CVA and vulnerable options pricing by correlation expansions (Q2241073): Difference between revisions

The vulnerable options are financial contracts that are subject to some default even concerning the solvency of the option's seller. In the present paper the plain vanilla credit value adjustment (CVA) is treated. Also it is shown that existing methodology can be extended to include some forms of the acronym \((X)\) value adjustment (XVA). The stochastic intensity approach for the time of default, when the investor might face either a total loss or a partial recovery of the investment's current value is employed. The asset price evolution is modeled as a solution of a linear equation that might depend on different stochastic factors. An approximate evaluation of the option's price by exploiting a correlation expansion approach is provided. The theoretical analysis of the paper is extended to include some further value adjustment, for instance due to collateralization and funding costs. Finally in the CVA case, the numerical performance of the developed method is compared with the one recently proposed by Brigo and Vrins and other authors. \par In Section 2 the general problem and setting are introduced. Here \([0,T]\) is a finite time interval, \((\Omega, {\mathcal F}, P)\) is a complete probability space, endowed with a filtration \(\{{\mathcal F}_t\}_{t \in [0,T]},\) augmented with \(P-\)null sets and made right continuous. The market is described by the interest rate process \(r_t\) determining the money market account denoted by \(\displaystyle B(t,s) = e^{\int_t^s r_u d u}\) and by a process \(X_t\) representing an asset \(\log-\)price. Some assumptions for the above model are presented. \par In Section 3 the details of the introduced market model are presented. the initial condition of the diffusion dynamics are introduced. \par In section 4 the convergence of the series in the case of a future contract is proved. Under some restriction on the parameters it is presented the corresponding model. Here the diffusion is considered as a Markov process. The price of any European defaultable derivative with integrable payoff is expressed as a deterministic function \(u(\cdot)\) of all initial data. In Theorem 1 the function \(u(x,\lambda, t,T,\rho)\) is approximated by means of a Taylor polynomial in \(\rho\) around 0. This Theorem shows the effectiveness of the developed approach. It is shown that this result can be extended to the European call options. In Proposition 1 it is shown that the defaultable European call increases with \(\rho\) in a small interval around \(\rho=0.\) \par In Section 5 it is shown that the correlation expansion for the more general market model developed in Section 3 can be extended to multi-factor models. Indeed the methodology remains the same and it is just a matter of handling more complex calculations that lead nevertheless to computable formulas. The details of the proposed method are presented. \par In Section 6 it is shown that under appropriate conditions, the main method of the paper can be extended to include several XVA's such a bilateral CVA, DVA (derb value adjustment), FVA (funding value adjustment) and LVA (liquidity value adjustment) due to collarization. \par In Section 7 the main features and results of the method of Brigo and Vrins are recalling. \par In Section 8 the author's method to compute CVA for a vulnerable option is compared with the method mentioned in the previous section and by using the Monte Carlo approximations as a benchmark. The obtained numerical results are graphically illustrated and discussed. \par The paper finishes with a small conclusion. Here the presentation of the CVA as an expectation of the derivative's payoff is discussed.