Gold predictions

All that glitters

May 25th 2008
From Economist.com

Rely on ratios at your peril

PLENTY of people believe in owning gold as a safeguard against hyperinflation or financial apocalypse. They see the rise in gold to more than $1000 an ounce earlier this year as signalling the debasement of paper money, particularly the dollar.

But what does the price of gold really tell us when measured against other assets? One particular ratio that might be of interest is the relationship between gold and the oil price (see chart). In other words, how many barrels of oil does it take to buy an ounce of gold?

According to Capital Economics, if we go back to the early 1980s, it has taken, on average, 16 barrels of oil to buy one ounce. Given that the current ratio is around 7, that suggests either that oil is very expensive or gold is very cheap. Indeed, if you believe in the latter, then gold could be heading above $2000 an ounce.

But should this relationship be constant? After all, most of the gold that has ever been mined is still in existence, but we are using up oil at a tremendous rate—and some believe global production may be at, or close to, its peak. In addition, oil is vital to the global economy; the main industrial use for gold is jewellery. So that suggests oil should be rising relative to gold over time.

David Ranson of Wainwright Economics, a gold enthusiast, reckons oil is rising, relative to bullion, at a trend rate of around 3.5% a year. As a result, the gold/oil ratio should fall over time. Indeed, Capital Economics reckons that the average ratio since 2002 is about 10, rather than the 16 that has pertained over a longer period.

That still indicates, with oil at more than $130 a barrel, that gold could soon cost $1300 an ounce. But, of course, there is nothing to say that the ratio has to be restored by gold rising; oil could be dear rather than gold cheap. A fall in the price of oil to $90-$100 a barrel would redress the balance.

Another relationship worth examining is that between the Dow Jones Industrial Average and the bullion price. If you go back to 1980, when gold briefly reached $850 an ounce, the relationship was around 1. But by 2000—the height of the dotcom boom—it took 40 ounces of gold to buy one “unit” of the stock market. (This requires the fiction of treating the level of the Dow as a dollar price). According to Alan Newman of the Crosscurrents website, the average ratio between the Dow and gold since 1975 has been just under 15. Currently, it’s almost bang in line.

Again, one needs to consider whether a fixed ratio is theoretically sound. After all, gold is an unproductive asset, virtually without a yield (it can be lent in the market at a cost). One would expect it to maintain its real value over time. But one would expect the Dow, as a claim on the corporate earnings and dividends of the American economy, to at least keep pace with economic growth over the long run. So one would expect the Dow to rise, relative to gold, over time. A historical average is thus of limited use.

Indeed, that is the problem of using charts in the first place. They can give investors an instant picture of what has happened in the past; academics are very dubious that they have anything to tell us about the future. It is very easy to find spurious statistical correlations. And even when there are sound theoretical reasons why a ratio should revert to the mean over time (such as profits as a proportion of GDP), it can take ages for the shift to occur. It is probably not wise to expect to make your fortune from trading the gold/oil relationship.