Streetwise Professor

Felix Salmon does me the kindness of responding in detail to my post on Coase Meets CDS.  Color me unpersuaded.  

Felix (if I may) strikes the pose of the Practical Man, juxtaposed to his portrayal of me as the Ivory Tower Pointy Head.  It is of course essential to keep the practicalities in mind, but  the Practical Man  (a) is often limited by what has been, and has a difficult time envisioning alternatives, and (b) often lacks a conceptual framework that helps sort through complex situations.  I respectfully suggest that Felix falls victim to both problems.  The comparative advantage of the Ivory Tower Pointy Head is that s/he–I, in this instance–can bring some conceptual order to the complexity.  Indeed, that’s the whole beauty of the Coase Theorem in particular.  By imagining an impractical world, it identifies the key practical aspects of a problem that deserve analytical attention.

Let’s focus on what is the issue at hand, and which Felix doesn’t IMHO adequately isolate, and instead in Practical Man fashion buries in a jumble of other empirically/observationally motivated considerations.  

The issue at hand is: What is the incremental contribution of CDS to the costs of financial distress?  Felix mentions the well-known transactions costs of the bankruptcy process.  But these costs were well known before CDS were a glimmer in some J.P. Morgan banker’s eye.  Relatedly, Felix mentions “very few people trade in [distressed credit]: basically it’s the province of a small number of distressed-debt funds who like to buy the debt cheap and then wade into invariably-arduous litigation.”  True.  But it was true before CDS.  It is true today even for distressed credits on which CDS are not traded, or for which the volume of CDS trading is de minimus.  Relatedly, Felix talks about how “arduous debt renegotiations are.”  True again, but the same remarks hold.  This does not address in any way the incremental contribution of hedged creditors to the arduousness of renegotiations.  So, these remarks are completely off point in responding to the issue at hand.  

The charge before the bar is that by undermining the incentives of hedged creditors to undertake part in value-enhancing pre-bankruptcy restructurings, the costs of these restructurings are higher when some creditors are hedged when they are not.  (Relatedly, it is possible that bankruptcy procedures themselves are higher when some creditors are hedged.)  Discussing the costs of workout and bankruptcy in general says absolutely nothing about this key issue.

When addressing that specific charge, the Practical Realities that Felix mentions are not particularly illuminating, because they are common to all dealings involving distressed credits.  

The relevant issue is the incremental cost of eliminating the putatively perverse incentive arising from hedged trades.  The very liquidity in CDS that Felix touts suggests that this incremental cost is not particularly high; the hedged creditor can reverse the CDS position.  If the creditor does not want to retain the risk exposure associated with the remaining debt that would result after reversing its CDS position, an EFP-style trade (a simultaneous transaction in the underlying credit and the CDS) with a party with an offsetting CDS position, to a first approximation (i.e., abstracting from basis risk) leads to an identical allocation of risk and the elimination of the putatively perverse incentive.  Win win.  Value created that can be shared.

EFP trades have been common features of derivatives markets since their beginnings (though they go by a bewildering variety of names).  Nothing new here.  Nothing complicated here.  Very Practical, in fact.  Practical Men have used them for a long time in many markets.  Nothing Ivory Tower about them.  Indeed, I can say from 20+ years of academic experience that 90+ percent of Ivory Tower finance people have never heard of an EFP.  I learned about it from being in the markets.  

Let’s stipulate that this isn’t being done, at least in quantity.  That means that the costs exceed the benefits.  This could be because the costs are high, or because the benefits are low.  Neither Felix nor any of the other disquisitions on this subject I’ve read present a convincing case that the benefits are high (because the empty creditor problem is particularly pernicious.  The absence of these sorts of trades is also consistent with the benefits of eliminating the empty creditor problem being small.  

Some of the very facts that Felix cites are consistent with the conjecture that the empty creditor problem isn’t that serious because workouts prohibitively costly when this problem is absent.  A lot of firms go into bankruptcy–meaning that pre-bankruptcy workouts are not universal.  Again, this was true before CDS, and is true for firms for which CDS hedging is not that material. This means that the transactions costs of the workout are sometimes prohibitively costly even in the absence of hedged creditors.  Indeed, I would argue that the relative rarity of workouts in situations not complicated by CDS suggests that the incremental contribution of CDS is likely to be small.  

Felix makes some lazy arguments.  About laziness, in fact: “Bondholders have discovered that if they hedge themselves in the CDS market, they can be lazy about such matters.  And there aren’t any easy ways to get a lazy bondholder do hard negotiating work.”  

Again.  If the empty creditor problem–now renamed the “lazy bondholder problem”–is that serious, then laziness is very costly.  There is value on the table that can be used to bribe the lazy to transfer their positions to somebody who is not lazy.  Indeed, the truly lazy can completely wash their hands of everything and pick up some spare change in the bargain.  Sounds tailor made for the lazy.  

There has always been the issue of control rights not being held by those best able to exercise them.  Insurance companies buy corporate bonds not because they have a comparative advantage in participating in workouts, but because the risk-return characteristics of the bonds meets their investing needs.  In the event of an issuer’s financial distress, the insurance company is unlikely to have the comparative advantage in participating in the workout.  The company could–and sometimes does–sell bonds to someone who can exercise the rights more efficiently.  But often it does not.  One reason for doing this is that the rights are not that valuable in the first place–due, in large part, to the arduousness of workouts even in the absence of hedging.  

In the end it’s all about a very basic economic proposition.  A very practical economic proposition that is lived out in markets every day, as it has been for millennia past.  That proposition is that resources flow to their highest value uses.  As modified by Coase, the “highest value use” takes transactions costs into consideration.  Those worrying about the lazy bondholder problem assert that CDS dramatically raise the transactions costs and impede the flow of control rights into their highest value use.  The case is not proven by a longshot.  

Felix objects to my starting the story at the end–i.e., when financial distress looms, rather than when the hedger initiates a CDS trade.  This objection is readily addressed, and indeed, extending the analysis in that direction strengthens my point.

First, that’s the natural starting point in addressing the empty creditor argument, precisely because that’s where Hu and others started it.  

Second, and more importantly, taking this into consideration forces one to consider the effects of CDS in their entirety–something I did in the last paragraph of my post.  The possibility that value will be destroyed due to perverse incentives in the event of financial distress will be priced into the CDS when a hedger initiates a position.  In particular, the costs arising from an inefficient resolution of a financially distressed firm will tend to increase the costs of protection accordingly; the protection seller, anticipating the possibility of making larger payouts in the event of financial distress on the named credit, will price that into the protection.  If consenting adults make the trade nonetheless, it means that the net benefit of the CDS transaction to them is positive even taking into account the possibility of a lazy bondholder problem.  Although it might be desirable to undertake actions that reduce transactions costs in the event of financial distress, there is no justification whatsoever for restricting trading of CDS based on fears of the empty creditor problem.  You may save on some costs, but only by imposing even greater costs in return.  You know that you are imposing even greater costs because you are preventing consenting adults from engaging in trades they deem to be mutually beneficial even given the possibility of the lazy bondholder problem down the road.  

This understanding motivates a very Practical Way of analyzing how to improve the CDS markets.  Focus on reducing the transactions costs of dealing in CDS and corporate bonds.  Avoid measures that constrain the ability of market participants to trade CDS–either to initiate, or reverse positions.  Sadly, to the extent the worries about the empty creditor problem contribute to the CDS are bad meme, they are invoked to support measures that would restrict CDS trading, thereby eliminating some mutually (and socially) beneficial trades.  As an example of the former policy, moves to improve standardization and fungibility would facilitate the development of EFP-style trades in CDS.  (Lack of standardization and fungibility make it costlier to deal with others than with the original counterparty, thereby limiting the possible universe of CDS trades.)  

A couple of final points.  Felix chides my CDS auction speedup proposal because it doesn’t address the pre-bankruptcy workout issue.  I make that very point in my post.  

And finally, Felix states that my original post is a “very long” one.  LOL.  That’s nothing, dude.  Want to see long?  Prowl around the site awhile and you’ll see long!