I returned home recently from another excellent “di Tella Summer Camp”. It’s no longer in the summer, it was not held in the Universidad Torcuato di Tella and it’s now called the Workshop in International Economics and Finance but the event is still going strong; this was the 19th year! Our extremely hospitable hosts were the Central Bank of Colombia and in the “policy panel” the vexed question of the nature of commodity prices raised its head again.[1]
Since oil accounted for over 50% of Colombian exports and the dive in prices implies a significant fall in public sector revenues, this was not surprising. Specifically, the question was whether this is a “permanent” or a “temporary” bust for oil. For a country such as Colombia it’s a vital question. If the shock is expected to be permanent a significant fiscal adjustment is required, while if it’s all temporary perhaps none at all.
Commodity Price Theory 101
Let’s step back for a minute. Many great minds have considered the nature of commodity prices; we certainly don’t need to reinvent the wheel. Keynes and Hicks discussed the relation between spot and futures prices,[2] Samuelson highlighted when such prices would be random walks, Muth devoted a section to commodities in his famous rational expectations paper and Deaton and Laroque’s two papers remain seminal reading. So here I mostly just summarize what such great masters have said about the topic.
Consider gold. It’s a highly storable commodity, it doesn’t degrade. Storage is relatively easy, so the current spot price and future spot prices (and futures prices) are all linked by intertemporal arbitrage. The current spot price summarizes virtually all you need to know. Gold prices, like stock prices, should follow something close to a random walk.[3] All shocks to the price are by definition expected to be permanent.
What about other commodities? Well, in normal times the same is true. But what does “normal times” mean? It means that stocks are freely available and that intertemporal arbitrage is possible. But times are not always normal. What if there is a big problem in supply or a big increase in demand such that availability is driven down to zero? This breaks the intertemporal arbitrage. Spot prices will rise and they may subsequently be expected to fall. In these circumstances, futures prices provide a lot of information, precisely because there is no intertemporal arbitrage. If the market expects the situation to revert in a year then the 12 month futures price will be well below the current spot price — the market will be in backwardation.[4]
Deaton and Laroque’s famous model for commodities derives a process for prices when there is uncertainty in supply, storage is generally possible but large shocks may provoke a stock-out.[5] The process is NOT a simple one and is highly asymmetric. Much of the time commodities follow something close to a random walk but with sharp booms and generally longer busts.[6]
Commodity Booms and Busts
What does this mean for the question posed above? Taken literally it means that all falls in the spot price are almost inevitably expected to be permanent. Price rises in normal times are also expected to be permanent but some in abnormal times may be temporary. Let me be very clear about price crashes. If ample stocks are available it is simply a logical inconsistency to say that there has been a large decline in prices but that they are expected to rebound quickly to where they were. This can never happen. If it were true, the spot price would not have fallen as much.[7]
Now, one could argue that the market got it all wrong. Market prices might be thought of as reflecting an average expectation and individuals may have different views. Thus, it would be logical to say, “I think the current large price falls are temporary because I think the market’s got it wrong.” Of course this begs the question, what do you know that the market doesn’t?
And what exactly does it mean that this price shock is expected to be permanent? Does it mean that prices will actually stay at current values? Certainly not. Commodity prices are very volatile; there are lots of shocks. And a shock doesn’t have to be something that has affected actual supply or demand today. If the market thinks the probability of a war in the Middle East in a year’s time just went up by 50%, the spot price of oil will react today. Markets are forward looking. But prices will move because of shocks, which by definition cannot be anticipated.
But wait a minute, this last boom lasted 10 years! Can you really call it temporary? True, the recent boom was exceptional both in its size and its persistence —even larger than the massive booms of the 1920s and 1970s and dwarfing the many mini-booms during the 20th Century[8]. But during virtually the whole period of the boom stocks were dwindling and stayed low. Some point to the role of commodity funds to explain the boom but I’m more of a fundamentals supply and demand guy. China was growing at an exceptional (and unsustainable) rate of 10.5% per annum from 2002-2012; as China slowed it was always in the cards that commodities would slump.[9] Ten years is indeed a long temporary boom but history tells us that supply eventually catches up. And it did. We reached record output levels for many commodities just as China’s demand started to wane.[10]
What Should Commodity Producers (and Importers) Do?
It’s actually not hard to decide if a commodity price shock is expected to be permanent or temporary. If ample stocks are available and the futures price is at a full carry, then any price change is expected to be permanent. If it’s a price hike, stocks are low, and futures are below spot prices, it’s expected to be temporary.
But remember: prices are extremely volatile and any projection including those in futures markets are extremely poor predictors. For example, the uncertainty surrounding oil these days is dramatic. The World Bank’s commodity price projections have oil prices rising 50% from January 2016 to January 2017. But looking at commodity option implicit volatilities the market thinks the uncertainty around any projection is huge. Just one standard deviation above implies a rise of 120% and one standard deviation below implies a fall of 15% over the same period.
Yet commodity exporting countries are forced to choose some commodity price to plan fiscal policy. The uncertainty is so great, exporting countries should focus at least as much on figuring out how best to manage the uncertainty as on attempting the impossible: projecting prices. Mexico has hedged its year-ahead budget forecasts for oil prices in recent years. So if a commodity exporter is forced to say the commodity price will be 100 in the next annual budget, then using hedging markets to guarantee revenues associated with that number is a sensible thing to do. A lower assumed number will generally lower the cost of hedging.[11] In fact, the best way to think about the exercise is not to try to come out with the “best prediction,” as even the very best prediction will almost certainly be bad. The best way to think about the exercise is to say, what is the most convenient price to assume for my budget and that I can hedge at a reasonable cost?
But hedging markets only go out so far; guaranteeing a price for more than 12 months is unrealistic for most markets. Another way to shift risks to others who may be in a better place to bear them is to issue commodity indexed debt. Such instruments should be seriously considered for commodity exporters.
Oil importing countries may also wish to hedge. Current low oil prices present a great opportunity to lock in prices by purchasing out-of-the-money call options. And if oil exporters sell oil indexed debt, oil importers are natural purchasers to hold in their reserves. Multilateral financial institutions could stand in between these trades to reduce performance risk and provide credit ratings appropriate for reserve holdings. Such innovations would help complete markets.
But as markets currently remain incomplete, hedging to shift risks to others has its limits. Countries should then also self-insure through stabilization funds. Hedging one-year-ahead budgets and employing stabilization funds to smooth the persistent shocks is a good compromise. The design of such funds demands careful thought and should be compatible with fiscal procedures and exchange rate objectives. The bottom line is that for any country that relies heavily on commodity prices, not seeking ways to manage such risks, is like driving a fast car on a very dangerous road with no insurance. Like bad car crashes, crashes in commodity prices have nasty, permanent effects.
[1] The policy panel was chaired by Hernando Vargas (Banco de la Republica, Colombia) and consisted of Andy Neumeyer (Banco Central de la Republica Argentina), Guillermo Perry (U. los Andes), Andy Powell (IDB), Augusto de la Torre (WB) and Charlie Vegh (Johns Hopkins).
[2] In “A Treatise on Money” (1930), Keynes suggested commodity markets would normally be in backwardation (futures prices below spot prices) but Hicks in “Value and Capital” (1939) argued otherwise – that the market might be in contango with futures above the spot price. A full contango is generally defined as the futures price being equal to the spot price plus the rate of interest plus the cost of storage.
[3] More generally they should follow a Martingale process where price innovations are independent. A random walk is a special case of such a process.
[4] There has been much discussion regarding the impact of commodities as a “new asset class” on this basic commodity theory. For example, the oil market at times has been in a so-called “super-contango” in which the futures price exceeds the spot price plus interest and storage costs. This situation is sometimes attributed to commodity funds that wish to have (long) exposure to oil through futures, perhaps as an inflation hedge, and thus bid up futures prices. It is also thought to give producers perverse incentives to produce less and store oil by keeping it in the ground. This was seen to be important in 2008-2011 (see for example http://ftalphaville.ft.com/2011/12/16/803831/the-curious-case-of-abnormal-backwardation/)
[5] Deaton and Laroque’s second paper introduces serial correlation in supply shocks. Uncertainty could also be in demand and switching it doesn’t change things too much. Dvir, E, and K S Rogoff (2010), The Three Epochs of Oil, Harvard University, Cambridge, MA. Unpublished. provides such a model.
[6] It’s hard to get longer-term cycles in commodities from these models. While they include rational storage, they do not include investment. Longer cycles may come from modeling investment behavior.
[7] And in general, if stocks are not ample, the price crash would not have occurred.
[8] In saying it lasted 10 years I am abstracting from the sharp fall in commodity prices due to the global financial crisis. This episode highlights the problem for the econometrician and the macroeconomist. If one looks at the historical spot price series, it looks like a sharp temporary bust but at the time expectations were not that at all. To formulate policy in a particular year the policy maker does not have the benefit of knowing the future, this highlights the problem of uncertainty considered further at the end of this blog.
[9] The risk of a fall in Chinese growth and a rebalancing of the Chinese economy and the impact on commodity prices were analyzed in Chapter 3 of the 2012 Latin American and Caribbean Macroeconomic Report. http://www.iadb.org/en/research-and-data/publication-details,3169.html?pub_id=IDB-MG-126 . The fall in oil prices was also due to supply shocks such as the decision by Saudi Arabia not to restrict production even at such low oil prices.
[10] Unfortunately such boom and bust cycles appear typical of commodity prices. See my recent VOX blog for further discussion on commodity booms and busts.
[11] Here I am thinking of an exporter buying an out-of-the-money put option or some type of collar arrangement where part of the upside of higher commodity prices is sacrificed to finance the purchase of the put.
Leave a Reply