An article in the Financial Times on Monday treated us to another take on the empty creditor hypothesis. Professors Marti Subrahmanyam (NYU) and Pablo Triana (ESADE) write that:
“Bond or loan holders using CDS to gain credit protection qualify as empty creditors: if the borrower gets into trouble, the CDS, if triggered, would cover any losses on the underlying position; if the borrower honours its obligation, the investment is made whole (minus the cost of the CDS protection).”
So what’s the problem? In the opinion of the authors, it’s this:
“Empty creditors are lenders (to a corporation or government) that cease to be concerned about whether the borrower fares well or poorly. Their interests are not aligned with those of other creditors, who prefer that the debtor does well, so that the debt is repaid…
“In a more extreme scenario, the empty creditor may benefit even more by “over-insuring” – purchasing a proportionately larger amount of CDS protection than the debt owned (there is no real limit on the amount of CDS “protection” investors can buy). Obviously, those who did not enter into CDS are not indifferent to bad news as they have a more asymmetric pay-off.”
Our research department wrote a piece that addresses many of these issues a few years ago. We won’t repeat the entire argument, except to highlight a few key points from that piece.
The first relates to whether CDS do indeed lead creditors to prefer bankruptcy over restructuring. It posits that if the ability to hedge using CDS tends to make restructurings less likely than a bankruptcy filing, the correlation between number of defaults and restructurings as a percent of defaults should be lower when CDS are available than when they are not. The paper stated:
“The data show that the correlation between number of defaults in a given year and restructurings relative to defaults in the same year is about 9 percent over the entire sample period. But restricting attention to the period of liquid credit default swap markets, which arguably began in 2003 with the publication of the 2003 ISDA Credit Derivative Definitions and the subsequent initiation of trading in the CDX and iTraxx credit indexes, the correlation jumps to 90 percent. While correlations within small data sets should be interpreted carefully, the correlation statistics presented here would not appear to support the empty creditor hypothesis, according to which the availability of credit default swaps would make restructurings less likely.”
But wait, there’s more:
“Further evidence comes from the list of restructurings that occurred during 2008 and the first half of 2009… During that time, twenty-one firms underwent out-of-court restructurings; credit default swap protection was available on eleven of them (52 percent). And of the restructurings that occurred during that period, four subsequently filed for Chapter 11 bankruptcy; of those four, two had liquid CDS available and two did not. Again, the evidence thus far does not appear to support the empty creditor hypothesis.”
The second point we would make has to do with what the FT article referred to as “over-insuring” and our paper described as “negative economic ownership.” Again, our paper stated:
“…one may reasonably question the plausibility of the second hypotheses on the basis of how the credit default swaps market treats distressed credit. If an investor were actually to try to build up a negative economic ownership position through overhedging, the strategy would be expensive and unlikely to yield a high return.
…[a]n overhedging strategy is likely to be profitable only if an unusually prescient hedger were to foresee accurately the failure of an investment grade company while the company’s credit default swaps still traded at a low spread. In such a case, the gain might be regarded as a windfall but would not lead to behavior that might affect the functioning of credit markets. And if the anticipated bankruptcy did not occur, the large hedge position could lead to large losses.
…Further, it is not clear how the investor would have been in a position to influence the likelihood of a bankruptcy, and thereby make a positive return more likely, other than by failing to support a restructuring if one were proposed. And as shown already, the evidence regarding restructurings does not support the contention that credit derivatives have had a negative effect on restructurings.”
A final thought: it’s interesting to note that a recent paper by Professor Subrahmanyam (“Does the Tail Wag the Dog? The Effect of Credit Default Swaps on Credit Risk”) also addresses the empty creditor issue in considerably more detail. The paper is the first empirical work to “formally address the empty creditor concern.”
There is nothing in that paper to change our view that there are many factors influencing the likelihood of an out-of-court restructuring even before considering the effect of hedging using credit default swaps. And beyond the evidence we have provided – which thus far does not appear to support the empty creditor hypothesis – the new evidence they offer is interesting but not conclusive, as there are so many other factors in play. There’s still plenty of doubt regarding whether the empty creditor hypothesis is valid or its impact is negative.