This blog post is not about whether CVA/DVA accounting makes sense or not; it is only about whether it makes sense to hedge the DVA. Modern accounting standards require derivatives and many other financial assets and liabilities to be stated at fair value. Fair value must take into account all characteristics of the instrument including the risk of non performance (default). CVA and DVA arise out of this fair value accounting.
CVA or Credit Value Adjustment accounts for the potential loss that the reporting entity would incur to replace the existing derivative contract in the event of the counterparty’s default (less any recovery received from the defaulting counterparty). It obviously depends on the probability of the counterparty defaulting and on the recovery in the event of default. More importantly, it also depends on the expected positive value of the derivative at the point of default – if the entity owes money to the counterparty (instead of the other way around), the counterparty’s default does not cause any loss.
DVA or Debit Value Adjustment is the other side of the same coin. It accounts for the possibility that the reporting entity itself could default. One could think of it as the CVA that the entity’s counterparty would need to make to account for the default of the reporting entity. It accounts for the potential loss that the counterparty would incur to replace the existing derivative contract in the event of a default by the reporting entity (less any recovery received from the reporting entity). It can also be thought of as the notional gain to the reporting entity from not paying off its liability in full. The DVA depends on the probability of the reporting entity defaulting, the recovery in the event of default, and the expected negative value of the derivative at the point of default.
DVA can also be applied to any liabilities of the reporting entity that are accounted at fair value and not merely to derivatives, but the logic is the same.
The application of CVA and DVA in valuing assets and liabilities on the balance sheet is perhaps the only logical way of applying fair value accounting of assets and liabilities in which non performance risk is material. But the accounting standard setters took another fateful and controversial decision when they mandated that changes in CVA and DVA be included in the income statement instead of letting it go straight to the balance sheet as a part of Other Comprehensive Income. As I said at the beginning, this blog post is not about the merits of this accounting treatment; I mention the accounting rules only because these rules create the motivation for the management of financial firms to try and hedge the CVA and DVA.
Hedging the CVA is relatively less problematic as it only increases the resilience of the firm under conditions of systemic financial stress. Counterparty defaults are somewhat less threatening to the solvency of the entity when there are hedges in place even if there could be some doubts about whether the hedges themselves would pay off when the financial world is collapsing. What I find difficult to understand is the hedging of the DVA.
The DVA itself is a form of natural hedge in that it produces profits in bad times. It is when things are going wrong and the world is worried about the solvency of the reporting entity that the DVA changes produce profits. One could argue that the profits are notional, but there is no question that the profits arise at the point in time when they are most useful. Hedging the DVA would imply that during these bad times, the (possibly notional) DVA profits would be offset by real cash losses on the hedges. A position that produces losses in bad times is not a good idea. Such positions have to be tolerated when they are intrinsic to the business model of the entity. What baffles me is why anybody would willingly create such wrong way risks purely to hedge an accounting adjustment.
The Modigliani Miller argument in capital structure theory (home made leverage) can be extended to hedging decisions (home made hedging) to say that hedging is irrelevant except when it solves a capital market imperfection. Bankruptcy costs are a major capital market imperfection that can make it advantageous to undertake hedging activities that reduce the chance of bankruptcy. In this framework, the only hedges that make sense are the ones that hedge large solvency threatening risks. The DVA hedge is the exact opposite. It produces large cash losses precisely at the point of maximum distress. For example, this Wall Street Journal story says that Goldman Sachs implements a DVA hedge by selling credit default swaps on a range of financial firms. The trouble with this is that these hedges will produce large cash losses when many other financial firms are all in trouble, and this is likely to coincide with troubles at Goldman Sachs itself. Far from mitigating bankruptcy risks, the hedges would exacerbate them.
The only way this makes sense is if investment banks think that losses during systemic crises can be pushed on to the taxpayer. If this assumption is correct, then DVA hedges work wonderfully to socialize losses and privatize gains!