For anyone who has been ultra-bullish on gold (NYSEARCA:GLD), the last month has undoubtedly come as a great shock. Some have claimed that it was a shell game all along. Although I have written about the behavior of gold in greater detail on previous occasions, I think a more comprehensive review of why the conventional views on gold prices are flawed (no matter what ideological or economic school you subscribe to) and where we should look for more reliable indicators would be timely. Much of this applies to silver (NYSEARCA:SLV), as well, which has also been battered since 2011 (down 50% now).
First, let's look at what are some of the most common myths about what drives gold behavior. I should say that I do not regard "myth" as necessarily a pejorative term, but I believe that a comparison of the relatively short history of gold trading (only four decades) with the broader history of the market can move us beyond the hyperbole that tends to overwhelm a more dispassionate view of what gold is all about.
1. Gold is manipulated, possibly illegally so.
This is the least compelling myth, but I would like to inoculate my article against this belief from the outset, simply because it is the one-size-fits-all explanation anytime gold behaves contrary to expectation. Every time gold fails to conform with one of the beliefs discussed below, someone will announce that this is definitive proof of foul play. But, I think the burden of proof must be greater than that. It is safe to say that governments, central banks, and financial institutions take an interest in precious metals, but on the principle that the markets constitute, in the end, an organic whole, any significant bout of manipulation would also have to be manifested in some other area of the market. If one could convincingly demonstrate that level of impact, a case might be made for a perversion of market forces, but without that, even if one could prove that an individual, an institution, or a cabal had managed to affect the price of gold in some fashion, I do not think that it is sufficient to demonstrate a significant level of manipulation.
2. Gold is a hedge against inflation.
It is not always clear what precisely is meant by this. Should gold be highly correlated with the general price level (e.g., CPI), or should it be highly correlated with the rate of change in CPI? My sense is that people tend to think of the latter relationship, although insofar as their argument is based on the notion of how gold ought to react to currency debasement (see Point 3), it would imply the former.
In fact, I think there is an element of truth to both notions. If you look at the following chart, you can see that one could make reasonable arguments for or against the gold-inflation connection, depending on the standard and time frame you selected.
(Source: Roy Jastram's The Golden Constant)
Once the market was allowed to determine the price of gold in 1971 and up until the dot.com crash in 1999-2000, one of the most satisfying fits was the correlation between the real price of gold and the rate of inflation.
But, that correlation appears to have broken down rather definitively since then, leading one to conclude that something fundamental has changed in the last decade, the inflation numbers are fudged, or the link between inflation and the real price of gold was somewhat incidental to begin with.
Over the long term, however, I think we can begin to describe the relationship between gold and inflation thusly: gold responds well to the persistent inflation associated with fiat monetary regimes (e.g., during the greenback era and especially since the end of Bretton Woods). Since the Nixon Shock of 1971, gold has performed well relative to consumer prices and--what may prove to be a more relevant benchmark--to other commodities.
At the level of the generational supercycle (say over a couple of decades), however, we have to qualify that. For reasons that will become somewhat clearer further on into this article, I think one can say that in periods of accelerating inflation (such as in the 1970s), gold (and silver) can be expected to outperform virtually every other asset class and most other commodities. In periods of low inflation or falling inflation, however, gold's record has been mixed. In the 1980s and 1990s, the price of gold fell along with the rate of inflation, but in the 2000s, inflation was relatively tame, yet gold rocketed upwards.
One way to fix that "problem" would be to reject the inflation data produced by the government. And, yet, appealing to the alternative numbers produced by ShadowStats creates problems of its own. According to them, the rate of inflation bottomed in the mid-1980s and then turned up after that. If that is the case, that would "explain" why gold rose in the 2000s, but it creates the new problem of why gold fell throughout the 1990s while inflation supposedly pushed upwards. That completely skirts the issue of why gold fell for so long whether inflation remained high and persistent (again, according to ShadowStats) or relatively low but persistent (according to the BLS) throughout the 1980s and 1990s.
Source: (Courtesy of ShadowStats)
So, from the two-dimensional perspective of how gold relates specifically to consumer inflation, if one feels comfortable waiting out bear markets in precious metals and commodities decades at a time, on the long-term horizon, an investor could argue that gold does very well under a paper currency and, as a consequence, could then remain unperturbed by crashes like the one we are experiencing now. The only people who could endure such bear markets with equanimity would probably be those who have gone out of their way to keep their exposure to gold to a minimum or perhaps to those who think of their wealth in intergenerational terms. Possibly also masochistic investors. For most others, I think, since gold has, over the long-term, been serially uncorrelated with the broader market, understanding generational swings in gold is a worthwhile pursuit.
In sum, there is a link between gold and inflation, but it is a somewhat roundabout one. Cluing in on the precise identity of that missing link does much to solve the problem. For now, although this is a slight over-simplification, I will note that the rate of inflation since the 1960s has tended to approximate the dividend yield, but also the spread between the dividend yield and other yields, especially the earnings yield.
3. The real price of gold is an inverse function of real interest rates.
In other words, when real interest rates rise, gold falls, and when rates fall, gold rises. This was suggested by Lawrence Summers and Robert Barsky in their paper [PDF] on Gibson's Paradox. Unfortunately, the connections between either the real or nominal prices of gold are difficult to substantiate. Barsky and Summers drew their conclusion on the basis of a sizeable portion of theoretical speculation combined with eleven years of evidence. Almost immediately after their paper was published in the mid-1980s, the correlation broke down. (If I understand correctly, GATA or its proxies pursued litigation against Summers and Alan Greenspan, among others, partly on the basis of this breakdown).
(Sources: Shiller, Jastram)
There is a hint of truth that can be excavated from Barsky and Summers' paper, however. By a rather convoluted route and in not so many words, they noted that the real price of gold was inversely correlated with nominal interest rates and equity yields (e.g., earnings and dividend yields) under the gold standard. Picking up from Robert Shiller, I believe, they also noted that the real rate of interest seemed to be inversely correlated with equity yields. It is not especially clear why Barsky and Summers chose to substitute the real rate of interest for equity yields, but they did. This would imply, however, that the real price of gold (as well as of other metals, they argued) should be positively correlated with real interest rates--the precise opposite of their conclusion.
As something of an aside, if it were true, as orthodoxy holds (as far as I can gather), that a) the real rate of interest is necessarily correlated with equity yields, b) the Fed has the power (in theory) to determine the real rate of interest, and c) that Shiller's P/E10 is necessarily correlated with equities, then d) real interest rates would be highly correlated with stock market behavior and e) the Fed would implicitly have the power to determine stock levels, a doubtful proposition, either theoretically or statistically.
In any case, like the less substantial (inverse) link between metals and real interest rates, the longer-standing (inverse) link between real interest rates and equity yields, already under increasing strain after the establishment of the Fed, collapsed shortly after they published their paper.
What remains, however, is the link between metals and equity yields, more on which in a moment.
Before that, the currency question.
4. Gold is inversely correlated with the value of the dollar.
This is really something of a variation on the previous two points. It assumes that virtually all changes in prices over the last few centuries are a result of the rise and fall in the value of the currency, whether it be gold under the gold standard or the dollar under the dollar one. One problem, however is how one measures the value of the dollar.
From the chart below, it is easy to see why one makes the connection between a weak dollar and strong gold, but a closer look would reveal that the bulk of the gold run of the late 1970s was accompanied by a strengthening dollar, while the weakening dollar of the mid-1980s only saw a bounce off the lows in gold.
If you use the dollar index, you are only comparing the dollar to other currencies, all of which could rise or fall in terms of purchasing power simultaneously. So, purchasing power itself would seem to be a better gauge of the potency of a given currency than would a currency index. A broad basket of goods and services such as PPI or CPI would therefore likely be a decent measure, so that the value of the currency is the inverse of the level of the general price level. But that really only brings us back to Point 2 above (i.e., that the price of gold is connected with inflation).
Some dogmatic gold bugs, I suspect, will argue that gold is naturally money and therefore that the value of a given currency is the inverse of the price of gold itself. Theoretically, that is an unassailable argument, so long as one can prove that gold is naturally money and behaves as such, but in terms of maximizing one's potential wealth, it is analytically useless, or at least so philosophical as to be applicable only in an oblique manner.
5. Gold is a barometer of fear.
In other words, gold is a function of risk-aversion. This argument recognizes the break with inflation and gold's habit of showing up in the middle of periods of maximum economic stress, such as in the late 1970s and then again in the years surrounding the most recent financial crises. Gold appears to fare well under periods of rising inflation or the threat of deflation. One notes, too, that over the last century, gold has often fared best during periods of equity weakness (e.g., the 1930s, the 1970s, and the 2000s).
This is perhaps the most satisfying thesis, but its greatest shortcoming is that it is almost impossible to measure or predict. During the top of the precious metals run-up in 2011, when everyone was talking about the break-up of the eurozone, at what point would one have known that risk and fear were finally exhausted? Is the current sell-off in precious metals a product of overwhelming euphoria?
Perhaps fear or stress is a "non-trivial" component of the price of gold, but it replaces a relatively difficult problem (understanding the price of gold) with an infinitely more difficult and intangible one (measuring and predicting financial abstractions centered on emotional states).
The overwhelming problem with any analysis of gold and, in fact, every major asset class is that we are trying to make our way through a monetary/financial system that is wholly unprecedented--or substantially unprecedented, if that's not a contradiction-- and which has created patterns of behavior that did not occur prior to the dollar standard. One must either come up with a new theory from a whole cloth or do what one can to extrapolate from historical patterns. Where possible, the first option is preferable. But, until a new theory is found, historical analysis probably remains our best shot.
Commodities and equity yields
As I've pointed out before, I believe that there is one historical pattern that goes a long way to explaining the behavior of gold, and that is the connection between commodity prices and equity yields. In slightly simplified terms, the "real price" of commodities has been highly correlated with equity yields for at least 140 years and, I believe, for the 140 years prior to that. (Although it is immaterial in the context of this article, I believe that under the gold standard, the nominal price level was the real price level). Once gold became freely traded after the fall of Bretton Woods, it also became highly correlated with equity yields, which have also become more closely connected with the rate of inflation (I have reduced this relationship to a formula here).
(Sources: Shiller, Jastram, and Stephan Pfaffenzeller)
Gold and silver also have a handful of special relationships with equity yields that make it easier for us to grasp the forces moving them, which I will summarize as follows:
1. Under the dollar standard (and in a complete inversion of the historical pattern of the gold standard), the real price of gold is highly correlated with equity yields. Also, the ratio of gold and silver prices to other commodities is correlated with equity yields. (In terms of commodity ratios, this might be more true for silver than for gold).
2. Gold and silver tend to have blow-off tops once equity yields peak, such as in 1974, 1980, 2008, and 2011.
(Sources: Shiller, World Bank)
3. Under the dollar standard, gold appears to slightly lag other commodities (notably oil).
4. Because oil is a leading indicator of equity yields and gold lags in the strength and timing of its 'response,' the ratio between oil and gold tends to indicate future momentum in equity yields and, therefore, the commodity complex.
Some of these characteristics are undoubtedly odd, and because they have only been around for forty to sixty years and the commodity market itself is so poorly understood, it is hard to say how enduring they will be, but so far, the current behavior of gold appears to be aligned with conditions we should associate with a bearish precious metals market under the dollar standard:
1. Equity yields have been falling (P/E ratios have been rising).
2. Oil peaked in 2008.
3. Industrial and agricultural commodities are significantly down from their 2008 and 2011 highs.
4. Unemployment appears to have topped out.
5. Precious metals (especially silver) are falling faster than other commodities (NYSEARCA:GSC).
As of this sell-off, gold appears to be behaving as it has for the last four or five decades, as a super-commodity linked to equity yields. Therefore, over the course of the next year or two, gold and silver will likely exhibit continued weakness. A bottom can probably not be called without a reversal in equity yields, the first inklings of which will probably be manifested in the energy complex, especially oil, and would require a doubling of the oil/gold ratio, if history is any guide. A sharp rise in the oil/gold ratio would suggest either a further collapse in the price of gold (to $700) or a surge in oil (to $160) or more likely a combination of the two (oil at $120, gold at $1000?). Without going into the nuanced relationship between the price of oil (NYSEARCA:OIL) and the yield curve here, one would generally expect an upturn in the former to be preceded by a flattening of the yield curve (although not necessarily an inversion). Although a downward blast in yields could get us there this year, it seems more likely that such a narrowing would have to come from upward pressure on short-term rates, a scenario that seems more likely next year than this year, to say the least.
Taken from another perspective, if we take it for granted that stocks (NYSEARCA:DIA) are currently behaving as if they were in a secular bull market, driving equity yields down, one wonders how long and far equity yields can fall. Although mathematically there is no upper bound to P/E ratios, one would imagine that there is a practical one, implying that equity yields would eventually have to bottom in the not-too-distant future (a couple of years?) and then rise, bringing commodities up along with them. The post-Bretton Woods system seems to have a bias in favor of lower equity yields (especially dividend yields), so it is possible that they could reach levels that have only ever been seen in the late 1990s, but how much lower could they realistically be expected to fall?
Even though current conditions are all confirming one another in the stock market's recent victories over gold and commodities, by extrapolating from the modern behavior of markets, it would not be surprising to see a return to bullish commodity conditions perhaps by 2016, assuming, I suppose, that everyone's capacity for surprise has been sufficiently diminished by now.
I would not touch commodities or precious metals with a ten-foot pole this year myself, and I would probably only switch to a five-foot pole next year, but a reasonable argument could be made that equity yields are already unsustainably low (on a very long-term basis) and on that view, as well as the relative strength shown by gold since the 1950s, gold might be attractive to long-term investors at these prices.
In sum, for as long as the world has markets, gold will probably have an important role to play in them, and it gives every indication of being as obedient to economic law as any other asset. It behaves much like the rest of the commodity complex, although it does seem to have a unique way of relating to both equity yields and inflation caused by paper debasement. Trying to weigh the various probabilities and how they might unfold in time, one could reasonably extrapolate, I think, from history that gold has a great deal of relatively short-term (over the next couple years) downside risk (I would pencil in $1000) but with an even greater degree of medium-term (over the next decade) and long-term (decades) upside potential. This story is not over by a long-shot.
Disclosure: I am long June Dow futures. I 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. I have no business relationship with any company whose stock is mentioned in this article.