Financial risk assessment: a prelude to active risk management at Dr. Reddy

Abstract

Dr. Reddy's is growing at a tremendous rate both organically and inorganically. The risk profile of the company is also growing at the same time. This is because of diversification into newer geographical areas (Mexico and Europe), increase in the portfolio size and highly polatile Forex and interest rate market, the variables that affect Dr. Reddy's portfolio the most. In order for Dr. Reddy's to focus on its core businesses, the need of the hour is to establish an active Risk Management system to prevent company from incurring staggering losses because of adverse market movement. The starting point for this is risk assessment, for which Value-at-Risk is the most researched and widely used tool ayailable.VaR is intuitively defined as the worst loss over a time horizon with a given confidence level . The important parameters in VaR estimation are time horizon, which we recommend to be 1 day for daily mark-to-market monitoring, and confidence level, taken as 99% to reduce frequency of loss exceeding VaR estimate. Another important characteristic to be determined for VaR estimate is the probability distribution for the return on portfolio, for which various methods are available namely Historical Simulation, Variance-Covariance, Constant maturity, and Monte Carlo Simulation . A snapshot of VaR estimate for Dr. Reddy's portfolio is detern1ined to be 38.6 Crores using Historical Simulation, the method that giyes the most conservative estimate. To use VaR for implementing active Risk Management, the target Ievel for the loss is determined in discussion with the management. If the VaR estimate exceeds tlus target level. treasury is mandated to hedge the risk, so as to maintain the VaR estimate within the target level. However, if it is within limit, treasury should have the autonomy to express theiropil11on about market movement through their investment decisions. However, VaR has got a few limitations as well. It does not give any indication of the "expected loss proyided loss is more than the VaR estimate", information that can be obtained from Conditional VaR. Hence VaR should be used in conjunction with Conditional VaR. Moreover, it is pretty cumbersome to estimate VaR on a day-to-day basis manually. Hence it is required to invest in software for VaR estimation.