An increasing proportion of large international banks' profit is being earned through proprietary trading. "Trading" is of course a euphemism for speculation, which, at one time (certainly when I joined banking 50 years back), was the dirtiest word in the banker's dictionary.

Traditionally, the banking business comprised intermediation and agency activities; in recent years, it has been rapidly moving to proprietary activities, as banks find the risks and rewards of such business far more attractive than traditional banking.

No wonder in several banks derivatives heads are reaching top management positions (Credit Suisse, Deutsche Bank, SoGen, BNP Paribas, etc.).

The accompanying table is based on data available in four major banks' annual reports for 2006 about their trading portfolios. Value at Risk (VaR) is a measure of the price risk inherent to the trading portfolio, and the numbers quoted are the average for one-day holding period and 99 per cent confidence level as disclosed in the accounts.

The capital charge for market risk prescribed by the Basle Committee on Banking Supervision is three times the value at risk based on a 10-day holding period, and has been estimated from the 1-day, published VaR numbers.

For a trading book, the biggest risk is market risk -- indeed the genesis of trading is to deliberately take market risk in order to profit from price movements. (The actual charge could be somewhat higher because of some technical issues, but not so much as to vitiate the conclusions.)

I have had to make robust assumptions while estimating the trading income: the reason is that no bank is separately giving expenses relating to the trading activity. I have therefore deducted from the income an amount equal to the proportion of total operating expenses to gross income: I believe this is a conservative assumption as, relative to income, trading activity is not very manpower-intensive -- although compensations are high.

On these assumptions, the net trading income, after expenses, amounts to several times the capital charge for market risk. The estimated numbers tie in quite well with the data published in the Bank of England's Financial Stability Reports.

The July 2006 report shows that the ratio of dealing profits as a multiple of VaR (holding period 10 days; 99 per cent confidence level), for major UK banks has been growing year-by-year. The median ratio was 25 in 2005, going up to 30 in 2006: this translates into something like an 800 per cent to a 1,000 per cent return on the market risk capital charge!

While the trading book does have a credit risk capital charge, this would be relatively small. Actually, operational risk capital charge may be somewhat high compared to market risk capital, but still leaving the RoE in triple digits.

If my assumptions and analysis are reasonably correct, the data lead to several issues and questions:

• Given the fact that an efficient and competitive market would balance risk and reward, is the risk being underestimated? Alternatively, is the BCBS capital charge for market risk too low?
• Is the risk being really borne by economic agents outside the trading system like the end-users of currencies, derivatives and bonds, while the profits come to the major dealers? End-users are also bearing the huge margins in ever more complex structures, which have little economic justification beyond the fat margin for the dealer.
• Are the major trading banks playing such an important role in the world economy as to deserve the outsize rewards?

To my mind, the positive side of the huge amount of speculative trading in currencies, bonds and derivatives is that it provides liquidity. The negative side is whether it adds to volatility of markets and therefore to the risks, and costs, of the real economy.

For, there is no question that the basic logic of institutional trading often adds to volatility. For example, given market efficiency, dealers are as likely to take wrong views as right and, therefore, stop loss reversals are an integral part of trading discipline. Thus, a trader holding a short position when the asset price is increasing would reverse it by buying the asset when a stop loss is hit, thus increasing demand and hence price.

Similarly, a player writing a call option on an asset would need to increase the delta hedge when the asset value starts moving up, further strengthening the trend.

If the banks make good money on trading, such profits are also reflected in the outsize bonuses dealers have been getting in recent years. If memory serves me right, Michael Lewis wrote in his Liars' Poker that in his third year as a trader with Salomon Brothers, he earned a bonus of a few hundred thousand dollars.

Wondering what socially useful work he was doing to deserve so much money, he quit Salomon and wrote his bestseller. (He perhaps earned millions in royalties for his books but feels that writing books is a socially more worthwhile activity than making speculative profits in financial markets.)

In today's world, Mr Lewis is obviously pass�: star traders in major banks perhaps earn bonuses in 8-figure dollars, and obviously consider themselves to be entitled to them, as the new Masters of the Universe -- notwithstanding that, as Satyajit Das wrote in Traders, Guns and Money, "often successful traders cannot explain themselves exactly how they made money": trading profits are as much a matter of luck as of skill -- and of selling complex instruments?