In 2017, Western Union acknowledged that it had aided and abetted wire fraud in violation of the Bank Secrecy Act. But rather than facing prosecution, the corporation — whose reported annual revenue was $5.3 billion in 2019 — entered a deferred prosecution agreement (DPA) with the United States. By agreeing to the DPA, Western Union avoided prosecution and agreed to pay $586 million to its victims, many of whom were elderly targets of consumer scams.

More recently, aerospace giant Airbus — which last year reported revenue of $70.5 billion — entered a DPA and avoided criminal prosecution for corruption in multiple countries. Officials accused Airbus of using intermediaries to bribe public officials and the company paid a $4 billion fine.

Then there’s General Motors. The automaker admitted that its faulty and long-hidden ignition problem killed at least 124 people in the span of a decade. But GM, which reported $137.2 billion in revenue in 2019, escaped prosecution by entering into a DPA in 2016 and paying a $900 million fine.

These cases are not outliers. In fact, they are typical of how major corporate crime is handled, with negotiated payments rather than prosecution, conviction and punishment. And the practice has become more common over the past decade: Syracuse University’s federal data clearinghouse TracReports noted that federal prosecutions of white-collar crime reached 9,412 in 1998, but the number had shrunk to less than 6,000 per year by 2017 and 2018. This is despite the fact that the economic costs of white-collar crime outpace the economic costs of street crime by 20-to-1.

DPAs are meant to reform a company, by increasing oversight and compliance to stop a company’s wrongdoing. But many critics contend that wrongdoers simply have their shareholders foot the bill for the required fines and move on with business as usual. As University of Virginia law professor Brandon Garrett told Vox, fines seem “like something of a worthwhile risk if you don’t always get caught.”

Many critics — including US Senator Elizabeth Warren — contend that fines aren’t enough, as they essentially allow corporations to pay money to break the law. (Warren has introduced legislation on the issue.)

Mihailis Diamantis, a corporate crime scholar at the University of Iowa, believes that forced rehabilitation is the only truly effective way to punish corporations and prevent further crime. It’s among the measures that he and colleague William S. Laufer, of the Wharton School of the University of Pennsylvania, assess in the 2019 Annual Review of Law and Social Science.

Knowable Magazine spoke with Diamantis about what’s wrong with the current approach and how it could be improved. This conversation has been edited for length and clarity.

Photo shows a man surrounded by media reporters with microphones and cameras pointed toward him.

French National Financial Prosecutor Jean Francois Bohnert speaks to the media about a deal with Airbus in a Paris courthouse in January 2020. Following an accusation of bribery, Airbus entered into a deferred prosecution agreement and agreed to a $4 billion fine, avoiding criminal prosecution in multiple countries.


When a corporation commits a crime, how does the justice system decide who gets blamed and punished?

There are many really embarrassing problems with our corporate criminal law. One is the law of respondeat superior. That’s the doctrine that helps courts determine when corporations are guilty of a crime. It basically says that corporations do whatever their individual employees do, so long as those employees are acting within the scope of their authority and intend to benefit their corporate employer. So, roughly speaking, a corporation commits a crime when, and only when, one of its employees commits a crime.

But there will rarely be direct evidence of a single person committing a crime at big, publicly traded corporations, so convictions are rare under this law. Even where evidence is available, it often shows that no individual did anything criminal, even when individuals collectively produced criminal harm. It seems that we need to consider all of a corporation’s actions as a unit, rather than those of any single employee, if we want to hold corporations accountable when they do things and there’s nobody to take the blame.

We also don’t have an adequate theory of what it means to punish a corporation. The two modes that people primarily use to think about punishment — retribution and deterrence — are in many ways not up to the task.

Bar graph shows fewer federal criminal prosecutions for white-collar crimes since 2012.
The number of federal white-collar crime prosecutions has declined each year since 2012.

Retribution requires the object of retribution to suffer. Corporations don’t suffer. They do not have emotions — they’re fictional entities. Deterrence imposes costs in a way that is intended to influence future behavior. But when you impose a criminal fine on a publicly traded company or large corporation, it’s paid by the shareholders. The shareholders, though, aren’t directing the corporation.

We have this fundamental division between ownership and control in the largest corporations. The fine hits only the owners, not the people who are in control, so the deterrent mechanism doesn’t work.

What are the alternatives?

We need to think more creatively about what it means to punish in this context. My preferred approach is forcible rehabilitation, to fix whatever organizational structures encourage, permit or facilitate misconduct.

Wells Fargo is an easy example. After Wells Fargo’s consumer accounts scandal, we knew that the corporation had at least three defects: a corporate culture that encouraged a lax attitude toward legal and ethical norms; a high-stakes compensation scheme tied to unrealistic quotas for employees; and broken reporting channels for addressing concerns about misconduct. Suitable reforms would include improved training to make legal and ethical norms more salient in the workplace; a different compensation scheme that de-emphasizes quotas and gives a role to quality control; and anonymous whistleblowing channels.

Other possible reforms are not necessarily tied directly to organizational defects but work in a more roundabout way. For example, tone from the top, from executives, turns out to be an important variable in determining a corporation’s predisposition to crime, so replacing senior management can often be an effective first step toward rehabilitation. Wells Fargo did this after it was caught fraudulently opening accounts for customers.

Forced rehabilitation would mean that we would stop fining corporations, as fines seem to be ineffective and unjust. As we increase the financial sanctions on corporations, we don’t get a corresponding decrease in corporate criminality. And fines tend to punish the wrong people, such as shareholders and creditors who haven’t done anything criminal. We need a mechanism in place that will translate criminal sanction into positive corporate action.

The question after rehabilitation: What more do we want to do? If we assume that we have fixed the corporation and it’s not going to engage in misconduct again, what further interest does the law have? Since we should expect no further misconduct from a reformed corporation, there is nothing to deter.

We humans do have retributive impulses that the law should attend to. But we should also hesitate before giving full voice to those in the corporate punishments we impose. Corporations are a cornerstone of our economy and society, so burdening them unnecessarily has serious costs. Furthermore, as I already said, the burdens of corporate sanctions tend to fall on innocent people.

You wrote that there’s no study of people who are victims of corporate crime. How can victims be brought into the rehabilitation process?

My colleague Bill Laufer has observed that we are missing an appreciation for the range and seriousness of corporate victimization. Without a subfield of corporate victimology, victims fall out of the picture. We thereby lose any hope of truly understanding corporate blameworthiness and of accurately linking legal consequences (like conviction and sentencing) to corporate culpability.

I maintain that rehabilitation is the best way to respect victims.

Who the victims are is usually clear, but not always. In the case of Wells Fargo, it was clear that its customers were victims. But sometimes — like in tax evasion, money laundering or insider trading — the individual victims are impossible to find. Society and the integrity of its financial mechanisms are sometimes the true victims.

When we know who the victims are, they have to be compensated for the wrong that was done to them, whether that’s in the civil system or through criminal statutes. But I don’t think that compensation alone pays victims full respect, especially not when the victims are impossible to compensate. The most significant statement the law can make to the victims of criminal misconduct is that what was done to them is wrong and won’t happen again. That’s a promise that only successful rehabilitation can deliver.

When we’re talking about criminal law, we want to say something more than, “There’s a price to doing wrong.” Instead, we should say, “We do not tolerate that kind of conduct.”

How have so many corporations avoided prosecution?

This is a result of the super-lenient corporate criminal justice system that has evolved. The basic established principles of corporate criminal law hold corporations responsible for employees’ actions. But criminal courtrooms are dangerous places for corporations to find themselves. The Department of Justice wants to avoid driving another massive company out of business. For example, the shuttering of accounting firm Arthur Andersen after the Enron prosecutions was a lightning-rod event, one the DOJ wants to avoid repeating. So the DOJ strives to keep large corporate criminals out of courtrooms.

These days, prosecutors and corporations usually negotiate deferred prosecution agreements (DPAs) to address suspicions of corporate misconduct outside of the judicial system. In DPAs, the idea is to resolve investigations rather than bring them to trial, and to institute internal corporate reform, akin to the changes we discussed with forced rehabilitation.

Graphic. Of some 41 million reports of corporate misconduct, only 2,171 cases reach prosecutors, with 237 prosecuted, 162 convicted and 30 settled by agreements.

Although reports show millions of US corporations involved in ethical misconduct, there’s a modest number of referrals for prosecution and convictions.

This practice has political benefits for the DOJ, but it cuts out the courts. DPAs allow corporations and prosecutors to negotiate the facts of the case in backrooms, rather than letting the facts be aired publicly. DPAs also typically allow corporations to escape admissions of guilt or wrongdoing. This shortchanges the important expressive value of the trial process and conviction, which is the formal public recognition of wrongdoing.

DPAs also cut courts out of sentencing. This leaves prosecutors in the position of undertaking any rehabilitative project to constructively sanction corporations. But prosecutors don’t have the expertise needed to adequately design agreements — they tend to use generic terms like “improve compliance,” instead of detailed plans for reform. If prosecutors do sometimes defer to experts — like corporate monitors, who review a corporation’s compliance practices and implement the general terms of the DPA — prosecutors still don’t have the expertise necessary to oversee and evaluate how well the rehabilitative process is going.

So prosecutors can put rehabilitative measures in place, but they don’t have the expertise to see them through?

Yes. And in DPAs, the rehabilitative measures are written in terms that are so abstract as to be meaningless. Businesses know a lot about compliance — they’re much more sophisticated than the prosecutors — and that’s why the prosecutors rely on the corporate criminals to rehabilitate themselves.

Prosecutors also have their own set of interests; this is where gamesmanship comes into play. Ideally, the prosecutor is after justice for victims and wants to hold the guilty accountable; that’s the public role of the prosecutor. But prosecutors are human beings who are looking for personal recognition, reputation and promotion within the department. What this means is that they want to get headlines, process cases and not have to hunker down on any individual matter.

Both prosecutors and corporations are interested in resolving cases quickly before going to court. The prosecution gets to move on to the next case faster, and the corporation gets a speedy resolution they can bring back to their investors, as if to say, “We’re moving past these suspicions, and we’ve got a game plan for moving forward.” The most salient terms of DPAs are what the prosecution wants — a fine, a big number they can put in a headline. Case resolved.

Primarily, both parties just want to move forward. You have an alignment of interests between two parties you might have thought should be averse to each other.

How could the system be shifted to ensure that a corporation changes in these situations?

We need injections of objectivity, expertise and consistency in the entire process.

One way would be to bring the courts back into the fold. If you could encourage prosecutors to bring more of their cases before judges so the judges have input — not only over whether the corporation is guilty of the alleged misconduct, but also at the sentencing phase when sanctions are imposed — you could inject more reliable reform. That’s not because judges are experts, but because they’re neutral. Ideally, judges don’t have this alignment of interests that the prosecution and corporations have.

Because judges have the capacity during sentencing to call in a broad range of experts, they would not have to construct plans of reform themselves. Instead, judges would rely on the prosecution, the corporation, regulators and perhaps even compliance experts to create a rehabilitation plan.

Prosecutors are really bad at following through on reform. In 2010, the Government Accountability Office found that there’s no evidence of the effectiveness of the Department of Justice’s use of DPAs and non-prosecution agreements. Courts can do better, because federal judges are appointed for life — there’s consistency in the process. Courts can continuously evaluate the progress of the sentence and appoint corporate monitors who aid them when they lack the expertise. They have the bandwidth to follow up.

You wrote that technology — such as blockchain, fintech, Big Data and artificial intelligence — could be a tool in disrupting the gamesmanship between prosecutors and corporations, as well as reducing corporate crime. How might this happen?

There are two significant ways that technology can disrupt the status quo. New technology can make corporate crime easier and allow corporations to commit crime with even more impunity. You can now do with an algorithm a lot of the things that you previously had to do with employees. And the law requires that in order to convict a corporation of a crime, you have to find the employee who committed the crime. But if you have an algorithm performing a job, it makes that task of identifying the responsible party more difficult.

In a lot of cases — especially when you’re using unsupervised machine learning — artificial intelligence may even end up doing things that nobody planned for it to do. For example, some firms in fintech are using algorithms to sort loan applications in a discriminatory way. This is the kind of misconduct our discrimination law was meant to address, but that corporations can now engage in with some amount of impunity because there’s not an employee acting.

Algorithms also have a lot of promise to positively disrupt corporate crime. Internally, technology could help corporations identify compliance failures sooner and more reliably. For example, algorithms could more effectively and cheaply carry out many of the audit functions currently delegated to employees. Externally, similar tools can help enforcement authorities identify misconduct. We already see that regulators can use machine learning to track stock-trading patterns and detect insider trading or other kinds of stock manipulation.

As these tools become more widely available, corporations will be able to engage in less expensive, more effective compliance measures. Potentially, corporations could produce compliance records and assurances that are verifiable by third parties — that’s where blockchain comes in. That technology can help generate records that are nearly impossible to manipulate. That would help address many problems. Corporations get cheaper compliance and visibility, and they can prove that what they’re doing is effective and legal.