After a lull due to the global financial crisis, commodity prices appear to be resuming their upward march. The post-crisis recovery has seen steady increases in food prices, and these price spikes have fed unrest across the Middle East and North Africa. These geopolitical tensions in turn fuel higher oil prices, and the cycle repeats itself.
Which naturally leads policymakers to wonder whether how long the recent runup in food prices is going to last, especially when it comes so quickly on the heels of the food price spikes of 2008. The potential causes of the current crisis have been well explored elsewhere. Rather than reiterate these issues, it may be worthwhile the extent to which the crisis is real: that is, how much of the crisis is due to the likelihood that the shocks to commodity markets are permanent and persistent, and how much of it is due to either financial speculation [*] or elements for which changes in policy---such as those related to monetary policy, biofuel policy, and trade policy---could credibly tame.
As discussed in an earlier post, QE2 has been accused of being a driver of the commodity price inflation (including in food). Since food commodities trade in global markets priced in U.S. dollars, and greater liquidity must (trivially) be reflected as higher prices (with monetary neutrality), monetary expansion by the Federal Reserve would certainly show up in the nominal dollar price of food. But if countries choose to operate a floating rate, then this purely monetary effect would generally be cancelled out by exchange rate movements (more precisely, by an appreciation).
Now, while developing countries may not actually maintain floating regimes, there are other currencies that do, in fact, float against the dollar, and these offer a way to parse out the monetary effect. [†] Doing so in terms of either the euro or yen (see figure) suggests that there is certainly an element of actual market pressure---from either the supply or demand side---that goes beyond excess liquidity alone. Although the runup in the series denominated in euros or yen is less pronounced than that in dollars, the recent increase is unmistakable: indeed, all three series there is a clear uptrend from the lows in early 2009, and the recent high in the JPY series is even higher than the highs attained in 2008.
Notes: FAO FPI index converted to alternative currencies using average monthly EUR and JPY exchange rates relative to the USD, normalized to unity in 1/2003. The FPI is an index of export share-weighted price quotations for 55 food commodities, based on the 2002--04 average, and measured in USD.
Alternatively, the FPI can also be "deflated" by the overall consumer price index.[‡] To the extent that excess liquidity can be expected to spill over into all classes of goods and services---and not only commodities---we would see an "inflation-adjusted" FPI move in lockstep with the unadjusted FPI. Intuitively, a divergence between the two series captures how much faster food prices are growing, relative to the overall price index. And as is evident (see figure), this divergence increased dramatically in the earlier 2008 crisis, and is once again diverging sharply. In fact, the gap between the two series in the summer of '08 is actually smaller than the gap this past December and January.
Notes: FAO FPI index deflated by the monthly U.S. CPI-U, normalized to unity in 1/2003.
These market pressures can also be seen in other classes of commodities, most notably in energy. Of course, the factors affecting food and energy may inherently be different, but there are some notable bridges that may account for some of the spillover effects. One important link is the role of biofuels (corn is both a food staple and the main feedstock in the production of corn-based ethanol). Thus, while it is true that the recent price spike is a little less broad-based than in 2008, corn prices have in fact risen dramatically, and corn has nasty habit of working its way into other commodity prices.
Here, however, the picture is a little more mixed. Energy has certainly risen sice the nadir in early 2009, but even so, its increase has been a lot less dramatic than the increase has been for food, with the index only halfway the highs achieved in the summer of 2008 (see figure). This is the case whether the index is measured in dollars or in pounds (although quantitative easing by the Bank of England may complicate a straightforward interpretation). Moreover, an Aussie-dollar denominated index would hardly have moved at all (again, the complication here is that the AUD is a commodity currency that is heavily affected by the boom in commodity prices more generally). Certainly, the current unrest in the Middle East may rapidly change this picture---especially if there is significant threat of a supply disruption from Saudi Arabia---but right now, it is hard to shrug off the notion that the market appears much more sanguine about the path of energy prices than it is about food prices.
Notes: The IMF fuel commodities index is an export share-weighted price quotations for petroleum, natural gas, and coal, based in 2005, and measured in USD.
Concerns regarding a rapidly changing oil picture aren't entirely misplaced. Oil shocks have historically brought on economic downturns, and while a temporary spike in oil prices may be shrugged off by a currently fragile (oil-importing) developed world, one cannot say the same for shocks that persist for six months or longer.[§]
Which brings us back to the question of how persistent we can expect these commodity price shocks to be. Here, we can marshall one piece of evidence which suggests that this time, commodity price shocks are different. In particular, food, energy, and metals indices appear to be moving more synchronously than ever before (see figure, top panel). This can be verified by taking rolling correlations of these series, which indicate that such synchronicity in the price cycles of all of these commodities is virtually unprecedented (see figure, bottom panel).
Source: World Bank staff calculations, from World Bank Commodity Price database.
Notes: Top panel: Monthly series for food, energy, and metals. Energy series for 1960--70, 1971--72, and 1973--75 are annual, biannual, and quarterly, respectively. Bottom panel: 36-month rolling bivariate correlations for the three series, smoothed with the 36-month moving average of each correlaation series.
Why might this be the case? There are reasons to believe that demand-side factors---especially growth in rapidly-industrializing emerging economy giants---may be giving rise to this commodity price pressure. While the jury is still out as to how persistent this phenomenon might be---historical price increases have typically been accompanied by supply-side responses---future developments may offer a clue as to whether such a hypothesis is indeed correct. If the correlation patterns we see above start to falter as emerging economies cool from their currently breakneck growth rates, then there may be reason to believe that demand-side pressure really is driving this current spurt in commodity prices.
In closing, it is important to point out that the discussion here about the reality of commodity shocks has centered on macro dimensions. Yet along another dimension, these commodity shocks are indeed very real. For many households in the developing world, these gyrations in global food and energy prices are more than frustrating fluctuations in the prices paid in restaurants, markets, and gas stations. With food accounting for a large share of consumption expenditures in the typical developing country household, a price spike in food may mean forgoing previously planned purchases, or even cutting back on food intake. As academics have repeatedly emphasized, the long-term human -capital consequences of even a transient food price spike can actually be remarkably persistent and severe. Pricey energy may mean that movement of goods is now more costly, which eats into already-thin profit margins for many developing country firms that rely on transportation to bring their goods to market.
Postscript: Recent work by Reuven Glick and Sylvain Leduc apply an event study approach to the issue, and find that while QE1 may have had a significant impact on commodity prices (whether direct or indirect), the effects of QE2 are decidedly more muted.
*. It is useful to remember that financial speculation, in and of itself, occurs primarily in futures markets and only induces real changes in spot markets when speculative expectations lead to actual stockpiling. This hypothesis is also inherently testable: such stockpiling must ultimate show up in inventory data.
†. Yes, of course exchange rates are determined by much more than just purchasing power parity (which only holds in the medium run in any case), and there may be incomplete pass-through. This naturally means that any exchange rate correction may be thrown off by a host of additional factors that may affect short-run exchange rate movements. Nevertheless, the concerted evidence for a range of floating rates (not just in the two reported ones, but even for other exchange rates) allows us to draw stronger conclusions.
‡. Deflating a price index with another is inherently tricky; after all, the FPI does not capture the production of actual goods and services (which a metric such as real GDP does), and the prominence of food items in the CPI itself also means that it is inherently tied to experienced inflation (as opposed to a real index of, say, stock prices). So the exercise does not, strictly speaking, produce a "real" index as much as an index that moderates the price changes in food with nonfood prices.
§. Of course, commodity price shocks are, technically, largely redistributive rather than purely contractionary: thus, while a spike in oil prices may lead consumers in oil-importing nations to cut back on spending on other goods, the additional income to oil-exporting nations will stimulate their respective economies. There may then be additional second-order effects, such as a recycling of this income through purchases of exports from oil-importing countries. Nevertheless, the historical evidence suggests that oil shocks tend to be contractionary at the global level.