had followed a substantial dip in the firm’s share price on worries that the Saudis might pull out of a deal to buy Eurofighters because of the SFO enquiry. The resulting sharp recovery of the firm’s share price as the threat eased was reversed a few months later towards the end of June 2007 when it was announced that the Department of Justice would be pursuing BAE under anti-corruption laws8 in the US. The dip, which reduced the firm’s market capitalisation by p. 28511 per cent, or more than £1 billion, was triggered apparently mostly on the strength of fears that the US investigation might prejudice BAE’s plans to expand its position as a defence supplier in the US. The US Congress had triggered the deal in the first place, some have argued, by their objections to the sale of F-15s to Saudi Arabia,9 although there does not seem to be any suggestion that the new investigation is in any way a product of pressure from Congress. A question remains as to whether the British Ministry of Defence will comply with demands from US investigators to see confidential documents relating to the case.

Obviously this is a complex case, but it does have some interesting lessons. It shows that

  • many stakeholders may be or become involved;

  • political and corporate interests may become entwined (possibly deliberately);

  • officials in one country may be reluctant to cooperate with their counterparts in another country;10

  • uncovering a payment or a payment agreement is by no means sufficient to succeed with a corruption prosecution;

  • players in the game seeing that the political and corporate stakes are high may, or may threaten to, increase them further as part of their strategy;

  • delaying tactics may be deployed by players with short decision horizons, to the frustration of others who want to eliminate uncertainty and bring matters to a conclusion.

7. Concluding remarks

As we noted at the beginning of the chapter, corruption is believed to be very widespread. At a UN conference on corruption in 2003, Daniel Kaufman, director of global governance at the World Bank Institute, told reporters that the cost of corruption represents 5 per cent of the world economy – or more than 1.5 trillion dollars a year. Such measures are first approximations and often refer to estimates of the size of the resource p. 286transfers taking place rather than to the resource costs at stake. But they do give an indication that corruption is likely a source of significant welfare loss in the modern world. The research on which they are based also has the merit of helping identify sectors towards which anti-corruption investigations or measures might best be targeted.

Wealth-maximisation motives may lead individuals to contemplate seeking to corrupt other individuals, organisations, political parties or whatever. Looking more widely, the same motives may also lead to efforts to bring about self-serving legislative or procedural change or changes in law enforcement priorities. In some respects the dangers attending this kind of high-level hijacking of the key instruments of state power are much greater than those associated with low-level petty bureaucratic corruption. The political safeguards against such efforts need to be robust. Development agencies have rightly pressed for ‘good governance’ as a precondition for providing aid to countries. But countries throughout the world are vulnerable to pressures from individuals and organisations to implement political and administrative change that serves these special interests at the expense of the wider community. Constitutional protections have an important role to play and institutions such as the media may provide greater transparency, which can inhibit such abuses or distortions. But neither device provides guaranteed protection. Politicians may be able to force constitutional change and the media will, in the final analysis, follow the dictates of their owners. This makes political office and/or control of the media an obvious target for those seeking to manipulate political institutions for their private benefit at the expense of others.

The economics literature has explored many dimensions of corruption. It ranges from straightforward empirical studies of the link between corruption and growth to highly abstract models of organisations and incentive structures. This diversity reflects the intrinsic complexity of the phenomenon. But it has not as yet been sufficient to provide a complete understanding of corruption or to inspire a truly effective effort to prevent it.

From the web site of the Anti-Corruption Commission of the International Chambers of Commerce.

The corrupt official may, of course, be able to pass most of her share on to her employers.

For details see: www.transparency.org/policy_research/surveys_indices/cpi. In addition to the CPI, the organisation also produces a Bribe Payers Index and a Global Corruption Barometer.

The World Bank Enterprise Survey includes a section on corruption: www.enterprisesurveys.org/ExploreTopics/CompareAll.aspx?topic=corruption.

Some officials of China’s State Food and Drugs Administration were recently sentenced to terms of between 13 years and life for taking bribes: Financial Times, 8 July 2007.

The 1997 Foreign Corrupt Practices Act, which was passed following Watergate.

These various developments are a summary based on a series of newspaper articles in the Guardian (‘Transatlantic Justice’, Leader, 27 June 2007; ‘Anti-bribery Watchdog Calls for Explanation of BAE Decision’, 5 January 2007; and ‘City Shrugs off Investigation at BAE as Earnings Soar to £1.2 b.’) and the Financial Times (‘BAE Faces Fines in US Bribery Probe’, 26 June 2007 and ‘Al-Yamamah: The Saudi Foreign Policy Connection’, 1 July 2007).

That is despite the UK being a signatory of the OECD’s 1997 convention on bribery.