Popularized in films like Limitless, legal smart drugs called Nootropics are becoming more and more prevalent in board rooms and on Wall Street.Keep reading »
The Fed has three basic functions: central banking, bank regulation, and calling down police brutality on Occupy Wall Street protesters. While the first function is getting all the attention today, the New York Fed’s blog is spending some time on the second. Specifically, they’re trying to figure out how bankers should get paid.
Optimal design of banker compensation is a thing that people like to think about, and that regulators like to regulate. We’ve talked about it before, and I’ve suggested that the right way to reward bankers is not to give them mostly equity or extra-levered equity, which encourages asymmetric risk-taking, but rather to give them exposure to their firm that roughly matches that of their main stakeholders. Which, for a bank, means basically various flavors of creditors. So a bank CEO whose net worth consists 20% of equity of his firm and 80% of unsecured debt of his firm, like Brian Moynihan, in theory has better incentives to do the right thing by bondholders, depositors and the financial system than someone who’s 100% in out-of-the-money stock options. And a banker who is paid in structured credit products that can’t be foisted on to clients has incentives … well, he’s an interesting case study at least.
I like the NY Fed researcher-bloggers because they’re pretty sober people who want to optimize banking regulation but don’t spend their time freaking out about stupid popular things like how CDS will kill us all, banning short selling, or just generally hating on bankers. So I’m pleased to see NY Fed researcher Hamid Mehran is with me on this whole comp thing:
Because stock- and options-based compensation is unlikely to curb the risk appetites of bank executives, Bolton, Mehran, and Shapiro in a 2010 New York Fed working paper proposed tying remuneration to a bank’s credit default swap spread. An institution’s CDS spread provides a market measure of the risk of a bank. Under our scheme, a high or increasing CDS spread would translate into a lower cash bonus, and vice versa. …
How would this scheme look in practice? CEOs could be required to write a given amount of CDS (or to buy swaps written by other insurers) for the duration of their employment contract. However, a more efficient policy might use money set aside by the bank at the start of the period. Deferred bonuses would then be reduced or increased under a pre-specified formula as the bank’s CDS spread deviated from the average bank spread: bonuses would be increased if the spread was below average and decreased if it was above average.
The paper itself is worth reading. Among other things, it suggests that their CDS-based comp scheme is better than just giving executives unsecured exposure to their banks by paying them in deferred cash comp, for reasons that I don’t really understand. (They seem to assume that banks have to set aside deferred cash in a pot at the time of grant, which, I mean, the whole point is that it’s unsecured, just promise to pay it later and if you don’t have the money tant pis.)
This bothered me a bit, because I am committed to the view that, in efficient markets, debtholders should view widening credit spreads and increasing probability of non-payment-at-maturity as equivalents. Thus the deep magic of DVA, where banks transmute a bad thing (widening own-credit spreads) into a creepy good thing (less chance we’ll have to pay off all these derivatives we wrote!). So, funding aside, a normal creditor of a big bank should be indifferent between writing CDS on that bank and owning that bank’s bonds. But I’m coming around to the Fed’s view, because of course the CEO is not a normal creditor: he got to be the CEO in part by being super optimistic about his bank’s chances. Making him mark his enthusiasm to market is probably a better idea than letting him discount the probability of nonpayment on his own.
Anyway, there seems to be broad agreement among, well, me and some Fed economists and not William Cohan, that giving bank executives a direct stake in the creditworthiness of their levered systemic firms makes more sense than just aligning them with shareholders, because bank shareholders are unsavory risk-loving characters just like bankers are. CDS is a natural proxy for that creditworthiness, and it’s a reasonably objective and liquid market-based proxy.
But there’s also this. Your model of a bank’s stakeholders could be 20% equity / 80% unsecured debt, in which case you could be a big fan of the Fed idea. Or your model could be 2% equity / 8% unsecured / 90% witchcraft and magic. The NY Fed, again, are sober types who are not believers in witchcraft and magic, and most of the time that’s probably a good way to look at the world. But the reason that lots of people like to freak out about bank compensation is that, every once in a great while, the fulcrum stakeholder is not a shareholder or a guy who holds an unsecured thing deliverable into CDS, it’s the financial system as a whole. When 90% of the money at risk in Morgan Stanley was the Fed’s,* Morgan Stanley wasn’t really working for either its shareholders or its bondholders – and its bonds, like its shares, looked like more or less at-the-money options to buy the firm from the vast mass of secured credit (Fed, repo, FDIC-insured, whatever you’ve got at your particular bank) that really owned the place.
CDS obviously reflects this sort of thing – Morgan Stanley CDS didn’t look so hot in September ’08 – but so do stock prices. Both are imperfect, conditional, convexity-laden reflections of the shifting risk profile of banks. Both probably under-reflect the risk to all stakeholders because in a really deep crunch the equity and unsecured are not where the bulk of the capital is at risk. If you wanted to really align bank CEO’s risk-taking incentives, you’d let the Fed empty their personal bank accounts to fund emergency loans to their banks.
Which I’m pretty sure would be a dumb idea. The lesson may just be that you can’t fully optimize comp to make bankers internalize risks. But you can improve from the current baseline. And this CDS-based comp thing seems like a decent attempt.
* Quoted without approval. I think his math and understanding of how the claims work are iffy but, y’know, the gist is not wrong and he’s not the only one thinking it.