Moral hazard and credit derivatives
Have you ever heard of “credit derivatives”? Most of us haven’t, but we should become familiar with them because they are poised to transform how we think about corporations. The basic idea is that banks and other holders of corporate debt can now spread this risk to other willing bearers using a variety of intricate financial tools. In the most basic flavor, a bank that holds a loan to a company on its balance sheet agrees to pay a quarterly fee to a third party (usually an insurance company, other bank, or hedge fund) in return for a make-whole payment by the third party in the event that the borrower on the underlying loan defaults. Called a “credit default swap”, this is nothing more than insurance against the reduction in value of the loan. The CDS market is currently over $26 trillion. That is right, TRILLION. Even more profound, however, is the rethinking we will have to do about our models of corporate governance and finance in a world in which debt starts to look more and more like equity (i.e., freely traded and held in diversified portfolios by dispersed individuals and entities). One potential issue with these transactions is the potential moral hazard it creates for borrowers.
Banks engage in these transactions because while they may be the most efficient originators of loans (because of relationships, experience, backoffice support, etc.), they may not be the most efficient bearers of the risk. This is in part because government-imposed capital adequacy requirements mandate that banks hold a certain percentage of risky loans in cash to prevent bank failure. Offloading risk via a CDS transaction is a way around these requirements. There are issues here about systemic risk and central bank policy, but I want to avoid these for now, instead focusing on moral hazard. (The role Basel II will have here is an interesting subject for latter discussion.)
In a working paper, David Skeel and Frank Partnoy discuss the risks and benefits of credit derivatives. One problem they highlight is the moral hazard that the offloading of risk by a bank creates for the borrower. The story goes something like this: banks serve a valuable monitoring role for certain borrowers that keeps managers from incurring a large debt and then following an overly risky or self-serving strategy; and risk sharing reduces the incentives of the bank to fulfill its monitoring obligations.
This result, however, is not an equilibrium solution, since it is difficult to imagine that sophisticated financial giants engaging in these transactions would act in ways that are so obviously inimical to their interests. So why is there no real moral hazard risk?
For one, the CDS market, which is the biggest credit derivative market (by far), consists almost entirely (95%) of borrowers that represent the largest 250 firms in the world. The bank monitoring role is much less for these firms because they are well followed and therefore information asymmetries between the bank and others are likely to be very low. In other words, it is unlikely that the bank will know things about GE that the hundreds of market professionals following GE don’t know. Banks are thus limited in their ability to shed risk by the underlying nature of the borrower. No one will buy the risk on opaque firms, since those are the cases where the bank may have an informational advantage, and we would expect a lemons market.
Of course banks don’t want to be limited in their risk sharing in this way, since it is precisely the riskier assets that they want to hedge. (Basel II, by the way, will encourage the down-market trend of credit derivatives by allowing banks to differentiate capital cushions based on the underlying riskiness of corporate loans, as opposed to applying a one-size fits all approach.) For firms where information asymmetries are large and the bank has information investors don’t about the quality of credits, there is a greater possibility of risk sharing leading to moral hazard problems. Again, however, such a solution would be out of equilibrium. How do banks and risk protection sellers solve this problem? One answer would be a ban on these transactions, as apparently the German government has done, but this may prevent Pareto-optimal contracting. Thankfully, there is a simple market solution.
Banks sell a portfolio of riskier loans (from the protection sellers’ point of view) to a special purpose vehicle (SPV) designed to hold these assets. The SPV, with the help of an investment bank and monitored by a trustee (from another large bank, typically), then divides the various seniorities of loans into tranches, which are sold off to different institutional buyers. These transactions are known as collateralized debt obligations (CDOs), and resemble similar products in home mortgages, credit card receivables, auto loans, and so on. The most senior (safest) are sold to insurance companies and regional banks, while the junior and equity tranches are sold to risk-seeking investors, like hedge funds. And here is the monitoring incentive solution: the originating banks routinely take the riskiest tranche, making sure that in the event of a default, they are the first to lose. Banks therefore have an incentive to monitor. I am trying to get data on exactly what positions banks take so that we can quantify the reduction in risk, and therefore any reductions in monitoring incentives, but it is possible that the transaction is risk neutral for the bank but still beneficial, since capital adequacy requirements can potentially be avoided even in that situation.
Whatever marginal decrease in monitoring incentive remains is likely cured by reputation. Banks, who are repeat players, rely on their reputation for not shirking or selling lemons. Buyers can also rely on the reputation of the SPV sponsor bank and the trustee bank to fulfill their monitoring obligations. (The market in each of these areas is highly consolidated and each of these market participants is a global financial player facing huge reputational penalties for bad acting.) Finally, borrowers by and large have no transparency into whether or not loans have been subject to these transactions, which makes any attempt to capitalize on a reduction in monitoring guess work at best.
My point is a simple one: moral hazard problems are recognized and (largely) solved via contract and product design. This contractual solution to the moral hazard problem is also used by banks in related areas, like loan sales and credit syndication. (Amir Sufi has a forthcoming paper in the Journal of Finance on this subject.)
We should expect the number and type of borrowers involved in these transactions to increase, not only because banks see the potential for risk sharing, regulatory relief, and fee generation, but also because the new capital adequacy rules will encourage it. The German reaction, if true, is likely an overreaction that is harmful for social welfare, and we should, at least at this point, resist similar calls for regulation.
This is but one of many issues raised by these financial contracts, and I plan to return to this subject again, coming back to talk about the implications for bankruptcy and workouts, insider trading, and other issues.