This document presents a methodology for estimating value-at-risk (VaR) for dynamic hedging. It involves calculating moments of the hedging error distribution by propagating matrices backwards on a stock lattice. Convex optimization is then used to estimate VaR from the moment information. The methodology is applied as an example to calculate VaR for a hedged European call option on a trinomial lattice stock model.
This document presents a methodology for estimating value-at-risk (VaR) for dynamic hedging. It involves calculating moments of the hedging error distribution by propagating matrices backwards on a stock lattice. Convex optimization is then used to estimate VaR from the moment information. The methodology is applied as an example to calculate VaR for a hedged European call option on a trinomial lattice stock model.
This document presents a methodology for estimating value-at-risk (VaR) for dynamic hedging. It involves calculating moments of the hedging error distribution by propagating matrices backwards on a stock lattice. Convex optimization is then used to estimate VaR from the moment information. The methodology is applied as an example to calculate VaR for a hedged European call option on a trinomial lattice stock model.
This document presents a methodology for estimating value-at-risk (VaR) for dynamic hedging. It involves calculating moments of the hedging error distribution by propagating matrices backwards on a stock lattice. Convex optimization is then used to estimate VaR from the moment information. The methodology is applied as an example to calculate VaR for a hedged European call option on a trinomial lattice stock model.
Download as PPT, PDF, TXT or read online from Scribd
Download as ppt, pdf, or txt
You are on page 1/ 13
VALUE-AT-RISK ESTIMATION
FOR DYNAMIC HEDGING
GROUP 6:Prashant Kumar
Parikhit Ghosh 1 (2012-14)
Concept of VaR and Dynamic Hedging
VaR is an attempt to providing a single number summarizing the total risk of the portfolio of financial assets. Banks regulators also use VAR in determining the capital a bank is required to keep for the risks it is bearing. Dynamic hedging involves monitoring a portfolio's position delta and readjusting this value as it deviates from a target number. 2
Methodology In this work, an efficient methodology for analyzing risk in the wealth balance (hedging error) distribution arising from a mean square optimal dynamic hedge on a European call option, where the underlying stock price process is modelled on a multinomial lattice. By exploiting structure in mean square optimal hedging problems, it is shown that moments of the resulting wealth balance may be computed directly and efficiently on the stock lattice through the backward iteration of a matrix. 3
Based on this moment information, convex
optimization techniques are then used to estimate the Value-at-Risk of the hedge. This methodology is applied to a numerical example where the Value-at-Risk is estimated for a hedged European call option on a stock modelled on a trinomial lattice.
Introduction Option pricing & hedging theory have been the core of modern mathematical finance since derivation of Black-Scholes formula. The key to their formula is that there exists a trading strategy which constructs a portfolio that perfectly replicates the payoff of a call (or put) option under the following two assumptions: the underlying stock price follows a geometric Brownian motion, trading may take place in continuous time.
The Black-Scholes analysis demonstrated that an option can be
created synthetically by dynamically trading in the underlying asset. However, continuous trading is never possible, and there always exists a hedging error, i.e., perfect replication is not 5 possible if the market is incomplete.
Objectives The objective of this work is to: Provide a new tool to analyze the risk in a dynamic
hedge when the market is incomplete.
Use option pricing and hedging theory to provide a theoretical value and hedging strategy for European call/put options.
The Wealth Balance of a Dynamic Hedge
The wealth balance of a hedged call option writer corresponds to the value of a portfolio consisting of writing and then hedging a European call option. Here it is proposed efficient methodology to estimate the VAR which involves the following two steps: 1. Under a mean square optimal dynamic hedging strategy, we show that the moments of the wealth balance distribution may be efficiently calculated using a backward recursion on the underlying stock lattice. 2. We then apply a convex optimization approach to compute upper and lower bounds on the VAR given the first m moments of the wealth balance distribution.
Since both the moment estimation procedure and upper and lower bound problems are efficient, we conclude that the resulting methodology provides a fast and effective algorithm for estimating the risk in dynamic hedges. 7
Optimal Hedging Strategy
We consider a hedging scheme which minimizes the mean square value of the wealth balance. In this paper, we associate this optimal initial portfolio value with the price" of the option, and assign C0 = 0. Under this price, the average wealth balance satisfies E (WN/ S0. 0) = 0 , and the objective function in MSOH becomes the variance of the wealth balance. It has been shown that the above price" can lead to arbitrage opportunities. 8
Backward Recursive Calculation of
Moments Here, a recursive algorithm to find moments of the wealth balance by solving the problem backwards on a multinomial stock lattice is proposed. Then the algorithm computes moments of any order efficiently with polynomial time computational complexity. To explain the idea, we first consider the second moment case, and then generalize to the m-th moment case. After that Algorithm and Computational Complexity Analysis is done.
Outcome: Exact computation of the VAR may involve an exponential order
computation and hence is generally computationally intractable. But, here the proposed methodology for calculating moments, is highly tractable since the moments are computed in polynomial time by propagating matrices Hn (n = 0,..,N). Although, the moment computation procedure for a European call option is explained, the same approach can be extended to other types of options, including many exotics and options with time optionality. The algorithm only requires a change in the boundary condition" corresponding to the appropriate option type, and proper discounting to account for the time value of different wealth balance cash flows. 10
VAR Estimation Under Moment
Constraints Here, convex optimization techniques to the problem of estimating the Value-at-Risk of the wealth balance WN under moment constraints is applied. In the optimization, the constraints are convex with respect to (x). Therefore, if X were fixed, the problem of minimizing through an appropriate choice of (x) would be an infinite dimensional convex optimization problem. Based on the results of Bertsimas and Popescu, the dual of this problem may be reduced to a semi definite programming problem which can be solved efficiently using interior point methods. The dual problem can be interpreted as a generalization of the well known Markov and Chebyshev bounds, and the solution to the dual problem is equal to the solution of the primal. 11
Numerical Experiments Here, the VAR in a mean square optimal hedge for a European call option is estimated. After computing moments of the wealth balance distribution, we first apply the semi definite programming approach to find hard upper and lower bounds on the VAR. These bounds are then improved by posing additional constraints. Here it can be concluded that the bounds obtained from moment conditions can be improved by using additional information such as the monotonicity constraint, and that the problem remains computationally tractable, providing a reasonable estimate of the VAR.
12
Conclusion First, a mean square optimal hedging problem was employed to determine an optimal hedging policy and optimal initial portfolio value. By exploiting structure in this problem, it is showed that moments of the resulting wealth balance distribution may be computed on the underlying stock lattice through the backward iteration of a matrix. Next, it is demonstrated that these moments can be used in conjunction with convex optimization techniques to estimate the Value-at-Risk in the wealth balance. Finally, this methodology was applied to a numerical example where the VAR was computed for a hedged European call option on a stock modelled as a random walk on a trinomial lattice. 13
1-Value at Risk (Var) Models, Methods & Metrics - Excel Spreadsheet Walk Through Calculating Value at Risk (Var) - Comparing Var Models, Methods & Metrics
The FundamentalThe Fundamental Theorem of Derivative Trading - Exposition, Extensions, and Experiments Theorem of Derivative Trading - Exposition, Extensions, and Experiments
1-Value at Risk (Var) Models, Methods & Metrics - Excel Spreadsheet Walk Through Calculating Value at Risk (Var) - Comparing Var Models, Methods & Metrics
The FundamentalThe Fundamental Theorem of Derivative Trading - Exposition, Extensions, and Experiments Theorem of Derivative Trading - Exposition, Extensions, and Experiments
