Antitrust Division Update: Trinko And Microsoft

Remarks by

J. Bruce McDonald
Deputy Assistant Attorney General
Antitrust Division
U.S. Department of Justice




Before the
Houston Bar Association
Antitrust and Trade Regulation Section

April 8, 2004

I always wondered what it was like on this side of the Houston Bar Association podium. After all those years begging other folks to stand up here, it must be payback time. Thanks to Alistair Dawson for the invitation.

This group has heard from its share of officials. Sean Royall, another Baker Botts alumnus and recently of the Federal Trade Commission, was here just a few weeks ago, telling tall tales about our sister antitrust enforcement agency. And Alison Smith can't stop talking about the DOJ in the Bush 41 administration.

Having heard from many officials over many years, you know that no set of remarks gets past the Justice Department censors without a few pages on increased criminal fines, merger statistics, and promises of reinvigorated enforcement of the Robinson-Patman Act. So I have a report on the Division's recent work. Then I'd like to talk about the Supreme Court's recent Trinko case and close with some remarks about Microsoft.

Criminal Enforcement

The Antitrust Division's highest priority is and will remain challenging criminal cartels. The Supreme Court in Trinko called collusion the "supreme evil" of antitrust, so we figure we're on the right track. The criminal enforcement program is indeed busy. Most of the statistics are up. We have record or near record numbers of grand juries open in domestic and international cartel cases, some of which involve very significant cartel activity. To take just one example, the recent Crompton plea involved a $50 million fine for participation in an international rubber chemicals cartel, and is merely the first case to come forward from an investigation that is likely to lead to others.

Prison sentences are the most powerful component of deterrence in cartel enforcement. Tell your clients, "You can go to jail." Both jail time and money fines are likely on their way up. Legislation is making its way through Congress that will increase criminal fines for corporations to $100 million and for individuals to $10 million. Maximum jail time increases to 10 years, bringing the antitrust penalties more in line with increased penalties for other white collar crimes. This is not behavior you want to encourage.

The Antitrust Division's leniency program remains the best method of cartel detection. Most of our criminal cases are developed directly or indirectly from a firm that sought leniency. The message of the program is, we'll go easy on you if you cooperate. The same federal legislation I just mentioned will make cooperation more attractive. A "de-trebling" provision limits damages in a follow-on civil action against a leniency program participant, arising out of the conduct for which it received leniency, to actual damages, as opposed to treble damages.

The cooperation requirement of the leniency program is serious. The Division recently kicked one participant out of the program for failing to abide by its terms. The Division also recently released a statement on the program's requirements.

Press Release, Justice Department Statement Regarding Antitrust Division's Corporate Leniency Program (Mar. 22, 2004).

The outcome of the Hoffman-La Roche v. Empagran case is important to criminal enforcement and the leniency program. Empagran involves the Foreign Trade Antitrust Improvements Act, which provides that the Sherman Act applies to non-import foreign conduct only if that conduct has "a direct, substantial, and reasonably foreseeable effect" on U.S. commerce and "such effect gives rise to a claim" under the Sherman Act. The plaintiffs are foreign corporations that were injured by the worldwide vitamins price fixing conspiracies. Those conspiracies certainly gave rise to some Sherman Act claims, but not to these plaintiffs' claims. These plaintiffs were injured as a result of the conspiracies' effects on markets outside the U.S. The question for the Supreme Court is whether the FTAIA requires that the plaintiff's claim arise from the U.S. effect of the anticompetitive conduct.

The United States will appear as an amicus in the upcoming Supreme Court argument. The United States' position is that the plaintiff's injury must result from the domestic effect of the conspiracy, and here we agree with the Fifth Circuit and disagree with the Second. One of our policy arguments is that foreign defendants are less likely to come forward in the Antitrust Division's leniency program if their confession may lead to paying damages in U.S. courts not only to plaintiffs injured in the U.S. but also plaintiffs injured anywhere.

Merger Enforcement

Turning to merger enforcement, the long-awaited uptick in merger filings has begun to tick up. As we approach the halfway point of FY2004, we've thus far seen 648 HSR transactions (1,273 filings), compared to 1,014 transactions in all of FY03 and 1,187 in FY02. So the clear trend is in favor of at least modest growth.

U.S. v. Oracle. Right now the biggest merger matter at the Division is the challenge to Oracle's attempted hostile takeover of PeopleSoft, two of the three players in the market for enterprise software. I will resist the temptation to argue the merits of our case here, and simply say that we very much look forward to the opportunity to present to the district court the evidence that caused us to believe this merger would lessen competition.

The Oracle judge is Hon. Vaughn R. Walker in the Northern District of California, familiar to some here from California electricity cases. He has put the parties on a schedule that would make the Galveston Division envious -- just over three months from filing to trial, scheduled for June 7.

The European Commission also is reviewing the Oracle/PeopleSoft transaction, and its decision is expected on May 11, unless it stops its clock to gather more information.

First Data/Concord. The proposed acquisition by First Data of Concord would have combined the two largest PIN debit networks, First Data's NYCE and Concord's STAR. These provide the network backbone for merchants to accept PIN-authorized debit cards. We argued that this was a relevant market separate from credit cards, signature debit cards, checks, cash, etc. The Division challenged the First Data/Concord merger, and it settled on the eve of trial in December on the same terms that the Division would have sought if successful at trial. Divestiture of First Data's NYCE network will ensure that merchants will continue to benefit from the existing competition between Concord's STAR and First Data's NYCE PIN debit networks.

U.S. v. Southern Belle Dairy. Another merger in litigation is U.S. v. Southern Belle Dairy, which challenges the already-completed merger of dairies in Kentucky, a transaction that was below the HSR $50 million threshold. Although the two dairies will be only partly owned by the same parent, Dairy Farmers of America, these dairies have a history of colluding (bid rigging under prior management) and coordinated effects are predictable. One of the defendant's arguments is that the Government is estopped from challenging a partial acquisition because it did not challenge another dairy partial acquisition a few years ago, in another market, and the DOJ issued a press release saying so. The Division's motion for summary judgment on that affirmative defense is pending.

Labelstock. Last summer the Division won at trial the action against UPM and MACtac, North America's #2 and #3 makers of labelstock, which proposed to merge. Although the two firms together had only about 25% of the North American labelstock market, they shared the market with #1 Avery, which had 70%. The court agreed with our prediction of coordinated effects and enjoined the transaction.

Several other recent merger challenges are worth discussing in great detail, if we had more time. Suffice it to note that there are examples of close cooperation and agreement with the EC, such as in Air France/KLM (airlines), Alcan/Pechiney (aluminum brazing sheet), and General Electric/Instrumentarium (medical equipment).

Civil Non-Merger Enforcement

The last item in my report is civil non-merger enforcement, which we will continue to emphasize before the merger wave hits shore again. Procedurally within the Division we have put in place a system that tries to keep civil non-merger matters on a faster track towards either enforcement or closing, something I am sure will be welcome news to friends in the private bar. We continue to put significant efforts into assuring compliance with our Microsoft consent decree. In the Dentsply case, while we were disappointed not to have prevailed in the district court, we will be pursuing an appeal on our Section 2 claim.

Having made my report on the Antitrust Division's recent work, I'll turn to Section 2 and the Trinko case.

Search for Standards in Section 2 and Verizon v. Trinko

It is not always easy to distinguish between unilateral anticompetitive conduct that violates Section 2 and competitive conduct that benefits consumers. But an important part of U.S. antitrust philosophy requires that we make this distinction. As Justice Scalia said in Trinko, "To safeguard the incentive to innovate, the possession of monopoly power will not be found unlawful unless it is accompanied by an element of anticompetitive conduct." If you believe that antitrust can protect competitive markets, but may not be so good at creating them, then to safeguard innovation, even dominant firms must be allowed to compete aggressively. We cannot demand, in the words of a recent European commentator, that firms compete "like gentlemen." Antitrust enforcement cannot be suspicious of the conduct of any firm, until there is evidence of exclusionary conduct.

Trinko v. Verizon is important because it prunes Section 2, which has grown to be a little unruly.

The "Struggle for Standards"

At its most basic, Trinko is about whether certain regulated conduct violates Sherman Act § 2. We all know that Section 2 requires proof of (1) creation or maintenance of monopoly (2) by anticompetitive means, that is, by exclusionary or predatory conduct. Addressing the conduct element, some courts have suggested tests that examine whether the allegedly exclusionary conduct would be economically irrational for the defendant but for the ultimate effect on competition. Others have considered it relevant that conduct that actually led to creating or maintaining monopoly was also improper for reasons extrinsic to the antitrust laws, such as false advertising or violation of regulatory duties.

But articulating the standard for exclusionary conduct in a useful way has been difficult. This difficulty reflects the first of what Judge Richard Posner describes as the two "struggles" of antitrust, the first being the "struggle for standards." In the Government's amicus brief in Trinko, we drew on precedent to articulate the standard, "[C]onduct is not exclusionary or predatory unless it would make no economic sense for the defendant but for its tendency to eliminate or lessen competition." Although the Court did not explicitly endorse any standard -- despite parading through a variety of monopolization issues -- I think decision has made a positive contribution to this struggle for standards.

Background

Quickly the background. Defendant Verizon is the incumbent local exchange carrier (ILEC) in New York, inheritor of the old Bell monopoly, like SBC here in Houston. The 1996 Telecommunications Act requires that ILECs sell unbundled parts of their local networks at cost to competitive local exchange carriers (CLECs), like Birch Telecom or AT&T. The goal of the Telecom Act is to undercut the ILEC monopolies by having the ILECs share their networks to give new competitors a toehold in the market. In essence, Verizon charges a wholesale price to AT&T, AT&T charges its customers a retail price, and Verizon provides access to its local network and service needed for AT&T's customers to talk on the telephone.

The FCC in an administrative proceeding asserted that Verizon breached its duty under the Telecom Act by inadequately providing service to the customers of its CLEC competitors, one of which is AT&T. Trinko is an AT&T customer in New York. In a follow-on lawsuit, Trinko alleged that Verizon had a scheme to discourage customers from becoming or remaining customers of CLECs like AT&T and claimed that Verizon's breach of its Telecom Act duties constituted a Section 2 violation.⁽¹⁾

The District Court dismissed the complaint for failure to state a claim. The Second Circuit reinstated the complaint in part, holding that Trinko may have stated an antitrust claim for monopolization under Section 2 by alleging that Verizon had not fulfilled its Telecom Act duties to AT&T.⁽²⁾

The Supreme Court's Decision

The basic question for the Supreme Court was whether a complaint that an incumbent had breached its Telecom Act duty to share its network with competitors states a Section 2 claim. (No.) But while reversing the Second Circuit, the Court said a lot more. I'll give you the law school outline of the majority's analysis, written by Justice Scalia, then make some predictions.

First, what effect did the Telecom Act have on the application of antitrust principles? One might argue that, because in the Telecom Act Congress created duties designed to promote competition, then violation of those duties should be enforceable with the competition statute, the Sherman Act. On the other hand, says Justice Scalia, the existence of a detailed regulatory scheme ordinarily suggests implied immunity for conduct covered by the scheme. Here, both arguments were precluded by the Telecom Act's antitrust-specific saving clause.

Second, did Verizon violate Section 2 under existing antitrust standards by failing to share its network with AT&T (adequately)? Section 2 requires exclusionary conduct. Is refusing to share exclusionary? The general rule is that even monopolists may refuse to deal with rivals.

Third, what about an exception to the general rule? Of course, the leading case on refusal to deal liability is Aspen Skiing. Justice Scalia distinguishes Aspen from the Trinko situation, concluding that Trinko does not fit within the Aspen exception and probably narrowing the future use of Aspen, which he described as "at or near the boundary of Section 2 liability."

The facts on which Justice Scalia focuses in distinguishing Aspen from Trinko are instructive in inferring what Section 2 standard the Court would endorse. The most interesting distinguishing factor is that the Aspen defendant's refusal to share -- refusal to continue offering a joint ski lift ticket for both its mountains and its competitor's mountain -- reflected its willingness to forego short term profits, presumably for long term gain with less competition. In contrast, Verizon had not before shared its network with rivals at all, before the Telecom Act required it. Similarly, while the Aspen defendant refused to sell to its competitor even at retail, Verizon's reluctance was to sell at cost. Furthermore, the Verizon product, unbundled elements of its local phone system, did not even exist before the Telecom Act required unbundled sales. All this led the Trinko Court to conclude that it could not follow Aspen to determine that Verizon's conduct was exclusionary.

Justice Scalia also noted that Verizon's conduct also did not fit the essential facilities doctrine more generally, which doctrine the Court again declined to recognize or reject.

Fourth, if the Telecom Act does not expand the Sherman Act, if the general rule is that a monopolist need not share, and if no existing exception applies, then what about a new exception?

Justice Scalia answers this question first by observing that antitrust analysis -- and crafting antitrust exceptions -- must consider the significance of regulation. Where there is regulation that can control the conduct, he asks, why extend antitrust liability to cover that regulated conduct? This makes sense.

Justice Scalia then proceeds to makes some remarkable statements about the value of regulation and the burdens of antitrust. (Keeping this statements in context makes them less startling: Justice Scalia is evaluating whether to make an exception to expand antitrust liability to cover conduct already covered by regulation.) He sings the praises of regulation but nearly condemns antitrust:

• "Where ... '[t]here is nothing built into the regulatory scheme which performs the antitrust function,' . . . the benefits of antitrust are worth its sometimes considerable disadvantages."
• "Under the best of circumstances, applying the requirements of § 2 'can be difficult' . . . Mistaken inferences and the resulting false condemnations 'are especially costly, because they chill the very conduct the antitrust laws are designed to protect.'"
• "'No [antitrust] court should impose a duty to deal that it cannot explain or adequately and reasonably supervise.'"
• "The Sherman Act . . . does not give judges carte blanche to insist that a monopolist alter its way of doing business whenever some other approach might yield greater competition."

Needless to say, the Court held that Trinko's allegations did not justify creating a new antitrust exception, especially because the Telecom Act had created a regulatory structure that specifically addressed Verizon's conduct.

The Court reversed the Second Circuit because Trinko's allegations do not state a Section 2 claim.

The Importance of Trinko

What does Trinko tell us about the standard for exclusionary conduct and the future of Section 2 enforcement? I have several suggestions:

• Trinko focuses on antitrust fundamentals.
• The case does not signal a new Section 2 standard, but implicitly endorses the "but for" test.
• In fact, there may be some retrenchment in Section 2 liability, pruning the tree.
• Complainants in regulated industries may be limited to regulatory remedies without resort to antitrust.

A return to fundamentals. In some ways, the Court's opinion is a return to a more focused application of the antitrust laws. The majority indicated that exceptions to antitrust's general rules are and should remain narrow. Justice Scalia wrote, "We have been very cautious in recognizing such exceptions, because of the uncertain virtue of forced sharing and the difficulty of identifying and remedying anticompetitive conduct by a single firm."⁽³⁾ This is especially so where the exception, such as imposing new obligations that require competitors to share resources, would undercut antitrust's more general goals of promoting competition and preventing collusion⁽⁴⁾ (which Justice Scalia calls the "supreme evil of antitrust"⁽⁵⁾).

Likewise, Justice Scalia says that broadening antitrust liability is less desirable where regulation is available and antitrust enforcement would undercut the regulatory scheme.

I think these tend to keep antitrust focused. Don't undercut or broaden antitrust where unwarranted.

No new standards. Trinko certainly does not announce a new Section 2 standard, and the Court did not explicitly mention the Government's recommended standard. But Trinko has made a contribution to the struggle for standards. The Court's Section 2 analysis focuses on the same facts that suggest a violation under the "but for" standard like that the Government proposed in its amicus brief. Recall the facts on which the Court focused, as it distinguished Aspen: whether the defendant forewent short term profits because it would profit in the long term after competition exited. From this and the fact Verizon won, one can infer the "but for" standard: "[C]onduct is not exclusionary or predatory unless it would make no economic sense for the defendant but for its tendency to eliminate or lessen competition."

By the way, one could read Justice Scalia's language, peppered as it is with colorful references to a defendant's "anticompetitive malice" and "dreams of monopoly,"⁽⁶⁾ as suggesting a standard for Section 2 liability that is based on the defendant's subjective motivation. Such a reading of the opinion probably is misguided. Although evidence of the defendant's intent can help the court determine the likely effect of a defendant's conduct -- as the D.C. Circuit found in Microsoft⁽⁷⁾ and the Supreme Court itself suggested in Aspen⁽⁸⁾ -- the focus should remain on the effect of the defendant's actions and not subjective intent.

Retrenchment? Certainly one may argue from the opinion that the Court has indicated that antitrust liability is not boundless, if not signaled future retrenchment.

First, the Court explicitly refused to endorse or reject the essential facilities doctrine, a signal that Justice Scalia may have had 4 votes but not 5 votes to reject it. But Justice Scalia did have the support to confirm what many suspected, that Aspen is "at or near the outer boundary of Section 2 liability."⁽⁹⁾

The Court also came very close to barring use of the essential facilities doctrine where the court order of sharing facilities would require complex judicial oversight.⁽¹⁰⁾ (When is a sharing order not likely to be complex?) If the doctrine remains viable at all, it is not as wide-ranging a doctrine as some courts have suggested, and it unquestionably will not be expanded much.

Second, in case there was lingering doubt, the Court buried monopoly leveraging as a distinct offense.⁽¹¹⁾ Proof of the mere "use of monopoly power" is not a substitute for exclusionary conduct,⁽¹²⁾ and the usual elements of Section 2 liability still apply: possession of monopoly power in the relevant market, and the wilful acquisition or maintenance of that power as distinguished from growth or development as a consequence of a superior product (etc.).

Regulated industries. The impact of the Court's decision will be most keenly felt in Section 2 enforcement in telecom and other regulated industries.

First, the argument that violation of a regulatory statute like the Telecom Act is ipso facto a Section 2 violation is dead. Both statutes seek competitive markets, but the likeness ends there. The Court observed that the Telecom Act "attempts to eliminate the monopolies enjoyed by the inheritors of AT&T's local franchises," whereas the Sherman Act "seeks merely to prevent unlawful monopolization." These are two quite different goals, and we agree with the Court's conclusion that it would "be a serious mistake to conflate" them.⁽¹³⁾

Of course, Trinko does not eliminate Section 2 liability in regulated industries. On the contrary, a monopolist certainly can violate Section 2 even in an area in which its conduct is regulated. Notwithstanding Trinko, the question is whether the conduct (which happens to violate a regulatory duty) also violates Section 2 according to pre-existing antitrust standards or, unlikely, under an antitrust exception. Proof of a violation of some other statute, by itself, will not and should not establish a violation of the Sherman Act. And neither the Telecom Act nor Trinko lessen the exposure of telecommunications industry participants to antitrust violation.

Second, the Court arguably signaled more ready acceptance of implied immunity for regulated conduct, at least when the regulators have some responsibility for competitive concerns. In Trinko an immunity argument was precluded because the Telecom Act includes an antitrust-specific saving clause. (The Antitrust Division is a chief proponent of antitrust saving clauses in federal legislation.) However, it is difficult to read the tea leaves here -- will the Court make finding implied immunity easier? -- because this was dicta, it was not explicit, and it did not undercut the Court's earlier instructions that antitrust repeals implied by a regulatory statute are strongly disfavored and should be found only when the regulatory scheme otherwise just wouldn't work.

False positives. Finally, it is worth noting the Court's deference to the unilateral business decisions of monopolists, especially on what, to whom, and at what price to sell.

"Mistaken inferences" is part of the Court's traditional concerns, fully discussed in the predatory pricing context of Matsushita and Brooke Group. The Trinko Court warns that "false condemnations" can result from such inferences, which can "chill the very conduct the antitrust laws are designed to protect."⁽¹⁴⁾ Such "false positives," the Court suggested, "counsel[] against an undue expansion of Section 2 liability," especially in an area that are "beyond the practical ability of a judicial tribunal to control."⁽¹⁵⁾

Simply put, courts and antitrust agencies are fallible, and using the antitrust laws as a kind of uber-regulation to force a monopolist in a regulated industry to share its resources with competitors "requires antitrust courts to act as central planners . . . a role for which they are ill-suited."⁽¹⁶⁾

Trinko focuses Section 2 where it can do the most good. This is part of what Judge Posner calls the "struggle for administrability." The Antitrust Division welcomes the Court's stated desire to balance what can sometimes be "the slight benefits of antitrust intervention" with "a realistic assessment of [the] costs" of such intervention.⁽¹⁷⁾ Just as much as regulation, overzealous antitrust enforcement can be anticompetitive.

Microsoft in Europe

Two weeks ago the European Commission announced its decision in its own Microsoft matter, which investigated whether Microsoft has unlawfully tied its Windows Media Player to its dominant Windows Operating System. Although the Division and our Brussels friends have a strong and positive relationship on most matters of antitrust policy, on this particular case we see things very differently. We are concerned about the apparent basis of the EC's chosen remedy and the serious potential divergence it represents.

Since 1996 the Antitrust Division has devoted substantial effort to a Section 2 case aimed at exclusionary conduct by Microsoft. The ultimate outcome reflected not only considerable analysis and enforcement judgments by the United States and many state attorneys general, but also thorough review of those judgments in the U.S. judicial system.

The Final Judgment that emerged from that process protects competition and consumers by preventing affirmative misconduct by Microsoft that would inhibit competition in middleware programs, like internet browsers or media players. Of course, almost every manufacturer of PCs installs Microsoft's Windows Operating System on its products. A key feature of the U.S. settlement is that Microsoft may not prevent PC manufacturers from substituting Microsoft middleware with a third party product nor prevent manufacturers from hiding the Microsoft middleware icon or other access to the Microsoft product; therefore, even if the software for Microsoft middleware is on the PC, perhaps to perform background functions, end users could be aware only of the third party middleware that the manufacturer has installed. The Division believed that this solution was effective and made it possible for rivals to compete for a place on manufacturers' PCs.

In contrast, the EC remedy requires that Microsoft offer for sale in Europe a version of the Windows Operating System that has the software code for Windows Media Player removed. Microsoft may also sell the bundled version. We think that this more aggressive remedy takes antitrust enforcement in the wrong direction.

• First, the decision reflects the fact that the standard for proving harm is much lower in the EU than in U.S. law. The EC is more concerned about false negatives, we about false positives.
• Second, antitrust remedies should do no harm, and the EC's "code removal" remedy may complicate software developers' efforts (and therefore increase costs) to write middleware to be used with the Windows Operating System, because third party programs often rely on the operating system software for certain functions.
• Third, it is a bad idea to impose liability on what may be more a product enhancement than a tie, and it is a very bad idea for antitrust enforcers to try to re-design a firm's product to better promote competition. This can chill innovation.

If the result is that a dominant firm simply cannot improve its product by adding features until that product becomes sufficiently inferior that its dominance is eroded, then the inconsistency with core principles of U.S. antitrust law is plain. Even if a workable standard could be advanced, the potential for anticompetitive and anticonsumer consequences would remain high, both in a U.S. system in which enforcement is amplified through private treble damages litigation and in an EC system where the threshold for finding "dominance" appears much lower. Again, our concern is that while certain competitors may well benefit from intervention, consumers and innovation ultimately may not.

Finally, given this significant prior U.S. effort in developing an effective remedy, it is in our view unfortunate that considerations of international comity and deference did not, in the Commission's judgment, carry sufficient weight to avoid the significant divergence that has now occurred. In a system of multiple enforcers, the divergence leads parties that can benefit from regulatory assistance to seek out the most restrictive regulator, and for global products the effects of that regulator's actions may affect all markets.

Thank you.

Endnotes:

1. Id. at 875-77.

2. Id. at 877.

3.

4. See, e.g., id. at 879 ("We have been very cautious in recognizing such exceptions, because of the uncertain virtue of forced sharing and the difficulty of identifying and remedying anticompetitive conduct by a single firm.").

5. Id.

6. Id.

7. United States v. Microsoft, 253 F.3d 34, 59 (D.C. Cir. 2001).

8. "[E]vidence of intent is . . . relevant to the question of whether the challenged conduct is fairly characterized as 'exclusionary' or 'anticompetitive' . . . or 'predatory.'" Aspen Skiing Co. v. Aspen Highlands Skiing Corp., 472 U.S. 585, 602 (1985).

9. Trinko, 124 S.Ct. At 879.

10. See Trinko, 124 S.Ct. at 883 ("An antitrust court is unlikely to be an effective day-to-day enforcer of . . . detailed sharing obligations.").

11. See Trinko, 124 S.Ct. at 883 n.4 ("The Court of Appeals also thought that respondent's complaint might state a claim under a 'monopoly leveraging' theory . . . . We disagree. To the extent the Court of Appeals dispensed with a requirement that there be a 'dangerous probability of success' in monopolizing a second market, it erred.") (internal citations omitted).

12. See id. ("[L]everaging presupposes anticompetitive conduct, which in this case could only be the refusal-to-deal claim we have rejected.").

13. Id. at 883.

14. Id.

15. Id. at 882-83.

16. Id. at 879.

17. Id. at 882.