I guess it just seems to me that you didn't answer (and didn't intend to answer) a much more interesting question: Should CDS be made to fit the definition of insurance, and then be regulated as such?Locrian is right that I didn't intend to address the normative question in my previous post, but I suppose I should. No, I don't think CDS should be made to fit the definition of insurance, and then be regulated as such. There are several key differences between CDS and insurance that make insurance regulations inappropriate for CDS. The most important difference is daily collateral posting—standard single-name CDS require daily collateral posting, which insurance contracts do not. To borrow an example from A Credit Trader: if you buy fire insurance on your home, and then one day your neighbor's teenage son takes up smoking (increasing the risk that your house will catch on fire), you can't demand that the insurance company post collateral to a specified collateral account to offset the increased risk of fire. In contrast, standard single-name CDS contracts require daily collateral posting to offset changes in the mark-to-market value of the CDS (that is, the CDS are collateralized). In other words, as the CDS spread rises—indicating that the reference entity is getting progressively closer to default—the protection seller has to post more and more collateral to offset the increased risk of default by the reference entity. (By the time a credit event occurs, protection sellers have usually posted most of the total payout in collateral.) Collateralization (which usually includes initial margin as well) is a common way to mitigate counterparty credit risk (i.e., the risk that your counterparty can't pay). Similarly, the strict capital requirements that insurance companies are subject to are just another method of mitigating counterparty credit risk—the purpose of strict capital requirements is to ensure that when it comes time to collect on your insurance policy, the insurer has enough money to pay you. Strict capital requirements are an appropriate method of mitigating counterparty credit risk in uncollateralized contracts like insurance, where the insurer doesn't have to post collateral on a daily basis to reflect changes in risk exposure. But as we've seen, standard single-name CDS are collateralized, which dramatically changes both the frequency and nature of payments. If standard single-name CDS didn't require daily collateral posting, and instead only obligated the protection seller to make a single lump-sum payment upon the occurrence of a credit event, similar to insurance contracts, then there would be a plausible case for regulating CDS as insurance. But collateralization, which is a feature of certain CDS trades (e.g., stand-alone synthetic bond trades), makes CDS fundamentally different from insurance contracts. Instead, standard(ized) CDS should be moved onto regulated central clearinghouses, which already require initial margin and daily collateral posting (known in clearinghouse jargon as "variation margin") as a matter of course. Clearinghouses also have the risk management and technological capacity necessary to calculate the proper amount of variation margin required from each clearing member on a daily basis, which insurance commissioners most certainly do not. I'll have to leave CDS on ABS/CDOs and the "insurable interest" argument for a later post, because, well, I just don't have time to get into those arguments right now. UPDATE: Commenter M expands on my point, and provides an excellent explanation of why daily collateral posting based on mark-to-market changes makes CDS a fundamentally difference product than insurance:
I think you leave an important feature of the difference between insurance and CDS out. Which is that insurance is only paying out in case of an event, and will not pay out in case nothing happens. You can have a view of the overall chance of certain events happening, but you can not buy and sell insurance depending on the change in circumstances. With CDS you can take a view on not just the actual event happening, which would result in a pay out, but also on a general deterioration of the credit environment or the improvement of the credit environment, allowing you to take a profit or avoid a greater loss without an actual default happening. For instance, if you buy and hold a bond of a company, in case the credit of that company goes down, the price of the bond will go down and you will lose out if you would want to sell the bond before maturity, even if the company does not actually default. However, if you would have seen it coming,and you bought a cds on the same company, for a fee you are protected to this change, not just the actual default. The change in creditspread is a source of income, a protection, much as an interest rate swap as a protection against changes in interest rates. That is what makes a CDS fundamentally different, and much more dynamic, than a insurance that is purely build around an event happening or not.Exactly. This is the distinction I had in mind when I discussed insurance law's use of strict capital requirements to mitigate counterparty credit risk. Strict capital requirements are appropriate for insurance because insurance contracts are essentially binary—either the insured event happens or it doesn't. Because there's so little interaction between the insurance company and the policyholder prior to the occurrence of the insured event, policyholders are much less likely to monitor the insurance company's ability to pay (i.e., its capital adequacy) on a day-to-day basis. That's why we need the government to impose strict capital adequacy requirements on insurance companies. But CDS are generally not binary, due to the daily collateral posting based on mark-to-market changes. So a difference regulatory regime is warranted.