Stolt Nielsen v.

United States:
The Justice Department’s Unconstitutional Pardons

VOLUNTARY TRADE REPORTS NO. 5

NOVEMBER 2005

Introduction

On November 22, 2002, the Wall Street Journal published an article describing a lawsuit
filed by Paul O’Brien against his former company, Stolt-Nielsen Transportation Group.
O’Brien had resigned as Stolt’s general counsel eight months earlier, and he was now alleging
wrongful termination. O’Brien said he was forced to resign after he told Stolt’s chief
executive, Samuel Cooperman, about possible antitrust violations at the company’s parcel
tanker business. O’Brien believed that Richard Wingfield, a senior Stolt executive, had
illegally shared customer information with two of Stolt’s competitors. Cooperman told
O’Brien that he would deal with the situation by strengthening Stolt’s internal antitrust
compliance program. But less than a month after he brought his concerns to Cooperman,
O’Brien was out as general counsel, and by November 2002, O’Brien believed that
Cooperman was an active participant in ongoing antitrust violations.

Within hours of the Journal article’s publication, the Department of Justice’s Antitrust
Division opened a criminal investigation of the parcel tanker industry. At that same time,
Cooperman and other Stolt executives met with John Nannes, a former senior official (and
acting chief) at the Antitrust Division, to solicit his advice on how to deal with O’Brien’s
charges and the DOJ investigation. After this initial meeting, Nannes did not know whether
Stolt had committed any criminal antitrust violations, but he believed that the company could
seek amnesty from prosecution under the Antitrust Division’s “Corporate Leniency Policy,”
which allows the first member of an alleged cartel (that approaches the government) to
receive immunity in exchange for incriminating the other members.

Attorneys like John Nannes are particularly valuable today, because criminal antitrust
investigations are more frequent and result in more severe sanctions. Since the DOJ revised
its initial Corporate Leniency Policy in the early 1990s, companies like Stolt increasingly find

Voluntary Trade Reports (ISSN applied for), number 5, is published biweekly by Citizens for Voluntary Trade
d/b/a The Voluntary Trade Council, Post Office Box 100073, Arlington, VA 22210. S.M. Oliva, editor.

©2005 by The Voluntary Trade Council. All rights reserved. This publication may be freely copied and
distributed, with attribution to the Voluntary Trade Council as author, for non-commercial purposes. For
commercial reprint permission, contact the Voluntary Trade Council, (703) 740-8309 or
info@voluntarytrade.org. Visit the Voluntary Trade Council’s website at www.voluntarytrade.org.
 Voluntary Trade Reports No. 5

themselves in what DOJ officials approvingly call “the race to the prosecutor.” The
Antitrust Division places a premium on maximizing guilty pleas and minimizing trials.

This approach, of course, has become common throughout criminal prosecutors’

offices, but the Antitrust Division’s particular method—the Corporate Leniency Policy
(hereinafter ‘CLP’)—has operated without much in the way of legislative, judicial, or
public scrutiny. The CLP, however, raises serious constitutional, ethical, and economic
questions that need to be addressed. This report will discuss these issues, using Stolt-
Nielsen’s case as an example of the CLP and its many troublesome aspects.

The Stolt-Nielsen Investigation

Two days after his first meeting with Stolt executives, John Nannes contacted James
Griffin, the head of criminal enforcement at the Antitrust Division. Griffin told Nannes
that an investigation of Stolt and its competitors had commenced. The two men,
onetime colleagues, then “exchanged a series of exploratory telephone calls during
which they shared vague pieces of information about possible collusive activity in the
[parcel tanker] industry,” according to court documents. Nannes apparently did not
disclose the identity of his client to the DOJ, only that he was representing a potential
target of the investigation. Based on these conversations, Nannes believed that the
Antitrust Division had not yet awarded amnesty to any of the other companies then
under investigation. Nannes also believed that Paul O’Brien’s termination and
subsequent lawsuit might impede Stolt’s ability to seek and receive amnesty, because
the CLP requires an eligible company’s management to cease any illegal antitrust
activity immediately upon discovery and promptly contact the DOJ.

On December 4, 2002, 12 days after the Antitrust Division opened its investigation,
Nannes held a formal meeting with Griffin and other DOJ attorneys. By this time the
DOJ knew that Nannes represented Stolt. Nannes was told, in no uncertain terms, that if
O’Brien had been fired for exposing Stolt’s antitrust violations, the company would be
ineligible for amnesty under the CLP. Nannes restated Cooperman’s earlier claim to
O’Brien that Stolt had already reformed its internal compliance policy in response to
O’Brien’s concerns. Nannes did not, however, represent that Stolt had ceased
participation in illegal antitrust activity in March 2002 or any other specific date.
Indeed, Nannes himself had not yet conducted an internal investigation to determine
whether Stolt had ever violated the antitrust laws. Instead, his primary objective on
December 4 was to reserve Stolt’s place at the head of the amnesty line, not ascertain the
facts or his client’s actual culpability.

-2-
 Voluntary Trade Reports No. 5

Thirteen days after the first meeting with Nannes, December 17, the Antitrust
Division gave Stolt a “marker,” which established that they were “first in line” to
receive amnesty under the CLP. Only after this marker was received did Nannes initiate
an internal investigation, which was kept secret at the DOJ’s request to prevent leaks to
the media.

Nannes’ investigation revealed that Stolt executive Richard Wingfield had, in fact,
exchanged customer lists with two Stolt competitors—Odfjell Seachem AS and Jo
Tankers B.V.—on at least four occasions between August 1998 and April 2001. Nannes
believed these actions violated the antitrust laws.

The line between legal and illegal conduct here is determined by the degree of
knowledge communicated between competitors. Stolt competes in the parcel tanker
market—it transports industrial liquids in deep-sea vessels. Stolt loses money if its
vessels are idle or empty. The same is true for Stolt’s competitors. This creates an
incentive for all parcel tanker companies to be aware of their competitor’s regular
shipping routes, and in some cases to discourage bidding for the same routes. If Stolt
successfully bids for one of Odfjell’s routes, then Odfjell will have empty space on its
vessels, which may cause Odfjell to bid on one of Stolt’s routes. The question, from an
antitrust perspective, is how the two companies determine what actions are in their
respective economic interests. If Stolt merely anticipated Odfjell’s likely reaction and
decided not to bid on one of its routes, regulators will often excuse this as “conscious
parallelism.” But if Stolt and Odfjell expressly told one another what routes that each
intended to bid on, that is condemned by the government as an illegal “restraint of
trade.”

On January 8, 1993, approximately three weeks after Stolt received its marker from
the Antitrust Division, Nannes made a formal proffer of information in exchange for
amnesty. Nannes told the Antitrust Division about Wingfield’s exchange of customer
lists. The Antitrust Division’s lawyers never asked Nannes about the specific date the
collusion between Stolt and its competitors ceased, and Nannes never offered that
information, assuming he himself knew.

One week after the proffer was made, Stolt and the Justice Department signed a
formal agreement whereby the government would not prosecute Stolt for any antitrust
violation it may have committed prior to that date, January 15, 1993. Stolt agreed to
provide “full, continuing and complete cooperation” with the Antitrust Division’s
ongoing investigation of Odfjell and Jo Tankers. The DOJ said the amnesty would
extend to any Stolt director, officer, or employee who cooperated with the investigation.

-3-
 Voluntary Trade Reports No. 5

Stolt gave the Antitrust Division the information it needed to prosecute Odfjell and
Jo Tankers. In November 2003, Odfjell pled guilty to violating the Sherman Act and
agreed to pay a $42.5 million fine. Jo Tankers entered a similar plea in June 2004 and
paid a $19.5 million fine. Several individual Odfjell and Jo Tankers executives also
entered guilty pleas, and some were sentenced to prison terms of less than one year.

In most cartel investigations, the story would end here. But in April 2003, less than
three months after Stolt signed its agreement with the DOJ, the Antitrust Division
suspended Stolt’s amnesty. (The amnesty was formally revoked on March 2, 2004.) The
Antitrust Division said that it had obtained new evidence—from a source that has never
been publicly identified—that proved Stolt did not cease its illegal activities until
October 2002, as opposed to March 2002, when Paul O’Brien first approached Samuel
Cooperman. As discussed above, however, neither John Nannes nor any Stolt executive
represented a specific date or time line to the Antitrust Division during amnesty
discussions. And the amnesty agreement covered Stolt’s actions before the date of the
agreement—January 15, 2003—not the date that its alleged collusion with Odfjell and Jo
Tankers ceased.

The Litigation

On June 24, 2003, Richard Wingfield was arrested and charged with criminal
antitrust violations. In February 2004, as the Antitrust Division prepared to indict Stolt
itself, the company filed a lawsuit in the U.S. District Court for the District of Delaware,
seeking an injunction compelling the government to honor its amnesty agreement. On
January 14, 2005, District Judge Timothy J. Savage ruled in Stolt’s favor, and enjoined
the DOJ from proceeding with its indictment. (The decision was reported at 352 F.
Supp. 553.)

Savage said the amnesty agreement was a binding contract. Because Stolt “waived
its constitutional rights and exposed itself to criminal liability,” due process required a
“judicial determination” of whether one party breached the agreement. In other words,
the DOJ could not unilaterally “suspend” or “revoke” amnesty without an independent
court finding that Stolt had made material misrepresentations or failed to live up to its
end of the bargain.

Savage rejected the DOJ’s view that Stolt could only seek relief after it was indicted
—that is, file a motion to dismiss rather than preempt the indictment altogether. Savage
said the indictment alone would irreparably damage Stolt’s reputation, and that the

-4-
 Voluntary Trade Reports No. 5

government’s interest would not be “significantly compromised” by dealing with

Stolt’s objections sooner rather than later.

Regarding the Antitrust Division’s reason for revoking amnesty—the dispute over
the date when Stolt ceased its illegal activity—Savage said that the DOJ, “as author of
the immunity agreement, bears the burden of precisely drafting it to reduce the chance
of ambiguity and confusion. Its inartful drafting cannot inure to its own benefit and to
[Stolt’s] detriment.” (Citations omitted.) Savage said the DOJ could not rely solely on its
own “tacit understanding” of Stolt’s representations when the agreement did not state
the matter with certainty. Savage said the DOJ could have simply made the amnesty
applicable before the date the illegal activity ceased rather than before the date the
agreement was signed.

The DOJ appealed Savage’s decision to the U.S. Court of Appeals for the Third
Circuit. On September 30, 2005, oral arguments were held before a three-judge panel
composed of Third Circuit Judges Samuel Alito, Jr., and Thomas Ambro, and James
Restani, chief judge of the U.S. Court of International Trade. The panel may issue a
decision at any time, although it is possible that only Judges Ambro and Restani will
participate. On October 31, President Bush nominated Judge Alito to be an associate
justice of the Supreme Court of the United States. He is expected to be confirmed in
January 2006.1

The Corporate Leniency Policy

The CLP is not authorized by the antitrust laws or any other act of Congress.
Rather, the CLP was the invention of former Antitrust Division head John Shenefield. In
1978, Shenefield simply announced that amnesty would be available to a company that
approached the DOJ before an investigation was underway. Amnesty would not be
granted automatically, but rather at the discretion of the Antitrust Division. This policy
did not produce a windfall of new prosecutions: From 1978 to 1993, only 17 companies
applied for amnesty, and only ten of those applications were granted, less than one per
year.

1 If the panel issues a decision before Judge Alito is confirmed, he could still participate. Since the
Justice Department has sought expedited review of this case, such a quick decision would not be out
of the question. Also, the losing party before the Third Circuit could seek further review in the
Supreme Court, in which case Judge (then Justice) Alito would face the decision of whether to recuse
himself. Recently appointed Chief Justice John G. Roberts, Jr., has said that he would not participate in
Supreme Court review of cases that he decided as a District of Columbia Circuit judge, and Judge
Alito would presumably follow the same policy, although it’s unclear whether that would include
cases that he heard oral arguments in but did not participate in the written decision.

-5-
 Voluntary Trade Reports No. 5

In August 1993, the Clinton administration’s Antitrust Division chief, Anne

Bingaman, announced a revised policy.2 Under the new CLP, amnesty would be
automatic for a company that approached the Antitrust Division before an investigation
commenced. This amnesty was not subject to the Antitrust Division’s discretion. But
even after an investigation started, a company could still receive amnesty at the
discretion of the Assistant Attorney General for the Antitrust Division, as was the case
with Stolt. In either case, amnesty could only be granted to the first company that
approached the government.

Amnesty agreements are non-public and not filed with any court. Although
amnesty recipients often acknowledge their participation in the CLP in the course of
routine corporate filings, the DOJ maintains total silence, even after the ultimate cartel
investigation is completed.

The revised policy has had the anticipated effect of spurring amnesty applications.
According to Antitrust Division officials, the post-1993 CLP has resulted in an average
of two applications per month, as opposed to the less than one per year under the old
policy. A majority of the DOJ’s criminal antitrust investigations now involve CLP
applicants. Scott Hammond, the Antitrust Division’s current head of criminal
enforcement, said in January 2005 that convictions obtained in CLP-assisted
investigations netted more than $2 billion in fines. Gary Spratling and D. Jarret Arp,
antitrust litigators at the firm of Gibson, Dunn & Crutcher, offered additional statistics
detailing the impact of the post-1993 CLP during a November 2004 presentation to an
international antitrust workshop in Australia:

On August 10, 1993, there were only three sitting grand juries in the
United States investigating suspected international cartel activity; on
August 10, 2003, there were 49. In August 1993 the highest criminal fine to
date relating to cartel conduct was $5 million; in August 2003 it was $500
million. Comparing the numbers of foreign defendants subject to criminal
prosecution in the United States for cartel violations shows similar
striking growth. From the passage of the Sherman Act in 1890 to August
1993, there were only eight foreign corporations and eight foreign
individuals – for a total of 16 – prosecuted for cartel activity in the United
States. By August 2003, however, there had been 72 foreign corporations
and 90 foreign individuals – for a total of 162 foreign defendants – subject
to criminal prosecution by the U.S. DOJ.

2 The text of the CLP is published at http://www.usdoj.gov/atr/public/guidelines/0091.htm.

-6-
 Voluntary Trade Reports No. 5

The dramatic increase in antitrust prosecutions post-1993 has led other countries to
adopt their own amnesty programs. According to Spratling and Arp, as of 2004 the
European Union and 24 nations had adopted, or were in the process of adopting, their
own amnesty or leniency policies modeled on the CLP.

Prosecutors: Discretion vs. License

The Antitrust Division considers the CLP a valid exercise of its “prosecutorial
discretion.” Recent administrations of both parties have claimed wide discretion in
enforcing the antitrust laws. In cases settled by consent judgments—which is more than
95% of all civil and criminal antitrust prosecutions—the Antitrust Division claims
absolute prosecutorial discretion when deciding to bring charges or establish the terms
of consent judgments.

Earlier this year, the Voluntary Trade Council criticized a consent judgment
between the DOJ and two sardine manufacturers. The Antitrust Division claimed that a
proposed merger of the two companies would violate the Clayton Act. VTC submitted a
public comment that questioned the economic basis of the government’s claims. In
response, the Antitrust Division said that neither the public nor the court overseeing the
consent judgment could challenge the government’s statement or interpretation of the
facts:

Comments challenging the validity of the United States’ case, or alleging

that it should not have been brought, are challenges to the initial exercise
of the United States’ prosecutorial discretion, which are outside the scope
of the [antitrust laws]. The purpose of the Court’s public interest inquiry is
not to evaluate the merits of the United States’ case, or to conduct a de
novo determination of facts and issues, because “[t]he balancing of
competing social and political interests affected by a proposed antitrust
decree must be left, in the first instance, to the discretion of the Attorney
General.” United States v. Western Elec. Co., 993 F.2d 1572, 1577 (D.C. Cir.
1993) (citations omitted). Courts consistently have refused to consider
“contentions going to the merits of the underlying claims and defenses.”
United States v. Bechtel, 648 F.2d 660, 666 (9th Cir. 1981).3

The executive branch views prosecutorial discretion as an essential element of the

constitutional separation of powers: That is, the courts may not exercise an inherently

3 Public Comments and Response on Proposed Final Judgment United States v. Connors Bros. Income
Fund and Bumble Bee Seafoods, LLC, 70 Fed. Reg. 15,886, 15,890 (March 29, 2005).

-7-
 Voluntary Trade Reports No. 5

executive function. But do the courts have to remain silent when the government obtains
a consent judgment based on demonstrably false facts or arguments? The text of the
Constitution demonstrates that the answer is no.

Prosecutorial discretion is not an intrinsic or mystical power akin to the “divine

right” once claimed by absolute monarchs. In the United States, all executive authority
vested in the President and his subordinates originates with Article II of the
Constitution. Section 3 of that article states that the President—and by extension all
executive officers commissioned by him—“shall take Care that the Laws be faithfully
executed.” The critical word in that clause is faithfully, because it contradicts the notion
of unlimited prosecutorial discretion. The clause constrains executive authority in the
same manner that the Necessary and Proper Clause of Article I restricts legislative
discretion. Both clauses require government agents to carry out their duties in
accordance with the entire Constitution, and not read specific enumerated powers so as
to vest unlimited power in the state. Faithfully, in constitutional terms, is a mandate for
the executive to keep its actions in context.

The traditional concept of prosecutorial discretion implies restraint rather than

unbridled activism. There are two basic elements of discretion: First, a prosecution must
arise from objective evidence. The prosecutor must find it necessary to bring a case
based on the evidence he has obtained. Logically, this means a prosecutor should
refrain from bringing charges that he cannot prove with the available evidence.

Second, a prosecution must conform to the entire body of constitutional and

statutory law, rather than reflect a narrow interpretation of a single statute. A
prosecutor must keep context. It is not a valid exercise of prosecutorial discretion to
enforce a statute in a manner that contradicts other laws or, more importantly, the
Constitution itself. Such an abuse of discretion would not be faithfully executing the
laws.

In recent years, both elements of prosecutorial discretion have been rejected by

prosecutors, who increasingly view their function as obtaining the maximum number of
convictions with the minimum number of actual trials. Numerous shortcuts have been
invented to improve judicial efficiency at the expense of individual due process. The
most common shortcut is the plea bargain. While plea bargaining is prevalent in the
federal judicial system, it does not fulfill the second element of valid prosecutorial
discretion—maintaining constitutional context. In a 2003 essay, Timothy Lynch of the
Cato Institute persuasively argued that plea bargaining was unconstitutional, because
prosecutors use sentencing leniency to coerce defendants into admitting guilt. In effect,

-8-
 Voluntary Trade Reports No. 5

the prosecutor is punishing the defendant who exercises his constitutional rights by
subjecting him to a harsher sentence if later convicted. “Plea bargaining rests on the
constitutional fiction that our government does not retaliate against individuals who
wish to exercise their right to trial by jury,” Lynch wrote.4

Paul Craig Roberts, currently of the Independent Institute, wrote in 2001 that plea
bargaining and the “breakdown of the prosecutorial ethic” have fundamentally
corrupted the judicial system:

Plea bargains undermine police investigative work. As few cases go to

trial, police have learned that their evidence is seldom tested in the
courtroom. Carelessness creeps in. The sloppier the investigation, the less
likely is the right person to be apprehended.

Prosecutors have found that plea bargaining greatly increases their

conviction rates and that they can coerce a plea by raising the number and
seriousness of the charges they throw at a defendant. Defendants are
swayed by the cost of a defense and by the realization that conviction at
trial on even one of the charges can carry more severe punishment than a
plea to a lesser charge. The sentencing differential alone is enough to
make plea bargaining coercive.

Plea bargaining originated as a way of meting out punishment in a timely

fashion. But as time passed, plea bargaining corrupted the justice system
by creating a fictional crime in the place of a real one. The practice of
having people admit to what did not happen in order to avoid charges for
what did happen creates a legal culture that, as it develops, eventually
permits prosecutors to bring charges in the absence of crimes.5

Many of Roberts’ objections to plea bargaining also apply to the CLP. As the Stolt
case demonstrates, both prosecutors and defense lawyers were less interested in
obtaining the facts then in achieving a quick resolution favorable to the government.
Stolt’s lawyer, John Nannes, did not conduct an internal investigation to determine
whether his client was guilty of a crime before first racing to the his former colleagues at
the Antitrust Division to secure a promise of amnesty. And that promise was given
without the prosecutors ever bothering to establish an accurate time line of events. The

4 Timothy Lynch, The Case Against Plea Bargaining, REGULATION, Fall 2003, at 21-24.
5 Paul Craig Roberts, From Blackstone to Bentham: Why Wrongful Conviction is on the Rise,
LEWROCKWELL.COM, March 31, 2001, http://www.lewrockwell.com/roberts/roberts4.html.

-9-
 Voluntary Trade Reports No. 5

government wanted incriminating evidence right away, and Stolt wanted to avoid the
expense of a potential trial. Stolt also wanted to avoid a criminal conviction that would
make it easier for the firm’s customers to seek triple damages in prospective lawsuits.6
And without amnesty, Stolt executives, like Richard Wingfield, could be charged
individually, subjecting them to substantial fines and multi-year prison sentences.

The key distinction between CLP-granted amnesty and a plea bargain, however, is
that the latter takes place with judicial sanction. Plea bargains must be reviewed by an
Article III judge, who is free to amend or reject a proposed sentence. A defendant must
also explain to the court’s satisfaction the nature of his role in committing the charged
offense. While judges may treat these duties as pro forma, the mechanism for judicial
review exists. This is not the case with the CLP, which treats amnesty as a purely
executive function.

Timothy Lynch argued that the government retaliates against defendants that reject
pleas and choose to exercise the constitutional rights to a jury trial and against self-
incrimination. That the courts don’t view this as coercive is illogical, Lynch explained,
because the government isn’t allowed to sanction defendants for exercising other due
process rights:

The [Supreme] Court has never proffered a satisfactory explanation with

respect to why the government should not be able to use its sentencing
powers to leverage the waiver of constitutional rights pertaining to the
trial itself. Can federal prosecutors enter into “negotiations” with criminal
defendants with respect to the exercise of their trial rights? For example,
when a person is accused of a crime, he has the option of hiring an
experienced attorney to prepare a legal defense on his behalf or
representing himself without the aid of counsel. Can a prosecutor induce a
defendant into waiving his right to the assistance of counsel with a
recommendation for leniency in the event of a conviction? Such
prosecutorial tactics are presently unheard of. And yet, under the
rationale of [Supreme Court precedent], it is not obvious why such tactics
should be constitutionally barred.7

The common theme described Roberts and Lynch, and demonstrated by VTC’s
recent experience with the DOJ, is that the constitutional principle of prosecutorial

6 A final judgment in a civil or criminal antitrust case is prima facie evidence of guilt in subsequent
antitrust proceedings. See 15 U.S.C. § 16.
7 Lynch, at 26.

- 10 -
 Voluntary Trade Reports No. 5

discretion has been replaced with the far-broader concept of prosecutorial license.
Prosecutors are no longer expected to conform their actions to objective facts or uphold
the entire Constitution as written. License means that prosecutors can engage in any act
designed to produce a criminal conviction—preferably without an actual trial—so long
as the courts (or perhaps Congress) do not expressly prohibit such conduct. Prosecutors
are no longer expected to restrain themselves, and thus the constitutional notion of
prosecutorial discretion is largely a historical relic.

The CLP, like plea bargains, is an exercise of prosecutorial license rather than
discretion. But unlike plea bargains, the CLP suffers from a different constitutional
defect. After all, Lynch’s objection to plea bargains—they subject defendants to
additional punishment for exercising their rights—is inapplicable to a CLP participant.
An amnesty recipient is not subject to any criminal prosecution or punishment.
Ironically, this total absolution is what renders the CLP unconstitutional.

The CLP and the Pardon Power

In treating CLP amnesties as acts of prosecutorial discretion, the DOJ seeks to

obfuscate the constitutional truth, which is that the Antitrust Division is granting full
(albeit conditional) pardons for alleged criminal acts. In the United States, a pardon is
never a valid exercise of prosecutorial discretion. Rather, it is an exclusive exercise of
presidential discretion.

Article II, Section 2 of the Constitution provides that the President “shall have
Power to grant Reprieves and Pardons for Offences against the United States, except in
Cases of Impeachment.” This is a non-delegable power that is not subject to any
legislative or judicial review. The President may, and does, receive advice from
executive branch officials in exercising the pardon power. The United States Pardon
Attorney, a subordinate of the Attorney General, traditionally reviews applications for
pardons or clemency, and prepares a recommendation for the President.8

Many people think of pardons and clemency in relation to capital offenses, where a
person has been sentenced to death. But the President may pardon a person for any
federal offense committed, even before trial or indictment. For example, in 2001,
President Clinton pardoned Marc Rich, a man who had been indicted for tax evasion
and related crimes. Rich had fled to Switzerland to avoid trial, and remained a fugitive
from the U.S. for nearly 20 years until Clinton issued a pardon the day before he left

8 The pardon power includes the grant of a full pardon and several lesser measures, including
reprieves, commutation of sentence, and remission of fines or restitution.

- 11 -
 Voluntary Trade Reports No. 5

office. Another example—and perhaps the most famous pardon in U.S. history—was
President Ford’s 1974 pardon of his predecessor, Richard Nixon, for all crimes that may
have been committed during his presidency.

Both the Rich and Nixon pardons were publicly criticized, and both Clinton and
Ford were accused of acting out of political self-interest. Nevertheless, neither pardon
could be legally challenged, a reflection of the absolute nature of the constitutional
pardon power.

The DOJ interchangeably refers to acts under the CLP as leniency and amnesty.
However, the former term implies an exercise of prosecutorial discretion—such as
seeking a lesser sentence due to mitigating factors—while the latter suggests an exercise
of the pardon power. Amnesty and pardon have been treated by the Supreme Court as
synonymous terms, as evidenced by Justice Stephen J. Field’s 1877 opinion for the Court
in Knote v. United States:

Some distinction has been made, or attempted to be made, between

pardon and amnesty. It is sometimes said that the latter operates as an
extinction of the offence of which it is the object, causing it to be forgotten,
so far as the public interests are concerned, whilst the former only
operates to remove the penalties of the offence. This distinction is not,
however, recognized in our law. The Constitution does not use the word
’amnesty;’ and, except that the term is generally employed where pardon
is extended to whole classes or communities, instead of individuals, the
distinction between them is one rather of philological interest than of legal
importance.9

Thirty-eight years later, in Burdick v. United States, Justice Joseph McKenna’s

opinion attempted to draw some distinction between amnesties and pardons, as well as
immunity:

This brings us to the differences between legislative immunity and a

pardon. They are substantial. The latter carries an imputation of guilt;
acceptance a confession of it. The former has no such imputation or
confession. It is tantamount to the silence of the witness. It is
noncommittal. It is the unobtrusive act of the law given protection against
a sinister use of his testimony, not like a pardon, requiring him to confess
his guilt in order to avoid a conviction of it.

9 95 U.S. 149, 152-153.

- 12 -
 Voluntary Trade Reports No. 5

It is of little service to assert or deny an analogy between amnesty and

pardon. Mr. Justice Field . . . said that ’the distinction between them is one
rather of philological interest than of legal importance.’ This is so as to
their ultimate effect, but there are incidental differences of importance.
They are of different character and have different purposes. The one
overlooks offense; the other remits punishment. The first is usually
addressed to crimes against the sovereignty of the state, to political
offenses, forgiveness being deemed more expedient for the public welfare
than prosecution and punishment. The second condones infractions of the
peace of the state. Amnesty is usually general, addressed to classes or
even communities,-a legislative act, or under legislation, constitutional or
statutory,-the act of the supreme magistrate. There may or may not be
distinct acts of acceptance. If other rights are dependent upon it and are
asserted, there is affirmative evidence of acceptance. If there be no other
rights, its only purpose is to stay the movement of the law. Its function is
exercised when it overlooks the offense and the offender, leaving both in
oblivion.10 (Citations omitted.)

The facts of Burdick merit recitation. George Burdick was a New York newspaper
editor who was called to testify before a grand jury investigating customs fraud. He
invoked his Fifth Amendment right against self-incrimination and declined to testify.
When recalled before the grand jury, the prosecutor handed him a pardon, signed by
President Wilson, absolving him of any criminal act that he “has committed or may
have committed” in connection with the case. Burdick refused to accept the pardon or
testify. He was held in contempt.

Upon review, a unanimous Supreme Court overruled Burdick’s contempt

conviction. The Court said that a pardon was a private act of the President, but that it
required the acceptance of the recipient, and that the courts could not force such
acceptance. As quoted above, a pardon differed from an amnesty in that the latter term
could be applied to more general acts—of the executive or even the legislature—of
forgiveness that did not necessarily require express acceptance by the beneficiary.

Burdick relied heavily on the Court’s 1833 decision in United States v. Wilson, where
Chief Justice John Marshall offered this explication of the nature and scope of the
pardon power:

10 236 U.S. 79, 94-95.

- 13 -
 Voluntary Trade Reports No. 5

A pardon is an act of grace, proceeding from the power intrusted with the
execution of the laws, which exempts the individual, on whom it is
bestowed, from the punishment the law inflicts for a crime he has
committed. It is the private, though official, act of the executive
magistrate, delivered to the individual for whose benefit it is intended,
and not communicated officially to the court. It is a constituent part of the
judicial system, that the judge sees only with judicial eyes, and knows
nothing respecting any particular case, of which he is not informed
judicially. A private deed, not communicated to him, whatever may be its
character, whether a pardon or release, is totally unknown, and cannot be
acted on. The looseness which would be introduced into judicial
proceedings, would prove fatal to the great principles of justice, if the
judge might notice and act upon facts not brought regularly into the
cause. Such a proceeding, in ordinary cases, would subvert the best
established principles, and overturn those rules which have been settled
by the wisdom of ages.

Is there anything peculiar in a pardon which ought to distinguish it in this

respect from other facts? We know of no legal principle which will sustain
such a distinction. A pardon is a deed, to the validity of which, delivery is
essential, and delivery is not complete, without acceptance. It may then be
rejected by the person to whom it is tendered; and if it be rejected, we
have discovered no power in a court to force it on him. It may be
supposed, that no being condemned to death would reject a pardon; but
the rule must be the same in capital cases and in misdemeanors. A pardon
may be conditional; and the condition may be more objectionable than the
punishment inflicted by the judgment. The pardon may possibly apply to
a different person, or a different crime. It may be absolute or conditional.
It may be controverted by the prosecutor, and must be expounded by the
court.11

All of these cases strongly suggest that the acts of “leniency” or “amnesty” tendered
under the CLP are, in fact, pardons. Consider the Stolt case. Knote and Burdick view
amnesty as something that may be granted to a large class of persons by the executive,
or even the legislature, for reasons of political expediency. Stolt was the target of a
specific DOJ investigation and provided testimony in exchange for remittance of all
criminal liability, a circumstance akin to (though different in outcome from) the pardon
offered in Burdick. Second, amnesty need not be accepted by the recipient under Burdick,

11 32 U.S. (7 Pet.) 150.

- 14 -
 Voluntary Trade Reports No. 5

but the CLP expressly required Stolt to report and confess its wrongdoing as a condition
of receiving “leniency.” Stolt and the DOJ also signed an agreement to codify the
amnesty and its conditions. Finally, Wilson makes it clear that a pardon is a private act
of the executive “delivered to the individual for whose benefit it is intended, and not
communicated officially to the court.” The DOJ’s position in the Stolt case has always
been that CLP acts are private acts of the executive branch that are not subject to any
judicial notice or review. If what Stolt received from the DOJ did not constitute a
pardon, then Wilson and its progeny have failed to properly interpret the pardon clause
of Article II.

Of course, the courts are not wrong here. The Department of Justice is the
malefactor. As Stolt demonstrates, the CLP is nothing more than a pardon scheme run
out of the Antitrust Division without the constitutional component of presidential
control. The Constitution vests the pardon power with the president exclusively. The
CLP claims this same power for the Assistant Attorney General in charge of the
Antitrust Division. Something isn’t right here.

In Federalist 74, Alexander Hamilton explained why the pardon power could not be
constitutionally exercised by any official (or officials) of the government except for the
president:

Humanity and good policy conspire to dictate, that the benign prerogative
of pardoning should be as little as possible fettered or embarrassed. The
criminal code of every country partakes so much of necessary severity,
that without an easy access to exceptions in favor of unfortunate guilt,
justice would wear a countenance too sanguinary and cruel. As the sense
of responsibility is always strongest, in proportion as it is undivided, it
may be inferred that a single man would be most ready to attend to the
force of those motives which might plead for a mitigation of the rigor of
the law, and least apt to yield to considerations which were calculated to
shelter a fit object of its vengeance. The reflection that the fate of a fellow-
creature depended on his sole fiat, would naturally inspire
scrupulousness and caution; the dread of being accused of weakness or
connivance, would beget equal circumspection, though of a different kind.
On the other hand, as men generally derive confidence from their
numbers, they might often encourage each other in an act of obduracy,
and might be less sensible to the apprehension of suspicion or censure for
an injudicious or affected clemency. On these accounts, one man appears

- 15 -
 Voluntary Trade Reports No. 5

to be a more eligible dispenser of the mercy of government, than a body of

men.

The President is detached from the day-to-day operations of the criminal justice
system, which the framers knew would give him a certain amount of objectivity in
weighing the merits of a pardon application. In contrast, Antitrust Division lawyers are
intimately connected with their potential amnesty recipients, and it is the Division
lawyers who stand to politically benefit from amnesty, as they usually gain evidence to
be used in prosecuting other alleged cartel members. Such bureaucratic provincialism is
the opposite of what Hamilton and the other framers intended in granting near-
absolute pardon power to the President.

Under a straightforward textual reading of the Constitution, then, it is impossible

for the Antitrust Division to grant any pardons without the decision and signature of
the President. Accordingly, the agreement executed between the DOJ and Stolt is
unconstitutional. What does that mean, however, for the government’s appeal now
before the Third Circuit?

The Estoppel Solution

There is good reason for the Third Circuit to overturn Judge Savage’s January 2005
decision enjoining the DOJ from proceeding with the indictments of Stolt and Richard
Wingfield. Jim Walden and Kristopher Dawes, attorneys at the firm O’Melveny &
Myers, wrote in March 2005 that Savage’s decision was essentially without basis in case
law:

Until Stolt-Nielsen, courts have generally refused to enjoin federal

prosecutions. “[I]n no case that we have been able to discover,” one circuit
court wrote, “has a federal court enjoined a federal prosecutor’s
investigation or presentment of an indictment.” Indeed, courts have
repeatedly rejected defendants’ attempts to enjoin federal indictments,
holding that a motion to dismiss an indictment under Rule 12(b)(1)
constitutes an adequate remedy at law. Courts have reached this
conclusion even where—as in Stolt-Nielsen—the movant sought to enforce
an immunity agreement with an injunction against an indictment.12

12 Jim Walden & Kristopher Dawes, The Curious Case of Stolt-Nielsen S.A. v. United States, THE ANTITRUST
SOURCE, March 2005, at 7.

- 16 -
 Voluntary Trade Reports No. 5

There is a substantial likelihood that the Third Circuit, with or without Judge Alito,
will follow precedent and find for the government in this case. Any further review by
the Supreme Court is unlikely. And that would be the correct decision if the DOJ had
not incited this case by granting Stolt an unconstitutional pardon in the first place.
Although the arguments raised in this paper have not been brought before the Court of
Appeals, it is important to address the situation now before the court given the
arguments presented above.

The amnesty agreement between Stolt and the DOJ is unconstitutional, but it is not
necessarily unenforceable. As Judge Savage noted, the DOJ “got the benefit of its
bargain” when Stolt provided evidence incriminating itself and other companies in the
alleged price-fixing cartel. In its brief to the Third Circuit, the DOJ disputed this
holding:

The court’s belief that the Division received the benefit of its bargain with
SNTG is irrelevant and wrong. It is irrelevant because the Division
bargained for an agreement that granted leniency only if SNTG satisfied
all of the pre-conditions in its Leniency Policy. In fact, as the Division soon
learned, SNTG was not eligible to receive leniency because its
involvement in the conspiracy continued for many months after the
conspiracy was discovered by O’Brien. And because SNTG was not
eligible to receive leniency, it, along with any of its culpable executives
that the Division should elect to prosecute, should be charged with a
crime and punished in accordance with the law like any other criminal.
Whether or not the Division received useful information from SNTG does
not change the fact that SNTG was ineligible to receive leniency, and it
assumed the risk that if leniency was denied or revoked, “any
documentary or other information provided by SNTG [or any SNTG
executive] . . . may be used against SNTG in any . . . prosecution,“ as the
Agreement expressly provides, and as [John] Nannes expressly
acknowledged.13

The DOJ claims that Stolt acted with unclean hands by deceiving the government
about its CLP eligibility. But it’s far from clear that the government was actually
deceived. The DOJ was aware of Paul O’Brien’s lawsuit in November 2002—indeed, the
government’s investigation commenced upon media reports about the lawsuit—and
from the outset the DOJ should have been suspicious of the timing of Stolt’s decision to
approach the Antitrust Division. The fact that John Nannes failed to commence an

13 http://www.usdoj.gov/atr/cases/f209100/209127.htm#17.

- 17 -
 Voluntary Trade Reports No. 5

internal investigation before securing an amnesty “marker” in December 2002 gave the
government a second clue that Stolt sat on its hands for several months before coming
forward.

That these signs were ignored suggests that the Antitrust Division knew it was
being misled by Nannes during the amnesty negotiations. Such an inference is just as
plausible as the DOJ’s claim that it realized it had been misled only after the agreement
was signed and new evidence came to the government’s attention. (It’s obviously
difficult to know for sure without seeing this evidence, which has been kept from the
public, though Judge Savage was clearly not persuaded by it.) The DOJ, therefore,
cannot claim to be a victim of Stolt’s deception, because the prosecutors were likely
aware of the deception and proceeded to grant amnesty (however unconstitutional)
with the intent of revoking it after Stolt provided useful evidence. Put another way, the
DOJ and Stolt may have lied to each other in order to have their cake and eat it too—
Stolt wanted an amnesty that it knew it wasn’t eligible for, the DOJ wanted easy access
to evidence without having to give up its ability to later coerce guilty pleas from Stolt
and its executives.

The basic issue here is whether Stolt’s transfer of self-incriminating evidence to the
Antitrust Division triggered any reciprocal duty on the government’s part. The DOJ
says that it did not. But the common-law principle of estoppel suggests otherwise.
Consider this explanation of estoppel in a 1992 essay by attorney N. Stephan Kinsella:

Estoppel is a principle of equity or justice which is invoked by a judge to

prevent, or estop, a party from making a certain claim, if the party’s prior
actions are in some sense inconsistent with making such a claim, and if
another relied on such prior actions to his detriment. For example,
suppose your neighbor hires a painter to paint his house, but the painter
mistakenly comes to your house and starts painting it. You see him doing
this, and realize the mistake the painter has made. But instead of stopping
him and telling him of his mistake, you wave at the painter and allow him
to finish, hoping, to get a free paint job. Later the painter asks you to pay
him. You refuse; he sues you for the price of the paint job. As a defense,
you claim that you did not have a contract with the painter, which is true.
At this point, however, the judge might say that you are estopped from
making such a claim (that you did not have a contract), because it is
inconsistent with your prior action (of letting the painter continue
painting your house), and because the painter in good faith relied on your
actions, to his detriment. You ’’will not be heard“ to claim there was no

- 18 -
 Voluntary Trade Reports No. 5

contract. Since you are prevented, estopped, from urging that defense,
you will lose.14

Stolt was the house painter who made a mistake. The Antitrust Division was the
misidentified homeowner who failed to correct the mistake and instead accepted the
unearned benefit. True, both Stolt and the DOJ intentionally misled one another, but the
estoppel principle still holds. Stolt is entitled to payment for its “paint job” in the form
of immunity from prosecution.

An objection to this proposal is that Stolt would be allowed to evade all criminal
charges while its two alleged co-conspirators have plead guilty and will pay substantial
fines. Does such an outcome uphold the principles of justice? Yes, it does. First, the
inequitable outcome is the ultimate result of prosecutorial misconduct, and it is far
more just to let a potentially guilty defendant off the hook than to reward the
government’s deliberate malfeasance.

Second, the underlying offense in this case is not a genuine crime—that is, a
violation of individual rights—but a political crime that is itself beyond the federal
government’s constitutional jurisdiction. The exchange of customer and bid information
among nominal competitors does not violate the rights of any individual or firm.
Nevertheless, the DOJ condemns such activities as a per se antitrust violation. The Stolt
case involves coordination among parcel tanker firms. Economist Dominick Armentano
has discussed the economic fallacy of per se condemnation in a related market:

The knowledge problem and the discovery principle with respect to price
coordination were exemplified in the needless controversy over the
expiration of the Reed–Bulwinkle Act (1948), which had generally
exempted the trucking industry from Sherman Antitrust Act jurisdiction.
The industry maintained that it required a partial continuation of that
exemption so that industry rate-bureau organizations could legally
continue to coordinate routes and prices for member truckers. Antitrust
enthusiasts maintained, on the other hand, that collective ratemaking
constituted horizontal price collusion, a judgment that would continue to
frustrate the welfare advantages associated with deregulation and
increasing competition in the trucking industry.

Clearly, the only way to discover whether collective ratemaking, on

balance, served shippers or not, was to permit the activity to continue in a

14 N. Stephan Kinsella, Estoppel: A New Justification for Individual Rights, REASON PAPERS 17, at 61-62 (1992).

- 19 -
 Voluntary Trade Reports No. 5

free and unregulated transportation market. Industry rate bureaus

performed a price- and route-coordination function in trucking for many
decades, and it is unclear, under conditions of free entry, why that activity
should be legally restricted or prohibited. If certain carriers employ rate-
bureau services and achieve efficiencies, then those carriers may gain
business and market-share relative to other carriers that price and route
independently. If, on the other hand, the costs to shippers exceed the
value provided by price coordination, the carriers may lose business and
market-share to carriers that price independently. In either case, there is
no ex ante logic for a per se prohibition.15

Because one member of an alleged cartel often receives amnesty under the CLP,
leading the other members waive their rights to trial and plead guilty, there is little
scrutiny of the economic arguments in favor of so-called “anti-competitive” behavior.
Indeed, the courts manufactured the per se rule to discourage such analysis even when
cases reach trial. But it must not be forgotten that antitrust is about politics, not
economics.

Finally, even without prosecuting Stolt, the DOJ has already achieved the
purported object of its investigation—the destruction of the alleged parcel tanker cartel.
Regardless of the specific date that Stolt ended its collaboration with Odfjell and Jo
Tankers, it did end more than three years ago. Furthermore, even if Stolt receives
immunity from criminal prosecution, it is still subject to civil lawsuits by the alleged
“victims” of the cartel. Proceeding with the indictments of Stolt and Richard Wingfield
could only result in additional criminal fines—removing capital from the private
market to the government’s treasury, which does nothing to improve competition—and
possibly a prison sentence for Wingfield, an act of retribution rather that also does
nothing to improve the market’s effectiveness. Indeed, in a normal cartel investigation,
the DOJ would consider the grant of amnesty to one cartel member, followed by the
conviction of all remaining members, to be an unqualified success. The Stolt case is the
exception to the Antitrust Division’s normative concept of justice.

Conclusion

Stolt-Nielsen’s error was not in collaborating with its competitors, firing its general
counsel, or even “misleading” the Justice Department. Instead, Stolt executives erred in
hiring John Nannes as their representative to the Antitrust Division. Nannes’ role was
not to serve the best interest of Stolt and its shareholders, but to facilitate the need of his

15 Dominick T. Armentano, ANTITRUST: THE CASE FOR REPEAL 92-93 (revised 2d ed. 1999).

- 20 -
 Voluntary Trade Reports No. 5

former employer, the Antitrust Division, to obtain Stolt’s confession and cooperation
before Nannes himself could ascertain whether any laws had been broken. Nannes is
part of a growing number of antitrust defense lawyers who considers winning the “race
to the prosecutor” a victory, rather than a failure of constitutional due process.

The Corporate Leniency Program reflects the breakdown of both criminal defense
and criminal prosecution. The Antitrust Division has employed the CLP as an
unconstitutional scheme that usurps the President’s authority to grant pardons and
reprieves. In handing out these unconstitutional—and non-public—pardons to
companies like Stolt, the DOJ weakens the rule of law. Modern antitrust enforcement is
concerned with manufacturing guilty pleas and collecting fines, not protecting markets
from the purported dangers posed by cartels. Competition will certainly not benefit
from the ongoing prosecution of Stolt, as it will not realize any benefits from the prior
prosecutions of Odfjell and Jo Tankers. In the end, the only beneficiaries of this
investigation will be the lawyers—both the government’s prosecutors and defense
attorneys like John Nannes, who will get paid regardless of what the Third Circuit
finally decides.

Additional Information

Anne M. Lockner, Amnesty and Immunity: How Contractual Interpretation Will Make or
Break Your Day (November 2005), available at
http://www.rkmc.com/Amnesty_and_Immunity_How_Contractual_Interpretation_W
ill_Make_or_Break_Your_Day.htm.

Jesse W. Markham, Jr., Antitrust Report: Department of Justice Amnesty Program Takes
Another New Turn (April 2005), available at
http://www.mofo.com/news/updates/files/update02014.html.

Letter from S.M. Oliva, Voluntary Trade Council, to Richard L. Huff and Daniel J.
Metcalfe, U.S. Dept. of Justice (October 7, 2005) (on file with author), available at
http://www.voluntarytrade.org/downloads/5I03_Appeal.pdf (Freedom of
Information Act appeal seeking access to Antitrust Division amnesty agreement with
Micron Technology, Inc.)

- 21 -