PrefBlog

After posting my review of his paper, A Structural Model of Contingent Bank Capital, I eMailed Prof. Pennacchi asking him about the political, regulatory and academic response to his paper and inviting him to comment further,

He very kindly responded and has granted permission to quote his reply:

I have presented my paper at the NY Fed, a Moodys-NYU Conference (both in NYC), and just recently at the International Risk Management Conference in Florence, Italy. I must say that I really haven’t received any negative comments on the paper. The reception has been quite good.

I understand your concern that if contingent capital (CC) converts at par, and bank assets follow a diffusion process (which, mathematically, means the value of the assets have a continuous sample path and cannot experience (downward) “jumps” in value), the paper concludes that CC will be default free. You are concerned that this would only hold if the new equity is sold immediately by the CC investors and that, during a crisis, new bank equity issues may have to be issued at a discount.

I have a couple of observations. First, looking at historical new equity issues during a crisis may not be fully relevant to an environment where CC is converted. This is because, historically, new equity issues during a crisis have occurred in a context where there is significant “debt overhang.” The discount occurs because issuing new equity makes the bank safer (less likely to default on its debt), thereby transfering value from the original equityholders to the bank’s debt holders (such as subordinated debt investors). Hence, under these conditions we would expect that the announcement of a new equity issue would result in a fall in the bank’s stock price.

But CC is different. Indeed, one of its advantages is that it reduces the debt overhang problem. When contingent capital is converted, there is a simultaneous wiping out of existing debt (the CC bond) replaced with new shares of equity (to the former CC investors). Hence, there is not the type of transfer of value from the original equityholders to debtholders (CC investors) as would be the case if new equity was issued without wiping out the claims of the bank’s subordinated debtholders.

Second, assuming that there would be no discount at conversion, then CC investors would receive $1 of stock for every $1 of par value at the new issue debt. At this point, the model assumes the debt has been paid off in full, so there has been no default. You question whether this is really default-free. It may not be default-free based on the CC bonds’ original maturity date, but it is default-free at the conversion date.

Note that the CC investors who are now stockholders could sell but they could also hold if they thought there would be temporary downward pricing pressure from others who sell. It could be that by holding on to their shares they would receive even more than their par value as of the CC bond’s original maturity date, which would be even better than holding a default-free Treasury bond. Of course they could also get less if the stock price declines. But the point is, the situation no longer becomes comparable to a default-free investment. I am taking the perspective of the bond being default-free as of the conversion date. As I state in the paper, the effective maturity date of this default-free security is uncertain. If you do not want to think of that as a default-free security, I have no problem with that perspective. However, the CC investors will get their par value (in stock if prior to maturity and cash if at actual maturity) at some date in the future, even if that date is not known ahead of time.

Third, and finally, I do not believe that bank assets follow a diffussion (no jump) process. The main, unique contribution of my paper is to value CC under the assumption that bank asset values (and stock prices) are likely to jump, especially jump downward during a crisis. So while my reasoning may differ from yours, what my paper shows is that CC will indeed be credit risky, not default-free. As my paper shows, one should expect that CC will have a positive credit spread when issued. I only compare my model to the diffusion (no jump) case to emphasize why jumps in asset value matter. However, less one thinks that such jump risks make CC a flawed product, my paper also goes on to show that CC is less risky than if the bank has, instead, issued a comparable quantity of subordinated debt. In summary, it is unrealistic to think that CC is default-free. However, conversion to equity when the bank’s condition has a moderate decline is actually a safety valve that relieves financial distress and protects CC investors relative to if they were sub debt investors. Because CC capital conversion reduces the bank’s leverage, it would make it easier for the bank to issue further new shares of common stock without experiencing much, if any, of a discount because there would be little overhang.

One policy recommendation from my paper’s results is that CC becomes safer (less credit risky) the greater is the value of original shareholders’ equity when conversion is triggered. In other words, it is best to set a trigger than can result in conversion when the bank’s original shareholders’ equity is only mildly depleted rather than have a trigger where conversion occurs at a very low value of original shareholders’ equity.

I hope this response helps to provide some intuition regarding the risk characteristics of CC.

Thank you, sir!

This entry was posted on Sunday, June 6th, 2010 at 5:19 pm and is filed under Contingent Capital. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.