Many happy returns? What determines the relative performance of stock market sectors?

Macroeconomic conditions are clearly important. Cyclicals tend to outperform as the economy improves; growth stocks do best when growth is poor; and financials benefit when interest rates are low and falling.

Fashion can also matter: it is hard to explain the dot.com craze in purely rational terms.

Ultimately, however, share performance should be driven by the underlying returns of the industry in question - probably best measured by the return on capital or the return on equity.

Happily, this is borne out by a new study from Mr Peter Oppenheimer, global investment strategist at HSBC. Looking at US equities between 1991 and last year, he found sectors with the highest returns on capital, by and large, performed best on the markets.

Pharmaceuticals, for example, with average annual returns on capital of 20 to 30 per cent, have beaten the US market by around about 10 per cent a year on average. By contrast, the paper industry, with returns of between 3 and 8 per cent, has under-performed by nearly 40 per cent a year.

The evidence suggests investors are pretty good at rewarding high returns irrespective of whether these returns are improving or deteriorating. What seems to matter most is the absolute gap between a sector's return on capital or equity and its cost of capital. In a low interest and low inflation environment, this makes eminent sense.

However, Mr Oppenheimer also found there was practically no relationship between the relative performance of sectors and changes in return on equity.

In other words, the market appears to be much less astute at spotting new opportunities (sectors with improving returns) and dangers (sectors with deteriorating returns).

Take telecommunications. Average annual relative out performance of 40 per cent in the 1990s is greater than can be explained by absolute levels of return. Indeed, after improving in the early years after deregulation, return on capital started to fall from 1996 as capital employed rose in response to higher returns. Despite the falling momentum, the market continued to reward telecoms companies for increasing their use of capital and the sector continued to outperform until spring of this year.

Arguably, investors should have been much quicker to notice the alarming fall in returns on capital and much less willing to fund the sharp rise in capital employed.

The likely explanation is that they believed the technologies on which the capital was being spent, such as third-generation (3G) wireless networks, would not only maintain future returns but significantly raise them. But as this year's fiercely competitive 3G licence auctions in Europe demonstrate, high returns attract competition.

It is a rare company that manages to build a sustainable competitive advantage that allows it to generate super-normal profits regularly, as Microsoft has.

For a sector as a whole, it is practically impossible.

Looking forward, all this provides some tentative pointers for future stock performance.

Among growth sectors, telecoms will continue to consume capital as the operators build out their infrastructure. In all probability, returns will fall further over the next year, making the sector relatively unattractive. The same goes for media, with its increased investments in digital technologies.

Other technology sectors, such as software, where margins are improving, should do relatively well.

For more mature industries, the best performers are likely to pursue consolidation and actually reduce their overall capital employed, by handing it back to shareholders.

Of course, even mature companies have to be careful not to ignore growth opportunities merely to raise their return on capital. But much of the old economy these days is focused on capital efficiency, just as the new economy, blatantly, is not.

Investors should pay more attention to this.