The Golden Constant
Below is an extract from a commentary originally posted at www.speculative-investor.com on 17th July 2005.
In last week's commentary we included a chart of the T-Bond/gold ratio and stated that the downward trend in this ratio over the past few years indicated that bonds had experienced a large decline in REAL terms and supported our view that bonds are in a secular bear market. This prompted a couple of subscribers to question whether it was appropriate to determine a market's real return by dividing its nominal price by the price of gold. We obviously think it is and today we'll take a shot at explaining why.
A critical point is that gold is always the same -- an ounce of gold today is the same as it was 5000 years ago and the same as it will be in 5000 years time -- whereas the US$ and all other forms of paper money are constantly changing. The US$, for example, might not exist in 20 years -- let alone 5000 years -- time, and even if it does exist it is certain to be a lot different than it is today. This is one reason why it makes a lot more sense to look at secular trends in terms of how the market in question performs relative to gold than how it performs relative to the dollar. But as far as determining real (inflation-adjusted) performances go, what makes gold better than any other unchanging physical quantity, or, for that matter, the many other methods that have been used over the decades to adjust for the effects of inflation?
One reason is that, as Paul van Eeden has shown, over the very long-term the US$ gold price eventually reverts to where it should be based on changes in the supply of US dollars. In other words, the fair value of gold can be calculated at any time based on the total quantity of US dollars in existence and history tells us that although the gold price will both overshoot and undershoot its fair value by wide margins it will eventually return to this 'fair value'. In effect, over the very long-term gold oscillates around a fair value and this fair value is determined by the total quantity of US dollars and the total quantity of gold.
The average commodity, on the other hand, does not behave in this way. One reason is that the price of every commodity except gold is strongly influenced by the current year's production and the real cost of production falls over long periods of time due to technological improvements. But in gold's case the existing aboveground stock is always so much larger than new mine supply that changes in the cost or quantity of annual production have almost no effect on the gold price. This characteristic results, in turn, from the fact that although gold might no longer be money in the strict meaning of the word (it is no longer the general medium of exchange), it is still accumulated as if it were money and it still trades as if it were a currency.
The above is an important part of the story, but it's not the whole story because if we just wanted to adjust nominal prices to account for changes in the money supply we could divide by M3 or whatever monetary aggregate we deemed appropriate. As a purely academic exercise adjusting nominal prices based on changes in the money supply might actually be OK, but as investors it is absolutely vital that we distinguish between those periods when inflation (money supply growth) is perceived to be causing a problem and those periods when it is not. Failure to do so will prevent us from identifying the secular trends in various markets because when it comes to the financial markets the old saying "perception is reality" is very applicable. And this is where gold comes into its own.
The gold price does not adjust in linear fashion to changes in the money supply and there can be very lengthy periods (10-20 years) when the gold price actually trends lower in the face of increases in the money supply. Instead, the gold price will tend to fall well below fair value when confidence in central banks is rising (when inflation expectations are falling) and then make a big catch-up move when confidence in central banks and the currencies they create embarks on a long-term downward trend.
The best way to demonstrate how changes in the gold price capture both changes in the supply of money (actual inflation/deflation) and changes in investor perceptions (inflation expectations) is via long-term charts showing stock market valuations -- the price/earnings ratio of the S&P500 Index, for instance -- and the stock market in terms of gold. The reason is that investors, as a group, will pay less for increases in corporate earnings that they perceive to be the result of inflation than for increases in corporate earnings that they perceive to be the result of real growth. Putting it another way, the more that profit gains are PERCEIVED to be due to a fall in the value of the money in which the profits are denominated, the lower the average P/E ratio will be. By the same token the greater the inflation problem is perceived to be the higher the gold price will become, meaning that long-term trends in stock market valuation levels should match long-term trends in the number of gold ounces it takes to purchase the stock market. Therefore, if it is appropriate to measure the stock market's REAL long-term trend by measuring its performance relative to gold then the below long-term charts of the S&P500 P/E ratio and the Dow/Gold ratio should look similar; which, of course, they do.
Now, we've used the Dow/Gold and S&P500/gold ratios for years to illustrate both the US stock market's real performance and its secular trend and no one has ever questioned our logic. This is perhaps because almost all of our readers believe that a secular bear trend is underway in the US stock market (we've found that people never dispute what we write when our conclusions agree with what they already believe, even if there are some flaws in our reasoning). However, if our logic is correct then it applies to all markets, not just the stock market.
In summary, price gains that are the result of a fall in the value of money are illusory so in order to determine a market's real performance and secular trend we need to find an objective way to adjust nominal prices to reflect changes in the value of money. Paul van Eeden's analysis of the long-term relationship between changes in the gold price and changes in the quantity of money confirm that the gold price adjusts to account for actual inflation (money supply growth) whereas the above charts confirm that the gold price also adjusts to account for changes in inflation expectations. When dealing with long-term trends this makes the gold price a good inflation indicator. In fact, it is by far the best inflation indicator that we know of.