London, 10 May 2011: Reuters

Spare a thought for the oil market's option dealers. While last Thursday's massive price fall will have frayed the nerves of futures market makers, it must have seemed far, far worse for many option desks.

At least profits and losses on a futures position are linear. Losses on a $12 oil price move will be four times those on a $3 move. Payoffs on option positions, however, are highly non-linear. Losses on a $12 price movement could be many multiples of the losses on a $3 shift. Price moves of $12 in a day are not supposed to happen -- at least not often. While the stress on futures books will have been high, the stress on non-linear option books will have been immense.

For a dealing desk trying to run a broadly price-neutral (" delta-neutral") option book collecting premium income from clients rather than speculating on the direction of prices, 4- and 5-standard moves present a severe risk management challenge.

With prices and option deltas shifting around so much, trying to manage risk exposure by re-hedging the book in the futures and option markets is tough for even the most experienced option writers.

RISK SYSTEM OVER-RIDES

Thursday's price move will have generated a blizzard of exception reports in the risk management systems of the major banks and hedge funds, as delta, gamma and volatility ("vega") all exceeded pre-set limits. There are two basic responses. Dealers can mechanically obey the risk management system and try to reduce their exposure. But in many cases that would have locked in losses. It would also have involved trying to short a market that was already falling rapidly.

Hedging option exposures in the futures and cash markets can often make a crisis worse -- as writers of equity options ("portfolio insurance") discovered during the 1987 stock market crash.

In fact, shifting deltas and the flurry of gamma-hedging almost certainly contributed to the record turnover and increase in reported open interest during Thursday's oil crash, and probably deepened the sell off, as option market-makers struggled to square up their positions as prices sank.

The other alternative is to ignore ("over-ride") the risk management system and try to ride out the storm in the hope prices will rebound and limit eventual losses if not eliminate them entirely. The strategy permits a limited excursion outside the pre-set parameters in the hope that it will prove temporary and less costly than trying to keep position within the limits at all times.

The downside is that it is also a (thinly disguised) directional bet on a price rebound. Knowing when to over-ride the formal risk management system is the secret of successful trading -- even if it does not generally find favour with the tidy minds of regulators.

MANAGING THE EXCEPTIONS

Rigid adherence to risk parameters is not an optimal strategy for dealing desks (or for markets). It would cause needless volatility precisely when markets most need calm. Perhaps the best analogy is with grid control in electricity transmission and distribution systems. Grid controllers balance power supply, demand and frequency across the system by bringing on or shutting down generating units.

If frequency deviates too far from the grid's design level (50 Hertz in the United Kingdom, 60 Hertz in the United States) it can produce cascading power failures. It is a matter of professional honour to keep the frequency steady. Grid controllers and operators compete to produce the most steady printout.

In practice, however, it is impossible and excessively costly to maintain frequency at exactly 50 or 60 Hertz, so systems are designed and operated to maintain frequency in an acceptable band (49.8-50.2 Hertz in Britain)

http://www.nationalgrid.com/uk/Electricity/Data/Realtime/Frequency/Freq60.htm

Even then this band will be breached occasionally. The "art" of grid management is to set the control limits well within safe levels and to operate the system within them so exceptions occur very rarely and even then can be dealt with safely.

Well run financial systems should operate in the same way. Risk limits are there to be broken from time to time but exceptions should be rare and carefully controlled. Limits need to be set conservatively so even when they are breached, there is still an addition margin of safety for the institution.

FLASH CRASHES MORE OFTEN

The oil price crash has worried some observers because it had no obvious cause. To use a weather analogy, it was a tornado that formed without warning, rather than a hurricane that could be identified and tracked. That arbitrariness is frightening for investors and hedgers who prefer slower moving, more predictable and more rational markets, leaving time to exit or cut losses.

But it is only the latest in a string of sudden and apparently arbitrary crashes, including stock market crashes in 1987, August 2007 and May 2010, as well as similar flash crashes in commodity markets such as raw sugar in February 2011. Flash crashes appear to be occurring more often. There are likely to be several causes for the increased short term volatility:

(1) The increasing volume of computer-driven algorithmic trading has dramatically accelerated the speed at which markets react to information (including price changes) compared with the old system of trading pits and telephone broking. Crashes that might once have occurred in slow motion over several days can now be compressed into a single session.

(2) In recent years, options have overtaken futures as the main vehicle for speculating on oil prices and hedging commercial exposure. The increasing popularity of optiontrading, structured products and other non-linear strategies compared with futures and swaps has added to fragility.

Large price moves are now more likely to beget more big moves as option writers move to re-hedge their exposures dynamically.

(3) Short-term speculative interest in commodity markets and equities has increased relative to the volume of underlying hedging and long-term "real money" investment. In normal times, heightened hedge fund and speculator involvement may add to liquidity. But in a crash that liquidity can disappear, and short-term speculators can become consumers rather than suppliers of liquidity, exacerbating the price move.

BIOLOGY RATHER THAN PHYSICS

Commodity markets, including oil, are best thought of as complex systems in which behaviour and subjective expectations matter as much as objective supply and demand fundamentals. They are characterised by powerful network effects and feedback mechanisms. Outcomes tend to be probabilistic rather than deterministic, and are highly non-linear.

Commodity analysts and economists are often accused of suffering from "physics envy." But biology and quantum mechanics may be a better analogy than Newtonian physics for the way in which markets behave and prices are determined.

Several observers have noted similarities between critical market states such as price crashes and biological phenomena such as epidemics and population collapses. Transmission of ideas and behaviour through the market network and the feedback mechanisms become central to understanding how prices move.

Biological analogies may explain why flash crashes appear to be happening faster and more frequently than before. In the same way increasing volumes of air travel have increased the risk and speed at which disease outbreaks spread around the globe, computer-driven trading, the growing popularity of non-linear strategies, and mounting volume of price-tracking purely speculative activity have all increased the potential for flash crashes to be transmitted through the market.

The dark side of market integration and financial complexity is growing fragility and risk of crashes. But market users and regulators may have to learn to live with it. No one would suggest abandoning air travel to limit the risk of influenza turning up in London. In the same way there may be no realistic way to prevent increasingly fast and integrated markets suffering periodic crash events.

Ends --

By John Kemp, Reuters market analyst – for Commodities Now. The views expressed here are his own.