The collectible premium
There are only three types of goods in an economy:The demand for every investment asset -- stocks, housing, commodities, etc. -- is a function of some combination of these factors. In housing all three demand factors are at work. Every asset that is not just a fixed monetary cash flow has a collectible premium that reflects, when rational, inflation expectations. (n.b. this is technical nomenclature that can have a slightly different meaning than the normal use of the words -- for example I, and presumably Devin, classify the enjoyment one can get from collecting baseball cards or wearing gold jewelry as part of their "utility" rather than their "collectible" nature, the latter being purely a matter of the asset's perceived or actual function as a store of value or medium of exchange).
1) goods providing direct utility ( a car, tv, chocolate )
2) collectibles ( a baseball card, diamonds, gold )
3) flows ( stocks, bonds )
It is often very difficult to guess what the factors of demand coming from direct consumption utility are (for example, is the sunny yet cool California coast and the wealthy and smart neighbors it attracts worth paying five times as much or is the coastal California price premium, as Devin suggests, primarily a collectible premium?). Sometimes it's also difficult to estimate cash flow. Thus it's often hard to estimate the collectible premium. Also, since the collectible (or inflation expectations) premium is little recognized, a wide variety of bogus explanations are often used to explain price movements that are actually due to changing inflation expectations: the most popular being dubious theories of changes in supply (e.g. "peak oil") or consumption demand, since demand for assets as inflation hedges is very under-recognized. Also common are theories that impute price changes due to changes in inflation expectations to purely irrational psychology, "technical", or "trend" factors based on the history of the price itself. (The history of housing prices from 1940 to 2005 made houses look like a lucrative and low-risk investment, for example).
Devin describes inflation expectations as causing a hunt for stores of value in which packs of investors irrationally change their focus from one market to another, causes bubbles and bursts. Since the collectible premium is often so hard to estimate, and is so little recognized, and so many investors do despite those warnings you hear invest based solely on past price history, I don't substantially disagree with this:
A stock market is in disequilibrium when people start trading stocks as collectibles rather than as flows. In other words, instead of buying a stock based on the hope of generating a return via dividends, they start buying a stock in order to sell it to someone else ( price appreciation)....In short, if long-term inflation expectations were zero, the prices of housing, stocks, commodities, etc. could be estimated from demand deriving solely from their cash flow plus consumption demand -- there would be no collectible premium. But since asset prices come with a difficult to estimate and fluctuating inflation expectations premium, this makes it far harder to judge whether an asset is over- or under-priced, leading to greater over- and under-pricing, i.e. seemingly irrational asset booms and busts.
When the government dilutes the money supply, people start searching for a replacement collectible to serve as a store of value. People end up buying stocks not based on dividend yields, but in order to trade for later at a higher dollar price. People buy houses not based on direct utility or as an alternative to paying rent, but in order to sell for later at a higher price. These goods start trading as collectibles, and thus are subject to the whims of the herd.
A couple of caveats:
(1) It's important to observe in the context of stocks that their cash flows can come from share buybacks and takeovers as well as dividends, so even with zero inflation expectations people would rationally invest partly for expected price rises in addition to dividends.
(2) There are a number of other sources of high uncertainty in asset markets, e.g. uncertainty over credit conditions, so that bubbles and bursts would not completely go away with zero inflation expectations. However the contribution of changing inflation expectations has, I believe, been the largest factor in most asset price movements since the 1970s.
Bubbles and bursts are based on high degrees of uncertainty about the future far more than on genuine mass irrationality. Of course, things like the Internet bubble seem quite irrational in hindsight, but if you keep in mind what was going on in the late 1990s -- a huge increase in subjective value due to new (to the vast majority of investors) Internet services such as e-mail, Web, search engines, and on-line price quotes -- it was not, without hindsight, terribly irrational to suppose that most of this value would be monetized as profit for the innovating companies. It turned out to be mostly unmonetized -- instead we got most of the foregoing plus shared music, Wikipedia, blogs, and other things of great value for almost free -- so that the Internet bubble collapsed and seems irrational in hindsight.
Bubbles and bursts due to uncertainty over technological innovation will undoubtedly occur again, but the greatest ongoing source of uncertainty in our markets is not regarding technology and monetization of innovation, but rather uncertainties about currencies, credit, globalization, and related political factors, out of which the uncertainty over currencies is usually the greatest.
My takes on last summer's commodity hysteria as it was happening can be found via here.
Devin has put these theories to good use:
since inflation helps build the bubble, even a small amount of deflation can cause rapid price collapses. The fall in price will be much greater than the fall in dividends. Late last summer I was looking at numbers, and noticed that real estate, equities, oil, and gold had all been down for two months straight. The only thing those goods have in common is the currency they are priced in. So I said, “Holy deflation, Batman!” and sold half of the index funds that I owned.I congratulate Devin and I do like the strategy suggested here, i.e. arbitrage between asset markets with irrationally different collectible premiums. This strategy does, however, assume that one can estimate the price derived from just supply and cash flow plus consumption demand, at least enough to be able to determine that there is a substantial collectible premium.
That said, I'd add (perhaps disagreeing with Devin) that in the large collapse in commodities since last summer, decreased expectations for future industrial consumption (due to rising expectations of a long-term worldwide recession) and probably a change in the rationality of the collectible premium of commodities were also major factors alongside the decrease in inflation expectations. Thus gold did not fall nearly as much as the industrial commodities. (It's also possible that this reflects some sort of security premium in a crisis of gold over other commodities, but despite all the gold coin ads to this effect I'm skeptical about that being a large factor). Of course, a change in consumption expectations will also tend to make stocks go down, due to the more direct cause of decreased profits leading to decreased cash flows from dividends etc. Commodities markets reflected informed expectations about future inflation and consumption more quickly than stocks, giving Devin his signal.
Gold and commodities prices have a good long-term correlation across business cycles. Industrial mineral prices do tend to vary more than gold within a business cycle due to changes in consumption expectation. (Of course these are not really predictable "cycles", but unpredictable effects of things like credit conditions, but "business cycle" is the unfortunate standard term in economics for this variation in overall credit and consumption). Another example of the collectible premium arbitrage strategy is to look at the oil/gold price ratio. Last summer this ratio was far too high (in hindsight, although at the time I was skeptical that this unprecedented ratio would last), reflecting a very high valuation of industrial commodities as collectibles. Recently it has much lower, reflecting much lower valuation as collectibles (probably irrationally too low), as well as lowered consumption expectations (the business cycle). In both cases one would have profited over the course of entire business cycle (and even in these two cases, luckily, over the short run) from betting against even more extreme deviations from the average historical ratio (or historical trend -- it's certainly plausible that oil and some other industrial minerals are being depleted faster than gold, or that there are long-term secular difference in their demand functions, but as the recent oil/gold ratio collapse to 1980 levels suggests, the secular depletion/demand gap between oil and gold is probably quite small, and almost certainly less than 2% per year over the long run).