It has been three years since the publication of my article "The End of an Era for Gold Investors." After publication on August 3, 2012, gold reached its year high within a few months at $1,780 and has since fallen 31%. It has lost 38% of its purchasing power.
I'll be the first to acknowledge that very few of my macro-based predictions ever turn out half this well, so maybe I just got lucky this one time, but someone should probably do a study to see if the number of negative comments an article receives in SA is a consistent predictor of accuracy. There were 229 comments, of which 216 were negative. I very much appreciate the handful of commenters who added additional, useful information, regardless of their opinion, but the vast majority of comments were just rude and added no information or insight that was of any use to anyone. Gold now trades at a 5-year low, and so far, no one has apologized.
Aside from the interesting potential indicator we might be able to derive from the number of negative comments in SA, I'm interested myself in figuring out if anything else has changed. In this follow-up, I'd like to review the factors I discussed in that article, namely inflation, exchange rates, real interest, credit default premiums, and growth in the Fed Balance Sheet, and then try to address a few others that were suggested by SA commenters. After that, I'll tell you how my model has performed, and what it is forecasting now.
I was actually surprised by the push-back on my statement that gold has no intrinsic value. This statement upset a lot of people, but it is actually self-evident, and the amount of controversy it generated is probably the most important reason so many people got the metal wrong. When Robert Balan argued essentially the same thing at a different time and a different place, he drew no fire at all.
Most of the critics understood "intrinsic value" to mean something very different than what I intended. As one who measures the value of assets for a living, I view intrinsic value as a price, stated in today's dollars, that the asset can reliably be expected to return to over time because of the discounted value of the income it produces. Gold does not produce income, and, therefore, by definition, has no intrinsic value. That's it. There is nothing to argue about.
Those who even have an inkling to argue about this in anyway are universally defining intrinsic value as "something worth having." They are not wrong, but they are having a different conversation, and it is a conversation not related to anything I was talking about in the article.
Since I have many other forms of insurance available to me, I am interested in purchasing this particular form of insurance if and only if I can obtain it cheaply. If I wrote an article about how to buy life insurance for the best price possible, I would not expect to get a lot of rude comments about why I should just always buy life insurance, even though, arguably, that is a much more valid statement than the one that I should just always own gold.
I argued that the annual change in the price of gold had little relationship with the annual rise in the CPI. This is very disappointing the search for an inflation hedge is the main reason that most people have any interest in gold in the first place. And between 2007 and 2012, there appears to have been a much closer relationship with inflation, but the scale of the linear chart deceives the human eye. Since we think it should be correlated, it looks correlated. Actually, the R2 of the annual change in gold prices and that of the CPI was just .17 between 1970 and 2012. If we count only the period between 2002 and 2012, the R2 actually falls, to just .158.
A significant amount of the feedback I received related to mistrust in the official CPI statistics. Most of these comments referred to the data produced by John Williams in his newsletter Shadow Government Stats. I should point out that most professional economists do not find any evidence to support Shadow Stat's contention that hedonic pricing adjustments in the calculation in the CPI account for anywhere near as much as a 7% reduction in reported inflation. I am aware that a lot of economists are often wrong about the economy, but that doesn't give me any reason to believe that I am smarter than they are. I still have to weigh the arguments of one economist against the arguments of another and try to figure out which one makes the most sense.
William's math is just very peculiar. Personal computers, microwave ovens, televisions, and other commodities for which hedonic models were more recently introduced have a combined weight of only about 1 percent in the CPI.
James Parsons points out in the Policy and Economy that if you believe the Shadow Stats version of inflation, then by extension, you believe there was never any bubble in the real price of housing in the United States. In fact, according to the Shadow Stats version of the CPI, real housing prices declined by 60% between 1980 and 2010. If you believe this, you have the burden of trying to explain what has happened to real housing prices in the subsequent period. That is a very heavy burden in my view.
Parsons further points out in a subsequent article that there are a number of institutions who publish their own inflation statistics, independently of the government. Google has their own price index. The Billion Prices Project sponsored by MIT is another. These institutions create their own index, and these indices are different enough from that of the government to demonstrate independence, but they are both far closer to the government's version than they are to the shadow stats version. So if you believe that the government is conspiring to mislead the public, you are also forced to conclude that MIT and Google are also lying to the public, and that only John Williams is telling the truth.
I could still be wrong, but it hardly matters. Neither Williams' approach to inflation nor anyone else's approach do anything to explain the drop in gold prices during the past three years.
Inflation, real or imagined, doesn't help us predict Gold prices.
I was actually so bold as to predict dollar strength. Unlike gold, currencies do have intrinsic values because you can hold them in the form of short-duration interest bearing securities, and when combined with relative inflation rates, you can make reasonably accurate long-term forecasts. As it turns out, the US dollar index is up 17% against a basket of other currencies.
Unfortunately, the dollar's rise does not does not explain very much of the decline in gold. The dollar's rise occurred almost entirely in the second half of 2014. Gold's decline was most precipitous in the first half of 2013, but has otherwise been fairly steady.
Real Interest Rates
During pretty much the same period as the gold decline, real interest rates on the 10Yr US treasury flattened, to -0.12%, from a trough of -1.99. I took a lot of flak for suggesting you should sell gold while at the same time claiming that negative real interest rates were not sustainable. The argument made by so many was that surely gold would be safer than bonds in such an environment. Let's see how that panned out. The price of the constant maturity 10Yr US treasury index is down 6.7% from that point in time, but interest income compensated so that the equivalent total return index is flat. If you held these securities in a taxable account, your down as much as 3% maybe, depending on your tax rate, but during the same period, gold is down 31.5%.
Still, the relationship between real interest rates and gold prices is tenuous at best, and without an ability to forecast real rates on a regular basis, it really doesn't matter.
Credit Default Rates
I argued that this factor wasn't really very important because the two times that credit spreads had spiked, and caused significant damage to other asset classes, gold did not act as a useful hedge, primarily because the dollar usually spikes during such periods. Maybe the default rates are needed to explain why gold doesn't perform even more poorly during such periods, but this doesn't amount to a reason for owning gold. I am still not moved to believe this indicator has any predictive ability.
Growth in Fed Balance Sheet
I showed how a 1373% rise in the federal balance sheet over a four year period had no visible impact on inflation. Since my original piece, Fed Reserve assets have grown another 25%, yet gold has declined. Europe's balance sheet has risen faster, and Japan's is certainly not shrinking. This factor simply has no explanatory power.
Some of the more thoughtful comments I received revolved around a concern that the model I was using was too dollar-centric. Some of these people didn't realize that I was trying to refute the model, not support it, but I don't think that inclusion of any non-U.S. factors would necessarily enhance it. It is true that gold prices are determined by buyers and sellers all over the world, but if we knew exactly what demand and supply would be, we wouldn't need a model. The purpose of the model is to predict supply and demand. We would need to have a factor that predicted global supply and demand that is significantly independent of all the other factors we have already used. Most foreign interest rates are not independent of the US rates, so in fact, US dollar based interest rates and credit spreads serve as a viable proxy for the multitude of interest rates and credit spreads all over the world. External factors that are not captured by these US-centric factors are often captured by the exchange rate.
A number of people mentioned debt, but the oxford model already includes the government balance sheet, and credit spreads, and real interest rates. Including any other factor related to debt would not add new information.
My Model Outperformed
If instead of trying to make gold conform to some preconceived notion of what we think it ought to be or do, we simply accept that it is a speculative instrument, we have a better chance of making sense of the moves as pure speculative waves of fear and greed. With gold, these waves have tended to be extremely long in duration. Perhaps because there are no real fundamental factors forcing the price back to any kind of intrinsic value, speculative waves feed on themselves and exhibit extremely long duration - probably longer than any other asset class.
This characteristic lends itself to pure technical analysis. Obviously, trend following methods will work better than mean-reversion methods. I don't divulge the exact parameters of my model, but almost anyone can construct a model using any number of moving averages. My most important contributions are 1) that I use gearing because that is the only way to beat a buy & hold strategy during bull markets, and 2) that I use shorts when bearish, just because I desire to make money during bear markets. Since the article was written, gold is down 31% and my model is up 18%. Currently, the model is still short gold.
Disclosure: I/we have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.
I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.