What do the remarkable swings in oil prices over recent months tell us about the state of the oil production, consumption, and the global economy? One would think a lot: rising prices signal weakening production, growing demand from consumers, and a relatively healthy global economy; falling prices reveal robust output, slow consumer demand growth, and, more broadly, a faltering global economy. Take the August oil price collapse: many observers of the world economy took it as a sign that the Chinese economy was stumbling. The remarkable behavior of oil inventories, though, suggests that recent price moves tell us much more about market sentiment and beliefs about the future, and considerably less about fundamentals, than one might imagine.
Economists typically argue that beliefs about the oil market, expressed through speculation, have no enduring impact on oil prices. If speculation drives oil prices above what current supply and demand fundamentals would imply, production will exceed consumption, and inventories will accumulate indefinitely. Conversely, if speculation drives oil prices below where they should be, inventories should draw down without end. Since we don’t typically see long, steady periods of inventory accumulation or depletion, speculation can’t do much to drive prices away from what production and consumption fundamentals imply.
So here’s a fact that should make you stop and think: the past twenty months have seen steadier U.S. inventory accumulation than any other twenty month period in the last sixty years. (All inventory figures here are taken from the EIA.) Specifically, eighteen of the last twenty months have seen combined U.S. stocks of crude oil and petroleum products grow, the only time this has happened during the sixty year period for which monthly data are available. (EIA hasn’t reported monthly data for August or September yet; I’ve interpolated those numbers from weekly figures.) If you broaden the lens and look for twenty-month spans during which inventories increase in at least seventeen (rather than eighteen) months, you’ll find only one more, the period from April 1979 to November 1980, when the Iranian revolution drove hoarding. (For the statistically inclined, the odds of finding such a streak in a random series of normally distributed numbers with the same mean and standard deviation as the actual series of inventory changes is very small.) You can do an analogous exercise with weekly inventory changes; the recent pattern is again unusual. Similarly, if you narrow or widen the window from twenty months, the results don’t change much.
One way to think about this is that, for most of the last couple years, the marginal buyer of oil has not been a consumer – it’s been someone who’s put the oil (or refined products derived from it) in storage with plans to sell or use it in the future. If that’s the case, changes in the oil price – even the spot oil price – tell us as much about what speculators think future supply and demand will look like as they do about what physical production and consumption look like now. For example, the best way to think about falling oil prices in August is probably that they told us that oil investors believed that future Chinese demand would be weaker than they’d previously expected, not that the price drop revealed anything about current Chinese consumption.
There are, to be certain, a bunch of wrinkles here. In particular, I’ve looked at U.S. data because it’s readily available; global data might reveal a different pattern. One also finds different patterns if one looks at crude oil only rather than crude and petroleum products.
The data also point to a basic puzzle: why have inventories accumulated steadily for so long when that’s rarely happened in the last sixty years? And, whatever the answer to that question is, what does it tell us about when the trend might reverse? That’s the likely subject for a future post.