[This is a comment I posted to Paul Krugman's blog. I'll suspect he'll never read it (literally -- his comments are read by factota, and whatever their screen is, I've never made it through to get published. But even if I did, there's a hundred comments on this topic, and I can't imagine he reads any that aren't by friends of his. I don't think I would. But nonetheless, I like this idea, I think it's smart, and I'd be interested in any suggestions of how to test it, and any criticisms of its logic.]
Dear Professor Krugman,
You've asked several times in recent posts if anyone could suggest a plausible mechanism by which future prices could drive up spot prices without inventories increasing:
http://krugman.blogs.nytimes.com/2008/05/13/more-on-oil-and-speculation/
I think I have one. I'm not sure how to test it; and I'm not sure it's what's happening; and I wouldn't be at all surprised if different dynamics were at work in different commodity markets. But this idea at least seems worth investigating. It bears some resemblance to roublen's comment:
http://krugman.blogs.nytimes.com/2008/05/13/more-on-oil-and-speculation/#comment-44132
and I think simply develops it further and makes its mechanics more explicit.
It's a model built on the 1979 oil price spike. As you probably know, the 1973 spike, although blamed on the oil "embargo" (which was in fact quite fungible), was mainly due to very tight supply constraints that had been suddenly further tightened, and to the overthrow of a price that had been artificially held down by collusion among the monopoly distributors, the so-called "7 sisters" oil companies. (Their collusion, which kept the world oil price absolutely level for years (kind of like the current "market" in cigarettes), was what kept everyone from noticing the tightened supply conditions until after the fact.) But the 1979 oil price spike was due almost entirely to panic buying. The Iranian revolution was preceded by a strike by oil workers and followed by a spectacular attack on Mecca, and the revolution itself looked like something that not only would suspend supplies but might spread. But in fact the actual supply disturbance lasted less than a month, and markets at the time were not unusually tight, and in normal times would have been able absorb it. What caused the 1979 price spike was cascade panic buying at every level of the oil market, from drivers waiting in lines for hours to top up their tanks, to countries leap-frogging each other to lock in long-term contracts for bilateral delivery. And it's this very last element I'd like to focus on.
In 1979, only a small percentage of oil actually traded on the Amsterdam spot market. It itself was largely a function of the 1973 spike; and it was the aftermath of the 1979 spike that finally developed the spot market out into the institutions we have today that actually largely govern the market price. But in 1973, the spot market was mainly a place where excess oil was traded, that is, oil that wasn't already bound up in long-term country-to-country delivery contracts.
What happened in the panic of 1979 was that the spectre of high prices staying high forever drove countries to try to get "better" deals for themselves -- better by comparison with the spot price -- by striking long term deals with producers. And since the market price was high, and it was a sellers' market, these long term contracts were for prices much higher than they would have been normally.
And that led to the second dynamic. Because every country was trying to do this, and each would arrive in a producer country only to see someone else's jet just taking off, there was a frantic bidding up of what were already high prices, and a willingness to make the contracts even longer term. Essentially producers were in the driver's seat.
And this led to the third dynamic, which is key for the current situation: all these long-term, large-scale, bilateral delivery contracts kept massive amounts of oil off the spot market -- thereby keeping the price high.
Now in this particular case, inventories also rose; and production rose massively in a relatively short time in response to the price movement; and there were strong cartel dynamics as well as cartel-cheating dynamics; and all of this had relatively little or nothing to do with the futures markets.
But I'm not the first one to suggest that what really kept oil prices up for the next 5 years despite the fact that there had never been a true supply shock was that it took that long for these long-term high-price bilateral delivery contracts to unwind. Most of the world was essentially stuck with contracts for high priced oil whether they liked it or not. And when everyone is stuck like that, it looks normal. And it was reflected in the spot price, since all this private placement kept production from ever reaching the openly spot market.
Thus, long-term panic buying could constrain the supply, and drive up the price, without leading to an increase in inventories, because all these long term contracts are essentially huge contracts for future delivery. And thus the "futures" market is driving up the up the "present" market even though it is dealing entirely with commodities that have not yet been produced.
I put "futures" in quotes because it wasn't actually a futures market in the technical sense we normally use that term. And that leads to the last point, which is also key in the current situation: all of this can theoretically happen in commodities that don't have functioning future markets or even (as in the case of oil in 1979) a fully functional spot market that governs the prevailing world price.
To put it in one sentence: long-term bilateral delivery contracts are future contracts where delivery is already accepted in advance. You can have a market in them even when there is no publicly traded futures market in the technical sense. They constitute a futures market that moves actual supply, and continues to constrain it when the future becomes the present, and delivery is accepted in real time. And it is normal in such a market for typical individual contracts to be huge relative to the market as a whole, and thus for each such struck price to be a major market mover.
But by now I think you can see the outline of the mechanism I'm proposing. If enormous amounts of money pour into the futures markets of various commodities (for classic bubble reasons: because nothing else is producing a high return; and because there is a new "theory" about how commodities are vital to diversification and are actually a path to safety rather than risk), we are all agreed that that will make future prices soar.
Now what if, in response to those future prices -- and. with all due respect, in response to the pronouncement of all respected economists that this isn't a bubble, but rather represents the real long-term situation of supply and demand -- countries engage in exactly the same leap-frog panic buying of long-term high-price contracts for bilateral delivery, thus taking lots of commodities off the spot market that otherwise would have reached it? Wouldn't this be rational behavior on their part in the economic sense of rational?
And couldn't that drive the real price higher in exactly the same way in did in oil in 1979? And lock in that increased price in the same way? And couldn't one, in this case, in fact blame speculative pressure for this price rise since it was the spectre of high prices in the futures market that stampeded people towards this bidding war of long term contracts and locked these high prices in? And couldn't such hysteria at least in theory spread to any commodity no matter how thin the tradeable market? Couldn't it in fact be even more effective precisely where tradeable markets were thin?
On this model, you could theoretically have exactly the same volume of supply and demand in the future, but traded at a much higher price, because if all of it was locked in at high prices and taken off the freely tradeable market, there would be negligible cheaper supply freely available.
I freely concede that in the real world, if there aren't solid production constraints, the high price should stimulate higher production above and beyond what is needed for all long-term contracts put together; and the spot market, initially constrained, should re-expand; and prices should eventually fall there precipitously; and this lead people stop making such high priced long-term contracts and to break the ones they've got.
But if the 1979 oil shock is any indication, that could take years to unwind. Especially in commodities where there aren't fully developed, transparent markets.
At any rate, that's my proposal for a mechanism, and I'd love to hear whatever you have to say in response.
Michael