The IMF currently holds about 3,000 tonnes of gold, worth some dollars 40bn ( pounds 26.7bn) at today's prices. The hoard earns no interest and benefits none of the fund's clients. Mr Clarke's suggestion, made at the Commonwealth finance ministers' meeting on Tuesday, was that the IMF sell bullion over several years and invest the proceeds - up to dollars 4bn - in securities issued by industrialised countries.
The interest from those investments would be lent to the world's most desperate economies - notably Uganda, Guyana, Honduras, Kenya, Laos, Mozambique and Sierra Leone - at rates as low as 0.5 per cent. The proposal would have a similar effect to the Trinidad terms, put forward in 1990 by John Major when he was Chancellor. The Trinidad plan wrote off pounds 3bn of bilateral Third World debt but did not cover those countries that relied primarily on the IMF for finance.
Where Mr Clarke's plan comes apart is in its effect on the price of bullion. The Chancellor cannot simply wave a magic wand, or ritually sacrifice a chicken, to suspend the law of supply and demand, no matter how worthy the cause. His scheme would at best stop the gold price rising as much as it would have otherwise. At worst, it could spark a wave of panic selling and the return of a bear market.
The gold market last turned bearish in December 1987 and mauled producers and speculators alike for more than five years. Prices fell from a high of dollars 499.75 an ounce to dollars 326.10 by 10 March 1993. But in the second quarter of last year, the price started to recover. It now stands at just under dollars 400 and market observers agree the fundamentals look strong. Even on the day of Mr Clarke's announcement, the price dropped only 55 cents to close at dollars 395.40.
There is a caveat, however. A survey of dealers by Dr David Gulley, dean of the business school at the American College in London, found their biggest worry was the possibility of central banks selling from their stockpiles, currently equal to 14 years' production.
Mr Clarke wants to release up to 300 tonnes onto the market. That is more than a seventh of 1993's mine production and almost double the average annual increase in supply since 1984. Even spread over several years, it can hardly be called an insignificant amount.
There is a reasonable chance that embarrassing headlines about the price of gold crashing could be avoided. For the last decade, the most persistent seller has been the Bank of Canada, but it has few reserves left. In the first half of this year, it sold less than 36 tonnes. Virtually all other official sources dried up, causing an 87 per cent drop in net supply from that sector. The Clarke plan would restore much of that flow, neatly cancelling a significant upward pressure on gold prices.
But the knock-on effect is potentially far more serious. If the IMF's sales were timed so that the nominal price of gold remained stable for several years, speculators would have less reason to invest in it. The implied threat of increased IMF supply and decreased investor demand is a fall in real gold prices. That could trigger other bullion owners to rush to the market at the first sign of trouble, possibly before the fund even gets there.
One big advantage of Mr Clarke's plan is that it would not cost rich countries anything. That does not mean it is a free lunch. The big losers would be gold producing countries, especially South Africa and Russia. Both are important to the West and both have a high potential for political instability. Increased gold prices would help them build stronger economies and, in turn, more reliable democracies. It is patently unfair for the rich countries to expect them to pay for reducing Third World debt.
Fortunately, the IMF's 175 members would have to vote 85 per cent in favour of Mr Clarke's proposal before it could be acted upon. The producing countries, possibly supported by those with large gold stocks of their own, should be able to block it.