64283 The World Bank Integrity Vice Presidency THE WORLD BANK 1818 H Street, N.W. Washington, DC 20433 Curbing Fraud, Corruption, and Collusion in the Roads Sector The World Bank Integrity Vice Presidency Curbing Fraud, Corruption, and Collusion in the Roads Sector June 2011 2011 The International Bank for Reconstruction and Development/The World Bank Integrity Vice Presidency The World Bank Group MSN U11-1100 Washington DC 20433 http://www.worldbank.org/integrity All rights reserved The �ndings, interpretation and conclusions expressed here are those of the author (s) and do not necessarily reflect the views of the Board of Executive Directors of the World Bank or the governments they represent. The World Bank cannot guarantee the accuracy of the data included in this work. Rights and Permissions The material in this work is copyrighted. Copying and/or transmitting portions or all of this work without permis- sion may be a violation of applicable law. The World Bank encourages dissemination f its work and will normally grant permission promptly. For permission to photocopy or reprint any part of this work, please send a request with complete information to the Integrity Vice Presidency at the address stated above. Contents ACKNOWLEDGMENTS v FOREWORD vii E X E C U T I V E S U M M A RY Ix I. INTRODUCTION 1 Development Impact of Roads 1 I I . W O R L D B A N K I N V E S T I G AT I V E F I N D I N G S 5 III. COLLUSION IN ROAD TENDERS 11 A. Evidence from Non-Bank Projects 12 B. Cartel Theory 12 C. Developed Country Experience 13 D. Effect of Collusion on Tender Prices 13 I V. F R A U D A N D C O R R U P T I O N I N C O N T R A C T I M P L E M E N TAT I O N 17 V. C O M B AT I N G C O L L U S I O N , F R A U D , A N D C O R R U P T I O N 19 A. Measures to Reduce Collusion 19 Punish cartelization severely 19 Create incentives for the exposure of cartels 20 Revise tendering rules 20 Require independent bid certi�cates 23 Retain an independent procurement evaluator 23 B. Measures to Reduce Fraud and Corruption 24 Strengthen the engineer 24 Hire a technical auditor 25 Engage civil society monitors 26 Develop accurate cost estimates 26 Check the wealth of key procurement agency of�cials 26 C. Longer-Term Capacity-Building Measures 27 Modernize the roads sector agency 27 Strengthen competition law enforcement 27 Curbing Fraud, Corruption, and Collusion in the Roads Sector D. Experimental Measures 27 Impose ceiling on bids 28 Use competitive negotiation 28 Contract out procurement 28 E. Issues for Consideration by Bank Operations Staff 29 Trade-offs between Transparency and Collusion 30 Subcontracting as a facilitator of capacity-building and collusion 30 Customizing measures to address fraud and corruption in civil works contracts 30 Developing expertise on cost estimating and detecting collusive bidding 30 Reevaluate current contract management form 31 Target enforcement on engineering �rms 32 Increase contingent of professional World Bank staff with road engineering expertise 32 Spend more on corruption prevention in projects 32 VI. CONCLUSION 35 REFERENCES 37 ANNEX 1: PROVING BID RIGGING ON ROADS TENDERS 45 ANNEX 2: REFORMS TO PUBLIC PROCUREMENT IN OECD COUNTRIES 47 Tables Table 1. Misconduct Cases in World Bank Roads Projects: Sanctions Imposed 6 Table 2. Misconduct Cases in World Bank Roads Projects: Sanctions Pending or Not Sought 7 Table 3. Estimated Cartel Overcharges 14 Table 4. Results of Audit of Zambian Roads Projects 17 Boxes Box 1. World Bank De�nitions of Misconduct 2 Box 2. Collusion and Cartels 8 Box 3. Ten Indicators of Collusive Bidding 12 Box 4. Publishing Cost Estimates: the Trade off Between Transparency & Collusion 21 Box 5. Combating Collusion by Changing the Procurement Process 22 Box 6. Certi�cate of Independent Price Determination 24 Box 7. Using Competitive Negotiation to Circumvent a Cartel: the US Experience 29 Box 8. Reducing Fraud and Corruption in Civil Works 31 iv Acknowledgments This report was authored by Richard Messick (INT) un- including Marc Juhel (TWITR), Gouthami Padam der the direction of Leonard McCarthy, Vice President, (TWITR), Hudayberdi Ahmedov (TWITR), Paul INT, and Galina Mikhlin-Oliver, Director for Strategy Bermingham (OPCOS), Bernard Becq (OPCPR), Maria and Core Services, INT. Valuable inputs were pro- Vannari (OPCPR), Moses Wasike (OPCFM), Richard vided by Anders Agerskov with support from Virginia A. Calkins (Consultant), Jack Titsworth (OPCFM), Pat Fatourou-Papanikolaou, Alba Struga, and Athene A. Rogers (OPCCS), Charles Kenny (Center for Global Vila-Boteler of INT’s Preventive Services Unit. Development), Luc Lecuit (EACTF), Carolina Monsalve (ECSS5), Bill Patterson (Consultant), Gaël Raballand Helpful comments and contributions were provided by (AFTPR), Donald Mphande (AFTFM), Tina Soreide colleagues across INT, including in particular Stephen (U4), Steve Burgess (EAPCO), Joel Turkewitz (EAPVP), Zimmerman (INTOP), Jonathan Shapiro (INTOP), Ahsan Ali (EAPPR), Cecilia D. Vales (EAPPR), Therese Simon Robertson (INTOP), Leonard Newmark Ballard (GSDPR), Antonio Capobianco (OECD), Ben (INTOP), Susan Hume (INTSC), Annie Yau (INTSC), Geiricke (ECSSD), Frank Fariello (LEGOP), Rowena M. and David Bernstein (INTSC), and numerous col- Gorospe (LEGOP), Laurence Folliot Lalliot (LEGOP), leagues in SDN, OPCS, the regions, LEG, and GSD, and Aneta K. Wierzynska (Consultant). Simon Robertson/World Bank Foreword From earliest times one of the strongest indicators of a be innovative and learn more systematically from our society’s development has been its road infrastructure, experiences and those of our development partners and or lack thereof. At its height, the Roman Empire reput- client countries. This report of the Preventive Services edly built the best engineered and most complex road Unit of the World Bank’s Integrity Vice Presidency network worldwide. The Old Testament also contains (INT) supports this effort by turning both the results of references to the ancient King’s Highway. Corduroy INT’s investigations and the experiences of developed roads were built in Glastonbury, England in 3300 BC and developing countries into practical advice about a with street paving going back to early human settlement range of measures in order to stem collusion in tenders around 4000 BC in the Indus Valley on the Indian sub- for roads contracts, and fraud and corruption in con- continent. Our history speaks roads. tract execution. Well planned, properly maintained, and safe roads are Einstein said, “We can’t solve problems by using the same critical for economic growth and overcoming poverty kind of thinking we used when we created them.� So we in developing countries. The roads sector has been a need to revisit past practices, drawing on the knowledge major target for development �nancing over the entire of those on the ground in client countries. history of the World Bank and remains important today. Between 2000 and 2010, the World Bank committed The report recognizes that conditions across borrowing close to $56 billion for road construction and mainte- countries differ signi�cantly, as they do in developed nance and expects to continue its active support for the countries, and that what works in one country may not roads sector in its client countries for years to come. in another. The measures we offer are not panaceas, or “cookbook� solutions. Diagnoses of the nature of the While roads projects supported by the World Bank problems are important in devising possible solutions. Group have had consistently positive development re- Our aim is to spur dialogue among all stakeholders on sults, dangers of fraud, corruption, and collusion plague how to improve the way the World Bank and its clients the sector worldwide. Though this is a problem for both do business in the roads sector. developed and developing countries it is much more costly in terms of opportunity costs and lost economic The bad news is that ensuring the integrity of roads growth for developing countries. Given the importance projects is a challenge for many developed and devel- of roads to the poor, this challenge is of special signi�- oping countries since fraud, corruption and collusion cance to the World Bank. historically prove resistant to easy treatment or simple solutions. The encouraging news is that the countries To help our clients safeguard their roads projects from that are committed to stamping out these problems can fraud, corruption, and collusion, the World Bank must draw upon the learning and successful experiences of Curbing Fraud, Corruption, and Collusion in the Roads Sector many others. The corrupt can be bested. Fraud can be safeguard an important driver of growth. The World thwarted. Colluding networks can be countered and Bank and other development partners stand ready to even broken. help. We want this report to be a living document, the breed- Robert B. Zoellick ing ground for new solutions, as we seek to protect and May 2011 viii Executive Summary Because an extensive, well maintained network of roads suggests that losses in Bank-�nanced programs are less is essential for economic development, road construc- than in those not subject to Bank oversight. Nonetheless, tion and maintenance projects have been a mainstay of for the developing countries of the world, any loss on a the World Bank’s lending portfolio since its founding. road project, whether funded by the World Bank or not, This long experience in the roads sector is reflected in is unacceptable. favorable project evaluations. The Bank’s Independent Evaluation Group reports that roads and other trans- This report explores how the World Bank and develop- port projects consistently score higher on measures of ing nations can reduce losses from collusion in procure- outcomes, institutional development, and sustainability ment and fraud and corruption in contract execution, than non-transport projects and the Bank’s Quality drawing on what INT has learned from its investiga- Assurance Group has found that roads projects are tions of Bank-funded roads projects, investigations and well-supervised. reports by borrowing country governments, and the experience of developed countries. The aim is twofold: At the same time, roads projects around the globe re- (a) to provide input into the World Bank’s review of its main plagued by fraud, corruption, and collusion. A policies and processes as part of the ongoing reform of Transparency International poll ranked construction its business model, and (b) to inform a broader dialogue as the industry most prone to corruption and a survey on ways to prevent collusion in procurement, and fraud of international �rms revealed that companies in the and corruption in contract execution in all roads proj- construction industry were more likely than �rms in any ects—no matter the funding source. other sector to have lost a contract because of bribery. World Bank-�nanced projects are not immune. Roughly The report begins with a review of the �ndings in 29 one-fourth of the 500 plus projects with a Bank-funded cases of misconduct in World Bank-funded projects. It roads component approved over the past decade drew follows with an analysis of the incidence of collusion one or more allegations of fraud, corruption, or collu- in procurement in non-Bank projects and estimates of sion; to date, the Bank’s Integrity Vice Presidency (INT) its impact on project price. It then examines measures has con�rmed allegations in 25 projects resulting in 29 developed countries have taken to attack collusion and cases of misconduct under Bank rules. suggests how they can be adapted to the environment in developing countries. Some steps will be the same The most common forms of wrongdoing in these 29 regardless of the country context. A country should have cases are collusion among �rms bidding on a project laws penalizing bid rigging, market division, and other and fraud and corruption in the execution of the re- forms of collusive behavior along with the commitment sulting contract. The Bank has controls to reduce these and capacity to enforce them. Other steps will depend forms of misconduct—procurement process reviews, upon the market conditions and other country-speci�c �nancial audits, and �eld supervision—and evidence circumstances and risks. Curbing Fraud, Corruption, and Collusion in the Roads Sector Some countries may wish to limit subcontracting or When collusion or corruption is systemic, change revise the rules governing how �rms qualify to bid on requires breaking the cycle of abuse by bringing in contracts. Other countries may decide that more signif- someone from the outside—a prosecution service, anti- icant changes in procurement procedures are required. corruption agency, competition law authority, supreme The report suggests that in considering such reforms, audit institution, or, in the case of a local government, trade-offs may be required to ensure that the values of the national government. If senior of�cials are involved, transparency, capacity-building through subcontract- introducing an outsider can be particularly challenging. ing, and other goals are pursued in a manner that does When corruption is deeply ingrained, short-term pallia- not inadvertently limit competition by facilitating tives, such as an independent procurement evaluator or collusion. technical auditor, may be the answer. More drastic mea- sures may also be required and the report reviews three: While preventing fraud and corruption during the ex- the use of bid ceilings, competitive negotiation, and ecution of a road contract should be everybody’s job, turning procurement over to an independent agent. the standard road contract used by the World Bank and most developing countries assigns this responsibility Not all corruption is systemic, and thus not all reforms to the consulting engineer. The engineer approves all require such signi�cant steps. In the World Bank- payment requests and change orders, ensuring in every supported Bali Urban Infrastructure Project, the cir- instance that the road is built according to speci�cations culation of tender notices to �rms in other provinces and that value for money is received. The engineer is defeated a local bidding ring. In the Philippines, civil thus the guardian of project integrity. In World Bank- society monitors uncovered corrupt schemes in a variety supported projects, however, INT has found instances of government contracts, and in the second phase of the where the engineer was asleep at the post and others National Road Improvement and Management Project, where the post was altogether deserted. Strengthening civil society groups will monitor all phases of the work. the engineer, changing the incentives faced on the job, or even retaining a second guardian to guard the The report suggests that, in addition to expanding proj- �rst guardian are some of the suggestions the report ect-level preventive measures, more attention should advances. be paid to project supervision, especially in high-risk environments, with a particular focus on veri�cation A need to appoint someone to guard the guardian is of cost estimates and the identi�cation of collusive bid- a sign of a systemic problem and INT’s �ndings echo ding. A review of the World Bank’s supervision strategy earlier reports by governments, NGOs, academics, and for roads projects may also be in order, something that donor agencies; collusion and corruption are sometimes might include ensuring that seasoned road engineers are deeply ingrained in the roads sector. The schemes may available to assist clients and enhance technical supervi- involve not only �rms but roads agency personnel and sion of the projects. even senior of�cials. In these later cases, the system feeds off itself. The higher the colluders raise the price, the None of the steps recommended are costless, but the more they can pay in bribes and kickbacks. The more losses from collusion, corruption, and fraud can be sub- they pay, the more they have to cheat the government stantial. This report seeks to spur a dialogue inside and to make a pro�t. The more corruption, the more all outside the World Bank on how to more effectively com- wrongdoers stand to gain. Thus all have a shared interest bat collusion, fraud and corruption and thus produce in business as usual. better development outcomes. x I Introduction The World Bank’s Integrity Vice-Presidency investigates signi�cant portion of the World Bank’s portfolio for misconduct in Bank-funded projects and advises World good reason: an extensive and well-maintained network Bank staff and borrowing country personnel on corrup- of primary, secondary, and feeder roads is critical for tion prevention measures. When INT �nds misconduct economic growth and poverty alleviation. As the Bank’s in a World Bank-funded project, the Bank can bar the transport strategy for 2008–2012 explains, “Because of �rms or individuals involved from bidding on future their high and diverse functionality and wide range of World Bank-�nanced contracts. It can also provide in- bene�ciaries, roads have become an essential component formation to national law enforcement authorities in the of all national transport systems, usually consuming country or countries where the misconduct occurred or the greatest proportion of public and private invest- where the companies or individuals reside for possible ment resources in both infrastructure and services� criminal prosecution. Its preventive unit distills investi- (World Bank 2008a, 48). A cross-country analysis done gative �ndings into thematic reports like this and other for the 1994 World Development Report con�rms the documents that it shares with World Bank staff and bor- importance of roads for development, �nding a strong rowing country personnel to help them reduce miscon- and consistent linear relationship between the extent of duct in future projects. a country’s road network and its level of development (World Bank 1994, 16). While documenting cases of misconduct, INT often learns of corrupt schemes prevalent in a country or Country-level studies also show the development impact across an entire industry. For example, INT’s investiga- of road construction. In rural India, road investment tion into the Philippine First National Road Improvement sharply boosted agricultural productivity and growth and Management Project revealed practices that inflated (Fan, Hazell, and Thorat 1999). In China and Thailand, highway construction costs throughout the nation. INT road investments contributed signi�cantly to growth in also found evidence of schemes involving bribery and both farm and non-farm output (Fan et al. 2000, 2002, siphoning of funds during contract execution in roads 2004), a �nding recently replicated in Uganda (Gollin projects in Bangladesh, Cambodia, India, Indonesia, the and Rogerson 2010). In Mexico, increases in investment Philippines, and Senegal. in roads led to a strong and positive increase in labor productivity (Deichmann et al. 2002). An analysis from the United States pointed to the steep decline in public Development Impact of Roads spending on road infrastructure as the likely cause of a fall-off in productivity in manufacturing in the 1970s In the period 2000–2010, the World Bank lent close to (Fernald 1999). $56 billion for road construction and maintenance— slightly less than 20 percent of the Bank’s total lending Roads projects are an important part of the World over the past decade. Lending for roads constitutes a Bank’s portfolio because, as the Bank’s Independent Curbing Fraud, Corruption, and Collusion in the Roads Sector BOX 1 World Bank De�nitions of Misconduct The World Bank debars any contractor found to have engaged in one or more of the following forms of misconduct on a Bank-funded contract: Corrupt practice: offering, giving, receiving or soliciting, directly or indirectly, of anything of value to influence improperly the actions of another party. Fraudulent practice: any act or omission, including a misrepresentation, that knowingly or recklessly misleads, or attempts to mislead, a party to obtain a �nancial or other bene�t or to avoid an obligation. Collusive practice: an arrangement between two or more parties designed to achieve an improper purpose, including to influence improperly the actions of another party. Coercive practice: impairing or harming, or threatening to impair or harm, directly or indirectly, any party or the property of the party to influence improperly the actions of a party. Obstructive practice: deliberately destroying, falsifying, altering or concealing of evidence material to the investigation or making false statements to investigators in order to materially impede a Bank investigation into allegations of a corrupt, fraudulent, coercive or collusive practice; and/or threatening, harassing or intimidating any party to prevent it from disclosing its knowledge of matters relevant to the investigation or from pursuing the investigation, or acts intended to materially impede the exercise of the Bank’s inspection and audit rights. Source: World Bank (2006a). Evaluation Group has observed, the poor are often their collusion in such projects is of special signi�cance to prime bene�ciary (World Bank 2007, 4). In Ethiopia, the World Bank. Evidence gathered by INT shows that access to all-weather roads reduced poverty by almost this impact can be quite substantial. In the Cambodia seven percent and increased consumption growth by Provincial Rural Infrastructure Project, collusion sharply 16.3 percent (Dercon et al. 1998). Ahmed and Hossain inflated construction costs. In Indonesia, the use of (1990) found that better road access by the rural poor substandard construction materials reduced the use- in Bangladesh increased household income from both ful life of a road and damaged the vehicles using it. wages and micro-business earnings. In rural Vietnam, According to trucking association representatives in the poor reported that the greatest bene�t they real- Bangladesh, poorly maintained roads halve the use- ized from improved access to roads was educational; ful life of members’ vehicles. INT also saw contractors children were able to attend school year-round (Songco fraudulently failing to comply with such essential safety 2002). An assessment of a World Bank-funded road features as lane markings, resulting in a sharply in- project in Morocco found that it not only boosted pro- creased risk of accidents. ductivity and encouraged the planting of higher-value crops but also improved access to health services and One of the challenges in preventing fraud, corrup- increased school attendance levels (Khandker, Lavy, and tion, and collusion in the roads sector is that there Filmer 1994). are so many ways they can seep into the process of designing, tendering, and managing construc- Because roads projects are especially important for tion contracts (Patterson and Chaudhuri 2007). The poverty reduction, the impact of fraud, corruption, and 2006 Project Appraisal Document for the Paraguay 2 Introduct i o n Road Maintenance Project identi�ed 36 areas at risk everything everywhere always, time and other resources of corruption in the design, planning, award, and are limited. The aim of this report is to help policymak- management of a roads contract and recommended ers prioritize oversight resources by identifying recur- monitoring 59 different indicators (World Bank 2006b, ring forms of misconduct in roads projects that cause 146–154). While in an ideal world borrowing coun- signi�cant harm and suggesting measures to reduce or try personnel overseeing roads projects would watch eliminate them. 3 II World Bank Investigative Findings A review of INT cases in the last ten years provides criti- single project or misconduct on separate projects may be cal insights into the nature of the problems that may treated as a single case. Thus, for ease of reference, cases arise in roads projects in terms of the various forms of in the table are separated by highlighting. Cases where fraud, corruption and collusion, and the World Bank’s proceedings are pending or where sanctions were not ability to detect, investigate, and sanction such mis- sought remain con�dential. Table 2 provides a general conduct. This data can inform and guide the reforms description of each of these 19. underway in the Bank’s business model and the related policies and processes. Given its limitations, however, The data in the two tables must be interpreted with care: this data cannot be used to extrapolate the scale of the problem. Moreover, many of the preventive measures The 29 cases arising from the 25 projects were being introduced in projects are relatively recent and opened on the basis of complaints INT received, not their impact and cost effectiveness require close observa- as the result of drawing a representative sample of tion and adaptation. the 540 projects with a road component approved during this period. Therefore, no inference about In the 10-year period July 1, 1999, to June 30, 2009, the incidence or degree of fraud, corruption, or col- INT found misconduct in 25 World Bank-funded roads lusion in the World Bank’s roads portfolio can be projects. Sanctions were imposed in ten cases and pro- drawn from these data alone. ceedings are pending in another �ve. Sanctions were not The fact that some countries have more cases than pursued in the remainder for one of several reasons: the others does not necessarily mean there is more government had already imposed effective penalties, the fraud, corruption, and collusion in their roads sec- evidence was too dated or was insuf�cient, or the sever- tor than in the roads sector in other countries. Cases ity of the violation did not merit the commitment of re- vary signi�cantly in scope and, as noted above, sources required to see the matter through the sanctions there are instances where misconduct on separate process. There were also instances when the pursuit of projects was lumped together in a single case and sanctions would have required revealing information others where more than one case arose from a single that was obtained in con�dence or that might put wit- project. nesses in jeopardy The data do not capture all attempts to corrupt the procurement process. Depending upon risk levels All cases that result in sanctions are in the public do- and national procurement capacity, a certain per- main and summaries are posted on INT’s website centage of contracts in every Bank-funded project is (www.worldbank.org/integrity). Table 1 lists the ten roads reviewed. Because roads projects are considered high cases by country with a brief description of the project, risk for corruption, the World Bank’s procurement its dates of operation, and the principal forms of mis- specialists review a large number of contracts in conduct found. More than one case may arise from a these projects and have uncovered efforts to falsify Curbing Fraud, Corruption, and Collusion in the Roads Sector TABLE 1 Misconduct Cases in World Bank Roads Projects: Sanctions Imposed Project Fraud in False Country Project dates Collusion implementation documents Bangladesh Third Road Rehabilitation and 1997–2005 Maintenance: One contract for supervision of road reconstruction. Third Road Rehabilitation and 1997–2005 Maintenance: One contract for design and supervision of feeder roads. Cambodia Provincial Rural Infrastructure: Seventeen 2003–2010 road rehabilitation contracts, total $8.9 million. India Andhra Pradesh State Highway: Two 1997–2004 contracts for road widening and strengthening of highways, total $91 million. Indonesia Sumatra Region Roads: Twenty-two road 1997–2005 rehabilitation contracts, ranging from $56,025 to $614,415. Second Sulawesi Urban Development: 1997–2002 One contract to refurbish roads in villages, $18,300. Second Sulawesi Urban Development: One contract to oversee design engineering work, $320,000. Kenya Urban Transport Infrastructure: A contract 1993–2005 * to build a GIS database of urban road inventory and condition survey, $2.7 million. Philippines First National Road Improvement 1999–2007 and Management: Two contracts to rehabilitate and upgrade roads and bridges, $33.2 million. Senegal Urban Development and Decentralization 1997–2004 Program: Two road rehabilitation contracts, $99,270 and $133,440. Urban Mobility Improvement: Three 2000–2008 contracts for road construction works. Key: violation substantiated; * violation reasonably suspected a bidder’s prior experience, �nancial strength, and instances of misconduct in World Bank-funded other quali�cations; the submission of fraudulent projects or remedies applied to address it. bid securities; and bidding patterns that suggest collusion. When misconduct is suspected, remedial Despite the caveats, these 29 cases do provide impor- action can be taken on the spot. INT is also noti- tant insights into misconduct in World Bank-funded �ed and depending on its priorities, may open an roads projects. They show �rst the different ways in investigation. INT data thus does not capture all which World Bank staff either discover or learn of the 6 World Ba nk Investiga tive Fi ndi ngs TABLE 2 Misconduct Cases in World Bank Roads Projects: Sanctions Pending or Not Sought Project Fraudulent False Region Project description dates Collusion Implementation documentation Africa, Works and Employment: One contract 2000–2007 Eastern for preparation of tender documents for Europe, paving three streets, two for technical Central studies, and one for road pavement Asia supervision, total $57,634. Transport Development: One road 2005–2010 rehabilitation contract, $7.5 million. Roads Improvement Project. Contract 2006–2013 to improve major highway, US$24 million. Municipal Development: Contract to 2002–2007 * rehabilitate four city streets, $727,000. East & Transport: Three contracts for 1998–2005 South Asia rehabilitating flood-damaged roads, each $2.5 million. Transport: One contract for supply and 1998–2005 installation of equipment, $128,700. Infrastructure Development Fourteen 1997–2007 contracts for repair of flood-damaged roads, $35 million total. Rehabilitation: Two road rehabilitation 2001–2005 contracts, $83,524 and $69,261. Road Improvement: Six contracts for 2001–2008 widening and strengthening highways. Regional Roads: Three maintenance contacts, ranging from $83,853 to $267,005. Urban Development: One contract to refurbish roads in villages, $13,700. Urban Development: One contract to refurbish roads in villages, $16,000. Urban Development: Contract for pedestrian road improvement, $120,000. Roads Infrastructure: A contract 2001–2009 for consulting services for project 2001–2009 preparation, $2.89 million. Regional Transport: One training contract and one design and supervision contract, $2.7 million total. Regional Transport: Two contracts to build two roads, $14.5 million total. (continued on next page) 7 Curbing Fraud, Corruption, and Collusion in the Roads Sector TABLE 2 Misconduct Cases in World Bank Roads Projects: Sanctions Pending or Not Sought (continued) Project Fraudulent False Region Project description dates Collusion Implementation documentation Highways Management: One 2002–2013 * consultancy contract for the development and implementation of a Central Roads Database System, a Bridge Management System and a Road Maintenance System, $2.5 million. Latin Rural Investment: Eleven road and one 1998–2006 * * America bridge rehabilitation contract, ranging from $30,000 to $300,000. Road Rehabilitation and Maintenance: 1998–2005 Fifty-four contracts for maintenance of roads, ranging from $6,200 to $47,000. Road Rehabilitation and Maintenance: 2006–2011 One contract for supply and transport of cobblestones, $2.7 million. Key: violation substantiated; * violation reasonably suspected. misconduct. In eight instances, the World Bank was BOX 2 alerted by competitors of the �rms sanctioned; in an- Collusion and Cartels other seven Bank staff discovered the misconduct; in �ve evidence was uncovered in the course of a �duciary Collusion refers to any combination or agreement—no matter how informal—among review jointly conducted by INT and regional staff; bor- sellers, to raise or �x prices or rig bids or to rowing country of�cials flagged suspicious activity in reduce output in order to increase pro�ts. three and the supervising engineer in two. Although the term cartel is often used when the collusive arrangement is a formal agreement, These data also show the types of misconduct most the economic effects of collusion and cartels often found in World Bank-funded projects. In the 29 are the same. In line with usage in many cases the three most common forms were: OECD countries, this paper uses the terms interchangeably. Collusion—bidders agreed among themselves who Source: OECD (1990). would win the bid (see Box 2). False documentation—typically, the submission of false documents to qualify to bid. grounds to believe it was present in four more projects. Fraud in the implementation of a contract—usually It also substantiated 11 instances of fraudulent docu- overbilling or undersupplying materials during con- ments and nine of fraud during contract execution. tract execution, often with the connivance of project overseers. Other forms of misconduct were less common. In a proj- ect in Asia INT uncovered evidence that of�cials of the As the tables indicate, in many cases more than one project overseeing the ministry had hidden interests in type of misconduct was substantiated; for example, in the winning bidder; kickbacks to career government em- the Cambodian Provincial Rural Infrastructure Project, ployees, elected of�cials, political parties, or some combi- INT documented all three. Across all 29 cases, INT sub- nation were alleged in several projects in South and East stantiated ten instances of collusion and had reasonable Asia and Latin America. Two World Bank staff skimmed 8 World Ba nk Investiga tive Fi ndi ngs funds from a project in Africa and were subsequently affecting roads projects in developing and developed dismissed and then prosecuted by national authorities. countries. Better understanding of these risks should enable the World Bank and its borrowers to detect and As the sections below demonstrate, INT’s �ndings address them more effectively. are consistent with the most common integrity risks 9 III Collusion in Road Tenders The World Bank’s mandate requires that it give “due at- of the contract value in exchange for contract awards. A tention to considerations of economy and ef�ciency� Kenyan informant said that “collusion was rife� in the when funding a project; its Procurement Guidelines nation’s roads sector, an allegation later con�rmed by the therefore require that, in all but a few narrowly cir- Kenyan Roads Authority and the Kenyan Anticorruption cumscribed instances, the contracts it �nances be let Commission (Government of Kenya 2007, 2004). After competitively (World Bank 2010a, 7). In roads projects, interviewing several �rms and government of�cials in competition most commonly takes the form of a one- Cambodia, INT investigators concluded that there were stage sealed-bid auction. The agency responsible for the strong indications that “a well-established cartel,� aided project prepares a description of the work required and and abetted by government of�cials, controlled the solicits bids from eligible �rms. Bids are kept con�den- award of roads contracts. In the Philippines, “Numerous tial until a speci�ed day, when they are opened in public witnesses independently informed INT investigators and the bidder offering the lowest price is declared the that a well-organized cartel, managed by contractors winner. When bidders have equal access to informa- with support from government of�cials, improperly in- tion about the proposed work and compete with one fluenced [Department of Public Works and Highways] another to win the tender, this method of awarding con- contract awards and set inflated prices on projects fund- tracts produces economy and ef�ciency (Milgrom 2004; ed by the Bank and others.� (World Bank n.d., 3) One McAfee and McMillan 1987). Indonesian respondent explained that “the Indonesian collusive system had been operating for 32 years, and Evidence gathered by INT, however, suggests that road many viewed the ‘free market’ system as counter to the contract awards are not always the result of competition. cultural norm of consensus and cooperation,� a state- For example, Bank-funded roads contracts require a bid- ment consistent with reports by Indonesia’s competition der to submit a bill of quantities, a document showing law authority (Soemardi 2010) and scholarly research the materials, equipment, and labor it expects to use to (Van Klinken and Aspinal (2011). build the road along with their costs. In a competitive market, a bidder calculates unit prices for each item on Besides these examples, some INT cases labeled “false the basis of its cost structure, estimates the amounts documentation� in the tables may be the result of col- required, and arrives at its bid price. But in a series of lusion as well. In a project in Eastern Europe, a World contracts in an Asian country INT found anomalies and Bank procurement specialist alerted INT to a pattern in inconsistencies in unit costs and totals for line items that the bids on a street rehabilitation contract that suggested showed that bidders had worked backwards from a pre- bid rigging. The cost �gures in the bids submitted by determined price. the only two �rms competing were virtually identical— down to the same typos in both. The only difference In an investigation in Bangladesh, evidence showed in the two bids was the total price: one was 1 percent that companies paid project of�cials up to 15 percent below the engineering cost estimate, and the other was Curbing Fraud, Corruption, and Collusion in the Roads Sector BOX 3 Ten Indicators of Collusive Bidding 1. Number of contract awards to a speci�c �rm 2. Project bid tabulations 3. Firms that submitted a bid later became a subcontractor on that project 4. Rotation of �rms that are the low bidder 5. A consistent percentage differential between the �rms’ bids 6. A speci�c percentage of the available work in a geographic area goes to one �rm or to several �rms over a period of time 7. A consistent percentage differential between the low bid and the engineer’s estimate 8. Location of the low bidder’s �rm versus location of the second and third low bidders’ �rms 9. Variations in unit bid prices submitted by a bidder on different projects in the same setting 10. Number of �rms that requested bid packages versus the number actually submitting a bid Source: Government of the United States (2004). 1 percent higher. While INT could not substantiate col- 2009). In Tanzania, a review by a former Prime Minister lusion in this case, it did �nd that the high bidder had disclosed an industry-wide cartel in the roads sec- provided a false bid security. When �rms have agreed tor (Government of Tanzania 1996). In 2005 Indian in advance which one will “win� the contract, the des- Deputy Government Secretary Sanjeet Singh told ignated losers frequently submit higher “cover bids� participants at an international conference that car- to camouflage the agreement (Khumalo, Nqojela, and tels in the roads sector operated in various Indian Njsane 2009) Further, because banks charge for issuing states (Singh 2005). A joint study by the Government a bid security, cover bidders often falsify the security to of Nepal, the Asian Development Bank, the U.K.’s save money. Collusion was also likely in a case in Latin Department for International Development, and the America in which three �rms that submitted low bids World Bank concluded that in recent years no tender on a contract were disquali�ed for reasons that INT in the Nepalese construction industry had been free suspected were aimed at keeping new entrants out, a of collusion (Government of Nepal 2009). A statisti- common strategy for preserving a bid-rigging scheme cal analysis of bids in road tenders by the Lithuanian (Lambert-Mogiliansky forthcoming). competition agency strongly suggested collusion among �rms there (Government of Lithuania 2008); a 2009 How common is collusion in roads projects? Neither the World Bank study of public procurement in Armenia data in INT �les nor information from any other source noted widespread reports of collusion in tendering can provide a de�nitive answer. But the INT �ndings, (World Bank 2009b); and in 2005 the Slovakia Anti- considered with the results of other case studies of the Monopoly Of�ce uncovered a cartel among road con- roads sector in developing countries, the experience in struction �rms (Government of Slovakia 2005). At the developed countries, and cartel theory, suggest that col- 9th Global Forum on Competition in 2010, the govern- lusion in roads projects in developed and developing ments of Columbia, Peru, Pakistan, and Turkey all re- countries is signi�cant. ported that cartels were operating in their roads sector (OECD 2010a). A. Evidence from Non-Bank Projects B. Cartel Theory Staff of the Overseas Development Institute reported evidence of an industry-wide cartel to �x prices on It is not surprising that cartels are common in the road roads contracts in Uganda (Booth and Golooba-Muteb construction industry in developing countries. Road 12 Collusion in Roa d Tender s construction and repair markets tend to be dominated forum that cartels operated in their roads and construc- by the same few �rms; the “product,� a road, is standard- tion industries (OECD 2008b). In 1992, the Dutch par- ized; prices are relatively insensitive to demand; entry is liament concluded that the entire construction industry often dif�cult, and market conditions are predictable. In in the Netherlands was cartelized (Van den Huevel addition, would-be competitors often exchange infor- 2006); in 2000 the Swiss Competition Commission con- mation about both past and future opportunities and cluded that the market for road surfacing in the north- develop ties through subcontracting, joint ventures, and eastern part of the country was controlled by a cartel membership in trade associations. The presence of any (Hüschelrath, Leheyda, and Beschorner 2009), and in one of these factors increases the likelihood of collusion. 2010 the Konkurransetilsynet, Norway’s competition au- When all are present, the probability of collusive behav- thority, �ned two companies for colluding on highway ior is very high (Grout and Sonderegger 2005). bridge maintenance tenders (Government of Norway 2011). Another indication that collusion continues to The awarding of contracts through public tenders ag- be a problem in developed countries is the work of the gravates the tendency toward cartelization in the sector. OECD. Over the past decade it has held �ve conferences To ensure that contracts are fairly awarded and corrup- and issued half-dozen papers on how to combat bid rig- tion risks minimized, both borrowing country govern- ging and cartelization in the construction sector. ments and Bank procurement rules require that tenders be conducted transparently. Yet, as explained below, disclosure of some kinds of information facilitates D. Effect of Collusion on Tender Prices collusion. The effect of a cartel is to raise prices above what they would be in a competitive market. An analysis of bids C. Developed Country Experience from the American state of Florida showed that collu- sion on highway contracts increased prices by 8 percent Collusion in the bidding for road contracts is a problem (Gupta 2001) and a similar study found prices in South for developed countries as well. The U.S. Department Korean highway construction markets to be 15 percent of Justice launched a vigorous effort in the late 1970s higher than they would have been without collusion to stamp out bid rigging in auctions for state highway (Lee and Hahn 2002). The Dutch parliament estimated contracts, bringing cases in 20 states that resulted in 400 that cartelization added as much as 20 percent to the criminal convictions, �nes of $50 million, and 141 jail price the government of the Netherlands paid on con- sentences between 1979 and 1983 (Flax 1983). Despite struction contracts (Van den Heuvel 2006), and col- these efforts, one-third of all Justice Department cartel lusion on construction contracts in Japan is thought prosecutions in the following four years were for bid to have raised prices anywhere from 30–50 percent rigging on state highway construction contracts (Joyce (Woodall 1996: 48). Surveying economic studies and 1989). Only in the 1990s did cartel prosecutions begin judicial decisions containing 1,040 estimates of cartel to decline, a trend of�cials attribute to both the imposi- overcharges, Connor (2009) found the median cartel tion of stiff penalties for collusion and changes in state overcharge was 25 percent. procurement laws to abolish publication of contract es- timates, public opening of bids, and convening of meet- These estimates are almost all drawn from cartels op- ings where all bidders can attend (Government of the erating in developed nations. What evidence there is United States 2008). from developing countries suggests the impact is even greater there. Using information from donor-funded The United States is not the only industrialized na- roads projects in 29 countries, Estache and Iimi (2008) tion where cartels are active in the roads sector. estimated that collusion can increase the per-kilometer Representatives of Denmark, France, Germany, Japan, cost for building a road by as much as 40 percent—from Sweden, and the United Kingdom told a 2008 OECD $0.5 million to $0.7 million. INT compared the winning 13 Curbing Fraud, Corruption, and Collusion in the Roads Sector TABLE 3 Estimated Cartel Overcharges have little reason to fear law enforcement authori- Road contracts ties. Bangladesh, Cambodia, and the Philippines, State of Florida 8% three countries where roads sector cartels have oper- ated, have no comprehensive anti-cartel legislation Republic of Korea 15% (Dabbah 2010). Even where an effective law is on the Tanzania 15–60% books, many developing countries have yet to create Philippines 20–60% institutions that can enforce it (Stewart, Clarke, and Sample 29 developing countries 40% Joekes 2007; Zoghbi 2009). All construction contracts – As staff in the Bank’s transport sector have ob- served, “government of�cials are often involved� in Netherlands Up to 20% the cartel (World Bank 2009a, 42). INT investiga- Japan 30–50% tors were told that foreign �rms wanting to bid All cartels 25% on roads contracts in Bangladesh were warned by See text for sources. a senior roads agency of�cial that they would be disquali�ed if they undercut the price local �rms had agreed on. In India, a senior of�cial reported that “road ma�as� of contractors, engineers, the local police, civil servants, “and last but not least local politicians� all conspire to keep prices on road bids on donor-�nanced roads projects in the Philippines contracts above market rates (Singh 2005); and in against engineering costs estimates and found a 30 per- explaining roads sector corruption in the state of cent variance; earlier estimates range from 20–60 per- Jharkhand, a civil society activist told the New York cent (Batalla 2000). Prices in Tanzania in the 1990s were Times that “the nexus of politicians, contractors found to be 15–60 percent above competitive prices and bureaucrats is very strong� (Polgreen 2010). (Government of Tanzania 1996); a 2003 investigation in In Uganda, “the tendering process has been turned Romania revealed that contractors conspired to mark into a business by politicians at the district to settle up the price of concrete used in road construction by 30 their economic problems. . . . [They] pressure evalu- percent (Oxford Business Group 2004); and a Turkish ation teams� to select certain contractors (Oluka government study showed that, thanks in part to cartel- and Ssennoga 2008). ization, road construction costs in Turkey were 2.5 times higher than in the United States (Gönenç, Leibfritz, and For a cartel to “succeed,� its members must (a) agree on Yilmaz 2005). who will “win� the tender and at what price, (b) curb “cheating� or undercutting the agreed price by individu- Cartel-set prices in developing countries are higher than al members, and (c) prevent nonmembers from disrupt- those �xed by cartels in developed countries for two ing the agreement by submitting a lower bid (Levenstein reasons. and Suslow 2006). Cartels rarely �nd permanent �xes to these problems. Some members cheat to boost short- Fear of prosecution moderates cartel overcharges in term pro�ts or new entrants succeed in submitting a developed countries. Members of a New York State winning bid. Even when the cartel is able to dictate highway bid-rigging ring counseled each other to who can bid and how much, there are often periods of limit excess pro�ts on tenders to 20–25 percent rath- instability during which the price to some customers is er than 40–50 percent. As one conspirator explained at or near the market price. But when government of- during a trial, “getting too greedy� might trigger �cials participate in the cartel, its durability is virtually an investigation (State of New York v. Hendrickson assured. They can dictate which member will “win� the Brothers Inc., 840 F.2d 1065 (2nd Cir. 1988)). By bid and at what price, rejecting bids that undercut the contrast, cartels in many developing countries often agreed price and refusing to permit non-cartel members 14 Collusion in Roa d Tender s to bid. Gambetta and Reuter (1995) reported that or- intimidation and violence and take a share of the cartel’s ganized crime families perform the same functions for pro�ts in return. The effect is the same as when govern- cartels in Sicily and New York: where family members ment of�cials enforce a cartel agreement: the long-term police compliance with the cartel agreement through stability of the cartel. 15 IV Fraud and Corruption in Contract Implementation The risk of misconduct in roads projects does not end TABLE 4 Results of Audit of Zambian Roads Projects with contract award. A winning bidder may fraudu- Percentage lently bill for work not done, materials not supplied, of contracts or both. Evidence INT gathered in a project in Africa Defect found in project affected shows fraudulent claims amounting to 15–20 percent Improperly sized aggregate 44% of the bid price. An INT analysis of two contracts let particles under a road project in Asia found that fraud may have Too much clay 75% inflated the �nal price on each contract by as much as 25 Aggregates did not meet 67% percent. INT substantiated misconduct during contract crushing strength performance in nine of the 29 cases shown in Tables 1 Base thinner than required 81% and 2 and suspected, although was unable to substanti- Surface dressing layers thinner 82% ate, its presence in several more. than required Cement content less than 100% Reports from Zambia suggest the scope of one form of speci�ed fraud—furnishing substandard materials during con- Concrete samples weaker than 50% tract implementation. Zambian contractors, engineers, required and government of�cials surveyed in 2008 reported that Source: Government of Zambia (2010). providing materials of lower quality than the contract called for was the single most “unethical� practice in the industry (Sichombo et al. 2009) and a 2010 audit of 18 Zambian roads projects jointly �nanced by the govern- of a form contract for construction developed by the ment and donors, shown in Table 4, con�rmed their International Federation of Consulting Engineers, view (Government of Zambia 2010). As the data there known by its French acronym “FIDIC� (Jaeger and Hök reveals, substandard cement was supplied in all projects 2010). The FIDIC contract provides that the govern- while in half the projects the concrete was weaker than ment agency issuing the contract will hire an engineer— required. INT found similar levels of fraud in a contract an individual, or, for large projects, a �rm—to oversee in Indonesia: the road was 40 percent thinner than the contract performance (World Bank 2010b, ¶3.1). The contract speci�ed and the contractor used 13 percent engineer must be expert in the design and construction less asphalt than required. of roads, for the FIDIC contract requires that he observe the work as it progresses, testing completed sections to For the construction of roads and other civil works, ensure they meet speci�cations, certifying the contrac- the World Bank requires borrowers to use a variation tor’s invoices, evaluating and passing on its requests to Curbing Fraud, Corruption, and Collusion in the Roads Sector vary from the original plans, and resolving conflicts investigators uncovered evidence that the engineer cer- between the borrower and the contractor (Ndekugri, ti�ed invoices for charges not covered by the contract. Smith, and Hughes 2007). If the engineer �nds that In Indonesia, engineers admitted they were bribed to the builder is supplying substandard materials or less ignore fraud, explaining that if they did not go along, lo- material than required, inflating invoices, or otherwise cal of�cials “in on� the fraud would refuse to hire them trying to milk the contract, he must refuse to certify the on future government projects. In a project in Africa, contractor’s payment requests. The engineer is explicitly INT received information that in return for approving responsible for the quality of the project and thus be- inflated invoices the engineer received 15 percent of the comes the implicit guardian of its integrity. amount overbilled. The practice is apparently wide- spread in that country; during the investigation INT Despite the engineer’s responsibility for project integ- learned that the builder had instructed its local af�liate rity, there is evidence that engineers have either failed to to “develop partnerships with local consultants,� so that spot fraud or corruption in project execution or become if they were appointed engineers on future projects, they willing participants. In a project in Latin America, INT would be sure to cooperate with similar schemes. 18 V Combating Collusion, Fraud, and Corruption This section describes a range of measures policymak- A. Measures to Reduce Collusion ers should consider to mitigate the risks of collusion, corruption and fraud in road contract procurement Punish cartelization severely and project execution. “One size does not �t all� is a staple of the development literature and one that holds To combat collusion and cartelization, countries should for both procurement rules and mitigation measures enact laws that make bid rigging, market division, and (Mariel 2003). Accordingly, in discussing the various other cartel-related behavior illegal. These laws need to recommendations, the report identi�es the risk pro�les contain “effective sanctions of a kind and at a level ad- and country contexts where they are most likely to be equate to deter �rms and individuals from participating appropriate. in cartels.� (OECD 1998) Deterring collusion can require more than criminal penalties; a European Commission The recommendations advanced range from modest White Paper argues that to effectively deter cartels, sanc- changes in procurement procedures to more funda- tions must give those harmed by cartel pricing, such mental, experimental measures that may be required as road users, the right to sue for damages (European where corruption is particularly entrenched. Some Commission 2008). country-level reforms, such as laws severely penalizing bid rigging or changes to public procurement rules, To enable effective enforcement, the anti-cartel laws can be put in place relatively quickly. Some project- may need to be supplemented with reforms to the laws level preventive measures, such as retention of inde- of evidence. Until recently, most developed country pendent watchdogs or strict scrutiny of procurement courts required direct evidence of an agreement to prove of�cials’ �nances, may take more time. Over the longer the existence of a cartel, a demanding standard often term the goal should be to build effective institutions interpreted to require testimony from one of the partici- to enforce anti-cartel laws and manage the nation’s pants. Contrary to that thinking, as Annex 1 explains, road network. But again, none of the measures de- once American and European courts had heard more scribed are meant to be adopted without close analysis cartel cases, they became more comfortable relying on of market conditions, the strength of national institu- indirect or circumstantial evidence. With what is now tions, the degree of political commitment to reform, known about the harm cartels cause (Transparency and other country-level factors. International 2009), developing country courts should Curbing Fraud, Corruption, and Collusion in the Roads Sector ensure that they do not make it too dif�cult for their case the winning bid was virtually identical to the esti- enforcement agencies to prove the existence of a cartel. mate—an almost certain sign of collusion. Because transparency in public procurement can fa- Create incentives for the exposure of cilitate collusion, agencies that enforce the competi- cartels tion laws in both developed and developing countries caution procurement staff to consider carefully what Even if courts accept circumstantial evidence, nothing information about a tender to release (Government of provides surer proof of a cartel than the testimony of a the Netherlands 2010; Government of El Salvador 2010; participant or witness to a bid-rigging scheme. Those Government of Brazil 2008; Government of the United with information about a bid-rigging ring should be Kingdom 2004, Government of France 2003, 2000). The encouraged to come forward. To do so, governments dilemma for policymakers is that the more they try to should consider granting immunity to witnesses willing reduce the risk of corruption through greater transpar- to provide credible evidence of a cartel. Consideration ency, the greater the risk of collusion. Because carteliza- also may be given to offering whistleblowers rewards tion is so prevalent in public tenders of all kinds, many commensurate with the savings realized from the break- OECD countries have revised their public tender rules up of a cartel. Allowing whistleblowers to share in the to reduce transparency in several respects. A list of the recovery can provide a powerful incentive for coming reforms different OECD countries have introduced is forward (Depoorter and De Mot 2005). contained in Annex 2. Members of cartels should also be given an incentive Policymakers in developing countries may wish to consid- to disclose the names of the other participants. The er such revisions as well to ensure rules governing public OECD (2003) recommends granting immunity to the tenders strike the right balance between transparency on �rst �rm or individual to reveal the cartel’s existence the one hand and the risk of collusion on the other. While and the World Bank itself encourages contractors to different economic conditions and different institutional reveal previous misconduct on World Bank-�nanced settings make it unlikely that any will adopt the OECD re- contracts. Recent research shows that such leniency form list wholesale, the accumulating evidence shows that programs not only are effective in revealing the exis- some changes are effective in a wide array of institutional tence of cartels but can also discourage their formation and economic settings. Following are some examples of (Miller 2009). procurement process changes to consider in appropriate situations. (Box 5 illustrates how changes to the procure- ment process helped combat collusion). Revise tendering rules To ensure fairness and reduce corruption in the pro- a. Bidder pre- and post-quali�cation curement process open, transparent procedures for the award of public contracts are recommended Road agencies understandably want to ensure that �rms (Transparency International 2006). The more transpar- bidding on a tender will have the �nancial strength ency, the more likely the contractor is fairly chosen and and technical capacity to perform the work if they win the less likely corruption will seep into the process. But the tender. Potential contractors are thus commonly disclosure of certain kinds of information may also required to “prequalify,� that is, to document their �- increase the risk that �rms will �x prices (Anderson, nancial and technical ability to execute the contract in Kovacic, and Müller 2010; OECD 2008b; Government of the event they win the bid. Indeed, for all major civil the United Kingdom 2004). The data in Box 4 provides works contracts, the World Bank’s Standard Bidding an illustration; cost estimates on 46 separate contracts Documents provide that only “exceptionally� and “with were publicly disclosed prior to tendering, and in every previous approval of the World Bank� can a post-award 20 Comba ting Collusion, Fra ud, a nd Cor r upt i o n BOX 4 Publishing Cost Estimates: the Trade off Between Transparency & Collusion In the name of transparency, many countries publish their engineers’ estimates of the cost of building a road and in its loan agreements the World Bank sometimes requires publication. When the market is competitive, publishing the estimates can produce lower bids (De Silva et al. 2008). Publishing the estimates also ensures that all bidders are on an equal footing, for companies with close ties to the roads authority often obtain the estimates “under the table.� Set against these bene�ts is the risk that publication of the estimate will facilitate collusion. When �rms are negotiating an agreement on a collusive price, cost estimate provides a target or focal point for their agreement (Knittel and Stango 2003). This effect is illustrated in an INT comparison of the estimated price against the winning bid on 46 contracts for road construction and repair let during 2009 and 2010 under a Bank-�nanced project in an Eastern European country. The chart below plots the differences in millions of U.S. dollars between the two. The red line is the estimate; the blue line, almost invisible because it tracks the red one so closely, is the winning bid. This degree of correspondence is unimaginable in the absence of collusion. Cost Estimates v. Winning Bids 5 4 3 2 1 0 1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 review of the winner’s capabilities be substituted for requirements on simple goods and small works in its prequali�cation (World Bank 2010b, v). At the same 2001 report on Indonesia’s procurement policies (World time, prequali�cation requirements can discourage Bank 2001, 20). A more recent review of World Bank- some �rms from bidding, and the fewer �rms that bid, funded roads projects in Africa recommended expand- the higher the winning bid (Estache and Iimi 2008; ing post-quali�cation to larger contracts (Alexeeva, Froeb and Shor 2005; Brannman and Klein 1992). The Padam, and Queiroz 2008, 41). Post-quali�cation was OECD (2008a) thus recommends that prequali�cation introduced into the Bali Urban Infrastructure Project in conditions be carefully drawn to ensure that quali�ed Indonesia and is being used in the second phase of the �rms are not excluded from the competition. National Roads Improvement and Management Project in the Philippines and the Northern Corridor Transport Policymakers may want to consider in at least some Improvement Project in Kenya. Post-quali�cation in- cases eliminating prequali�cation all together in fa- creased the number of bidders on contracts in the Bali vor of a post-quali�cation review of the winner’s project, and the early results from Kenya are promising. quali�cations. The World Bank recommended that On all three Kenya tenders for which post-quali�cation Indonesian of�cials consider scrapping prequali�cation was used, the tenders attracted three or four quali�ed 21 Curbing Fraud, Corruption, and Collusion in the Roads Sector BOX 5 Combating Collusion by Changing the Procurement Process: The Bank’s Experience with the Bali Urban Infrastructure Project World Bank staff became suspicious when only three bids were submitted for one of the �rst contracts on the Bali Urban Infrastructure Project. Suspicions were heightened when, despite wide variations in labor and materials prices on the bidders’ bills of quantity, the prices submitted by all three were within 0.02 percent of the engineer’s estimate. When additional investigation con�rmed the existence of a bid-rigging cartel, the Bank made a number of changes to the procurement process to increase competition: Procurement notices were widely publicized in both national and provincial papers in prominent place and in large typefaces. Local authorities’ attempts to limit eligible bidders to local �rms were rebuffed. Bidders’ quali�cations were evaluated after, rather than before, the tender. Mandatory participation in pre-bid meetings, which had given colluders an opportunity to agree on prices and intimidate �rms not part of the ring, was ended. A complaint handling mechanism was introduced that allowed contractors and community members to anonymously report fraud, collusion, corruption, and intimidation. The impact of the changes was dramatic. As the table below shows, bids dropped from amounts virtually identical to the engineer’s estimate to amounts 35–40 percent less. Overall, the project team estimated savings of 15–30 percent on contracts let post-changes. Bids for $50,000 Contract: Best Three Bids as Percentage of Engineer’s Estimate Original Post-changes 98.9% 58.0% 99.7% 67.6% 100.0% 68.0% bids, more than the average when prequali�cation was by discouraging participation from small �rms that can required. More tellingly, the winning bids were below build one or two 50 km segments, but lack the experi- the engineer’s estimates, rare in Kenyan road tenders. ence or �nancial strength to build a 500 km road. One way to balance the competing interests is to tender the larger 500 km project as ten 50 km contracts, but allow b. Bid package design larger �rms to combine segments in their bids and even submit a single bid for the entire road. Knowing that Procurement of�cials often have signi�cant discretion smaller companies are competing on shorter segments, when deciding how to let a road construction project. the large �rm will have an incentive to “sharpen its pen- A project to build a 500 km road might be tendered as a cil�— that is, cut its price —to win the contract. At the single contract or divided into two contracts of 250 km same time, knowing that large international �rms can each or ten contracts of 50 km each. Different packages bid on a package can deter local �rms from rigging bids have different competitive effects. While larger pack- among themselves. ages encourage interest from international �rms and are subject to the more rigorous international competitive Besides allowing �rms to bid on one or more prede- bidding procedure, they also can reduce competition termined segments, the tender might also allow them 22 Comba ting Collusion, Fra ud, a nd Cor r upt i o n to offer to build segments of their choosing. As is Require independent bid certi�cates sometimes done in World Bank-funded projects, bid- ders could submit a bid on the condition that the total Successful prosecution of a road construction cartel re- award will not exceed a speci�ed amount. The bids on quires showing that members actually agreed to rig bids the various components of the project would be opened on a tender. Although, as Annex 1 explains, many courts sequentially. Once a �rm’s speci�ed limit is reached, its now accept circumstantial evidence of collusion, prov- bids would not be considered on the remaining compo- ing collusion can still be dif�cult and time-consuming. nents. Sequential bidding provides incentives for �rms By contrast, it is relatively easy to show that �rms traded to bid on more projects without worrying about taking price lists, shared cost data, or exchanged information on more work than they can handle (Allen, Culkins, and about the bids they intended to submit, practices that Mills 1988). Mixing up the “menu� of contract offers in the U.S. Supreme Court has held are anticompeti- these ways makes it harder for �rms to agree beforehand tive (United States v. Container Corporation, 393 U.S. on who will win what. 333, 337 (1969)) and that the European Commission has recently said should be considered a restriction of competition (European Commission 2010). Thus, c. Pre-bid meetings and subcontracting one approach to easing a prosecutor’s burden is to (a) require �rms to submit a certi�cate that they did not Pre-tender meetings should, whenever practical, be communicate with one another and (b) make falsi�ca- limited to one �rm at a time. As the author of the �rst tion of the certi�cate a serious crime. To prove a viola- economics textbook warned, “People of the same trade tion, all the prosecution would then have to do is show seldom meet together, even for merriment and diver- that �rms communicated. Box 6 describes the elements sion, but the conversation ends in a conspiracy against of such a certi�cate; model certi�cates are available the public or in some contrivance to raise prices� (Smith in English (Government of the United States 2007), 1937 [1776], 128). While one-on-one meetings increase French (Government of Canada 2010), and Spanish the risk that a procurement of�cial will provide a favored (Government of El Salvador 2010). �rm with con�dential information or otherwise tilt the procurement process, safeguards can be introduced to minimize this risk. An outsider can attend, or video re- Retain an independent procurement cordings or transcripts can be made and circulated. evaluator Subcontracting can also facilitate collusion, for it can INT has found evidence in some cases that those re- be a way of dividing the pro�ts realized from bid rig- sponsible for policing the tendering process condoned, ging. Testimony in a criminal prosecution of collusion in or even encouraged, collusion. Where this risk is present, roads contracting in Oklahoma revealed such a scheme. introduction of an outsider into the tender evaluation Competitors of the Boce Company allowed it to win a process can reduce that risk. For example, in the second tender “without having to �ght,� and in exchange Boce phase of the Philippine National Roads Improvement and agreed to subcontract all the work in one region to a com- Management Project an independent procurement eval- petitor (United States v. Metropolitan Enterprises, Inc., 728 uator has been hired to work alongside the Department F.2d 444 (10th Cir. 1984)). Countries with a large number of Public Works and Highways procurement of�cials. of capable �rms should consider banning subcontract- The terms of reference provide that the evaluator must ing among competitors altogether or, as the January 2011 develop “speci�c systems to identify or detect indica- European Commission Green Paper on procurement tors of corrupt practices in the bids, including collu- reform suggests, barring subcontracting by �rms which sion, price-rigging, fraud, obstruction or coercion.� participated in the tender (EC 2011). If subcontracting is (Government of the Philippines 2007a) Although the permitted, data should be kept and analyzed periodically evaluator cannot veto the highway department’s deci- for any signs that suggest collusive arrangements. sions, he must regularly report his �ndings to the World 23 Curbing Fraud, Corruption, and Collusion in the Roads Sector BOX 6 Certi�cate of Independent Price Determination A Certi�cate of Independent Price Determination requires the bidder to warrant that: The prices in the bid have been arrived at independently, without any consultation, communication, or agreement with any other bidder or competitor relating to (a) prices, (b) the intention to submit a bid, or (c) the methods or factors used to calculate the prices offered. The prices in the bid have not been and will not be disclosed by the bidder, directly or indirectly, to any other bidder or competitor before bid opening (in the case of a sealed solicitation) or contract award (in the case of a negotiated solicitation), unless otherwise required by law. No attempt has been made or will be made by the bidder to induce any other �rm to submit or not to submit an offer for the purpose of restricting competition. The statement is made under the provisions of a law that imposes stiff penalties for lying in the statement �led. The bidder can be prosecuted, if the only evidence is that it disclosed bid prices to its competitors or attempted to convince its competitors to rig bids. The evidence needed to prove a violation of the Certi�cate of Independent Price Determination is signi�cantly less than that needed to prove an illegal agreement. Source: Government of the United States (2007). Bank and other partner organizations. Lessons from as well, or in lieu of, its roads agency counterpart: the the Philippine experience will be used to guide future head of the agency, the transport minister, or the chief arrangements. prosecutor or head of an anticorruption agency. With World Bank-supported contracts, it should include the World Bank itself. A suspicion of corrupt or fraudulent B. Measures to Reduce Fraud and activities should be highlighted in a covering note or Corruption executive summary. Strengthen the engineer Policymakers should also examine the utility of (a) creating incentives for the engineer to expose fraud As the discussion above showed, the engineer in a FIDIC and corruption, (b) penalizing engineers that fail works contract is the �rst line of defense against fraud to detect either, and (c) severely sanctioning those and corruption. Whenever a roads agency uses this or who participate in fraudulent or corrupt schemes. any contract that grants similar powers to the engineer, Sanctions could range from repayment of fees to �nes the engineer’s role in combating corruption should be and stiff prison terms. Because the engineer enters into made explicit and measures taken to help him discharge a relationship of trust with the borrower, which he his responsibility. Thus, for example, the common prac- betrays if he participates in corruption, harsher penal- tice of hiring the engineer after the contractor has begun ties than those levied on other participants may be work, which immediately puts the engineer in a “catch- warranted. up� mode, should be discouraged. How engineering services are procured may also merit Road construction contracts should expressly require review. Should price be the only factor as it is often so the engineer to immediately report any activity that now? Or should selection follow a two-step process that suggests fraud or corruption. Because line managers in focuses on “quality� �rst (including past general per- roads authorities are sometimes participants in corrupt formance, and success in deterring or rooting out fraud schemes, the engineer should send the report to others and corruption in particular), and price second, for 24 Comba ting Collusion, Fra ud, a nd Cor r upt i o n those who have met the quality requirements. Would it would increase the contract price by 15 percent or more, make sense to adopt a point system that factors in qual- and (f) conduct a comprehensive completion review of ity and price? What criteria could help ensure an objec- all civil works and of the highway department’s supervi- tive evaluation of the engineer’s quality? sion of each contract (Government of the Philippines 2007b). Different ways of determining the engineer’s fees should also be explored to ensure that all incentives, Even the threat of a technical audit can reduce corrup- including the fee structure, are consistent with the tion. In a �eld experiment conducted as part of the engineer’s quasi-�duciary role. For example, where Bank’s Kecamatan Development Program in Indonesia, the risk of corruption during contract performance one group of villages participating in a nationwide road is particularly high, would it make sense to agree to a construction program was told beforehand that their combination of a �xed fee for basic work and an hourly projects would be audited and all projects were subse- rate for certain kinds of tests and inspections relating quently audited. In a second group, audits were neither to integrity risks? What safeguards could be introduced threatened nor conducted. The difference between into such arrangements to avoid unnecessary testing amounts claimed on the contractors’ invoices and the and veri�cation procedures to simply increase the fee? amounts actually spent was on average 8 percent less in These and similar issues should be examined with a those villages that were subject to audit than in those view of strengthening the engineer’s role in helping de- villages that were not (Olken 2007). tect and address fraud and corruption during contract implementation. The challenge when hiring a technical auditor is ensur- ing that this second guardian remains a faithful guard- ian, serving the interests of the borrower rather than Hire a technical auditor being drawn into a scheme to cheat it. While profes- sional norms and the auditor’s character provide one Where there is a risk that the engineer will be drawn guarantee of faithfulness, creating a powerful economic into a circle of corrupt actors, it can be minimized or incentive for the auditor to remain honest provides eliminated by retaining another overseer to “guard the more assurance. This can be accomplished by fostering guardian� (Hurwicz 2007). The classic guardian of the economic conditions that handsomely reward honesty guardian in a roads project is a technical auditor. Unlike and severely punish its absence. Auditors who perform a a �nancial auditor, whose review is con�ned to the job well should be paid well and those who succumb to �nancial statements and supporting documentation, corruption �ned and imprisoned. a technical auditor periodically inspects the project to ascertain that the materials and labor provided “were More important than changing the cost-bene�t cal- appropriate to their intended purpose and were deliv- culus for a single job is creating a market in which ered in the quantity, quality, and location or disposi- those who perform well will enjoy a steady stream of tion speci�ed� (Patterson and Chaudhuri 2007, 181). remunerative work and those who don’t, won’t—in A technical auditor will be hired for the second phase short, a market where the discounted present value of the Philippine National Roads Improvement and of future revenues exceeds the immediate pro�t real- Management Project. The terms of reference provide, ized from the one-time acceptance of a bribe. In many among other things, that the auditor will (a) investi- markets, from the long-distance trade in commodities gate the quantity and quality of site surveys at com- in the Middle Ages to the sale of consumer appliances pleted works, (b) review the audit support provided to in modern times, the bene�ts of a good reputation the project, (c) review all supervision reports on the and the harm from a poor one have deterred brib- contracts carried out under the Bank’s international ery and other types of short-term, opportunistic competitive bidding rules, (d) test on-site the quality of behavior (Greif 2006; Klein 1997). The key in every contractors’ materials, (e) audit all change orders that instance is seeing that information about an auditor’s 25 Curbing Fraud, Corruption, and Collusion in the Roads Sector performance is widely circulated to future employers, a provided by the Australian Aid Agency. It is co-�nancing role the World Bank could assume as part of its knowl- the project with a $1.1 million grant supporting the edge-sharing work. It is also important that employers Network and other civil society monitors (World Bank consider reputation when hiring an auditor—an ap- 2008b, 68). proach that may, as it did in the U.S., require revisions to public tender rules (Kellman 2002). Develop accurate cost estimates Engage civil society monitors Critical to evaluating bids are reliable, independent, current estimates of the projected cost of the contract. Nongovernmental organizations (NGOs) or community Cost estimates should be prepared using the same level advocacy groups can be watchdogs, too. For civil society of detail the industry uses, and should reflect what the groups the challenge is less remaining independent of procuring authority is willing to pay for performance the corrupt scheme than having the resources and ex- of the contemplated work. The U.S. Department of pertise to effectively monitor a roads project. Although Transportation advises that estimates should be within the Indonesian research found that community-level 10 percent of the low bid for at least half of the projects. monitoring did not prevent corruption in roads projects “If this degree of accuracy is not being achieved over . (Olken 2007), Philippine NGOs have enjoyed a good . . one year, con�dence in the engineer’s estimates may deal of success in monitoring procurements and un- decline� (Government of the United States 2004). Where covering corruption in the health and education sectors there is a history of cartelization in the roads sector, care (Ramkumar 2008, 52–61). Indeed, in one case a local must be taken to ensure that current estimates do not group, albeit one whose membership included a civil reflect past instances of overpricing (Feinstein, Block, engineer, discovered the use of substandard cement in a and Nold 1985). provincial road construction project (Cadapan-Antonio 2006–07, 656–657). Check the wealth of key procurement An important complement to engaging civil society agency of�cials monitors is requiring the publication of the contract and related documentation such as audit reports, More than 100 World Bank client countries require fund disbursement schedules, and project perfor- designated civil servants and elected of�cials to peri- mance reports. As Kenny (2010) argues, publication odically submit statements showing their income and of such information raises the threat of scrutiny by assets (Messick 2009). These statements can be used the media and civil society, thus reducing the likeli- by enforcement agency personnel or civil society to hood of bribery and other corrupt payments. The compare what is reported about the values of homes, Construction Sector Transparency Initiative, a joint the number of cars owned, and so forth, with what DFID-World Bank initiative, �nanced a working pa- real estate and automobile registries show and what per that itemizes the information that should be dis- visual inspections and interviews with neighbors and closed (CoST 2009). friends reveal. The Philippine Center for Investigative Journalism conducted such “lifestyle checks� on mid- A number of well-resourced NGOs are monitoring level of�cials of the country’s tax service, �nding that the second phase of the Philippine National Roads many had signi�cantly understated the value of their Improvement and Management Project and one, the homes or lied about the number of cars they owned Transparency and Accountability Network, has pro- (Batino 2003), and an Indonesian magazine recently duced a sophisticated guide to monitoring procurement reported that the country’s Auditor General had accu- contracts for roads and other civil works (Cerna 2009). mulated enormous undeclared wealth during a career Funds to support this monitoring arrangement are in the tax agency (Tempo 2010). Countries without 26 Comba ting Collusion, Fra ud, a nd Cor r upt i o n income and asset declaration laws should enact them; Strengthen competition law enforcement those with such laws should ensure that key procure- ment agency personnel are covered and the laws vigor- Road construction and maintenance is just one of many ously enforced. industries in which law enforcement authorities have discovered cartels. Over 300 cartels in industries as diverse as ready-mix concrete, vitamins, �ne art, snow C. Longer-Term Capacity-Building removal, and intravenous solutions have been unearthed Measures in developed countries (OECD 2003). Like roads sector cartels, these others have also caused enormous damage Modernize the roads sector agency by raising prices, furnishing shoddy goods, and corrupt- ing government of�cials. To detect and prosecute cartels, While independent watchdogs, civil society monitor- the OECD (2003) recommends that law enforcement ing, and the other short-term measures discussed in this authorities have in place the following powers: report offer ways to defeat the collusion and fraudulent schemes INT found most prevalent in Bank-funded The power to grant leniency to cartel participants roads projects, no solution will succeed in the long run willing to give evidence against other members without a modern, professional, and capable entity to The ability to conduct unannounced visits to mem- manage a nation’s road network. To be effective, the bers’ of�ces to review documents and electronic road agency must have appropriate powers, skills and evidence resources, and operate within an effective framework The authority to take oral testimony from members’ of accountability, internal controls, and performance employees for use in criminal and civil proceedings measurement. It must have the authority and capacity The right to use listening devices and other special to carry out its regulatory, planning, �nance and invest- investigative measures to collect information and ment, coordination, and management responsibilities; evidence it also must have the technologies, equipment and ef- �cient and transparent processes and procedures to Building an institution that can wield these powers re- enable ef�cient use of resources allocated to the roads sponsibly and effectively takes time, but the damage car- sector, including effective management of risks. Helping tels in any sector do to the economy and polity of devel- a country build such an entity begins with a candid oping nations argues for giving priority to strengthening assessment of the authority’s weaknesses and identi�ca- the entities that enforce competition law. A number of tion of measures needed to address them. The second organizations provide technical assistance to competi- phase of the Philippine National Road Improvement tion law authorities; the United Nations Conference and Management Project is a good example of such ef- on Trade and Development sponsors peer reviews of forts. As reflected in the Project Appraisal Document, enforcement efforts and hosts an annual meeting of the loan budgets $6 million to support strengthening competition law agencies from developing nations measures for the Department of Public Works and (UNCTAD 2010) and the OECD, the World Bank, and Highways, including the introduction of new busi- other multilateral and bilateral agencies also furnish ness processes, a more robust internal audit staff, and various forms of technical assistance. improved �nancial management (World Bank 2008b, 52–56). To achieve sustainability and move to scale in helping client countries build and maintain sustain- D. Experimental Measures able road networks with minimum losses to fraud, corruption and collusion, it is essential that the World Where the risks of fraud, corruption and collusion are Bank and its partners support the countries’ efforts in particularly high, traditional reform measures, such building strong, effective and accountable roads sector as training staff, modernizing facilities, and upgrading institutions. information and communications technology, will not 27 Curbing Fraud, Corruption, and Collusion in the Roads Sector by themselves be effective (World Bank 2008c, 32). To agency staff is trained on price estimation techniques address the issues, innovative, creative, and less con- and bid variance analysis, and (d) a system is in place to ventional steps may be required. This section describes monitor and compare bid prices against the estimate. three such measures for consideration in high-risk situations. Use competitive negotiation Impose ceiling on bids Where roads cartels are particularly entrenched, poli- cymakers might experiment with a form of competi- The Philippines is experimenting with a cap on bids; for tive negotiation. The procuring agency chooses a �rm each contract the procuring agency calculates a maxi- it believes quali�ed to build the road in question and mum price. With roads and other infrastructure con- negotiates a price. It could be lump sum, cost-plus, or tracts, implementing regulations specify in detail how some combination. If the �rm is not interested or is un- the maximum price is to be calculated (Government of willing to accept the price offered, the agency goes on to the Philippines 2003). Bids over the budget are rejected another �rm. There is a risk of corruption in the form and if, after two rounds of bidding, no company has of favoring one contractor over another or negotiating submitted a price equal to or less than the maximum, too high a price. To help address such risks, competitive the agency then “directly negotiates a contract with negotiation should: (a) be limited to clearly de�ned situ- a technically, legally and �nancially capable supplier, ations, (b) subject to appropriate safeguards, such as a contractor or consultant.� (Government Procurement prior short-listing of �rms based on speci�c criteria and Reform Act, Republic Act 9184, § 43(e)) (c) adopt clear and transparent objectives against which to conduct negotiations with each short-listed �rm to There are risks to this approach. Absent genuine compe- obtain the best proposal. tition, the ceiling price almost certainly puts a floor on the bid price, something that appears to have happened in Japan when road cartels were in existence (McMillan Contract out procurement 1991). In addition, cost estimates on road contracts can be unreliable because market conditions change, engi- When state capacity is especially weak and the involve- neers can make errors in the estimation process, and ment of high-level political of�cials in procurement there is always the possibility that corruption will creep widespread, a foreign company can be retained to ad- into the process. Moreover, if no company bids at or minister the entire procurement process from project below the ceiling price and the procuring agency decides identi�cation to design, tendering, and contract man- to go forward with the procurement, it will have to enter agement. This is a broader application of the indepen- into direct negotiations with a �rm, an action which, as dent watchdog approach described above, and the same explained below, creates its own set of problems. concerns about ensuring that the procurement agent is genuinely independent and that reputation mechanisms An analysis of the early experience with the Philippines’ are in place apply—only on a much larger scale. use of bid ceilings is expected to appear in 2011. While in the meantime the World Bank is not permitting the Hiring an independent procurement agent is not with- Philippines to use ceiling prices on Bank-funded con- out its problems, however. Newly hired agents face steep tracts let under its international competitive bidding learning curves that often delay procurements and na- procedures, the Bank has agreed to permit the practice on tional agencies displaced by the agents can lose interest contracts using national competitive bidding procedures in, and ownership of, the projects handled by the agents so long as four conditions are met: (a) the bid documents (Ali and Moss 2010). The biggest drawback with inde- are freely and easily accessible, (b) the ceiling price is pendent procurement agents is that they can undercut based on the engineering cost estimate, (c) the procuring efforts to build local procurement capacity. 28 Comba ting Collusion, Fra ud, a nd Cor r upt i o n BOX 7 Using Competitive Negotiation to Circumvent a Cartel: the US experience The American military used competitive negotiation to circumvent a cartel in the Republic of Korea in the 1970s. The military regularly tendered for goods and services from Korean �rms and its rules required that all but the very smallest contracts be let competitively. Procurement of�cers encountered the same problems evident in the roads sector today, collusion coupled with corruption underpinned by a culture that eschewed competition. At �rst procurement of�cials sought to overcome these problems through stricter enforcement of the Republic’s competition laws. Cartels were in�ltrated and evidence of collusion was turned over to South Korean prosecutors. Local procurement staff caught furthering cartel activities were prosecuted. Although a few cartels were broken up, the results were disappointing; cartels regrouped and worse, in several cases, informants were murdered. Frustrated with the lack of progress, senior procurement staff turned to negotiated procurements, similar to the two-stage tendering often used in tight construction markets (Davis and Dornan 2008). The procurement of�cer chose a �rm he believed capable of doing the work and invited it in to negotiate a deal. A �xed price might be negotiated or sometimes, as in the case of two-stage tendering, the �rm would work on a cost- plus basis. The results exceeded expectations; prices were 10–15 percent less than those under “competitive� bids, and the kickbacks and violence associated with “competitive� tenders disappeared. The key to the military’s success was its “clean,� independent procurement personnel. Procurement staff stood outside the network of collusion and corruption that authorities sought to defeat, administering the rules evenhandedly and with integrity, thus acting as a de facto independent procurement agent. Source: Martin (1983). World Bank experience with independent agents in independent agent to run the customs agency and to Southern Sudan and Cambodia shows two pitfalls train national staff. to avoid when retaining an independent agent. In Southern Sudan the agent did not �eld suf�cient staff to provide the training required (Price Waterhouse E. Issues for Consideration by Bank Coopers 2008) while in Cambodia the agent’s terms Operations Staff of reference omitted capacity building (Ali and Moss 2010). Bank experience with independent agents in The four objectives underlying the World Bank procure- customs in Angola and Mozambique, however, illus- ment policy—transparency, fair treatment, capacity trate the advantages when these problems are avoided. building, and competition—are sometimes in tension. In both countries the customs function was contracted Policies that advance transparency, fair treatment, or out, with a deadline for turning responsibility back capacity building can undercut competition; on the over to the government. Corruption was sharply other hand, those that further competition may inad- reduced in the short run and over the long term na- vertently compromise one or more of the other three. As tional capacity was built (Mitchener and Maurer 2010; the World Bank reviews its procurement policy, it would Mwangi 2004). Common to both efforts was not only be important to consider trade-offs between these a clear understanding on the deadline for handing objectives, adapted to the speci�c country’s risks and back responsibility, but also suf�cient resources for the circumstances, including the state of competition, the 29 Curbing Fraud, Corruption, and Collusion in the Roads Sector capacity and performance of the responsible agencies, risk of corruption. Limiting the pre-bid conferences to such as highway authorities, the effectiveness of anticor- one �rm at a time, while requiring that each meeting be ruption and competition laws, and the track record of attended by an independent party and include a video- the prevention and enforcement authorities. The follow- recording or a meeting transcript, is a good example of ing discussion covers some issues that emerged in the an alternative that can be considered in high-collusion context of this review that may merit consideration as environments. Having an independent evaluator certify part of the reform. that the lowest price was chosen is another. Trade-offs between Transparency and Subcontracting as a facilitator of Collusion capacity-building and collusion As discussed above, while open and transparent pro- Allowing less experienced local companies to subcon- cedures for the award of public contracts help ensure tract with experienced international �rms gives them the fairness and reduce corruption in the procurement chance to learn new techniques and build domestic ca- process, disclosure of certain kinds of information may pacity. However, when losing bidders are permitted to be- also facilitate cartelization and price-�xing. (Anderson, come subcontractors to the winning �rm, subcontracting Kovacic, and Müller 2010; OECD 2008b). The most can be a way colluders pay one another off for sticking clear-cut example is the requirement that the name of together. As suggested above, where the risks of collusion each bidder and the amount of the bid be publicly dis- are high, consideration should be given to prohibiting closed. Publishing all bids received both eliminates the subcontracting with losing bidders or at least monitoring risk that a corrupt of�cial will accept a high-priced or subcontracting patterns to identify collusion risks. nonconforming bid and reassures �rms submitting bids they are being treating equally. But as Stigler (1964) ex- plains in a classic article on collusion, cartels are under Customizing measures to address fraud constant threat of breakdown from secret price cuts by and corruption in civil works contracts a member seeking to expand business at the expense of the other members. How can colluders protect against Faced with the risk of fraud and corruption in World an outbreak of competitive pricing? Bank-funded civil works contracts, the Bank’s procure- ment specialists have developed various mitigation “The system of sealed bids, publicly opened measures, summarized in Box 8. The effectiveness and with full identi�cation of each bidder’s price replicability of these measures merit further evaluation and speci�cations, is the ideal instrument and discussion by the World Bank’s transport sector and for the detection of price-cutting. There ex- procurement specialists as part of the reform process. ists no alternative method of secretly cutting prices (bribery of purchasing agents aside). Our . . . prediction, then, is that collusion will Developing expertise on cost estimating always be more effective against buyers who and detecting collusive bidding report correctly and fully the prices tendered to them� (48). To prevent collusion and corruption in roads projects, accurate cost estimates and the ability to spot collusion Where the risk of collusion is particularly high, thought are essential. The World Bank should consider becom- should be given to achieving the goals of transparency ing a center of excellence for both, creating a cadre of in alternative ways. Such alternatives should be designed experts on each topic who can follow developments in a manner that maintains public con�dence in gov- in the �eld, train country counterparts, and step in ernment institutions and processes and addresses the when country capacity is weak. With its Road Costs 30 Comba ting Collusion, Fra ud, a nd Cor r upt i o n Knowledge System, a database of historical information of contract management originated in 19th century on roadwork costs per kilometer, the World Bank has England and in the 1950s spread to developing coun- taken the �rst step with cost estimating. An easy �rst tries where it seemed well suited to their needs (Lyon step for identifying collusive bidding would be to begin 1995). The uncertainties in building public works analyzing bids submitted on projects. A number of tests in the then largely unknown settings in developing have been developed to determine whether bids were countries created signi�cant risks, ones that could not arrived at independently and they can be programmed be speci�ed, let alone allocated by detailed contract using standard statistical packages (Bajari and Ye 2003; language. Much had to be left to work through on the Porter and Zona 1993). The World Bank should ensure ground as the project progressed, creating the possibil- that �rms bidding on Bank-funded projects submit the ity that the construction of roads and other critically data necessary to conduct these tests in machine-read- needed infrastructure would be stalled as the contrac- able form. The investment required to build on these tor and the government squabbled over who was re- �rst steps would be minimal and the potential payoffs— sponsible for what unforeseeable event. Exacerbating with a projected lending program of $7–8 billion for the tension, the builder was inevitably from a devel- FY11—enormous. oped country and possessed a high degree of technical knowledge, while the developing country client had little. A strong, technically competent engineer, inde- Reevaluate current contract management pendent of both (and, importantly, with the power form to mediate their disputes and so keep the project on track) provided a workable solution. Like the FIDIC model on which it is modeled, the World Bank’s works contract form makes the engineer As developing countries gained experience and exper- the central �gure in contract administration. This form tise with infrastructure construction, however, they BOX 8 Reducing Fraud and Corruption in Civil Works 1. Ensure accurate cost estimates (quantities and, more importantly, unit rates) to exclude the up-front inflated padding that serves as a reference cover to hide the high bid prices downstream. 2. Be sure the bill of quantity is correct to minimize variations in the downstream implementation of unit rate/ad-measurement contracts. 3. Encourage, where feasible, a lump sum output-based approach for tendering and contract implementation to reduce the possibility of downstream quantity variations during contract implementation. Provide concurrent training of borrower staff and private sector contractors in the application of the lump-sum output-based approach in the procurement and implementation of civil works contracts. 4. Include contract provisions that provide an incentive for contractors to deliver cost savings at the end of the completed contract (�nal completed contract price vs. initial contract award price) — for example, bonuses or a percentage of the cost savings. 5. Closely supervise construction supervision, preferably through external international engineering �rms, along with independent technical audits and a higher level of quality checks by the Bank during project supervision. 6. During project supervision, the Bank or its appointed auditors should randomly check the contractor’s and supervisory consultant’s �nancial records, applying in practice the provision in the Bank’s Procurement Guidelines, standard bidding documents, and standard forms of contract, which allows the Bank to undertake such audits. Source: World Bank procurement staff. 31 Curbing Fraud, Corruption, and Collusion in the Roads Sector saw less need for a powerful engineer; in response, the contract for the same length of time (Government of World Bank and other international �nancial institu- Tanzania 2004). These steps will help to create a market tions have progressively modi�ed the FIDIC contract where only honest engineers prosper. to strengthen the government’s control of the engineer. Whereas the engineer once independently determined whether a contractor’s invoice was in order and there- Increase contingent of professional fore should be paid, that is no longer the case. Likewise, World Bank staff with road engineering the current version of the FIDIC contract used by the expertise World Bank gives the government the power to replace the engineer at any time with no real input from the While World Bank supervision efforts now stress �nan- contractor. cial and �duciary controls, despite their usefulness these efforts do little to detect malpractice and the practical While the move away from a powerful, independent en- impacts of corruption in the realization of the works. gineer was prompted by many factors, project integrity Since third party technical audit solution will remain does not appear to have been one. With the growing expensive and impractical in many projects, a simpler recognition of the harm from fraud and corruption remedy to start tackling this issue is to strengthen the in road works, the development community should professional technical capacity of the Bank’s project reevaluate the way roads contracts are managed. Is a teams. This means maintaining a suf�cient number of weakened engineer overseen by a sometimes-corrupt seasoned road and highway engineers. At the design agency the best guarantor of project integrity? Are those stage these professionals can detect potential weaknesses forms of project management that assign the engineer’s or omissions and help make bidding documents more responsibilities to different entities more likely to reduce reliable with less room for interpretation or deliber- corruption? Should the engineer be more independent ate misconception. At the construction stage they will of government? Advances in the economic study of con- know where to look and what to probe when supervis- struction contracts (Chakravarty and MacLeod 2006) ing road construction or rehabilitation. Combined with and the accumulated experience from different forms the �duciary controls, this approach would provide a of construction contract management (e.g., Kluenker much-improved protection against corruption in proj- 2001) provide a wealth of information for considering ect implementation. such issues. Spend more on corruption prevention in Target enforcement on engineering �rms projects If the engineer’s role in project integrity is to be Combating corruption requires not only technical strengthened, then corrupt engineers must be severely skills, competence, and commitment, but also re- sanctioned. INT will do its part by targeting engineers sources. An example of good practice is the second in its investigations of misconduct in Bank-funded road phase of the Philippine National Roads Improvement and infrastructure projects. Whenever it is determined and Management Project, which earmarks $7.54 million that an engineer has been involved in corruption, that for anticorruption activities, of which $1.14 million is engineer should be debarred for a lengthy period and part of the Bank loan and $6.40 million a grant from borrower countries should be urged to cease hiring him the Australian Aid Agency. Although these amounts on non-Bank-�nanced projects. Tanzania already does may seem high, between the loan and its own funds, so. Its procurement law provides that any �rm that has the Government will spend $240 million for roads been debarred “by a foreign country, international or- through the project. Given the importance of corrup- ganization or other foreign institutions on grounds of tion prevention to the ultimate success and sustain- fraud or corruption� cannot compete for a government ability of the roads projects and programs, the World 32 Comba ting Collusion, Fra ud, a nd Cor r upt i o n Bank should review its supervision strategy for the roads implementation support; (b) establishing a trust fund sector that looks at various factors, including budget to �nance independent procurement oversight; and (c) and skills. In conducting such review, options to be con- grouping audits, review, and supervision work for mul- sidered may include: (a) reallocating resources towards tiple projects to achieve economies of scale. 33 VI Conclusion As this report has shown, fraud, collusion, and cor- A review of procurement policies to address areas ruption in roads projects wreak enormous damage on that may constrain borrower authorities, the World developing countries. Roads cost more to build than Bank, and its staff from taking appropriate action they should, do not last as long as they ought to, and An assessment of whether changes are needed in the the corruption proceeds can pollute a nation’s political current model for preparation and supervision of system. The aim of this report has been to help reduce roads projects and the relative roles of government these losses by sparking a dialogue among policymakers authorities, the engineering profession, World Bank and stakeholders inside and outside the World Bank on staff, and civil society representatives developing solutions to these problems. This dialogue An evaluation of experience to date with building should include the following elements: effective public works institutions in borrowing countries. A robust assessment of the impact and cost- effectiveness of different mitigation measures INT is ready to work with its operational colleagues in applied in different countries (including under the Sustainable Development Network, the regions, and World Bank-funded projects in Kenya, Indonesia, Operations Policy and Country Services and with gov- Philippines and others that include robust mitiga- ernment counterparts, the private sector and civil soci- tion measures) ety to advance this dialogue. References All Web sites were visited between June 15, 2010, and April at: http://www.oecd.org/dataoecd/12/1/44456320. 15, 2011. pdf. Batalla, Eric C. 2000. “De-institutionalizing Corruption Ahmed, Raisuddia, and Mahabub Hossain. 1990. in the Philippines: Identifying Strategic Developmental Impact of Rural Infrastructure Requirements for Reinventing Institutions,� in Bangladesh. IFPRI Research Report No. 83. Paper prepared for Conference Sponsored by the Washington, D.C.: International Food Policy KAS and the DLSU-YCEA, Makati, Philippines. Research Institute. Available at: http://books. Available at: http://unpan1.un.org/intradoc/groups/ google.com/books?hl=en&lr=&id=jLxjsR2OZr public/documents/apcity/unpan013117.pdf 8C&oi=fnd&pg=PA9&dq=Developmental+im Batino, Clarissa S. 2003. “4 BIR, BOC Of�cials Fail pact+of+rural+infrastructure+in+Bangladesh Lifestyle Checks: Mismatch in Salary, Wealth &ots=EMIZboQq7_&sig=ApXRxFYJIs1nP3Y7 Uncovered,� Philippine Daily Inquirer. July 21. NxJvyaMuB7M#v=onepage&q&f=false Bajari, Patrick, and Lixin Ye. 2003. “Deciding Between Alexeeva, Victoria, Gouthami Padam, and Cesar Competition and Collusion,� Review of Economics Queiroz. 2008. Monitoring Road Works Contracts and Statistics 85(4): 971–989. and Unit Costs for Enhanced Governance in Sub- Booth, David, and Frederick Golooba-Muteb. 2009. Saharan Africa. Washington, D.C.: World Bank, The Political Economy of Roads Reform in Uganda. Transport Sector Board. Available at: http:// London: Overseas Development Institute, siteresources.worldbank.org/INTTRANSPORT/ Working Paper 307. Available at: http://www.odi. Resources/336291–1227561426235/OBEC_ org.uk/resources/download/3829.pdf. BecasyOportunidades_BancoMundial.pdf Brannman, Lance E., and J. Douglass Klein. 1992. Ali, Ahsan, and Shaun Moss. 2010. “Use of Independent “The Effectiveness and Stability of Highway Bid- Procurement Agents: The Cambodian Experience.� rigging,� chapter 4 in David B. Audretsch and John Presentation to World Bank Fiduciary Forum, S. Siegfried, eds., Empirical Studies in Industrial March. Organization: Essays in Honor of Leonard W. Weiss. Allen, Gary R., Donald Culkin, and Cheryl Mills. 1988. Dordrecht, the Netherlands: Kluwer Academic Legal and Economic Aspects of Competitive Market Publishers. Behavior. Charlottesville, Virginia: Virginia Cadapan-Antonio, Myrish T. 2006–07. “Participation of Transportation Research Council. Available at: Civil Society in Public Procurement: Case Studies http://tug.cloverleaf.net/PDFFiles/virginia.pdf from the Philippines,� Public Contracts Law Anderson, Robert D., William E. Kovacic, and Anna Journal 36(4): 629–665. C. Müller. 2010. “Ensuring Integrity and Cerna, Toix. 2009. How to Monitor Public Procurement: Competition in Public Procurement Markets: A Procurement of Goods and General Services. Dual Challenge for Good Governance.� Available Manila: Transparency and Accountability Curbing Fraud, Corruption, and Collusion in the Roads Sector Network. Available at: http://www.tan.org.ph/ “The Impact of Public Information on Bidding index.php?option=com_docman&task=cat_ in Highway Procurement Auctions,� European view&gid=42&Itemid=81 Economic Review 52(1): 150–181. Chakravarty, Surajeet, and W. Bentley MacLeod. Estache, Antonio, and Atsushi Iimi. 2008. Procurement 2006. Construction Contracts (or: “How to Get Ef�ciency for Infrastructure Development and the Right Building at the Right Price?�). Bonn: Financial Needs. Washington, D.C.: World Bank, Forschungsinstitut zur Zukunft der Arbeit, Policy Research Working Paper 4662. Discussion Paper No. 2125. Available at: http://pa- European Commission. 2011. Green Paper on the pers.ssrn.com/sol3/papers.cfm?abstract_id=902820 Modernization of EU Public Procurement Policy: Connor, John M. 2009. Price-Fixing Overcharges, 2nd Towards a More Ef�cient European Procurement ed., West Lafayette, Indiana: Purdue University. Market. Brussels: COM (2011) 15. Available at: Available at: http://www.agecon.purdue.edu/ http://eur-lex.europa.eu/LexUriServ/LexUriServ. staff/connor/papers/PRICE%20FIXING_ do?uri=COM:2011:0015:FIN:EN:PDF OVERCHARGES_FULL_TEXT.pdf European Commission. 2010. Guidelines on the CoST [Construction Industry Transparency Initiative] Applicability of Article 101 of the Treaty on the 2009. The Working Paper Series: A Summary. Functioning of the European Union to Horizontal Available at: http://www.constructiontransparency. Cooperation Agreements. Brussels: Of�cial org/ResourceCentre/Publications/ Journal of the European Union, 2011/C 11 /01. Dabbah, Maher M. 2010. International and Comparative Available at: http://ec.europa.eu/competition/anti- Competition Law. Cambridge: Cambridge trust/legislation/horizontal.html University Press. European Commission. 2008. White Paper on Damages Davis, Cecily, and Peter Dornan. 2008. “The Rise and Actions for Breach of the EC Antitrust Rise of Two-Stage Tendering,� International Rules. Brussels: COM(2008) 165. Available at: Construction Law Review 25: 511–517. http://ec.europa.eu/competition/antitrust/actions- Deichmann, Uwe, Marianne Fay, Jun Koo, and Somik damages/documents.html V. Lall. 2002. Economic Structure, Productivity, Fan, Shenggan, P. Hazell, and S. Thorat. 1999. and Infrastructure Quality in Southern Mexico. Government Spending, Agricultural Growth and Policy Research Working Paper Series 2900. Poverty: An Analysis of Interlinkages in Rural India. Washington, D.C.: World Bank. Available at: IFPRI Research Report No. 110. Washington, D.C.: http://www-wds.worldbank.org/external/default/ International Food Policy Research Institute. WDSContentServer/IW3P/IB/2002/11/01/0000949 Fan, Shenggan, P. Hazell, and S. Thorat. 2000. 46_0210180511041/Rendered/PDF/multi0page.pdf “Government Spending, Agricultural Growth, Depoorter, Ben, and Jef De Mot. 2005. “Whistle and Poverty in Rural India. American Journal of Blowing,� Supreme Court Economic Review 13, Agricultural Economics 82 (4): 1038–1051. 2005; George Mason Law & Economics Research Fan, Shenggan, L. Zhang, and X. Zhang. 2002. Growth, Paper No. 04–56. Available at SSRN: http://ssrn. Inequality, and Poverty in Rural China: The Role com/abstract=622723 or doi:10.2139/ssrn.622723 of Public Investments. IFPRI Research Report No. Dercon, Stefan, and Pramila Krishnan. 1998. 125. Washington, D.C.: International Food Policy Changes in Poverty in Rural Ethiopia 1989– Research Institute. 1995: Measurement, Robustness Tests and Fan, Shenggan, S. Jitsuchon, and N. Methakunnavut. Decomposition. WPS/98–7. Oxford: Center 2004. The Importance of Public Investment for the Study of African Economics. Available for Reducing Rural Poverty in Middle-Income at: http://www.bepress.com/cgi/viewcontent. Countries: The Case of Thailand. DSGD Discussion cgi?article=1071&context=csae Paper No. 7. Washington, D.C.: International Food De Silva, Dakshina G., Timothy Dunne, Anuruddha Policy Research Institute. Kankanamge, and Georgia Kosmopoulou. 2008. 38 Refe rences Feinstein, Jonathan S., Michael K. Block, and Frederick Government of France. Conseil de la Concurrence. 2003. C. Nold. 1985. “Asymmetric Information and Avis n° 03-A-05 du 25 avril 2003 relatif à un projet Collusive Behavior in Auction Markets,� American de décret réformant le code des marchés publics. Economic Review 75(3): 441–460. Paris. Available at: http://www.autoritedelaconcur- Fernald, John G. 1999. “Roads to Prosperity? Assessing rence.fr/user/index.php the Link between Public Capital and Productivity.� Government of France. Conseil de la Concurrence. 2000. The American Economic Review 89 (3): 619–638. Avis n° 00-A-25 du 20 novembre 2000 relatif à Flax, Stephen. 1983. “The Crackdown on Colluding un projet de décret portant réforme du code des Roadbuilders,� Fortune, October 3. marchés publics. Paris. Available at: http://www. Froeb, Luke M., and Mikhael Shor. 2005. “Auction autoritedelaconcurrence.fr/user/index.php Models,� chapter 9 in Hakrider, John, ed., Government of Kenya, Kenyan Anticorruption Econometrics: Legal, Practical, and Technical Issues. Commission. 2007. Examination Report into the Chicago: American Bar Association. Systems, Policies, Procedures and Practices of the Gambetta, Diego, and Peter Reuter. 1995. “Conspiracy Roads Sub-Sector. Nairobi, KACC, Directorate of Among the Many: The Ma�a in Legitimate Preventive Services, July. Industries,� chapter 5 in Fiorentia, Gianluca, and Government of Kenya, Ministry of Planning and Sam Peltzman, eds., The Economics of Organized National Development. 2004. Proceedings of the Crime. Cambridge: Cambridge University Press. First Kenya-South Africa Roundtable Discussion Gönenç, Rauf,Willi Leibfritz, and Erdal Yilmaz. on Science & Technology Partnerships Related to 2005. Reforming Turkey’s Public Expenditure Infrastructure. Nairobi. August 31–September 2. Management. Paris: OECD Economics Available at: http://www.buildnet.co.za/akani/2004/ Department Working Papers No. 418. nov/pdfs/Kenya_SA_Roundtable_Proceedings.pdf Available at: http://www.oecd.org/of�cial- Government of Lithuania. 2008. “Contribution,� chapter documents/displaydocumentpdf?cote=ECO/ in OECD 2008b. WKP(2005)5&doclanguage=en Government of Nepal. 2009. A Review of Cases of Golllin, Douglas, and Richard Rogerson. 2010. Collusion and Intimidation in Public Procurement: Agriculture, Roads, and Economic Development in A Way Forward to Mitigate Risks of Collusion and Uganda. National Bureau of Economic Research. Intimidation. Katmandu. Working Paper 15863. Available at: http://www. Government of the Netherlands, Netherlands nber.org/papers/w15863. Competition Authority. 2010. Bid Rigging: Government of Brazil, Departamento de Proteção Detecting and Preventing Collusion in Procurement. e Defesa Econômica da Secretaria de Direito Government of Norway. 2011. Annual Report of Econômico do Ministério da Justiça Esplanada dos Norwegian Competition Authority 2010. Oslo: Ministérios. 2008. Combate a Cartéis em Licitações. Konkurransetilsynet. Available at: http://www. Brasília. Available at: http://www.comprasnet.gov. konkurransetilsynet.no/en/2010/ br/banner/seguro/Cartilha_Licitacao.pdf Government of the Philippines, Department of Government of Canada. 2010. Truquage des Offres: Public Works and Highways. 2007a. Terms of Sensibilisation et Prévention. Bureau de la Reference, Reform Agenda 007–02, Independent Concurrence. Available at: http://www.bureaudela- Procurement Evaluator. Available at: http://www. concurrence.gc.ca/eic/site/cb-bc.nsf/fra/02601.html pegr.org.ph/pegr/fckuploadfolder/�le/RA007–02%20 Government of El Salvador, Superintendencia de TOR%20IPE_22%20Jan%202009_RPT_V2.pdf Competencia. 2010. Detectando y Previniendo Government of the Philippines, Department of Public Ofertas Colusorias en las Compras Públicas. San Works and Highways. 2007b. Terms of Reference, Salvador. Available at: http://www.sc.gob.sv/up- Reform Agenda 007–05, Independent Technical loads/Manual_ofertas_colusorias.pdf Audit. Available at: 39 Curbing Fraud, Corruption, and Collusion in the Roads Sector http://www.pegr.org.ph/pegr/fckuploadfolder/�le/ Greif, Avner. 2006. Institutions and the Path to the RA007–07%20TOR_%20Internal%20Control- Modern Economy: Lessons from Medieval Trade. Audit_RFT_12Jun08%202.pdf Cambridge: Cambridge University Press. Government of the Philippines. 2003. Implementing Grout, Paul A., and Silvia Sonderegger. 2005. Predicting Rules and Regulations of Republic Act 9184, Cartels. London: Of�ce of Fair Trading. Available Annex A. Available at: http://www.comelec. at: http://www.oft.gov.uk/shared_oft/reports/comp_ gov.ph/invi2bid/2009/RA%209184_IRR-A%20 policy/oft773.pdf amended2_3rd%20edition.PDF Gupta, Srabana. 2001. “The Effect of Bid Rigging on Government of Slovakia, Antimonopoly Of�ce. 2005. Prices: A Study of the Highway Construction Cartel Agreement on Highway Construction. Industry,� Review of Industrial Organization, 19(4): Available at: http://www.antimon.gov.sk/391/2097/ 451–465. cartel-agreement-on-highway-construction.axd Hurwicz, Leonid. 2007. But Who Will Guard the Government of Tanzania. 2004. Public Procurement Guardians? Lecture Delivered upon Receiving Act of 2004, Section 57. Available at: http:// Nobel Prize, Oslo. Available at: http://nobelprize. www.auwsa.or.tz/policies_regulations/ org/nobel_prizes/economics/laureates/2007/hur- PROCUREMENT%20ACT%202004.pdf wicz_lecture.pdf Government of Tanzania, Presidential Commission of Hüschelrath, Kai, Nina Leheyda, and Patrick Beschorner. Inquiry Against Corruption. 1996. Report on the 2009. “Assessing the Effects of a Road-Surfacing State of Corruption in the Country. Dar Es Salaam. Cartel in Switzerland,� Journal of Competition Law Government of the United Kingdom, Of�ce of Fair & Economics 6(2): 335–374. Trade. 2004. Assessing the Impact of Public Sector Jaeger, Axel-Volkmar, and Götz-Sebastian Hök. 2010. Procurement on Competition. London: Of�ce FIDIC: A Guide for Practitioners. Heidelberg: of Fair Trade. Available at: http://www.oft.gov.uk/ Springer. shared_oft/reports/comp_policy/oft742a.pdf Joyce, Jon. M. 1989. “Effect of Firm Organizational Government of the United States, Delegation of the Structure on Incentives to Engage in Price United States to the Competition Commission. Fixing,� Contemporary Policy Issues 7(October): 2008. Note by the United States. Paris: OECD, 19–33. June. Available at: http://www.ftc.gov/bc/ Kellman, Steven. 2002. “Revising Federal Procurement,� international/docs/Construction_US.pdf Public Procurement Law Journal 31(4): 581–622. Government of the United States, Department of Kenny, Charles. 2010. Publishing Construction Contracts Justice. 2007. Public Procurement: The Role of and Outcome Details. Washington, D.C.: World Competition Authorities in Promoting Competition. Bank, World Bank Policy Research Working Paper Available at: http://www.ftc.gov/bc/international/ 5247, March 2010. Available at: http://charleskenny. docs/pubproc.pdf blogs.com/weblog/2010/04/publishing-construction- Government of the United States, Departments of contracts-and-outcome-details.html Transportation and Justice. 2004. Guidelines on Khandker, Shahidur, Victor Lavy, and Deon Filmer. Preparing Engineer’s Estimate, Bid Reviews and 1994. Schooling and Cognitive Achievements of Evaluation. Available at: http://www.fhwa.dot.gov/ Children in Morocco. World Bank Discussion programadmin/contracts/ta508046.cfm#s6 Paper 264. Washington, D.C.: World Bank. Government of Zambia. 2010. Report of the Auditor Available at: http://www-wds.worldbank.org/exter- General on the Road Development Agency for nal/default/main?pagePK=64193027&piPK=64187 the Period January 2006 to September 2009. 937&theSitePK=523679&menuPK=64187510&sea Lusaka. Available at: http://www.scribd.com/ rchMenuPK=64187283&theSitePK=523679&entity doc/33544699/2006–2009-Auditor-General-Report- ID=000009265_3970311124059&searchMenuPK= on-Roads-Development-Agency 64187283&theSitePK=523679 40 Refe rences Khumalo, Junior, Pamela Nqojela, and Yongama McAfee, R. Preston, and John McMillan. 1987. “Auctions Njsane. 2009. Cover Pricing in the Construction and Bidding,� Journal of Economic Literature 25(3): Industry: Understanding the Practice Within a 699–738. Competition Context. Pretoria. Paper prepared McMillan, John. 1991. “Dango: Japan’s Price Fixing for Fourth Annual Competition Commission, Conspiracies,� Economics and Politics 3(3): Competition Tribunal and Mandela Institute 201–218. Conference on Competition Law, Economics Messick, Richard E. 2009. Income and Assets Declarations: and Policy in South Africa, Pretoria, September. Issues to Consider in Developing a Disclosure Available at: http://www.compcom.co.za/ Regime. Bergen, Norway: Chr. Michelsen Institute. fourth-competition-law-conference/ Available at: http://www.cmi.no/publications/ Klein, Daniel B. 1997. Reputation: Studies in the publication/?3396=income-and-assets-declarations Voluntary Elicitation of Good Conduct. Ann Arbor, Milgrom, Paul. 2004. Putting Auction Theory to Work. Michigan: University of Michigan Press. Cambridge: Cambridge University Press. Kluenker, Chuck. 2001. “Risk vs Conflict of Interest Miller, Nathan H. 2009. “Strategic Leniency and Cartel – What Every Owner Should Consider When Enforcement.� American Economic Review, 99(3): Using Construction Management,� CM eJournal, 750–68. January. Available at: http://cmaanet.org/�les/ Mitchener, Kris, and Noel Maurer. 2010. “Renting risk_vs_conflict.pdf Institutions to Combat Corruption.� Available at Knittel, Christopher R., and Victor Stango. 2003. “Price http://chartercities.org/blog/99/renting-institutions- Ceilings as Focal Points for Tacit Collusion: to-combat-corruption-kris-mitchener-and-noel- Evidence from Credit Cards,� The American mauer Economic Review 93(5): 1703–1729. Mwangi, Anthony. 2004. “Mozambique,� chap- Lambert-Mogiliansky, Ariene. Forthcoming. ter 5 in De Wulf, Luc, and Jose B. Sokol, eds., “Corruption and Collision in Procurement: Customs Modernization Initiatives: Case Studies. Strategic Complements,� chapter in Susan Rose- Washington, D.C.: World Bank. Ackerman, ed., The International Handbook on the Ndekugri, Issaka, Nigel Smith, and Will Hughes. Economics of Corruption. 2007. “The Engineer under FIDIC’s Conditions Lee, In Kwon, and Kyungdong Hahn. 2002. “Bid-Rigging of Contract for Construction,� Construction in Auctions for Korean Public Works: Contracts Management and Economics 25(July): 791–799. and Potential Damages,� Review of Industrial Organization for Economic Cooperation and Organization 21(1): 73–88. Development [OECD]. n.d. Guidelines for Fighting Levenstein, Margaret C., and Valerie Y. Suslow. 2006. Bid Rigging in Public Procurement: Helping “What Determines Cartel Success?� Journal of Governments to Obtain Best Value for Money. Paris: Economic Literature 44(1): 43–95. OECD. Lyon, M. Beth. 1995. “The Role of the Consulting OECD. 2010a. 9th Global Forum on Competition, Engineer in Developing Country Construction Paris, February 18–19. Available at: http:// Under FIDIC Form Contract,� Law & Policy in www.oecd.org/document/34/0,3343, International Business 26(1): 273–294) en_40382599_40393118_44060194_1_1_1_1,00.html Mariel, Hector A. 2003. “Government Contracts Under OECD, Directorate for Financial, Fiscal, and Enterprise Argentine Law: A Comparative Law Review,� Affairs, Competition Committee. 2010b. Fordham International Law Journal 26: 1716–1753. Collusion and Corruption in Public Procurement, Martin, George. 1983. And Never the Twain Shall Meet: Introductory Note by the Secretariat, Paris, Cross-Cultural Conflict in the Administrative OECD. Available at: http://www.oecd.org/datao- Process, U.S. Military Procurement in Korea. ecd/38/33/44511981.pdf Bloomington: International Development OECD. 2008a. Fighting Cartels in Public Procurement. Institute, Indiana University. Paris: OECD. 41 Curbing Fraud, Corruption, and Collusion in the Roads Sector OECD. 2008b. Policy Roundtables: Construction Industry. Budget Project. Available at: http://www.inter- Paris: OECD. nationalbudget.org/library/publications/index. OECD. 2006. Prosecuting Cartels without Direct Evidence. cfm?fa=view&id=3520 Paris: OECD. Sichombo, B. et al. 2009. “The Need for Technical OECD. 2003. Hard Core Cartels: Recent Progress and Auditing in the Zambian Construction Industry,� Challenges Ahead. Paris: OECD. International Journal of Project Management 27: OECD. 1998. Recommendation of the Council 821–832. Concerning Effective Action against Hard Core Sing, Sanjeet. 2005. “Fighting Corruption in Developing Cartels. Paris: OECD. Available at: http://www. Countries: Dimensions of the Problem in India.� oecd.org/dataoecd/39/4/2350130.pdf. Paper prepared for Good Governance, Institutional OECD. 1990. Glossary of Industrial Organization Integrity and Human Resources Management for Economics and Competition Law. Paris: Road Administrations, Session 2, Integrity in the OECD. Available at: http://www.oecd.org/datao- Road Sector, World Road Association. Available ecd/8/61/2376087.pdf at: http://publications.piarc.org/en/seminars/semi- Olken, Benjamin A. 2007. “Monitoring Corruption: nars_05/warsaw_October05.htm. Evidence from a Field Experiment in Indonesia,� Smith, Adam. 1937 [1776]. An Inquiry into the Nature Journal of Political Economy 115 (2): 200–249. and Causes of the Wealth of Nations. New York: Oluka, Pross Nagitta, and Francis Ssennoga. 2008. Modern Library, available at: http://www.econlib. Tackling Corruption in Public Procurement: Case org/library/Smith/smWN.html of Local Government in Uganda, Paper pre- Soemardi, Tresna. 2010. “Interface Between Competition pared for 3rd International Public Procurement Authorities, Competition Regulations, and Other Conference,G3. Available at: http://www.unpcdc. Regulatory Frameworks and Institutions in org/media/13246/tacklg%20corrup%20in%20 Infrastructure Service.� Report prepared for Multi- pp%20loc%20govt%20in%20uganda.pdf year Expert Meeting on Services, Development Oxford Business Group. 2004. Emerging Romania. and Trade: The Regulatory and Institutional London. Dimension. Geneva, UNCTAD, March 17–19. Patterson, William D. O., and Pinki Chaudhuri. 2007. Available at: http://www.unctad.org/templates/Page. “Making Inroads on Corruption in the Transport asp?intItemID=5353&lang=1 Sector Through Control and Prevention,� chapter Songco, Jocelyn A. 2002. Do Rural Infrastructure 5 in Campos, J. Edgardo, and Sanjay Pradhan, Investments Bene�t the Poor? Evaluating Linkages: eds., The Many Faces of Corruption: Tracking A Global View, a Focus on Vietnam. Policy Vulnerabilities at the Sector Level. Washington, Research Working Paper Series 2796. Washington, D.C.: World Bank. D.C.: World Bank. Available at: http://ideas.repec. Posner, Richard A. 2001. Antitrust Law. Chicago: org/p/wbk/wbrwps/2796.html University of Chicago Press, 2nd ed. Stewart, Taimoon, Julian Clarke, and Susan Price Waterhouse Coopers. 2008. Monitoring Agent Joekes. 2007. Competition Law in Action: Report to the MDTF-S Administrator, Final Experiences from Developing Countries. Ottawa: Report. January. International Development Research Centre. Polgreen, Lydia. 2010. “Right-to-Know Law Gives India’s Available at: http://www.idrc.ca/uploads/user- Poor a Lever,� New York Times, June 28. S/11781215481Competition_Law.pdf. Porter, Robert, and Douglas Zona. 1993. “Detection of Stigler, George J. 1964. “A Theory of Oligopoly,� Journal Bid-Rigging in Procurement Auctions,� Journal of of Political Economy 71(1): 44–61. Political Economy 101: 518–538. Tempo. 2010. “Hadi’s Holdings: State Audit Head and Ramkumar, Vivek. 2008. Our Money, Our Responsibility: Former Director-General of Tax, Hadi Poernomo, A Citizen’s Guide to Monitor Government Failed to Disclose All his Assets to the KPK.� June Expenditures. Washington, D.C.: International 23–29. 42 Refe rences Transparency International. 2009. Countering Cartels World Bank. 2009a. Deterring Corruption and to End Corruption and Protect the Consumer. Improving Governance in Road Construction and Berlin: Transparency International. Available Maintenance. Washington, D.C., World Bank, at: http://www.transparency.org/publications/ Transport Sector Board. publications/policy_positions/pp_7_2009_cartels World Bank. 2009b. Republic of Armenia: Country Transparency International. 2006. Handbook for Curbing Procurement Assessment Report. Washington, Corruption in Procurement. Berlin: Transparency D.C.: World Bank, Report No. 49975-AM. International. Available at: http://www.trans- Available at: http://imagebank.worldbank.org/serv- parency.org/publications/publications/other/ let/WDSContentServer/IW3P/IB/2009/10/05/00033 procurement_handbook 3037_20091005002008/Rendered/PDF/499750ESW UNCTAD [United Nations Conference on Trade and 0AM0P1IC0disclosed01011191.pdf Development]. 2010. Competition Law and World Bank. 2008a. Safe, Clean, and Affordable . . .: Policy Web Site: http://www.unctad.org/Templates/ Transport for Development. Washington, D.C., StartPage.asp?intItemID=2239&lang=1 Transport Sector Board. Van den Heuvel, Grat. 2006. “The Parliamentary World Bank. 2008b. Project Appraisal, Phase 2 of the Enquiry on Fraud in the Dutch Construction National Roads Improvement and Management Industry: Collusion as Concept between (APL) Program. Washington, D.C.: World Bank, Corruption and State-corporate Crime,� Crime, April 15. Available at: http://www-wds.worldbank. Law & Social Change 44(2): 133–151. org/servlet/main?menuPK=64187510&pagePK=64 Van Klinken, Gerry and Edward Aspinall. 2011. 193027&piPK=64187937&theSitePK=523679&enti “Building Relations: Corruption, Competition tyID=000334955_20080428082019 and Cooperation in the Construction Industry,� World Bank, Independent Evaluation Group. 2008c. chapter in Edward Aspinall and Gerry van Using Knowledge to Improve Development Klinken, eds., The State and Illegality in Indonesia. Effectiveness: An Evaluation of World Bank Leiden: KITLV Press. Economic and Sector Work and Technical Woodall, Brian. 1996. Japan Under Construction: Assistance, 2000–2006. Washington, D.C. World Corruption, Politics, and Public Works. Berkley: Bank. Available at: http://www-wds.worldbank.org/ University of California Press. external/default/WDSContentServer/WDSP/IB/200 World Bank. n. d. Redacted Report: National Roads 9/05/22/000333037_20090522014837/Rendered/PD Improvement and Management Program (Phase F/485970PUB0know10Box338915B01PUBLIC1.pdf 1). Washington, D.C.: World Bank, Integrity Vice World Bank. 2007. A Decade of Action in Transport: Presidency. Available at: http://siteresources.world- An Evaluation of World Bank Assistance to the bank.org/INTDOII/Resources/INT_redactedre- Transport Sector, 1995–2005. Washington, D.C.: port_Philippines.pdf The World Bank, Independent Evaluation Group. World Bank. 2010a. Guidelines: Procurement Under World Bank. 2006a. Expansion of Sanction Beyond IBRD Loans and IDA Credits. Available at: http:// Procurement and Sanctioning of Obstructive web.worldbank.org/WBSITE/EXTERNAL/ Practices. Washington, D.C., July 28. Available PROJECTS/PROCUREMENT/0,,contentMDK:200 at: http://siteresources.worldbank.org/INTDOII/ 60840~menuPK:84282~pagePK:84269~piPK:6000 Resources/SanctionsReform06.pdf 1558~theSitePK:84266,00.html. World Bank. 2006b. Project Appraisal, Republic of World Bank. 2010b. Standard Bidding Documents: Paraguay, Road Maintenance Project. Washington, Procurement of Works & User’s Guide. May. D.C.: World Bank, July 24. Available at: http://im- Available at: http://siteresources.world- agebank.worldbank.org/servlet/WDSContentServer/ bank.org/INTPROCUREMENT/Resources/ IW3P/IB/2006/08/08/000090341_20060808104830/ Works-4–07-ev2_May2010_v2.pdf Rendered/PDF/36421.pdf 43 Curbing Fraud, Corruption, and Collusion in the Roads Sector World Bank. 2001. Indonesia Country Procurement Zoghbi, Valentina. 2009. “Strategic Priorities of Assessment Report: Reforming the Public Competition and Regulatory Agencies in Procurement System. Washington, D.C.: World Developing Countries,� chapter 3 in Mehta, Bank. Available at: http://imagebank.worldbank. Pradeep S., and Simon J. Evenett, eds., Politics org/servlet/WDSContentServer/IW3P/IB/2001/04 Triumphs Economics? Political Economy and the /27/000094946_01041707211320/Rendered/PDF/ Implementation of Competition Law and Economic multi0page.pdf Regulation in Developing Countries. New Delhi: World Bank. 1994. World Development Report 1994: Academic Foundation. Infrastructure for Development. New York: Oxford University Press. 44 Annex 1 Proving Bid Rigging on Roads Tenders It does not have to be made in writing; no in rigging by, for example, submitting a cover bid or formalities are necessary, and no contractual agreeing not to bid. Other forms of direct evidence sanctions or enforcement measures are re- would include the testimony of those who witnessed the quired. The fact of agreement may be express rigging, such as clerical or support staff of the compa- or implicit in the behavior of the parties. nies involved, or �rms that were invited to rig the tender European Communities v. F. Hoffman-la but declined. Documents disclosing some or all of the Roche AG, 4 C.M.L.R. 22, 37 (2003). details of the bid rigging would be another form of di- rect evidence. As the European Court of First Instance observed in the above excerpt from the Vitamin Cartel case, collu- Indirect evidence is generally broken down into two sive agreements come in many forms—written or oral, categories. There is �rst economic evidence showing informal or formal, express or implied. Parties to such the market is not competitive. In the case of road con- agreements generally go to great lengths to conceal their tracts, it would consist of evidence demonstrating that existence, and so, particularly when an agreement is conditions make it likely that bidders do not compete informal or implied, establishing its presence in a legal for tenders. As the discussion in this report showed, in proceeding can be a challenge. With courts and competi- the roads sector in most countries a plethora of this tion agencies hearing an ever larger number of allega- type of evidence will likely be available: the product is tions of collusive agreements, however, some common highly standardized; prices are inelastic, that is, insensi- principles have emerged for proving collusion, a conver- tive to changes in costs; and a few �rms dominate the gence furthered by a rich cross-national dialogue and an market. When coupled with a system of open public expanding body of comparative law scholarship. tendering, the economic case for collusion is very strong. Additional economic evidence of collusion in particular All jurisdictions distinguish between direct and indirect cases would include (a) bids signi�cantly in excess of evidence of collusion. Direct evidence is testimony de- costs, (b) �rms with excess capacity, or other economic scribing, or documents showing, a collusive agreement. incentives to bid, declining to do so, (c) the market Indirect, or circumstantial, evidence consists of facts shares of the large �rms remaining stable over time, and and circumstances from which an administrative body (d) a pattern of winning bids showing �rms taking turns or a court of law can infer the existence of a collusive “winning� over time. agreement. No matter how strong the economic evidence, courts In cases of bid rigging, direct evidence would include the and competition law agencies almost always require testimony of one or more individuals who participated some additional evidence to �nd collusion. The reason is Curbing Fraud, Corruption, and Collusion in the Roads Sector that the structure of some industries alone can produce characteristics of the particular market and the type noncompetitive conditions, what is termed “oligopolistic of collusive arrangement alleged. One fairly exhaustive interdependence,� even without a collusive agreement. list is in OECD 2006. Examples of plus factors in the This interdependence is typically found in markets roads sector would include (a) bids that are identi- where a few �rms manufacture a homogeneous prod- cal in all or almost every respect except price, (b) an uct and prices are inelastic and publicly posted or an- econometric or statistical analysis showing that the nounced. In these markets, it is in each �rm’s long-run bids were not prepared independently, (c) the submis- self-interest to maintain supra-competitive prices, and sion of fraudulent bid securities by well-established if all �rms recognize this, an agreement not to compete �rms, (d) oral or written communications about plans may not be necessary. The use of a �rst-price, sealed-bid to bid or the amount of a bid, (e) agreements on sub- auction to award road construction and maintenance contracting, (f) the purchase of bidding documents by contracts makes oligopolistic interdependence in the �rms that did not bid, and (g) communications and roads sector unlikely. On the other hand, there can be meetings just before a tender is due. circumstances—such as when engineering costs esti- mates are disseminated or the names of all bidders and As in any factual determination, the evidence must be the amounts each bid are revealed—under which �rms considered as a whole. Credible direct evidence of bid in the roads sector might be able avoid competing with- rigging is often suf�cient to show collusion. In its ab- out a collusive agreement. sence, the economic and noneconomic evidence will be weighed together. In the roads sector, where the eco- For this reason, some evidence of an agreement will nomic evidence of the absence of competition is likely be useful to assure the fact-�nder that collusion to be strong, the plus factor or factors presented may not is present. Such additional evidence is commonly need to be as probative as they would have to be when termed a “plus factor,� and courts and commentators the economic evidence is more problematic (OECD have identi�ed various types, depending upon the 2006, Posner 2001). 46 Annex 2 Reforms to Public Procurement in OECD Countries Banning pre-bid meetings with more than one po- Keeping bidders’ identities con�dential and not lim- tential supplier. iting their number unnecessarily. Limiting communications between bidders during Requiring bidders to disclose all communica- the tender process. tions with competitors and to sign a Certi�cate of Using negotiated tenders and framework agreements Independent Bid Determination. when collusive behavior persists. Banning subcontracting in appropriate circum- Using a ceiling price only if it is based on thorough stance and at a minimum requiring bidders to dis- market research and engineering estimates and of- close in advance if they intend to use subcontractors. �cials are convinced it is very competitive. Ensuring Banning joint bids when appropriate as they facili- it is kept it con�dential. tate communication and pro�t splitting among bid Taking precautions when using industry consultants riggers. to conduct the tendering process; ensuring they have Subjecting external consultants to a reporting re- not established working relationships with indi- quirement if they become aware of improper com- vidual bidders. petitor behavior or any potential conflict of interest. Whenever possible, requesting that bids be �led anonymously (e.g. consider identifying bidders with numbers or symbols) and allowing bids to be sub- mitted by telephone or mail. Source: OECD n.d.