THE STASH JULY 9, 2009
-
Read Later
READ LATERAvailable only to subscribers. SUBSCRIBE TODAY
-
Listen
ARTICLE AUDIO
- Font Size
Not according to a Federal Reserve paper released today. The study, written by George Korniotis, looks at a number of different factors that could suggest speculators' influence on commodity prices, but finds little evidence that any of them have been at work.
First, we should see a break-down in the correlation of price changes between tradable and non-tradable commodities once speculators started trading commodity futures in larger volumes at the beginning of this decade. (The theory being that speculation wouldn't affect the prices of non-tradable commodities):
I test the comovement hypothesis by studying the time patterns of metals with and without futures contracts. Over the period 1991 to 2008, I find that the correlation of metal price growth rates was consistently positive and did not decrease after 2000.
Second, Korniotis thinks that the big jump in nearly all commodity prices earlier this decade can be explained by the surge in global economic growth:
I use the world GDP growth rate to capture world economic activity. I find that world growth rate is positively correlated with metal price growth. Also, similar to metal price growth rates, world growth started to steadily rise after 2002. Therefore, accounting for world growth reduces the statistical significance of the structural break in metal spot prices.
(If you're unfamiliar with the term, the easiest way to think about a structural break is to imagine a smooth looking chart of stock prices or population growth, etc. which all of a sudden jumps to a new level. It typically means that something about the state of the world has changed. Here's a nice example.)
Third, Korniotis finds that there's no relationship between the returns of the S&P Goldman-Sachs Commodity Index--which he argues is a proxy for profits earned by investors--and commodity price changes. And fourth, Korniotis tests the theory that commodity suppliers could be affecting prices by hoarding inventories and then entering into futures contracts with speculators:
My analysis finds no evidence of physical hoarding. In particular, inventory growth is negatively correlated with price growth. Also, this negative relationship is present even after 2002.
This first chart from the paper shows the quarterly annualized growth rates in spot (i.e., current rather than futures) prices for traded and non-traded metals. The shaded area is the time period when prices accelerated upward:
And this next chart shows the annual correlations across all metals (solid line) and between traded and non-traded metals (dashed line) and that they have remained fairly constant over time:
While this is all circumstantial evidence, it should give the CFTC at least some pause before going ahead with stricter limits on speculative activity. On the other hand, there are some very suspicious signs right now that hoarding is influencing oil prices (and which are too recent to appear in the Korniotis paper). From Paul Krugman:
This time, however, oil inventories are bulging, with huge amounts held in offshore tankers as well as in conventional storage. So this time there’s no question: speculation has been driving prices up.
Now, “speculation” isn’t a synonym for “bad”. If the underlying assumptions that seem to have been driving oil markets were right — namely, that a vigorous recovery is just around the corner, and demand will shoot up soon — then it would be perfectly reasonable to accumulate oil inventories right now. But those assumptions are looking less reasonable by the day.
Since this is such a contentious issue these days, it's also worth pointing out that the paper doesn't necessarily represent the views of the Fed and Ben Bernanke, just the author.
--Zubin Jelveh
4 comments
The link that Krugman gives for his "bulging" statement says the opposite, that oil is flowing out of storage. Also, it's a pay for view site, which gets covered up before you finish reading it. Hmmmm.
- AMVHuck
July 9, 2009 at 3:51pm
AMVHuck, the Krugman blog isn't a pay site but you do need to set up a (free) account on the NYT site to access it. I'm looking at Krugman's blog post right now, didn't have to pay but I do have an freebie account with the NYT site
- tnmats
July 9, 2009 at 4:28pm
AMVHuck: Um, that is because the article says that oil is flowing out of storage right now, at this very moment, after being stored for months. There is literally no storage left, so they can't store anymore. This has been known for quite some time (I read an article about a month ago saying that at current storage rates they would run out of storage in roughly a month). The article linked to says that that oil is now being released because the price is at a high price. In other words, the speculators are selling off the stored oil at higher prices - precisely what Krugman is claiming. Now, this should make the price of oil drop in the near future, but that does not mean that the current runup in the price was not due to manipulation/speculation.
- rhorath
July 9, 2009 at 6:10pm
Charts.
Like statistics, they tend to be built around the facts rather than from them. At least the ones that revolve around social, political and economic agendas.
Even if all the numbers can be verified empirically, the chartmakers can always manipulate the truth by what they leave out....or by what they choose to emphasize....or by who they go to for information....or by what the numbers are said to "mean".
Charts used in venues like this are often very different from charts used in venues frequented by physicists, chemists, mathenaticians, geologists etc.
Why? Because charts used by natural science almost always focus on material relationships. Either the numbers or the correlations reflect the natural world as it is or they don't. With social science, on the the others, charts are often used not just to say what is true but what ought to be true.
In other words, they are often used to plug or unplug particular ideological assessments of human moral and political values. Hard science doesn't debate whether gravity or hurricanes or rock formations or human DNA is behaving as it ought to be, only as it does or does not.
What make charts used by economists all the more problematic is the manner in which the numbers and the correlations combine both empirical and non-empirical data.
For example, suppose you display a chart that shows laizze-farie capitalism to be a much more productive engine for producing commodities. This may well be empirically sound. But it does not demonstrate that laizze faire capitalism is the most rational economy...let alone the most ethical. It shows only that if your goal is to produce lots and lots of commodoties it is the one to pick.
But this does not factor in all the other variables a particular human community might choose to stress in interacting. Variables like justice, liberty, equality, opportunity, fairness, compassion.
After all, how do you depict this on a chart? How do you integrate or reconcile the numbers on a chart with the sheer complexity of human emotional and psychological reactions?
You can't, of course.
george walton
- iambiguous
July 10, 2009 at 12:10am