On the Fallacy of Speculation-Driven Commodity Price Spikes
Update: Mark Thoma has also has nice piece linking to the new report by the Federal Reserve.
Kay MacDonald over at Big Picture Agriculture has a nice post on the role of speculation in world commodity prices.
Kay is absolutely right: commodity prices--both oil and food--appear to be overwhelmingly about fundamentals and not about excessive speculation. Most economists I know and respect who study this stuff agree with Kay (and me) about this. Bubbles may sometimes happen in some markets. But I think bubbles are vanishingly rare in markets of consumable, non-durable commodities.
A lot of good people have tried to explain this and have offered evidence: Jim Hamilton at UCSD, Scott Irwin at UI, Mark Thoma at UO, Paul Krugman, and many others, not to mention yours truly. Just consider for a moment that, when prices spike, people are still buying and consuming most of this stuff. Inventories only account for a small fraction of one year's production.
Also, during the vast majority of the times when prices spike, inventories decline. It's impossible to have a bubble with declining inventories. Now, it could make sense for both prices and inventories to spike at the same time. That can happen if people anticipate an imminent spike in demand or an imminent decline in supply (e.g., a possible oil export embargo). While this situation might be hard to distinguish from a bubble, I think in most historical examples the evidence suggests that such expectations were reasonable when they happened.
Below is a graph Nam Tran, one my graduate students, recently put together recently for food. It shows inventories and the caloric-weighted average price of corn, soybeans, wheat and rice, the four biggest food crops in the world that comprise roughly 75 percent of the caloric base. Note that production is growing rapidly over time, so that, relative to the size of the market, inventories (storage) have actually declined markedly since the mid 1980s. It seems pretty clear to me that prices are first and foremost a function of inventories.
Kay MacDonald over at Big Picture Agriculture has a nice post on the role of speculation in world commodity prices.
Kay is absolutely right: commodity prices--both oil and food--appear to be overwhelmingly about fundamentals and not about excessive speculation. Most economists I know and respect who study this stuff agree with Kay (and me) about this. Bubbles may sometimes happen in some markets. But I think bubbles are vanishingly rare in markets of consumable, non-durable commodities.
A lot of good people have tried to explain this and have offered evidence: Jim Hamilton at UCSD, Scott Irwin at UI, Mark Thoma at UO, Paul Krugman, and many others, not to mention yours truly. Just consider for a moment that, when prices spike, people are still buying and consuming most of this stuff. Inventories only account for a small fraction of one year's production.
Also, during the vast majority of the times when prices spike, inventories decline. It's impossible to have a bubble with declining inventories. Now, it could make sense for both prices and inventories to spike at the same time. That can happen if people anticipate an imminent spike in demand or an imminent decline in supply (e.g., a possible oil export embargo). While this situation might be hard to distinguish from a bubble, I think in most historical examples the evidence suggests that such expectations were reasonable when they happened.
Below is a graph Nam Tran, one my graduate students, recently put together recently for food. It shows inventories and the caloric-weighted average price of corn, soybeans, wheat and rice, the four biggest food crops in the world that comprise roughly 75 percent of the caloric base. Note that production is growing rapidly over time, so that, relative to the size of the market, inventories (storage) have actually declined markedly since the mid 1980s. It seems pretty clear to me that prices are first and foremost a function of inventories.
Update: Here is a plot for a follower who does not agree with my characterization of inventories of oil. It shows US inventories less the strategic petroleum reserve and oil prices. Inventories were stable to declining when oil prices spiked. When the financial crisis hit in October 2008, growth prospects fell through the floor immediately, and then inventories started to grow. This looks like textbook commodity prices to my eyes. The data can be downloaded from the Energy Information Administration (http://www.eia.gov/).
Michael,
ReplyDeleteThanks! Wonderful graph by your student!
You might check this out, and comment on it, also:
http://www.nakedcapitalism.com/2011/10/links-101611.html
Yves is one who will not give this up, she doesn't feel she's convinced.
These changes in inventory do not explain the radical increase in commodity price volatility in recent years.
ReplyDeleteWhile there are some studies saying that speculation has no effect on prices, I believe there are many more, from respected institutions, saying that excessive speculation is a problem. I draw your attention to this list of more than 80 such studies:http://www2.weed-online.org/uploads/evidence_on_impact_of_commodity_speculation.pdf
Also, a recently released study showing that commodity index trading is especially detrimental to commodity markets around the time of their roll, driving prices higher and to be more volatile:
http://www.bettermarkets.com/reform-news/new-better-markets-research-report-shows-wall-street-driving-food-fuel-prices
Dave,
ReplyDeleteWell, I can't debunk all of those articles. But I can see that a lot of them are simply documenting more trading in commodity futures. I'd interpret that as the arrow of causality going from fundamental to higher and more volatile prices to more trading in commodities.
Others take quotes completely out of context. Indeed, some of those papers actually argue *against* the idea excessive speculation is driving everything.
These kinds of prices spikes--and ones potentially much, much larger--do occur naturally in storable commodity markets. This is the basic nature of these markets. We will have often have long stretches of low and not-so-volatile prices, but with occasional periods of very high and volatile prices.
Besides, we can see the fundamentals here. Demand is growing like crazy with growth in China and other emerging economies. And supply is really slowing down. We're just having hard time producing enough oil and food to keep prices as low as they used to be.
It's called scarcity.
Michael, Thanks for your thoughtful response.
ReplyDeleteI understand commodity markets to be finely tuned with traditional speculators helping by going long and short depending on their opinions of supply and demand conditions.
But investments in these new commodity indexes and ETFs are almost exclusively long and those investment decisions are not made on the basis of supply and demand conditions in each commodity market, but on portfolio allocation reasons.
So my question is how can hundreds of billions of dollars, almost exclusively betting long for long periods of time, NOT mess up the functioning of these markets?
This study shows how commodity index rolls mess up markets: http://www.bettermarkets.com/reform-news/new-better-markets-research-report-shows-wall-street-driving-food-fuel-prices
Dave,
ReplyDeleteI think more people are investing in commodity indexes because it's a good diversification strategy. But the speculative market understands all of this very well, and will have a strong profit motive to speculate in a manner that will effectively work to equilibrate the market nicely.
Look, I'm not ideological about this stuff. I think speculation can sometimes get out of hand and cause some problems. But I don't see any signature evidence of that in non-durable commodities. Furthermore, I think it's harder for things to get out of hand with non-durables--mainly because people are buying and consuming that vast majority of the stuff. Contrast corn and rice with durables like gold, houses, and stocks, where prices appear to stray far from fundamentals. With non-durables, markets will punish bad speculators very quickly.
Hamilton is pretty convincing on this, but he leads us to the question: why does the entire commodities structure seem to move up or down as a group? Inventories do not seem to answer the question. In oil, in 2007-2008 prices moved up despite growing inventories and with the futures market in steep contango.
ReplyDelete(Also, for crude oil stocks, if you include the SPR, as of last week were at 57.5 days cover, or 16% of the year, which I wouldn't regard as a "small fraction.")
freude,
ReplyDeleteI'm less familiar with oil futures markets. But I believe the point in time to which you are referring in 2008 was relatively short. Inventories were rising from a very low level.
Today inventories are higher and prices are much lower.
There is certainly a complex dynamic here. But I don't see the tell-tale sign of a bubble. That would be continually increasing prices feeding ever rising inventories in anticipation of ever increasing prices. We just don't see that. When news hits the market about greater inventories prices fall fast. Unless there's a middle East scare
Micheal -- And I come here because I'm not that familiar with ag markets, but the price rise from 2007-2008 was not short, nor were inventories rising from a very low level.
ReplyDeleteRefer to http://www.eia.gov/dnav/pet/hist/LeafHandler.ashx?n=PET&s=MCRSTUS1&f=M and you will see that in the late 2000s crude stocks were historically-speaking high (though global--not US--demand was rising more rapidly.) Total stocks (crude + products) demonstrate a small decline in inventories in 2007 and 2008, but they remained at elevated levels throughout. http://www.eia.gov/dnav/pet/hist/LeafHandler.ashx?n=PET&s=WTTSTUS1&f=W Note that prices post the 2008 crash recovered even as inventories rose to unprecedented levels.
Verlenger gives the explanation that demand for ultra-low sulfur diesel was leading the price of crude up ... http://www.eia.gov/dnav/pet/hist/LeafHandler.ashx?n=PET&s=WTTSTUS1&f=W, but then earlier Verlenger himself blamed the contango and the financialization of oil for the rise in prices in tandem with rise in stocks from 2003-2006 in a piece called "How Wall St. Controls Oil" which is no longer online, but which I can email you if you're curious.
In any case, it still begs the question of why all commodities go up in lockstep, given that inventories for all were not at depressed levels, nor are they now, if I understand correctly.
freude bud: You have to look at inventories relative to consumption. It's silly to look at them relative to the 1930s. You didn't mean that, did you?
ReplyDeleteLook: everyone was producing flat out as fast as they could in 2007-2008, and that picture you linked to shows, very clearly, that inventories were not spiking.
No one has presented a shred of evidence of oil hoarding. So all that oil was being consumed at those high prices.
Where's the bubble? You don't have a story that holds together. None of the commodity bubble people do.
We need evidence of hoarding, withheld production or inventory increases to even begin to entertain the idea that speculation caused anything. Even that would not be sufficient to prove a bubble. But it is necessary.
http://www.reuters.com/article/2012/03/02/us-fed-banks-commodities-idUSTRE8211CC20120302
Deletewell, to begin with, i'm not a commodity bubble person.
ReplyDeletebut, you wrote that inventories were declining ... they clearly weren't insofar as oil goes. you wrote that inventories were a negligible portion of the year, and, insofar as oil goes, that is also clearly not the case. why would commodities rise and fall as a group? inventories don't explain that.
it's not ideological, merely a question of fact.