1 Introduction

Multilateral development banks, or MDBs,Footnote 1 are responsible for billions of dollars in development financing each year. The projects financed by MDBs are diverse, and often include infrastructure and public service components. Private companies that work in these sectors therefore place a high premium on being selected to perform contracts under MDB-financed projects and find that forced exclusion from such contracts may be devastating to their reputation and revenue. As such, administrative remedies that MDBs may apply to corrupt behavior, especially when imposed on firms in a coordinated fashion, can exert a strong influence on companies in industries that rely on publicly financed projects.

Many of the MDBs operate an internal administrative sanctions system that provides the institutions with an opportunity to impose administrative measures on companies and individuals found to have engaged in corrupt conduct while competing for or executing projects financed by the MDBs. These sanctions may range from a letter of reprimand to a public and permanent debarment of the accused company or individual. Over time, the types of sanctions applied by MDBs against corrupt or fraudulent actors have reflected a cross-institutional tendency of these systems toward transparency, innovation, self-assessment, and coordination.

This paper considers the background and stated goals of the MDBs' present approaches to administrative sanctions, the current practices with respect to sanctions adopted by the MDBs, and what lessons might be learned – and questions asked – as the MDBs continue to tackle the problem of corruption through sanctions of private companies and individuals. Professional communities concerned with integrity in international development are urged to remain sensitive to lessons learned from these systems, given the broad scope of affected businesses and persons, and the extent to which administrative sanctions become part of an increasingly sophisticated and formalized process.

2 An overview of the MDBs’ sanctions systems

The MDBs that operate internal sanctions systemsFootnote 2 employ largely parallel administrative methods designed to investigate a range of misconduct in development, including corrupt conduct;Footnote 3 temporarily exclude (aka suspend) accused companies or individuals from engagement in institutionally-financed projects; and provide a closer quasi-adjudicative review of the evidence of misconduct.Footnote 4 This quasi-adjudicative review culminates in either the imposition of a final sanction or the termination of the temporary suspension imposed on accused firms and individuals. The process, although internal and administrative, is adversarial and accompanied by formal procedures. Importantly, with the exception of settlements, the sanctions process requires an MDB’s investigative department and the company/individual accused of misconduct to present competing evidence and arguments to an impartial decision maker or independent panel within the relevant MDB. The systems generally function in a two-tiered manner.Footnote 5 The evidence is first quickly reviewed by a first-tier officer outside of the investigative unit to ensure no apparent error in designation of the accused party and confirm that the evidence is sufficient for a finding of defined misconduct.Footnote 6 The first-tier officer, upon reviewing the evidence presented, may disagree with the investigators and decline to accept “the case” or agree with the investigators and, depending on the severity of his/her recommended sanction, may suspend the accused party from engagement in Bank-financed projects.Footnote 7 Where the first-tier officer agrees with the investigators that, on a preponderance of the evidence, a sanctionable practice did take place, he/she issues a “Notice” to the accused party, which provides the accused with a full account of the accusations submitted by the investigators, the evidence presented in support, and the sanction recommended by the first-tier officer at that time.Footnote 8

The accused may provide brief arguments and evidence to request the first-tier officer to revise his/her assessment of liability or sanction. In addition, the accused may ask for a second full de novo review of the matter by the second tier of the sanctions system – a panel of independent judges termed a “sanctions board,” “sanctions appeals board/committee,” or “enforcement committee,” depending on the institution. This second-tier panel reviews the case and issues its own final judgment with respect to liability and sanction; often the parties have the opportunity for additional written submissions and oral hearings. Some of the MDBs require the second tier decision-makers to communicate their findings in reasoned decisions that may be publicly released in full or in summary format.Footnote 9

Distinctions between these systems do exist, primarily with respect to procedural steps constituting each process, but these differences do not detract from the overall common approach: an adversarial system that presents evidence from both parties to at least one impartial decision maker and follows procedural rules regarding submissions, scope of evidence, and other like matters.

3 Goals and purposes

The MDBs’ present transparent use of administrative sanctions as a formal tool against corruption and other harmful practices is a relatively new approach in the development community that represents an historic shift from a system that had given less public recognition to the problem of corruption in development and humanitarian assistance.Footnote 10 Prior to the development of MDB-specific sanctions systems, international financial institutions worked within an environment of acknowledged tension between (i) the common assumption that corruption was the cost of doing business in complex and risky environments and (ii) the institutions’ interest in disbursing funds in an efficient and effective manner that furthered development goals. This tension was memorably confronted during the 1996 speech by then-President of the World Bank Group, James D. Wolfensohn. In it, President Wolfensohn described corruption as a “cancer” that must be targeted by the development community generally and institutions like the World Bank in particular.Footnote 11 That speech was followed by a series of administrative policy revisions and internal measures, including a formal anti-corruption strategy within the World Bank Group that continues its development today.

As all the MDBs engaged in parallel internal administrative reviews to target the risk of corruption in funded projects, their sanctions systems grew and developed to provide for a broad scope of available sanctions and a progressively more structured approach to selecting the appropriate sanctions with respect to each accused firm or individual. Coordination among the MDBs on this point is notable, and has continued for a number of years. In 2006, the MDBs concluded a “Uniform Framework for Preventing and Combating Fraud and Corruption,”Footnote 12 followed up with “General Principles and Guidelines for Sanctions,”Footnote 13 and in 2010, adopted a “Cross Debarment Agreement”Footnote 14 that extended application of certain sanctionsFootnote 15 to potential beneficiaries of projects funded by all five signatories.

In the two decades since President Wolfensohn’s speech, certain patterns in how the MDBs approached sanctions of culpable entities have emerged.

3.1 Sanctions are not punitive

The individual MDBs have consistently suggested or expressed directly that their sanctions are not unidimensional punitive measures.Footnote 16 The World Bank Group’s Sanctioning Guidelines describe sanctions as driven by a dual purpose: (a) exclusion of corrupt actors from access to Bank Group financing and (b) deterrence.Footnote 17 The World Bank Group’s Sanctions Board has observed that sanctions represent a “protective and deterrent measure.”Footnote 18 The African Development Bank ("AfDB") and the European Bank for Reconstruction and Development (“EBRD”) similarly tie the purposes of their respective sanctions systems to the goal of ensuring that institutional resources and assets are used only for their intended purposes and underscore the importance of cooperative and corrective actions by culpable parties.Footnote 19 Internal guidance documents used by each of the MDBs reflect an array of sanctions and sanctioning factors that reward attempts to preempt misconduct and incentivize meaningful rehabilitation among accused companies and individuals.Footnote 20 Specifically, the documents avail the various decision-makers within the system with opportunities to allow for conditional early release from debarment or conditional non-debarment of culpable entities.Footnote 21 Notably, one condition applied with some frequency by both tiers of the various sanctions systems within the MDBs has been the implementation or improvement of an integrity compliance program within a company accused of misconduct, or any companies controlled by a culpable individual.Footnote 22 The sanctions systems’ procedural documents do appear to anticipate such use of conditionalities in sanctions and sometimes go as far as to define the manner and order in which fulfillment of such conditions is to be assessed.Footnote 23

3.2 Sanctions are administrative, albeit applied through a quasi-adjudicative process

Second, the MDBs have commented on the administrative nature of sanctionsFootnote 24 and their function as institutional business decisions,Footnote 25 rather than the results of external legal findings. This relates to an important distinction between the anti-corruption and anti-fraud measures used by the MDBs and those adopted by sovereign states, which often require a higher standard of proof for a finding of culpability and form part of a fully adversarial and adjudicative process.Footnote 26 In contrast, the MDBs’ sanctioning decisions are a product of similar but non-identical systems that often involve independent actors and several tiers of decision makers,Footnote 27 but also leave room for operational and management considerations and are procedurally less rigid.Footnote 28

3.3 Sanctions systems seek to afford reasonable transparency and due process

Third, the MDBs have placed a premium on establishing and operating sanctions systems that function according to published frameworks, provide a measure of due process to accused companies and individuals, are sensitive to international norms, and reach non-arbitrary, evidence-based outcomes. These goals appear to be largely consistent with the recommendations set out in the report authored by Richard ThornburghFootnote 29 (aka the “Thornburgh Report”), which provided initial commentary on the World Bank Group’s early sanctions system: “[t]he structure that the Bank has imposed upon the process is designed to . . . demonstrate an exemplary degree of fairness and regularity.”Footnote 30 One example of the effort to attain these goals is the World Bank Group’s evolving approach to the final decisions of its Sanctions Board. Whereas early decisions by the World Bank Sanctions Committee and later, the first decisions of the Sanctions Board, were neither available to the public nor subject to specific and public content guidelines, the current sanctions framework requires that Sanctions Board decisions set out the full reasoning behind the Sanctions Board’s conclusionsFootnote 31 and that both Sanctions Board Decisions and determinations of the Suspension and Debarment Officer be published in full when issued.Footnote 32 Other MDBs similarly ensure that the frameworks and procedures governing their sanctions processes are availableFootnote 33 and – even where full decisions are not published – the institutions follow a general practice of providing context in sanctions cases and publishing the basis of a debarment.Footnote 34

4 Application and choice of sanctions

4.1 Fixed-period or permanent debarments versus conditional release from sanctions

A review of entities debarred by one of the MDBs over the past decade reveals a possible transition from fixed or permanent debarments for firms and individuals to a larger proportion of sanctions that provide some condition for release.Footnote 35 Certain caveats must be set out in this observation. First, the MDBs’ determinations are fact- and case-specific, so distinctions between sanctions applied over time may be due more to the circumstances of the individual cases than to institutional sanctions preferences. Second, the data set of “MDB sanctions” is incomplete and inconsistent: the institutions tend to report information relating to currently-ineligible entities rather than historic figures relating to sanctions cases, and do so with varying levels of detail. Third, the data set is uneven among the MDBs, which presents obvious difficulties for attempts at trend assessment. The World Bank Group’s sanctions system has, so far, seen the greatest number of cases and has sanctioned hundreds of entities since 2007. In contrast, the African Development Bank Group has the youngest formal sanctions system, adopted in 2012, and has sanctioned fewer entities to date, not counting settlement agreements.Footnote 36 All that said, an institutional sensitivity to rehabilitation of sanctioned entities is apparent from the underlying frameworks. The Sanctions Procedures or equivalent documents applicable to each MDB provide for conditionalities to release from sanction.Footnote 37 Several MDBs identify the sanction of debarment with conditional release as a “base” or “baseline” sanction that serves as a starting point for sanctions calculation.Footnote 38 Some of the MDBs, particularly the World Bank Group, operate an independent office to assess compliance with sanctions-related conditions.

4.2 Financial sanctions and remedies

The MDBs’ respective sanctions procedures provide for various types of financial sanctions as possible administrative remedies against culpable entities or individuals. The World Bank’s Sanctions Procedures identify “restitution” as a type of available sanctionFootnote 39 and the World Bank has used restitution as a component of a more complex sanction in the past. Footnote 40 Sanctions Procedures of the African Development Bank Group and the Inter-American Development Bank provide for a broader scope of financial remedies, including restitution.Footnote 41 Financial penalties have appeared as the sole outcome in several settlement agreements between the African Development Bank and companies accused of sanctionable conduct.Footnote 42

Financial remedies are notable in that they reach beyond ineligibility and a restriction of a culpable entity’s work opportunities. Financial remedies can more visibly eliminate or minimize the profitability of sanctionable conduct and would appear to be less susceptible to evasion than public declarations of debarment. As a signaling tool, such remedies also may provide a measurable consequence for misconduct and communicate effectively the impact of a finding of liability for a culpable company or person. Furthermore, financial remedies can help meaningfully respond to certain misconduct by redirecting misused funds to entities or communities that may typically bear the long-term costs of corrupt or fraudulent schemes. Finally, financial remedies underscore the MDBs’ positive role and contribution as stakeholders that address sanctionable schemes in development and as facilitators of recovery from such misconduct.

Financial remedies, notwithstanding their positive potential, also carry costs and questions that range from mechanics of mandatory payments to the optics of such remedies themselves. The first challenge in implementing a financial remedy is identifying a connection between the misconduct and the anticipated payment so as to ensure that this payment is indeed “remedial” and can facilitate recovery. Sanctionable conduct that may have a most natural link to financial penalties may include false billing or the receipt of bribes in a way that financially harms the project or a constituency. Second, and related to the first, the remedial payment ought to be quantifiable with some specificity. This question may be rendered more challenging by the fact that the MDBs’ sanctioning guidelines generally speak to the extent of debarment to be imposed, and do not appear to include direct analysis of how mitigating or aggravating factors may impact a financial remedy. Third, the institution must appropriately identify the recipient of the payment and mechanics of the transaction in a way that is relevant to the misconduct and likely to produce a timely and effective resolution. Unexpected complications may be introduced when a planned recipient is a sovereign state or the culpable party wishes to pursue an installment agreement through an MDB or other third-party stakeholder. Financial remedies that do not adequately address such concerns may be difficult to implement or may inadvertently communicate a punitive intent behind the sanction – contradicting the original purposes of rehabilitation and risk-reduction expressed by these institutions.

As intergovernmental financial institutions consider progressively more instances of misconduct, they will build a body of sanctions work – not only under the umbrella of exclusion and debarment, but also in the category of financial remedies. As the family of sanctions cases illustrates at this point in time, restitution in and of itself has not, to date, shown itself to be a particularly popular independent sanction. Perhaps this is so because of some of the considerations listed above or perhaps it is because of a more basic threshold concern – implementation of restitution requires cooperation and participation of the culpable party and may be difficult to impose as a standalone sanction by an intergovernmental organization without the powers of a sovereign state and without a surrounding framework of incentives to ensure that companies comply with a restitution order.

4.3 Formal decision-makers vs negotiated agreements

Settlement agreements often allow for a more prompt resolution of a case, because they are frequently based on a more cooperative process that has fewer requirements for adversarial and procedurally rigid conduct.Footnote 43 That is, the accused party and the MDB’s investigative unit agree to certain facts of the case and agree to an appropriate sanction, without seeking the decision of a first-tier evaluation officer or a higher-tier panel of judges on the merits of that case. However, in spite of these time savings, a settlement agreement process is also typically less transparent. Although some institutions have in place specific parameters and internal procedures to guide sanctions decisions reached through settlement agreements, and many settlements are accompanied by public press releases,Footnote 44 the institutions do not typically disclose (i) the facts and evidence related to each case, (ii) the analysis that leads to a specific sanction, (iii) or the specifics of any conditions applied. At the same time, settlement agreements are achieved with a lower transactional cost for both the institution and the accused parties as the need to pass through different formal decision makers according to rigid timelines is obviated. This dichotomy with respect to costs and transparency/formality may be a valuable and balanced approach in select cases.

5 Implementation of sanctions

A final note on application of sanctions is the structural divide within each of the MBDs’ individual systems between (a) determination and application of sanctions on one side and (b) the implementation of selected sanctions on the other. Decision makers within the sanctions system – evaluation and debarment officers on the first tier, and sanctions appeals judges on the second tier – are generally restricted to the task of reviewing evidence presented against accused parties and arriving at sanctions appropriate under the circumstances of each case. The question of implementing sanctions – verifying asserted compliance with conditions for release, screening firms and individuals for potential exclusions, and extending the application of sanctions to successors and assigns – stands outside of that system.

The implementation-related tasks have grown in complexity as the sanctions systems themselves have become more sophisticated, and as the decision-makers have applied sanctions of increasing diversity to a growing number of culpable companies and persons. Departments responsible for screening of partners for ineligibility increasingly have to keep track of named affiliates or successors, and the structure of specific sanctions imposed. In the future, screenings by MDBs may also have to include previously undisclosed beneficiaries and other connected entities. Assessment of a sanctioned entity’s compliance with any conditions for release co-exists with a vibrant corporate compliance industry and methodologies that would seem to render MDBs’ compliance assessments and determinations increasingly complex. At present, the various sanctions systems frameworks have focused on the “front end” of sanctions; “sanctions systems” discussed publicly are presumed to include the investigative and adjudicative components. However, the growing body of sanctioned entities paired with the increasing sophistication of these systems themselves necessarily produces an effect in terms of implementation. The anti-corruption community may therefore look forward to an increase in the volume of institutional work relating to implementation of sanctions as the systems continue to mature.

6 Impacts and costs

6.1 Impacts on markets and development

Setting aside the above analysis of institutional sanctions processes, it is useful to keep in mind the business and development environment within which they exist. Although sanctions are imposed on specific entities based only on defined evidentiary records, the sanctions can also be expected to produce an aggregate effect on markets, economies, and development goals. A frequently asked question is whether exclusion of individual firms may reduce market competition by progressively subtracting more and more players. A competing hypothesis is that exclusion of firms that engage in misconduct in fact improves market competitiveness and even size, by lowering corruption or other misconduct-related barriers to entry. There has so far been little research into the aggregate effect of sanctions on supply markets, although this may prove useful in terms of informing the MDBs’ and other similar organizations’ strategies with respect to sanctions.

6.2 Cost and practice of applying compliance conditions

A large part of the “customization” of sanctions discussed in this paper is reflected in the types of conditions for release from debarment, or the conditions required to avoid debarment altogether.Footnote 45 Because of sanctions frameworks’ structures, almost all conditions imposed by the MDBs may be to some extent customized at the discretion of the officer or office assessing condition compliance. The MDBs’ respective Sanctions Procedures or equivalent documents set out the range of acceptable conditions in broad terms, such as “implementation of corporate compliance or ethics programmes, integrity and/or corporate controls,”Footnote 46 and compliance with “certain remedial, preventative or other measures.”Footnote 47 Where an institution’s sanctions framework identifies a decision-maker to confirm compliance with conditions, it again does not elaborate or add specificity to the scope of conditions or the scope of that decision-maker’s review. Instead, it appears that some discretion remains with the compliance officer, who must determine what specific actions constitute compliance with conditions in each case and who may even impose additional reporting or audit requirements.Footnote 48

Verification of compliance is a costly and complicated task, as it involves fact-finding and decision-making without control over the volume of that work program. On the one hand, the institutional compliance officer may wish to implement an analytical approach that acknowledges and parallels sophisticated corporate compliance programs. On the other hand, commitment to a thorough review must be balanced against the need for timely determinations in cases where sanctioned entities are waiting to be assessed and released. MDB sanctions frameworks may eventually address this tension directly as the number of sanctioned entities subject to conditions for release or non-debarment continues to grow.

7 Routes for continued development

Taking into account the growth of MDB sanctions systems – both in volume of impact and in sophistication of their procedural provisions – certain future developments may be useful in helping to keep the focus of sanctions systems on effective rehabilitation.

First, the task of annual reporting may in the future be treated more as a necessity than as an optional feature of sanctions systems. At present, the MDBs do seem inclined toward a joint or coordinated position with respect to sanctions, as evidenced by the “Uniform Framework for Preventing and Combating Fraud and Corruption,”Footnote 49 “General Principles and Guidelines for Sanctions,”Footnote 50 and the “Cross Debarment Agreement.”Footnote 51 However, this unification of principles is not yet reflected in institutional publication or reporting practices – some institutional units produce annual or periodic reports, but they are not easily comparable and apply different scopes of review and analysis. At the same time, the institutions produce valuable data that – even if reported in an aggregate fashion – may inform important questions in the anti-corruption and development arenae and fuel further academic research into the scope and impact of administrative remedies.

Second, and perhaps depending on the first, MDBs that operate sanctions systems may in the future begin to assess the role of tertiary development-related factors on sanctions. Such tertiary considerations may not relate strictly to culpability, but may include issues relating to development needs, markets, and impact. Of course, this consideration will only be possible when such factors themselves can be reliably measured and assessed, and may or may not lie with the same decision-makers as the evaluation officers and sanctions appeal board members that exist within the system today.

Third, greater articulation and possibly harmonization of principles specific to the implementation of sanctions, including assessment of compliance with conditions, would seem to be a natural step towards ensuring more effective rehabilitation. The institutional frameworks may try to provide broad but consistent guidance as to who assesses compliance with conditions, how compliance is assessed, and what timelines apply.

Fourth and finally, the institutional sanctions frameworks of the MDBs may soon be ripe for substantive commentary on evasion. As more entities have been subject to sanction, the investigative and screening units may now have ample data to provide input on typical methods of evasion and possible policy solutions. In order for sanctions to serve their purpose, exclusion must be effective when applied. Efforts to prevent evasion must evolve regularly and take into account common practices.

8 Conclusion

MDB sanctions systems are a microcosm of a larger anti-corruption movement within the fields of development and humanitarian assistance. They are also a key tool to help ensure development effectiveness and accountability, as they represent an institutional (and cross-institutional) commitment toward both ethical business practices and transparent due process for the accused. These systems have grown more robust and have gained impact through their choice to coordinate and harmonize where appropriate. Through their recent evolution, the MDBs’ sanctions systems have also revealed a central tendency toward rehabilitation of culpable actors by way of tailored exclusion, often with conditional return to business. A view of the future development of MDB sanctions systems through this lens of rehabilitation via exclusion, can and perhaps should inform the sanctions systems’ future goals, risks, and priorities.