Analyzing Settlements in Corruption Cases: A Primer

As prosecutions for bribery and other corruption crimes have ramped up around the world, so too has a practice common in the United States that is now spreading: resolving criminal cases short of full trials.  A prosecution can be cut short three ways.  The first is through a plea bargain.  The defendant admits guilt at some point between the investigation and the rendering of a final verdict.  As the term implies, the admission is the result of a bargain, the defendant receiving something in return, most often a lesser sentence.

The second and third ways are through non-prosecution and deferred prosecution agreements.  With the former, the prosecution does not charge the accused with a crime even though it has sufficient evidence to do so; in the latter, a charge is filed but immediately set aside.  Like a plea bargain, non-prosecution and deferred prosecution agreements are the result of an agreement between the accused and prosecutors.

The American practice of settling a criminal case short of a trial has always had its critics.  With an increasing number of countries adopting similar practices, several for the explicit purpose of resolving corruption prosecutions, the concerns about settlement heard in America, along with ones peculiar to corruption cases, are now circulating in a larger international community (examples here, here, and here).  For a paper on developing countries and settlements, I summarize the literature on how to analyze settlements.  It appears below.  I believe it robust enough to apply to any country, but would like to hear readers’ comments.

The sine qua non of a settlement policy that advances the public interest – i) punishment of a criminal wrongdoer and ii) the deterrence of future crime – is a well-functioning criminal justice system.  A criminal defendant will only bargain away the right to a trial if he or she (or “it” in the case of a corporate entity) fears conviction if the case goes to trial.  That fear rests on the accused’s belief that:

  1. investigators have discovered sufficient facts to prove a violation,
  2. the government is willing and able to prosecute the case, and
  3. the court can and will accurately determine the facts and correctly apply the law.

If the defendant believes the criminal justice chain cannot reliably perform each of these tasks for whatever reason (lack of capacity, scarcity of resources, corruption, lengthy delay), there no reason for the defendant to accept a plea bargain or an out-of-court settlement. A trial ending in an acquittal is the preferred outcome.

Likewise, if the defendant is 100 percent certain that a trial will result in a conviction, it is to the defendant’s advantage to agree to any settlement that results in a punishment less than what the court would impose were the case tried.  Indeed, it would even pay to agree to accept the punishment the court would apply upon conviction to save on legal fees, the greater adverse publicity a trial could produce, and the other consequences of a trial resulting in a finding of guilt.

While a defendant’s decision to settle depends solely on minimizing the expected punishment (meaning all adverse effects from prisons and fines to harm to reputation and the like), the prosecution’s decision depends upon whether settlement will further the public interest through punishment and deterrence.  The first factor is how confident prosecutors are in the case.  Suppose the prosecution is 100 percent certain that a trial would result in a conviction.  If it is, it has no reason to agree to any settlement that would result in less than what the court would impose at the end of trial.  So, as a first approximation, if prosecutors are certain the outcome of a trial would be a fine of $10,000, they will not accept a settlement offer less than $10,000.

But the prosecution must consider two other factors: the costs of taking the case to trial and the effect of any settlement on the behavior of other firms.

Take the cost issue first.  Trials, particularly of complex corporate crimes like bribery, are costly.  Lawyers and investigators must put in long hours over weeks if not months; witnesses and documentary evidence located and brought to court; and finally, there is the court’s time spent on the case.  If the case settles short of a full trial, these resources can be used to pursue other cases.  Settlement thus further the public interest by freeing up resources for use elsewhere.

Settlement Calculations I:Prosecution
Trial outcome $10,000
Cost to try case $1,000
Government net $9,000
Settlement $9,200
Government cost $0
Government net $9,200
Settlement $9,000.01
Government cost $0.0
Government net $0 .01

So how close must the settlement be to the trial outcome for settlement to be in the public interest?  Take an example shown in the table above where the outcome of the trial would be a $10,000 fine and the cost of trial would be $1,000.  Using these numbers, the government would realize $9,000 after expenses.  Now suppose the defendant settles for $9,200.  Assuming for simplicity it costs the prosecution nothing to settle, as the table above shows, the government’s net would be $9,200. Hence, the government comes out ahead $200 by settling for less than the full amount it would win at court.  Indeed, as the table also shows, a settlement of $9,000.01 would still be in the public interest for the government would realize one cent more than it would realize if a trial were held.

But this calculation ignores a second, critical factor: the effect a settlement would have on the behavior of others.  To see this effect, one fact needs to be added to the example and one changed.  The added fact: suppose the charge is bribery and the defendant made $9,000 by paying the bribe. The changed fact: the prosecution’s cost of trying the case is $2,000. The second table looks at the settlement calculation from the defendant’s point of view with these facts.

If paying a bribe (which for simplicity we will assume was very small) produces a return of $9,000, it would be in the defendant’s interest to settle for $8,500, for it would be $500 ahead after i) paying the bribe, and ii) settling the case.  It would also net the government $500, for with a $2,000 trial cost, it nets only $8,000 by going to trial and winning a $10,000 fine.

Settlement Calculations II: Defendants
Benefit bribery $9,000
Settlement $8,500
New benefit $500
Benefit bribery $9,000.00
Settlement $9,000.00
Net corp. benefit $0.0
Benefit bribery $9,000.00
Settlement $8,999.99
Net corp. benefit $0.01

Yet despite the savings the government would realize from this settlement, it is not in the public interest.  Paying a bribe that returns a profit of $9,000 and only having pay $8,500 when prosecuted would not discourage bribery. The prosecution should therefore refuse to settle for this amount even though focusing on costs alone it would be better to settle than try the case.  Taking account of the benefit to the defendant changes the calculation for the prosecution.  Here, settlement would only advance the public interest if it were equal to or above $9,000.   Any lower amount and the defendant comes out ahead, even a settlement of $8,999.99, which would save government just less than $1,000, leaves it better off.

While the examples illustrate the factors both the defendant and the prosecution must consider when deciding whether to agree to settle the case or go to trial, they simplify the calculus in one critical, all-important respect.  The examples assume that:

  1. the outcome of the trial, conviction with a fine of $10,000,
  2. the government’s cost to litigate the case, and
  3. the deterrent effect of a fine or settlement

are all known with absolute certainty.  None of these figures are of course.  Rather, each is based on a judgment of the probabilities.

Settlement Calculations III:  Uncertain Outcomes
Trial outcome $7,500
Cost to try case $1,000
Government net $6,500
Settlement $7,499.00
Government cost $0
Government net $7,499.00
Settlement $6,500.01
Government cost $0.0
Government net $0 .01

To see how this further complicates matters, take the example above where, instead of being 100 percent certain the trial would result in a fine of $10,000, the prosecution believes it is only 75 percent likely (thus a 25 percent chance of acquittal).  Using the standard conventions of probability, the amount the government would expect to receive from a trial is now $7,500.  The table shows the calculations if the trial cost remains $1,000.  A trial nets the government $6,500. Any settlement between $7,499.99 and $6,500.01 now leaves it better off than if it tries the case. (Benefit to defendant is left out for simplicity.)

In the real world, it is not only the government which cannot be sure of the outcome of a trial.  The defendant cannot be sure either, and thus its calculations of the outcome of the trial and whether it will have to pay a $10,000 fine are also probabilistic.  There are three possibilities:

  1. the defendant’s estimate of the probability of conviction equals the prosecution’s;
  2.  it is lower, for example the defendant believes it is only 10 percent likely it will be fined $10,000 if trial is had, and
  3. it is higher; the defendant believes (say) it is 90 percent likely it will be fined $10,000 if it elects to go to trial.

If both the prosecution and the defendant think the chances of conviction are 75 percent, ignoring other factors a settlement between $7,499.99 and $6,500.01 is highly likely, for both are better off than if there is a trial.  The government nets more than the expected amount of $6,500 and the defendant pays less than the $7,500 it believes it would have to pay upon conviction.

Settlement odds increase if the defendant believes its chances of prevailing at trial are less than what the government believes.  If it thinks it is 90 percent likely it will be fined $10,000 at trial, its expected penalty is $9,000 (90% x $10,000) and it will thus be willing to pay as much as $9,000 in settlement, increasing the range of possible settlements from the lowest the government would accept, $6,500.01, to $9,000, the maximum to which the defendant would agree.

On the other hand, if the defendant believes it only 10 percent likely it will have to pay a fine, settlement will not occur.  The defendant’s maximum expected lost is $1,000 (10% x $10,000 fine), and a settlement is only attractive to it if the amount is $1,000 or less, a number far lower than the prosecution’s calculations of a settlement that would be advantageous to it.  (Parenthetically, one explanation for the high percentage of settlements in the United States are extensive pre-trial investigations in both criminal and non-criminal cases.  As a result, lawyers on both sides of the cases tend to converge on the likelihood of success at trial.)

Research in the United States shows that models similar to that sketched here can predict whether a case will settle or go to trial.  But only roughly, for as is obvious from the discussion above, these models assume away many factors that in a real case affect the outcome.  One is whether the prosecution expects to gain something from a settlement, say information about other wrongdoers.  A settlement that under the calculations above would leave a defendant profiting handsomely might still be in the public interest if it led to many other cases.  An especially confounding issue is the amount of the settlement necessary to deter future wrongdoing.  One research paper suggests that in corruption cases fines must be far greater.  Even more challenging is deciding the amount required to deter corporate criminality.  The extant research offers little guidance.

A settlement may either advance or retard the public interest.  It can free up resources to allow more cases to be brought or it can simply let the guilty off easy.  Judging it requires a careful evaluation.  This summary is meant to help in that process.  It draws from work by American scholars as well as the author’s personal experience negotiating settlements.  Reader comments are warmly solicited.

Sources upon which this note draws:

Elder, Harold W. “Trials and Settlements in the Criminal Courts: An Empirical Analysis of Dispositions and Sentencing,” Journal of Legal Studies 18(1): 191 – 208, Jan. 1989.

Landes, William M. “An Economic Analysis of the Courts,” Journal of Legal Studies 14(1): 61 – 107, Apr. 1971.

Levi, Michael and Nicholas Lord. “White-Collar and Corporate Crime,” chapter 32 in Alison Liebling, Shadd Maruna, and Lesley McAra, eds., Oxford Handbook of Criminology, 6th ed., Oxford: Oxford University Press, 2017.

Schelling, Thomas C. The Strategy of Conflict. Oxford: Oxford University Press, 1963.

Shavell, Steven. Foundations of Economic Analysis of Law. Cambridge: Harvard University Press, 2004.

Simpson, Sally S, Melissa Rorie, Mariel Alper, Natalie Schell-Busey, William S. Laufer and N. Craig Smith. Corporate Crime Deterrence: A Systematic Review. Oslo: Campbell Systematic Reviews 2014:4,


3 thoughts on “Analyzing Settlements in Corruption Cases: A Primer

  1. Thank you for this interesting post. A couple of thoughts…..
    1. You have not mentioned declinations. These are as important as NPA’s and DPA’s especially in view of the new presumption in favour of a company that satisfies the standards set by the DOJ under the new FCPA Corporate Enforcement Policy (since November 2017). 

    2. ‘The sine qua non of a settlement policy that advances the public interest  etc…..A criminal defendant will only bargain away the right to a trial if he or she (or “it” in the case of a corporate entity) fears conviction if the case goes to trial.’  – My view on this is that this is too narrow a lens. Also important for the advancement of the public interest is a means of ensuring that laws are actually enforced especially where the machinery of enforcement has been compromised by corruption. If the case never even gets to trial or, even where it does, it gets stuck in a corrupt process, there is equally nothing to fear by the alleged offender. A lot will depend on the context of this argument. In a country with well developed institutions and good governance, yes your summation is quite relevant. However in a society where corruption has come institutionalized, settlements (with no trial, no conviction) may provide a way of bypassing compromised institutions. This is also in the public interest.

    3. ‘If the defendant believes the criminal justice chain cannot …….A trial ending in an acquittal is the preferred outcome.'(?)  I think what you mean here is that the defendant will take its chances on the trial process.

    4. ‘Likewise, if the defendant is 100 percent certain that a trial will result in a conviction, it is to the defendant’s advantage to agree to any settlement that results in a punishment less than what the court would impose were the case tried……adverse publicity… etc.’.   It is also important from the perspective on not being debarred from future public contracting. This is worth emphasizing even if it falls under ‘other’ because it is an important consideration for corporations.

    5. ‘If the case settles short of a full trial, these resources can be used to pursue other cases …. ‘ Plus, the deep pockets of the corporation are happy to appeal the decision  all the way to the top, if need be, adding to the cost issue, as well as, the fact that the time between the bribery transaction and the final decision (after all appeals have been exhausted and IF there is a conviction) can be more than a decade in some jurisdictions.

    6. ‘So how close must the settlement be to the trial outcome for settlement to be in the public interest?  etc ….’ I find this too narrow a band of factors in determining cost and gains. The fact that settlements are based on prior behaviour, including the introduction of an effective compliance program, voluntary disclosure and a very high level of co-operation that comes at great cost to the corporation (and before any discovery of the alleged violation is made by the authorities) is not included. The collateral cost (damage) of criminal conviction of the corporation on innocent third parties, employees, shareholders, the general public is a cost not factored into this summation. The cost of the investigative work being quasi-privatized and executed by the company itself also not factored in.

    This cost/gain summation does not, in my view, reflect these broader aspects of settlements. 

    It’s a good discussion to be having. Thank you for this,

    • Thanks very much for your thoughtful comments. They helped a great deal in revising a paper on settlements and developing states that I will post Wednesday. In the meantime, in some thoughts on your comments.

      1. For analytical purposes I think it safe to consider a “declination” (a decision to decline to prosecute a case) a type of a non-prosecution agreement. I realize that in some jurisdictions (the U.S. for one), they are slightly different animals.

      2. But in societies where corruption has become institutionalized, won’t the courts or the prosecution or the police, some or all, be compromised too? And if one or more is, what confidence do you have that a trial will produce a conviction? Indeed, isn’t it almost certain to result in acquittal? And won’t a string of acquittals lead to more corruption as others see they can get away with it?

      3. Yes, and in particularly weak criminal justice systems the defendant isn’t taking a chance. Acquittal, or its equivalent – decades long delay — is a certainty.

      4. I should have made clear that I was using the term “punishment” to include all the adverse consequences of a conviction. I have in the revised paper.

      5. Good point but in the U.S. appeals are more common in cases of individuals. Going to jail frightens them (one fainted in our office when told jail was a possibility) and they will fight to the bitter end. Fining a corporation only harms the shareholders and employees (in most cases) so the execs are not willing to put up a fight. And by the time the case has gone through a trial, the lawyers have 1) earned enormous fees and 2) both the exec and the lawyers are sick of the matter.

      6. These factors can all be added in; the same principles apply when they are included but it makes the presentation more complex and I think obscures a critical point: a settlement can be in the public interest and still benefit both the prosecution and the defendant. So long of course as the settlement contains, in the words of the OECD Antibribery Convention, “effective, proportionate and dissuasive criminal penalties.”

  2. Pingback: Analyzing Settlements in Corruption Cases: A Primer | Matthews' Blog

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.