DVA is the controversial twin of CVA. The acceptence of DVA as a mark-to-market item on derivatives positions should be no more controversial than the acceptance of CVA. If one accepts that in a trade between institution A and institution B, institution A will mark-to-market CVA charges on their positions, then for there to be any agreement between the two parties regarding the mark-to-market or settlement value of the position, institution B  will need to mark-to-market DVA charges. DVA is essentially CVA seen from the other side of the trade.

The more controversial issue is whether or not to include DVA in CVA/Value Adjustment VaR calculation. Under Basel III, DVA is excluded, and for a few good reasons. Let us again examine the regulatory CVA formula :

Let us examine the CVA and DVA charges on trade, from the perspective of Institution A, having exposure to institution B.

The associated charges on the trade will be . Institution A receives CVA as compensation for institution B’s credit riskiness, while institution A pays DVA to compensate institution B for it’s credit riskiness.

The value adjustment VaR charge will arise due to changes of the credit spreads (and loss given default, etc.) across scenarios. is the credit spread associated with institution A, is the exposure of insitution B to a default by institution A, etc. We will focus on changes in the credit spreads and (though the same principle applies to changes in ). If the values go unhedged, then a lower value adjustment VaR will be realised if the trade is executed with counterparties with highly correlated credit spreads. This encourages systemic risk, as institutions have incentive to trade with correlated counterparties rather than diversifying their credit exposure.

Alternatively, if CVA VaR is reduced by hedging CVA (through CDS etc.), then the question becomes how to hedge DVA. This is non-trivial. At present the most popular method of hedging DVA is to hedge with CDS on proxied counterparties with highly correlated credit spreads.  One would hedge DVA to avoid mark-to-market losses in the event that the institutions creditworthiness improves. The problem with hedging this risk with a proxied CDS is that an institution has control (to an extent) over it’s own creditworthiness. If it manages to improve its creditworthiness, but credit proxies do not do likewise, the hedge is ineffective, though it is an effective hedge for systemic credit upgrades.

These are some points in favour of excluding  DVA from the value adjustment VaR calculation. A counterargument is that there is a high degree of unavoidable systemic risk between financial counterparties, and that excluding DVA from the value adjustment VaR calculation leads to a higher capital charge than is justified, or requires institutions to hedge risks which are already, in some sense, “hedged”. Either way, it is another point on which institutions may wish to diverge from the regulatory prescription, and take an independent approach to credit adjustment risk management. The sullen institution may need to simply accept the capital charge as a model risk charge.