Monday, 20 October 2003

Copyright Insurance: Further Discussion

This article continues the discussion on copyright liability insurance, proposed (separately) by P2P Fund and Steven Wu of LawMeme. Ernest Miller posted some criticisms of the idea, to which I responded. Ernest has posted his reply, highlighting the weaknesses of my response.

Ernest points out that the insurance policy I suggested would be unusual: it would cover settlements but not judgments in copyright infringement suits. This may seem bizarre, but it is necessary for two reasons. First, it is an attempt to dodge the required disclosures under Rule 26 of the Federal Rules of Civil Procedure (FRCP). If the policy does not cover judgments, it does not fall within the literal language of the rule. If push comes to shove, a court might see it for what it is (an attempt to dodge the rule) and assess sanctions for failure to disclose. I would not recommend this litigation strategy for most people, but someone interested in setting precedent (EFF? ACLU?) might be interested.

Second, Ernest mentions that most people would want an insurance policy that covers them for judgments as well as settlements. I do not think I argued to the contrary in my article. However, I believe that some people would make this trade-off to make the system workable and to keep prices reasonable. A culpable file sharer would be liable for at least $750 per song after judgment. Since RIAA is currently suing only people who (allegedly) share at least 1,000 copyrighted songs, an adverse judgment on all counts would incur liability of at least $750,000. Any commercially viable insurance product would require high premiums to cover such payouts — especially considering the self-selection problem and risk factors I discussed in my previous article. Few, if any, people would be willing to pay enough for the insurance to make it commercially viable. All that said, RIAA has filed over 450 file-sharing lawsuits this month. It cannot possibly intend to take them all (or even most) to trial and judgment. Settling the vast majority quickly is an integral part of RIAA's litigation strategy, and most of those settlements have been for amounts between $2,000 and $15,000. An insurance market can cover payouts of this size without exorbitant premiums, so buying settlement insurance is a reasonable counter-strategy for most people.

The next point is one that we should perhaps ask of an experienced litigator. I suggested that settlement insurance would be outside of the disclosure requirement in FRCP 26 because it would not involve a "judgment." Ernest responded that a settlement is really an "offer for judgment" under FRCP Rule 68. My understanding is that an offer for judgment under that rule is a unique species of settlement offer. The rule requires the "losing" party after trial to pay the "winning" party's legal costs that are incurred after an offer for judgment if the eventual adjudicated judgment is nearly the same as the offer. The rationale is to encourage parties to settle and to economize judicial resources. Certainly, other forms of settlement offer exist. One common settlement structure entails voluntary dismissal of the complaint, mutual release of all claims, and (usually) some money or property changing hands. There is no reason for a settlement to require a judgment in many cases. (Notable exceptions are class actions and cases where extended judicial oversight of some aspect of the settlement is required.) That is the beauty of settlements being contracts — they can be whatever the parties want them to be.

Finally, I believe Ernest misunderstands my argument about Federal Rule of Evidence (FRE) 411 and the admissibility of evidence of liability insurance to prove wrongful conduct. The Rule expressly forbids the admission of such evidence for this purpose. When I wrote that evidence of insurance is admissible for other purposes such as proving ownership, control or knowledge, Ernest jumped on knowledge as an element of wrongfulness. This argument has some intuitive appeal, since knowledge is an element of willful copyright infringement. However, in the context of FRE 411, "knowledge" means something different — it generally means that a person knows what is being done with her property.

Proving ownership, control, or knowledge would likely be done in a case where the defendant argues that RIAA has mistaken her for someone else. The case of Sara Ward is illustrative. RIAA sued this 66-year-old "computer neophyte," alleging that she made over 2,000 songs available for sharing through Kazaa. Later, it came to light that most of the songs she is alleged to have shared would appeal to someone in a much younger demographic group — e.g., "I'm a Thug," by the rapper Trick Daddy. Furthermore, she owns only a Macintosh and no PC, and there is no Macintosh version of Kazaa. She appears to have compelling defenses against the lawsuit. However, if RIAA later discovers that she purchased copyright insurance, this would be strong evidence that she really does own a PC or that she knows someone is trading files through her ISP account, or that something else fishy is going on. This is what the FRE mean by "ownership" and "knowledge." Other media stories in recent weeks have highlighted the problem of children sharing files without their parents' knowledge. These parents might defend a lawsuit based on a lack of knowledge, and evidence that they purchased copyright insurance would eviscerate such a defense.

I see no difference between the proposed copyright insurance and any other kind of liability insurance. I can buy copyright insurance to defray the cost of defending myself against a lawsuit without admitting wrongdoing, just as I buy car insurance to protect myself against the improbable event that I negligently cause an accident or that someone else hits me and has no insurance to cover my injuries. I have never driven carelessly, and no would-be plaintiff could introduce my insurance policy as evidence of carelessness. When a doctor purchases malpractice insurance, he does not use it as an excuse to deliver substandard care or deliberately harm his patients — he uses it to defray the costs of defending himself against rare lawsuits by disgruntled patients and even rarer payouts for settlements or judgments. Similarly, I trade many files via Gnutella that are not copyrighted, that I have permission to distribute, or whose copyrights I own. Occasionally, for fun, I rename files as "NOT [songname].mp3" in a deliberate effort to confuse RIAA. This practice may well land me in court some day, if RIAA does not believe those files are really "NOT" what they are named. Additionally, with the cases of obvious mistaken identity that have come to light in the last two weeks, there is even more reason for innocent file sharers — who are not infringing copyrights — to worry.

Copyright liability insurance would be designed with this audience in mind. Are they likely to be its biggest consumers? Probably not. Is that a bad thing? I am not sure where I stand on that.

Posted at 9:39:30 PM | Permalink
| Comments (6)
Trackback URL:
Topics: IP
Email this entry to:

Your email address:

Message (optional):


Fingerman responds:

Ernest points out that the insurance policy I suggested would be unusual: it would cover settlements but not judgments in copyright infringement suits. This may seem bizarre, but it is necessary for two reasons. First, it is an attempt to dodge the required disclosures under Rule 26 of the Federal Rules of Civil Procedure (FRCP). If the policy does not cover judgments, it does not fall within the literal language of the rule. If push comes to shove, a court might see it for what it is (an attempt to dodge the rule) and assess sanctions for failure to disclose. I would not recommend this litigation strategy for most people, but someone interested in setting precedent (EFF? ACLU?) might be interested.

I remain skeptical that one can dodge the required elements of Rule 26 with this ruse. However, even if one does dodge the required disclosure under Rule 26, that does not mean they don't have to disclose. It just means that disclosure isn't automatic. To the extent that the public policy behind Rule 26 mandates automatic disclosure of insurance policies, I would expect that Rule 26 would encourage disclosure of insurance-like policies that cover "settlements" in response to a specific discovery request. You could fight the discover request, I suppose, but I'm not sure that would win you many points with a judge or the plaintiffs.

And who exactly would litigate the issue anyway and on who's behalf? Which raises the issue of whether it is proper for representatives of this "insurance" company to represent the interests of those it insures. Normally, an insurance company's lawyers are allowed to represent the insured in court because it is assumed that the insurance company and the insured have interests that are quite similar - the lawyers for the insurance company will fight as hard as if they were directly employed by the insured. Why is that at all the case here? If anything, the lawyers for the "insurance company" in this case have every incentive to press for a judgment of any sort, thus eliminating any requirement for the "insurance company" to pay out. I wouldn't want a lawyer with such a mismatch of incentives to represent me ... heck, I'm not even sure it is ethical.

Dan next agrees with me that insurance for judgments would cost too much. He notes that the RIAA would prefer to settle claims rather than litigate them. However, the point for the RIAA is not to litigate or settle, but to discourage people from file sharing - the current settlements are part of that strategy. To the extent that permitting this insurance business to operate would undermine this strategy, the RIAA will push forward to litigation where necessary. In such a case, if you are buying insurance because you are worried about the consequence of a lawsuit, the fact that the RIAA can push forward with litigation to "judgment" (at their discretion) leaves you as unprotected as if you didn't have any insurance whatsoever. So, why would people pay for this protection that is no protection whatsoever?

Dan then notes that there are a variety of settlements and not all of them will be considered "judgments." This is true, and I should have been more clear. The point is that the RIAA can basically insist that the defendant make a Rule 68 offer or no settlement, and a Rule 68 offer is a "judgment." What is a defendant going to do? The RIAA says, "make us a Rule 68 offer of $3,000 and this case goes away, otherwise we take you to trial where you will be liable for a minimum of $750,000." If the defendant is so concerned about such an outcome that they are taking out insurance in the first place are they going to say "no"? What would be the point? And why wouldn't the RIAA insist on a Rule 68 offer? Because otherwise the defendant will insist on going to trial?

Knowledge and Willfullness

Dan argues that evidence of liability insurance can't be brought in as evidence of willfullness under FRE 411. True, but it can be brought in for a number of other purposes (such as knowledge), including those that would more likely result in a finding of willfullness. Dan insists that the knowledge provision applies only to knowledge of what is being done with a person's property. I disagree, I believe that it can be used to show knowledge that file sharing programs can infringe copyright. Furthermore, even if that isn't permitted, it certainly would be permitted to rebut the credibility of a defendant who denied knowledge that file sharing programs can infringe copyright. Basically, buying insurance will significantly reduce your ability to claim "innocent infringement" and make it much more likely that willfullness will be found.

One of the first questions the RIAA's lawyers will ask defendants who don't settle is whether or not they knew that P2P could be used to illegally file share. If they say, "yes", you are well on your way to a finding of willfullness. If they say, "no", then the evidence of insurance can be instroduced to rebut or cast doubt on their veracity.

Dan postulates a defendant who sometimes names files "NOT [songname].mp3" in order to trick the RIAA into an attack. Bad, bad idea, if you are actually file sharing files that are copyrighted by the RIAA. According to the RIAA, they are actually downloading songs from the individuals they target in order to verify that a representative number of songs are actually copyright violations. If this is this case, and you are also sharing files simply to taunt the RIAA, you are definitely looking at a finding of willfullness, whether you buy insurance or not.

If the RIAA sues someone by mistake, and they can prove it, then the insurance evidence may not be relevant. But, really, are people so afraid of being mistakenly sued by the RIAA that they are going to buy insurance that only covers settlements?

Posted by Ernest Miller: Tue, 21 Oct 2003, 5:51:24 AM


Maybe you can tell me if the following twist changes this whole situation: suppose P2PFund has no idea who its customers are - and this is not a theoretical point, there is a technical way to achieve exactly that.

So P2PFund cannot be forced to disclose anything, since all it knows is a bunch of account numbers that cannot be tracked back to real people.

It would seem to me that this approach completely prevents the possible "sue the people with money" strategy by RIAA. However - when the subpoenaing phase is behind us, RIAA has a user in its sights, and this user knows that she is the P2PFund customer (but P2PFund is not technically capable of confirming or denying it, since it has no clue!) - so can this user be somehow forced to disclose her relationship with P2PFund? I mean, let's say she says 'no' - what could be the possible penalty for her 'forgetfulness'?

Posted by Serguei Osokine: Tue, 21 Oct 2003, 9:26:00 AM

Responding to Ernest:

At this point, I think we have highlighted the major problems with the insurance proposal. I suggested a few things that I thought might alleviate those problems, in an effort to explore whether such a system is workable. You have shown convincingly that my suggestions are far from complete answers. You did an especially good job of making me less sure of my FRE 411 argument than I was yesterday. I still think the basic idea has merit if people sufficiently expert in copyright, insurance and litigation resolve the problems we addressed. I admit that I could be wrong in the end. But I hope you agree that the insurance idea was at least interesting to begin with.

Posted by Dan Fingerman: Tue, 21 Oct 2003, 10:17:25 AM

Responding to Serguei:

This is one question that Ernest and I have been discussing. The Federal Rules of Civil Procedure have a process called "discovery." Discovery is a mandatory part of litigation, in which the parties are required to share information with one another. Common discovery tools include depositions (where the parties ask questions of one another, under oath, with a court reporter present), interrogatories (written questions that require a written response), and requests to produce documents (which the other party is required to copy and turn over). There is also Rule 26, which requires parties to disclose some insurance information immediately, so that the other party does not have to ask for it via a discovery mechanism. Rule 26 is designed to make litigation more efficient. Since every litigant is going to ask for that information in the course of discovery anyway, everyone saves time by simply handing it over at the beginning.

To answer your other question, yes, there may be penalties for parties' refusal to turn over information during discovery or disclosure. The court may impose sanctions ranging from holding a party in contempt (fines, jail time) to deeming that party to have admitted the fact that the other party sought to discover. A court would have a hard time believing that someone merely "forgot" about an insurance policy, if it finds out through some other means that the person is "lying." If RIAA can establish sufficient doubt about the veracity of the defendant's answers, there may even be grounds for RIAA to get a hold of the defendant's bank and/or credit card records in an effort to prove a payment to P2P Fund.

These are hard problems for a defendant to work around. That is why I tried to come up with a creative way to keep the insurance policy outside the language of the automatic disclosure rule. Ernest has made clear that this is extremely difficult and risky. We may never know if it can work until some brave person bites the bullet and tries it. If the first person fails, someone else may be brave enough to try again. But these are awfully big risks to take, since the downside is a judgment of at least $750,000.

Posted by Dan Fingerman: Tue, 21 Oct 2003, 10:42:30 AM

Dan, thank you for an answer!

Two quick comments and a question:

1. "...the downside is a judgment of at least $750,000" - that's exactly why the user might want to keep this information secret until the settlement is done. She'd probably prefer to pay a $2,000 settlement (paid by P2PFund anyway) than to risk RIAA "going for the jackpot" and refusing to settle for less than the full balance of the P2PFund bank account.

2. "...there may even be grounds for RIAA to get a hold of the defendant's bank and/or credit card records in an effort to prove a payment to P2P Fund"

- there might be ways to hide that, either. The same payment anonymizer mechanism can work both ways. But I completely agree - this is something that would be more prudent to avoid.

And now, the question: P2PFund is *not* the insurance business, and has little hope of becoming one, because of all the insurance regulation requirements. And as this discussion shows, it probably would be better off if it does not even try. Does this make any difference for the court? Or the judge can say that P2PFund is de facto insuring people and thus should be subject to all the relevant discovery laws?

Posted by Serguei Osokine: Tue, 21 Oct 2003, 11:46:48 AM


Despite what you might guess from my postings here, I really know very little about the laws and regulations that govern insurance. From a litigation perspective, however, I suspect that P2PFund's self-identification as "not an insurance company" would make little, if any, difference.

The rule that we have been discussing, Rule 26(a)(1)(D) of the Federal Rules of Civil Procedure, requires each party to disclose "any insurance agreement under which any person carrying on an insurance business may be liable to satisfy part or all of a judgment." If P2PFund has nonprofit status or takes some other measures to keep itself outside the federal and state laws that govern insurance, there is room to make the following argument: P2PFund is not "carrying on an insurance business," therefore the defendant is not required to disclose the insurance policy because the rule requires disclosure only if the insurer is "carrying on an insurance business." This is very similar to the argument I made earlier, when I said that the insurance policy could be designed to pay out for settlements but not "judgments" because the rule mentions "judgments" but not "settlements."

Both scenarios fall outside the literal language of the rule, but a court might still sanction a defendant for failing to disclose the insurance policy. This whole exercise is, after all, an attempt to avoid the spirit of the rule. Whether our argument succeeds may well depend on something as capricious as how sympathetic the judge is.

Posted by Dan Fingerman: Tue, 21 Oct 2003, 6:02:01 PM

Powered by Movable Type