He’s your guy when stocks are high
But beware when they start to descend.
It’s then that those louses go back to their spouses –
Diamonds are a girl’s best friend.
Anita Loos, whose character Lorelei Lee sings those lyrics in Gentlemen Prefer Blondes, didn’t have a PhD in economics but her observation on the inverse relationship between investment returns and the value of precious metals (or diamonds) was ahead of its time.
In 1985, about 60 years after the creation of Lorelei Lee, Lawrence Summers, then a Harvard economics professor, and Robert Barsky, another Harvard economist, published a study of gold prices and real returns that is well worth revisiting.
Mr Summers had a difficult time as president of Harvard, a post he is giving up at the end of this week. He will go back to being a professor there, where he has done great work in the past.
An example of that is “Gibson’s Paradox and the Gold Standard”, the study referred to above, the first version of which was published in 1985. It seeks to explain why the price level and nominal interest rates move together when monetary theory says interest rates should move with the rate of change of those levels. Mr Summers and Mr Barsky used data from the period of the gold standard along with real returns, interest rates, prices, and gold prices during the post-gold standard period.
What Mr Summers and Mr Barsky found was that the price of gold moved inversely with the real returns people can earn on capital. “The willingness to hold the stock of gold depends on the rate of return available on alternative assets. We assume that the alternative assets are physical capital and bonds.” This makes intuitive sense; gold is, over time, a way to preserve capital, not increase it. If you can earn high rates or profits you will be induced to sell gold and invest in productive capital.
While Mr Summers and Mr Barsky went on to other topics, others updated their study over the years. Peter Palmedo, who runs Sun Valley Gold, a gold asset portfolio management company, did a regression analysis of gold against the S&P 500 monthly returns from 1994 to the end of 2002 and found a negative correlation of 94 per cent.
So, between 1994 and 2000 large capitalisation stocks earned annual returns of 15.6 per cent, while gold earned an annual return of -7.6 per cent. During the last gold boom, between 1970 and 1980, gold earned an annualised return of 19.9 per cent, while large cap stocks brought in 0.2 per cent, not enough to keep the Lorelei Lees in diamonds.
I called Mr Summers to discuss his old paper, and the 1990s real returns results. “These movements you cite were a vindication of the paper Barsky and I wrote,” he said. “It stands to reason that high real returns were associated with a period when gold prices fell, or the subsequent decline [post-1999] was associated with an increase in the price of gold.” Indeed, the gold price bottomed in September 1999 and US stock prices peaked in March of 2000.
According to Mr Palmedo’s analyses, if real returns on capital drop only to their long-term average of 5 per cent, instead of overshooting to the low side, gold’s real returns will be more than 6 per cent. If returns drop to 2 per cent (still above those of the 1970s), gold’s real returns rise to 12 per cent or better.
There is one anomaly that shows up in the application of the Summers/Barsky model over the 1990s. While the real returns on risk capital fit the model very well, real interest rates in the late 1990s were below the predicted trend.
The equity return numbers do make sense. Mr Summers and Mr Barsky wrote that the real rate of return is “subject to shocks. These shocks reflect changes in the actual or perceived productivity of capital.” In the mid to late 1990s that productivity rose due to the application of computing and communications technology, as well as the rapid increase in developing world productivity.
Real interest rates in the late 1990s, however, were held down below where they should have been. Alan Greenspan’s Fed, along with its counterpart central banks, kept real rates lower than they should have been to offset the effects of financial crises, the Asian flu in 1997, the LTCM/Russia crisis of 1998, and the Y2K non-crisis of 1999/2000. So real interest rates during this period did not track real returns as they should have.
This has been turned into a conspiracy theory by goldbugs. Ignoring the tight negative correlation between equity returns and gold, they have focused on the low interest rates, concluding that they illustrate a suppression of the gold price. Reg Howe, a leading goldbug and a partner in Golden Sextant Advisors, wrote in 2001 that: “The historical evidence adduced by Barsky and Summers leaves but one explanation for this breakdown of Gibson’s Paradox: what they call ‘government pegging operations’ working on the price of gold,” adding that “Barsky and Summers underscore the futility of trying to control [long-term rates] without also controlling [gold prices].”
I went to see Mr Barsky to discuss this interpretation. He has not been close to Mr Summers in the intervening decades but dismissed the notion that the former Treasury Secretary was carrying out some conspiracy to suppress gold. “The ‘government pegging operation’ we referred to was the gold standard itself. By definition, there wasn’t a “free market in gold” during the period of the gold standard. This wasn’t a policy prescriptive paper at all. In any event, the causality we identified in our paper was: real rates of return drive real interest rates and the gold price. The gold price doesn’t drive real interest rates.”
There was, many would argue, a policy failure from the late 1990s to the end of the Greenspan era: monetary policy was too loose, and that is leading to serious trouble. But it didn’t have anything to do with gold price suppression.
While the work of Mr Summers, Mr Barsky, and Mr Palmedo does not cover what the professors refer to as “high frequency” fluctuations in the gold price, the thesis does steer investors in a useful direction: years of low return on risk capital go with years of high returns on gold.