What’Swap: Accounting for Financial Innovation

Robert Waldmann

has become interested in Credit Default Swaps. I’m not just wondering whether they played a major role in destroying the world financial system or were just along for the housing bubble, MBS, CDO ride. I also wonder whether the acronym for the plural should be CDS or CDSs or even CDSes. I’m glad to see I’m not the only one.

But more even that that, I wonder why credit default insurance is called a credit default swap when the contract is as asymmetric as a financial contract can be. I have no information on the history of the term and will just speculate. I assure the impatient reader that this post is not totally pointless (as far as I know) and leads to a practical proposal for regulatory reform.

After the jump, I will argue that the main motivation for the invention of new financial instruments was to evade relax capital requirements, that this relaxation was reasonable in the case of interest rate swaps, that it was unreasonable in the case of credit default swaps and that credit default swaps should be recorded in balance sheets as if they were interest rate swaps (which they are in disguise).

I will start with interest rate swaps, which are clearly swaps. In an interest rate swap contract to entities agree to exchange a constant times interest paid on one bond during the term of the interest rate swap contract for a constant times interest paid on another bond during that period. For old timers, it is as if they were buying and selling coupons not bonds. Why would they do that ?

First it is possible to construct an interest rate swap using a portfolio of older assets, bonds, bond futures, short positions on bonds and short positions on bond futures. Instead of buying interest paid on a bond this year I can buy the bond and sell short the bond a year from now.

My first claim is that the new instruments exist, because if they appear in balance sheets total assets and total debt are smaller numbers — that is for accounting and regulatory reasons. This is not the most common explanation for the existence of interest rate swaps. The explanation is that corporate bonds are not liquid, that is, the market for them is thin (or illiquid in financial operator speak). This means that if I send huge orders for bonds and short orders for bond futures to the double auction market, I will pay a huge price for the bonds and get a low price for the futures. Thus it is better for each of two firms for them to negotiate a bilateral deal with an agreed price and quantity. This gets us as far as “swap” but doesn’t explain why these contracts are written as exchanges of interest payments and not as exchanges of bonds and bond futures. So far, it seems that it would make no difference.

If the firms are banks with binding capital requirements it makes a huge difference. A huge long position in a bond is a huge asset and a huge short position in the futures is a huge liability. A position recorded as a portfolio of bonds and bond futures would imply that a huge amount of wealth must be set aside to satisfy capital requirements. If it is recorded as a much smaller position in interest paid on bonds, then banks are not required to have so much idle capital to satisfy capital requirements.

In this case the new accounting is reasonable. The risk in holding a long term bond for, say, a year is almost entirely in its price at the end of the year. So long as the issuing firm doesn’t default in the course of the year, (nominal) interest payments are safe. The the long position in the bond and the short future position are an almost perfect hedge. The two sources of risk cancel. It would make no sense to evaluate the total riskiness of the position at the sum of the two risks as if two almost perfectly correlated positions were uncorrelated.

Now with an acute sense of how existing regulations have nothing to do with portfolio theory or 20th century risk management, regulatory authorities were willing to the innovative accounting.

The next step is almost logical. If interest paid on a bond is scheduled interest (a known constant) minus losses due to default, one could rephrase an interest rate swap as a credit default swap. I pay you the interest paid on Bond A and you pay me the interest paid on bond B = I pay you a constant (which can be negative) plus losses due to default on B and you pay me losses due to default on A. Now that we already have the constant, there is no reason to make the positions in both bond A, so an interest rate swap becomes two credit default insurance contracts. I think this is why credit default insurance is called a credit default swap.

OK so still there is no change in possible financial transactions. CDSs like interest rate swaps are redundant assets which can be created out of portfolios of bonds, futures on bonds and short positions of them. So what is the point ?

Well the accounting innovation has become an accounting innovation squared. As the interest rate swap meant that the value of the bond at the end of the term no longer appears as an asset, the CDS means that scheduled interest payments no longer appear as an asset. The CDS appears in balance sheets at its fair market value, not as a large position in a bond and a short position in cash (or debt of the firm which bought the credit default insurance if its required payments are spread out over time). By introducing interest rate swaps and CDSs into accounting, firms have managed to rewrite immensely huge positions in bonds etc to merely huge positions in interet rate swaps to merely tens of billions in CDSs.

Now this second bit of innovative accounting is totally unreasonable. The part which no longer appears in accounts, the scheduled interest rate, is not an apparent source of risk which is hedged. It is the safe part. Writing a CDS is a way of bearing all the risk with a very small number recorded as the value of liability.

In the case of interest rate swaps, firms had a way to hedge risk which was not automatically recognized as such by accountants and regulators. So they changed the accounts so that large apparent risks which cancelled didn’t appear.

With credit default swaps, firms found a way to describe the exact same transaction so that the numbers on their balance sheets were different and so that their required capital was smaller.

Now it is fairly easy to argue financial market innovation is socially useful, but few people would argue with a straight face that we need more innovative approaches to accounting. However, it appears that many people were willing to argue that innovative accounting *was* innovative finance. They convinced Gramm and Clinton went along (who was his treasury secretary at the time ?).

Now to me the reform is obvious. My proposed regulation follows.

People can trade what they want, but accountants must write accounts based on standard assets. New assets can be accepted into accounting only once their risk to value ratio is determined by the Basel III standing committee (Oh and the USA participates in Basel III). CDSs are not standard assets and CDS positions must be rewritten as interest rate swap positions.

An exception to the above shall be allowed. If a firm really insists on putting an unrated asset into its accounts, then it can. However, the number written as “liabilities” must be the present value of the maximum conceivable payments on its position and the number written as “assets” must be the present value of the minimum possible payments it will receive.

Believe me, bankers will find a way to express most new assets as portfolios of standard Basel III acceptable assets.

I admit that, under this plan, an authenticly genuinely new non-redundant instrument will be rated as extremely risky until diplomats and bureaucrats are convinced that it isn’t. I consider that a feature not a bug.