Stock exchanges are planning to shift to risk-based margins based on the Value at Risk (VAR) model. In news reports and TV talk shows expectations are being raised about the difference VAR can make in management of settlement risk at the exchanges. Let’s discuss what is meant by VAR margin, how it is different from the present margining system, and can it fix risk management.
A stock exchange takes margin from stock brokers so that in case a broker defaults in payments or delivery of securities, any loss incurred on squaring defaulter’s positions is off set by margin. There are two types of margins, initial and mark to market (MTM). Initial margin is what a broker has to deposit prior to taking a position. MTM margin represents day end losses, and it is usually submitted after close of trading.
If initial margin is 10 per cent, then a broker would first deposit Rs10 of cash or securities before buying a security of Rs100. If by day end, price of that security falls to Rs95, then broker has to submit Rs5 as MTM margin. The rationale behind collecting initial margin is that a broker could default without having submitted MTM margin. You can think of initial margin as an estimate of maximum MTM margin.
VAR helps the exchange determine the rate of initial margin based on volatility of securities. A typical VAR margin statement states the maximum percentage price change that a security may experience with a specified level of confidence. Simply put, if VAR for a security is 10 per cent, then it means that exchange can be 99 per cent confident that maximum daily price change this security could experience is 10 per cent. If exchange takes Rs10 as initial margin, then it should be safe if by the time of squaring of positions and liquidation of margin, price of security purchased by a defaulter falls to Rs90.
VAR is supposed to replace the current slab system. Under the slab system exchanges club together all outstanding positions and charge a certain rate. As the size of positions crosses specified thresholds, the margins rate jumps up. At KSE, in ready market, the percentage rate starts from five and goes up to 25 while in futures, it starts from 7.5 and goes up to 20.
There are three important differences between VAR and the current slab system (i) Under VAR, brokers would pay lower margin on less volatile securities and vice versa, whereas in the slab system, all securities are subject to equal margin rate regardless of their volatility. (ii) VAR margin should increase with increase in volatility whereas in slab system margin rates increase with the size of a broker’s outstanding positions. (iii) VAR is a relatively objective system and it is used by a number of international exchanges whereas slab system is highly subjective and obsolete. Due to its superiority over slab system, VAR margins should be a welcome development. However, we need to be mindful that impact of VAR margins on overall risk management in this stock market is unlikely to be substantial and if and when it would be implemented also remains uncertain.
VAR only deals with initial margin which is but one of several risk management measures being used, such as MTM margin, clearinghouse protection fund, price limits etc. Other measures of risk management in our market are also subject to various deficiencies compared to international best practice.
For instance, here we are using the narrowest price limits anywhere, whereas developed exchanges usually use index-based trading halts with relatively wide or no price limits. There are some measures of risk management used internationally that are altogether missing in our market, such as securities borrowing and lending mechanism to manage delivery defaults.
Even within initial margin, there are issues that VAR does not deal with. Let’s take two examples.
(a) form of margin (cash or securities) is critical to its usefulness. VAR margin assumes liquidity, that is, defaulter’s positions could be squared and margin securities sold within a single day. Due to lack of depth in our market and concentration of exposure and margin in the same securities, this is very difficult, if not impossible to achieve.
(b) Some brokers do not take adequate margin from clients and use lower margins as a means to attract clients. Large institutional investors are also not known for depositing margin to their brokers. To meet margin calls by the exchange, some brokers pledge securities of inactive clients against their proprietary positions and against positions of active clients.
VAR models use complex statistical assumptions. Decision makers may not understand these things well and put more confidence in the system than is justified. Even if we ignore these assumptions on the grounds that these are also used in international exchanges, there are some technical issues in implementation of VAR in this market.
(a) Estimates of daily volatility for securities cannot be relied upon due to narrow price limits. Price limits also affect MTM margin because this margin is calculated based on closing prices, which cannot move beyond the limit. There is no objective way to measure how much further price of a stock would have moved in a single day if there were no price limits.
(b) Standard VAR calculation methods assume adequate liquidity in underlying security. If a security is thinly traded, then for days its price may not change, showing low volatility but in a default scenario it could be very difficult to sell.
Our market lacks depth and we have seen that in volatile times, buying interest could disappear even in volume leaders. Since 95 per cent of traded value happens in just about 30 securities, VAR would have limited application in any case. Subjective adjustments would have to be made to conventional VAR models to adjust for price limits and lack of liquidity in most securities.
Our track record on implementing ideas is unenviable. Recent examples of snail pace implementation are CFS regulations and new future contracts. Efforts to introduce VAR margins also date back to 2002 but we have seen little progress. Some stock brokers are welcoming VAR because they are under the impression that it would necessarily reduce margin and thus increase their ability to trade.
According to the VAR proposal that has been tabled, margin should be based on (a) netting of positions at client level instead of broker-level (b) elimination of netting across settlements (c) elimination of netting across securities to calculate MTM margin (d) extension of time horizon beyond one day and (e) increase in the proportion of cash. These measures may be well justified but they would most likely increase the margin burden. Broker support for VAR would evaporate and either this proposal would be spice-down or put on the back burner.
Achievements in risk management have often been exaggerated in the past. March crisis has shown that despite years of improvements, risk management system was too weak to cope with broker defaults. Due to political fall outs, defaults had to be prevented through a bailout package which included bending of rules, special Badla sessions, off-system inter-broker sale of positions, consortium buy out deal, bank financing etc. There is a fear that potential impact of VAR could be exaggerated too.
In sum, while VAR system is better than the slab system, realities on the ground demand that we greet any news about such initiatives with cautious optimism. Indeed, it would be far better that instead of piece meal improvements, the entire system for management of settlement risk is re-engineered in a transparent and consultative manner, to bring it in line with international best practices. If we shy away from dealing with the bigger issues in risk management and be content with marginal improvements, then come next crisis, bourses would once again have to rely on a bailout package.
Stock exchanges are planning to shift to risk-based margins based on the Value at Risk (VAR) model. In news reports and TV talk shows expectations are being raised about the difference VAR can make in management of settlement risk at the exchanges. Let’s discuss what is meant by VAR margin, how it is different from the present margining system, and can it fix risk management.
