We compare two different bilateral counterparty valuation adjustment (BVA) formulas. The first formula is an approximation and is based on subtracting the two unilateral Credit Valuation Adjustment (CVA)'s formulas as seen from the two different parties in the transaction. This formula is only a simplified representation of bilateral risk and ignores that upon the first default closeout proceedings are ignited. As such, it involves double counting. We compare this formula with the fully specified bilateral risk formula, where the first to default time is taken into account. The latter correct formula depends on default dependence between the two parties, whereas the simplified one does not. We also analyze a candidate simplified formula in case the replacement closeout is used upon default, following ISDA's recommendations, and we find the simplified formula to be the same as in the risk free closeout case. We analyze the error that is encountered when using the simplified formula in a couple of simple products: a zero coupon bond, where the exposure is unidirectional, and an equity forward contract where exposure can go both ways. For the latter case we adopt a bivariate exponential distribution due to Gumbel to model the joint default risk of the two parties in the deal. We present a number of realistic cases where the simplified formula differs considerably from the correct one.
Completing CVA and Liquidity: Firm-Level Positions and Collateralized Trades
目前实施的双边CVA具有违反直觉的效果,即从自己不断扩大的CDS利差中获利,即在实践中增加违约风险。Morini & Prampolini(2010)提出的CVA和流动性的统一图景有助于理解这一点。然而,在实际执行中有两个重要的遗漏,它们来自同一来源,即在通常的头寸保持系统中没有登记的头寸。第一个遗漏是公司层面的头寸,这些头寸在公司违约时会改变价值。商誉就是一个例子,它是资产负债表上的一个项目,通常在违约时减记为零。另一个例子是公司股权。第二个遗漏涉及抵押头寸。当这些头寸在未来出现资金短缺时(这是一个积极的概率),它们将需要无法使用头寸本身获得的资金。这些或有未来融资头寸通常不会记入任何头寸保持系统。我们在这里展示如何包括这两种类型的头寸,从而帮助完成CVA和流动性的统一图景。对于一个特定的大型复杂的金融机构,在保守的假设下,盈利$2.5B from spread widening we show that including Goodwill would have resulted in a $ 4B损失。虽然我们不能对其抵押衍生品投资组合进行类似的评估,但我们计算了一系列掉期的融资成本和自身违约的CVA,并发现CVA在示例中是一个积极的贡献。