Although nobody much believes in the efficient markets hypothesis any more, its lesser known cousin the CMH appears to go from strength to strength. CMH stands for the "costs matter hypothesis", and was first promulgated by Jack Bogle, the combative founder of Vanguard. It asserts that the one certain element in fund performance is the bite that fees and other costs will take out of the returns that the investor's money earns.

Like death and taxes, which are anything but hypothetical, fund costs are one of nature's few sure things. Mr Bogle likes to call them the "croupier's take". As in the casino, they are capable of imposing a serious drag on performance. Costs are irrecoverable and compound over time. In their Global Investment Returns Yearbook, Dimson, Marsh and Staunton estimate the long-run cost of holding equities through funds to be in the region of 3 per cent a year. At that rate costs will eat up 27 per cent of your starting capital over 10 years and almost half the long-run annualised historical equity risk premium.

As certain as fund costs are, the paradox is that measuring exactly how big a bite costs take out of a fund's performance is anything but. Meaningful standardised information on the true cost of fund ownership, as the FT regularly points out, is hard to come by. The current regulatory requirements on funds to publish total expense ratios and deceptively mild reduction in yield figures (whose significance no investor I have ever met seems to understand) fall well short of full transparency.

Mr Bogle and other proponents of the CMH point out that, among other hidden cost items, TERs do not include the impact of direct and indirect transaction costs, an important omission in an era when the portfolio turnover of the typical fund has fallen to less than a year. The "true" cost of owning an actively managed fund can therefore be much higher than its reported TER suggests.

David Fuller's view This is an informative column. Funds can be expensive, as this site has discussed on earlier occasions.