The anticorruption community owes the American Economic Association and Nicolás Campos, Eduardo Engel, Ronald D. Fischer, and Alexander Galetovic a debt of gratitude. The AEA for publishing their article “The Ways of Corruption in Infrastructure: Lessons from the Odebrecht Case” and making it available free to non-members (here). The four Chilean scholars for showing how much can be learned when a command of the literature on corruption is coupled with a careful, painstaking study of a single case.

In 2016, the Brazilian engineering and construction company Odebrecht admitted in a settlement with American, Brazilian, and Swiss authorities (here) to bribing 600 officials in 12 states either to secure contracts to build roads, powerplants, and other large infrastructure projects or to agree to raise the contract price during construction of the project. Information the authors pieced together from the settlement documents show the company grossed $3.3 billion in profits from paying $788 million in bribes.  These numbers confirm the obvious: the returns from infrastructure corruption are enormous, and significant resources should be devoted to preventing it.

Digging deeper into the massive amount of paper the several prosecutions of Odebrecht and its executives have generated, the authors report other findings that are not so obvious.

• Guestimates in academic writings claim bribes on infrastructure projects run anywhere between 5 and 20 percent of construction cost and sometimes even higher. The authors show Odebrecht’s bribe payments were a tiny fraction of the project’s final cost, generally less than one percent.
• At the same time, the profits Odebrecht earned from bribing made all the difference to the company’s bottom line. “Most of the profits Odebrecht made during the period [2004 – 2014] were due to bribing.
• Bribery at the contract award stage focused on manipulating the subjective criteria used to evaluate bidder to exclude or disadvantage rivals.  This was especially true when bidders submitted only preliminary designs and the subjective evaluations consisted of judging the different bidders’ proposals for completing the design.
• Bribes paid to renegotiate the contract during its execution are especially costly. In the Dominican Republic, the company paid $1.6 million to increase the price of an aqueduct from the initial price of $161 million to $250 million. (More on how Odebrecht conspired with DR leaders to cheat its citizens here.)
• For a sample of contracts where a change in the initial contract price was the result of an honest, unforeseen event, the average price increase is 11 percent.  Where Odebrecht paid a bribe to in return for a price increase, the average increase was 85 percent.

Lessons the authors draw include:

• International competition for infrastructure contract pays off. “The combination of competition and some transparency at the tendering stage limits the discretion of public officials.” In his plea agreement a Peruvian official revealed that Odebrecht reduced the bribe it had agreed to pay the country’s president because he could not force the technical agency in charge of bidding to change the tendering documents.
• Whenever possible, eliminate subjective criteria for determining the winning bidder. (As GAB readers know, in infrastructure contracts, this means substituting design-bid-build contracts for design-build ones.)
• Given the enormous opportunity for abuse when renegotiating a contract during construction, more resources should be devoted to assessing the reasons for the renegotiation and the price agreed to. The authors recommend the terms of any renegotiation be reviewed by an independent panel and any significant change to the original contract be subject to a competitive tender that excludes the winner of the initial contract.

There is far more to the authors’ analysis than what this brief summary presents, and the entire article is must reading for anyone working on infrastructure corruption.

The article is must reading too for another group of concerned with corruption, those looking to write something useful about it. The topic is a challenging one to study, for much corrupt behavior is hidden from view. As a result, too much of what is written is based on guestimates and uninformed conjecture. What the authors have shown is that there is reliable, important data on the public record that can be mined to generate important insights about corruption. Let’s hope their Odebrecht article inspires others to take advantage of the data turned up in other large cases. There are unfortunately many to choose from.