Drivers of the gold price


28 Aug 2012 by Jim Fickett.

Valuing gold is neither as simple nor as complex as some would make it out to be. Fundamentally, for everyone except a few idealists, gold is money, and its long-term value is determined by, in essence, exchange rates. The price is presently high according to current exchange value, but perhaps not according to anticipated exchange value. To me this suggest that one's hedges against inflation should include some gold but also some commodities and equities that are relatively less expensive.


First let's get a few irrelevancies out of the way

  • “Gold is nonproductive and a lousy investment”. It is true that farms produce food (and expenses), apartments produce rent (and expenses), and companies produce profits (and losses), while gold produces nothing. But productive or not, the desirability of gold is deeply rooted in the human psyche (Many people seem to want physical gold, in hand), and the price goes up and down, so volatility can produce returns.
  • “Mine production is up (or down) this year”, or “The cost of production is rising (or falling)”. It doesn't matter. Mining only adds about 1-2% to the existing stock of gold each year, so even if nothing were mined, it wouldn't change the market much.
  • “Gold preserves value” or “gold doesn't really preserve value”. Both are often claimed, and usually the person making the claim has chosen the time period to make his/her point. In fact, gold more or less preserves value on average, over the long term, and in the short term it goes up and down a lot. It definitely matters when you get in and out. ( has a 600-year price chart.)
  • “If we have hyperinflation, the price of gold will skyrocket.” True, but not good enough. The price of everything will skyrocket, and taxes will attach a payload of lead to many rockets. One needs a more quantitative analysis than that.

Fundamentally, what drives the price of gold is people's desire to hold something of real value. In India it is simply part of the culture to see gold as the natural store of savings. In China and the US currently, the desire for gold is at least partly a reaction to financial repression (negative real interest rates engineered by the government), which erodes savings in the bank. In Europe it is fear over the return of national currencies, likely to be accompanied by big devaluations. Globally, there is a general sense that many debts will never be paid, and currencies may be inflated. Overlaying the fundamentals, there is always, of course, crowd momentum, speculation, and trader volatility, but the fundamental driver of the price is a desire for a store of value.

If you just want some insurance, some emergency money to help you survive in case of an apocalypse, the price doesn't, perhaps, matter too much, and a few ounces might make a big difference (Sizing an insurance position in gold).

To invest in gold, with the hope of a good return, one needs to somehow quantify the real value in gold that people are willing to pay for. And because gold is money, it all comes down to comparing gold to other money.

Two main valuation approaches

One approach that has served me well is to look at the record, over several decades, of the purchasing power of gold, that is, its inflation-adjusted price in dollars, or the gold-to-constant-dollar exchange rate (Historical prices of precious metals). That analysis suggests that gold is currently quite expensive, as its inflation-adjusted price in dollars is more than twice the 40-year average. This approach, then, suggests that there is no margin of safety in the current price of gold, and one should be cautious about buying in at current levels.

To be quite correct, I should have said “purchasing power for the American consumer”, rather than “purchasing power”, above. The dollar may be the single most important currency in the world, for now, but it is not the only one. So some people track the exchange rate of gold against other currencies, or against a basket of currencies. Any of these approaches will suggest that the current price is somewhat high. (Some also like to compare the gold to silver price ratio over time. I find this approach to be of limited value. It is true that the two tend to rise and fall together, to some extent, and so have at least some tendency to have a fixed price ratio. But silver is more of an industrial metal than gold is, and mine supply matters much more (see, for example, Silver supply and demand update for 2010), so the comparison is, in my view, less useful than a comparison to currencies.)

Gold is, in one very important respect, different from all other major currencies – since the centuries-long quest for the philosopher's stone was ultimately unsuccessful, governments cannot create more gold at will. Hence gold is often held as a hedge against the loss of value in other currencies, and so its value may anticipate future exchange rates.

Peter Bernholz, in his monumental study of inflations throughout history, showed that the money supply typically rises earlier and faster then the price of better money (including gold), which rises earlier and faster than the prices of goods and services (Bernholz, Monetary Regimes and Inflation, 4). In other words, buyers of gold see the money supply rising, and buy in anticipation of future inflation.

So an alternative to the basic approach of looking at gold's current purchasing power is to look at gold's price discounted by the money supply, as a possible indication of its future purchasing power. Rich Toscano of Pacific Capital Associates wrote this approach up nicely in a blog post, and I reviewed the approach in The gold-dollar exchange rate suggests gold is fairly valued. This approach suggests that gold is approximately at fair value when discounted by the money supply, and history suggests that its real value will rise further if high inflation takes hold.

(Note that again these data are in terms of dollars. A global analysis would be useful but, lacking that, we can note that many governments besides the American one are also creating large amounts of money.)

Bottom line for current decisions

Taking the two main valuation approaches together, then, suggests

  • Looking at gold's current purchasing power, it is expensive, but
  • Looking at the (quite likely) scenario of high future inflation, it may be fairly valued for future conditions and,
  • Given historical precedent, there may not be a much cheaper entry point before the future inflation runs its course

If you are more concerned about not overpaying, perhaps you should weigh more heavily the current exchange value, which does not depend on guesses about the future. If you think high future inflation is very likely, and would really like to own some gold as part of your portfolio, then the money supply argument suggests that now may be as good a time as any to get in.

Personally, I come down somewhere in the middle. I have about 5% of the portfolio in gold, but do not want to go higher than that. All commodities provide imperfect inflation protection (Commodities provide less inflation protection than you might hope), and tax on fictitious nominal gains makes the problem much worse. So one wants not just an inflation hedge, but an inflation hedge bought at a good price. On the whole I agree with Jeremy Grantham, that many commodities and equities can provide decent inflation protection, and some of them can still be bought cheaply (natural gas stocks, for instance). (For relevant quotes from Grantham see Some practical investment thoughts from Jeremy Grantham and Grantham fall letter).

Footnote -- market direction is not value

As with equities, most discussion of gold in the news is not about value, but about the direction of the market, i.e., is it going to go up or down tomorrow? Trying to guess the direction is something a value investor should not spend too much time on, but some idea of forces shaping the direction is sometimes useful. The two main rules bandied about are

  • When fear increases, gold rises, and
  • When real interest rates are negative, gold rises

The first rule is hard to make quantitative, or to apply consistently. Worries over Europe seem not to increase the dollar price of gold on a short time scale, perhaps because traders rush into and out of Treasuries based on the level of fear in the market (Rebalancing gold).

The second rule does have some validity (Gold and short-term interest rates):

Out of 13 years in which 3-month T-bill rates were less than the inflation rate (counting one borderline year with a real rate of -0.1%), gold rose in 12. This suggests the gold price is rather unlikely to fall any time soon.