Systemic risk: prevention and cure
From my ivory tower, this has been a great year for the study of financial crisis. Numerous experiments have taken place, which give us opportunities to understand the financial system. Three of these fall under the banner of "systemic risk": the BSE payments crises of June, and the recent problems at UTI and the US hedge fund Long Term Capital Management (LTCM). Systemic risk is the term employed to convey a breakdown in the financial system, going beyond one company.
As an illustration, consider the BSE payments crisis. This works in two steps. Initiation: A few brokerage firms with large positions default. Contagion, or "domino effect": Some brokerage firms default because they are denied the payments they expected to receive from the clearinghouse. Other brokerage firms fear that if they pay in funds due from them, they might not get the securities that a bankrupt clearinghouse owes them; this generates incentives for them to not pay, thus exacerbating the crisis.
When a payments crisis takes place, the integrity of the financial system is violated in the sense that people who have paid money fail to get their securities or people who have supplied securities fail to get funds. The exchange clearinghouse is a powerful center of transmission from which one failed firm can hurt the entire system.
Every scenario for systemic risk has these key features: initiation of a few firms going bankrupt, and a transmission mechanism through which the fallout can hurt bystanders. Notice the contrast between the financial sector and manufacturing in this regard. If (say) Maruti goes bankrupt, there is no way in which this can impose a negative externality upon the system.
With UTI and LTCM, there is no exchange in the picture. The systemic risk posed by these entities derives from sheer size. LTCM had positions in excess of $50 billion, and its failure could have driven atleast a few of its counterparties (in off-exchange trades) bankrupt. The total assets with UTI amount to over 10% of India's equity market capitalisation. A run on US-64 would impose two kinds of externalities:
- It would have a long-term impact upon the trust placed by investors in the financial system,
- In the absence of an index futures market, UTI would have no alternative but to sell on the cash market. The cash market, today, lacks the liquidity to absorb such sales. This would have a serious impact upon the allocative function of capital markets.
This emphasises the interplay between (a) the size of UTI and (b) the liquidity of Indian markets. If UTI were smaller, and if the markets were more liquid, then large redemptions would be less fearsome. Similarly, if markets were more liquid, an unceremonious closeout of LTCM's positions would have been possible.
It is fashionable these days to criticise all bailouts, and bailouts which are insurance policies for well--connected incompetents are obviously unsavoury. However, we cannot deny that there is a legitimate tradeoff. Consider LTCM, where it is easier for us to be detached. If there is no bailout, the LTCM failure imposes externalities upon innocent bystanders, and could possibly generate a crisis in the entire system. If a bailout is organised, it encourages large institutions to be reckless in the future, knowing that help will arrive when failure is imminent.
I believe that the truth lies somewhere in between. Government needs to find a way to balance the objective of protecting systemic integrity versus the pitfalls of "crony capitalism". There is no recipe through which this balance can be found. The only answer, for us as a country, is to build a government staffed with individuals of the highest calibre and immune to political interference, who can make judgements on such problems on a case--by--case basis. This is extremely hard, but there is no shortcut through which this can be avoided.
On a more cheery note, there is a lot that we can do to prevent such crises. Some of the ingredients of a policy framework which would diminish the possibility of such outcomes are:
- Securities exchanges should be built to the highest standards of reliability of the payout. NSE's clearing corporation is an important role model here. Individual bankruptcies have taken place, but they have never snowballed into systemic risk.
- The LTCM crisis is a reminder of the intrinsic strengths of exchange-traded derivatives when compared with off-exchange derivatives. India's fixed income and currency markets still lack the institutions of an exchange.
- When a crisis takes place, remedial actions by the regulator should ensure that the same problem should not recur. We can recall three `hot spots' with BSE payments problems in the recent three years, which suggests that SEBI's reactions have been inadequate.
What is the remedy for US-64? It should ultimately become just another scheme operating under SEBI's policy framework for NAV computation, NAV based pricing, portfolio disclosure, etc. This may be infeasible in the short run, but it is the only viable destination. In this sense, it is a major step forward for UTI to have gone public with the problems in US-64. We understand the need for NAV-based pricing much better today than was the case in 1964; there is no need to cobble along with the mistakes of the past indefinitely.- UTI and LTCM pose risks to the system through a combination of size and market illiquidity. There is an obvious mismatch between the size of SBI and UTI and the liquidity of our markets as of today. Capping the size of financial entities would help address the first, improving market liquidity would address the next. As markets get more liquid, larger intermediaries would work well.
Market liquidity (and systemic integrity) on the equity market would improve through (a) depository adoption, (b) derivatives trading, and (c) complete transition into rolling settlement. An anonymous electronic debt market would improve the liquidity of large bond portfolios, such as US-64 or SBI. UTI and SBI would be the greatest beneficiaries from a movement towards these four goals.- Policy discussions should explicitly worry about the once-in-a-few-years events where market integrity is tested. When badla was resuscitated in late 1997, it appeared to work well for a few months before exchange mechanisms were unable to cope with a burst of volatility. Such bursts of volatility are unpredictable. It may take months or years before they come. Systems are only tested in times of market volatility; short track--records of functioning under normal conditions should not generate confidence.
The best designed regulatory framework will fail to generate safety if it is not backed by effective enforcement. This takes us back to the slow process of building government agencies staffed with "individuals of the highest calibre and immune to political interference".
Back up to Ajay Shah's media page