If you believe that passive management is the same as indexation you should be prepared to think again. At any rate, a distinction is made in a paper by Mr Alan Brown, chief investment officer of State Street Global Advisors, the world's second-biggest tracker managers, behind Barclays Global Investors.

At one level, Mr Brown is making an intellectual defence of Mr Bill Sharpe's Capital Asset Pricing Model, which implies that holding a market portfolio will maximise the expected return, on a risk-adjusted basis. At another, it reads a little like a cry from the heart of a passive manager being buffeted chaotically by the technicalities of index structures.

At the heart of the argument is what he describes as the "certain metric" of tracking error. If a global fund can report just, say, 20 basis points of tracking error against a global index, great reassurance is given to investment clients. But, as with many apparently precise measurements in an uncertain world, it is achieved at a price.

As index-tracking and benchmarking grow increasingly important there are accompanying distortions. Take the S&P 500, for example. Some 200 stocks moved in (and out) during the 1990s. The average price gain for additions was 7.5 per cent and the negative impact on deletions was 4 per cent. You can argue that 3.5 per cent is the net benefit to companies of being in the index, but that amounts to a cost for investors.

Dimension Data caused significant distortions when entering London's FTSE 100 Index last October.

Some managers paid a heavy extra price by waiting for the initial closing price. This week, ironically, although perhaps predictably, Di Data left the Footsie after just 11 months - it stays in the All-Share.

Mr Brown highlights the "absurdity" of the Vodafone/Mannesmann deal in early 2000, when British managers - who usually have a strong domestic bias - were required to raise their weighting in Vodafone sharply, whereas if the deal were structured the other way around, they would have had to cut it. Mr Paul Woolley, of active manager GMO, has calculated that all passive investors would be better off if Vodafone had never existed.

State Street Global Advisors argues that long-term investors such as pension funds should not be obsessed with the ultra-close tracking of indices but should seek to be integrated with the global market. This means, for a start, scrapping the domestic bias.

For investors in small countries almost all equities held should be listed elsewhere (for US investors the international exposure, at present, would be no more than 45 per cent, but greatly more than is now common). Also, various other top down preoccupations with sectors, or with distinctions between developed and emerging markets, should be abandoned.

Other tests should be applied in judging passive managers. Are they able to minimise turnover? Do they avoid unnatural prices at times of poor liquidity?

Mr Brown does not reject the benefits of a "reference benchmark" such as one of the global indices. But he claims a loser methodology would match, say, the MSCI World Index to within a tracking error of about 50 basis points. Long-run annual cost savings would be 15 basis points, representing a 50 per cent ratio of enhanced return to extra risk.

So, State Street is seeking to avoid being a passive victim of hedge funds, index arbitrageurs and other opportunists. Not that it is opposed to the improvements being made to the global indices, notably through free float adjustments. Some 30 per cent of MSCI-linked clients have switched to the provisional indices being introduced before full 2002 rebalancing. But Mr Brown objects to the need for "knee-jerk reactions" to index changes. He says that passive funds should be providers of liquidity and should be driven by their own cash flows, rather than be squeezed by weightings changes.

The question, however, may be whether clients will be ready to relax tight controls. They would be required to accept more complicated metrics and adopt more elaborate governance principles. This looks rather like another case in which clients would rather be precisely wrong than approximately right. The attractions of spurious precision are enormous.