Monday, 20 October 2003
Copyright Liability Insurance: A Response
Steven Wu of Lawmeme has suggested a system of copyright liability insurance to protect file sharers from the risk of being sued by RIAA. The newly-organized P2P Fund is designed to serve a similar purpose, folded into a compulsory-licensing system. Ernest Miller has responded, arguing that such insurance is a bad idea. I find the idea intuitively appealing, and I think Ernest's criticisms are answerable. That said, I have reservations about the suggested implementation of the insurance scheme that must be resolved before it becomes workable.
As Steven Wu explains, self-insurance (where each person pays his own way) can be prohibitively expensive. If a sufficient number of file sharers buy into an insurance market, each one can pay a small premium and be protected. He compares this to patent infringement liability insurance, which is "professional liability insurance for manufacturers, users and sellers accused of infringing a patent holder's rights."
1. Critical Mass & Self-Selection
Steven acknowledges the first major problem, critical mass, but he underestimates its size. An insurance company lives or dies by the number of people it can attract as customers. Steven treats this as merely a problem of getting the word out to file sharers ("people will join it out of self interest, which is always more reliable than altruism: the price is worth the peace of mind."). However, he gets to this point through a general economic analysis of insurance, comparing the economic advantages that an insurance market holds over self-insurance. Unfortunately, he ignores the concomitant problem of self-selection. The people most likely to buy into an insurance plan are those most likely to need its payout — in this case, the most active file sharers (which I shall call "heavy users" of file sharing networks). RIAA has stated publicly that it will at first sue the heaviest users, people who share over 1,000 copyrighted songs. "Light users," those who share a smaller number of songs or who share none at all, are less likely to pay into a system from which they are less likely to need a payout.
Self-selection works against the viability of an insurance system, but it is not insurmountable: the system can account for varying degrees of risk. Life insurers do this routinely, charging different premiums or offering different payouts based on such factors as age and smoking, which affect a person's risk of developing health problems. Similarly, the copyright liability insurance policy could peg each person's premium to the number of files he shares, the number of hours per week he is logged onto the network, and other factors that affect the risk of attracting RIAA's ire. The identity of the insured's ISP may also be a valid risk factor. While Verizon and SBC have challenged DMCA subpoenas, others have complied without a fight.
Heavy users would pay higher premiums, giving them a disincentive to buy insurance. On the other hand, this scenario gives light users a strong incentive to buy insurance. Light users would be footing a smaller percentage of other people's legal bills and paying a small monthly premium, so they may be likely to join. This, in turn, will raise the insured base and permit the insurer to charge lower rates to heavy users after a time. The net effect of these incentives is hard to predict in advance, but the insurance industry has good economists and actuaries working for it. Someone should be able to pin down a good set of risk factors and prices.
2. Discovery & Disclosure
Ernest argues and Steven acknowledges that RIAA will likely change its litigation strategy if such an insurance system arises. I agree. However, finding out who is covered may not be as easy for RIAA as Ernest suggests.
Federal Rule of Civil Procedure (FRCP) 26 requires automatic disclosure of some insurance policies. Rule 26(a)(1)(D) requires disclosure of "any insurance agreement under which any person carrying on an insurance business may be liable to satisfy part or all of a judgment which may be entered in the action or to indemnify or reimburse for payments made to satisfy the judgment." The language of the rule plainly contemplates insurance policies that will pay judgments against a litigant, not settlements. Most insurance policies will pay out for both, because they are designed to insure against types of claims that are frequently resolved by either judgment or settlement. In this particular market, however, RIAA has repeatedly stated to the press that it wants to settle its claims against file sharers, and no claim against an individual file sharer has yet gone to trial, let alone to judgment. An insurance policy could be written to reflect this reality and insure against settlements but not judgments, thereby sidestepping FRCP 26. If there is caselaw interpreting FRCP 26 as requiring disclosure of all insurance policies (and not just those insuring against judgments), I did not find it in a brief search.
3. Insurance as Evidence of Willfulness?
Ernest further argues that the purchase of a copyright liability insurance policy is "likely to be considered evidence of willful infringement, which" would permit RIAA to seek higher statutory damages. I think he is wrong for two reasons. First, he does not explain why copyright insurance would be viewed with more suspicion than any other form of insurance. Do courts view doctors' and lawyers' purchase of malpractice insurance as evidence of negligence or willfulness? I have yet to see evidence of open season on patients and clients. Second, Federal Rule of Evidence (FRE) 411 prohibits the admission of evidence of liability insurance for this purpose. The rule reads, in relevant part, "Evidence that a person was or was not insured against liability is not admissible upon the issue whether the person acted negligently or otherwise wrongfully."
Under FRE 411, then, RIAA could not use the proposed copyright liability insurance against file sharers in court to prove willful infringement. The file sharers therefore have a colorable argument that insurance is irrelevant to the cause of action, should RIAA ask for that information in discovery. This is a valid ground to object to questions in depositions, documents requested for production, or facts requested for admission. The insurance may be relevant (and therefore discoverable) for other purposes if the defendant makes particular arguments — e.g., if he denies having shared files, owning a computer, or some other fact that is material to RIAA's claim. In such a case, RIAA would be able to use the insurance to prove ownership, control, or knowledge — but still not wrongfulness.
Once RIAA learns through discovery that a defendant has purchased an insurance policy, it can use that information for whatever purpose it likes during settlement discussions, even if a court would bar it from using it for the same purpose at trial. Therefore, RIAA could learn of and use the insurance policy against defendants who deny certain aspects of RIAA's case but not against defendants who admit infringement. This is a problem that remains to be solved before this idea goes mainstream.
So what should we make of copyright infringement insurance? The basic idea appeals to me. Would I buy it? Sure, if the price is right and these problems are worked out. The problems are real, although not as insurmountable as Ernest implies. This debate is a good first step to solving them.