Everybody seems to agree that there is more volatility in commodity markets. But it is not clear whether everybody is talking about the same. Let me focus on food commodities to elaborate the point.
The recently-leaked report for the G-20 is an example of how the issue of volatility is being approached (see http://www.boell.org/downloads/G20PriceVolatilityMarch3version-1(1).pdf). Under the title “What is volatility?” it goes to explain the issue as follows: “In a purely descriptive sense volatility refers to variations in economic variables over time. Here we are explicitly concerned with variations in agricultural prices over time…” Then, the Charts included in the document show real (I did not find what the deflator was) and nominal prices in levels for some commodities and a chart with the standard deviation of monthly price inflation for a selected period (it does not say whether they have been annualized or not). If the last chart with the volatility measure follows the studies presented in the references of that paper, then it has been calculated as the standard deviation of a series constructed as:
Ln(Pt) –Ln(Pt-1), where T is defined in months, P refers to prices in levels and Ln is the natural logarithm.
We all know that differencing a series in Ln is a proxy for the rate of change of the variable in levels; therefore here Ln(Pt) –Ln(Pt-1) is a proxy for monthly nominal inflation. Then the standard deviation of that series is taken as the measure of volatility (it can be annualized by multiplying by the square root of 12). It has been argued that this measure is better than other potential metrics because it avoids the issue of how to define the trend (see for instance Gilbert and Morgan, http://www.bis.gov.uk/assets/bispartners/foresight/docs/food-and-farming/drivers/dr18-food-price-volatility.pdf).
However, I would argue that leaving out the discussion of the trend is missing a key component of what we should be looking at. The next graph shows the monthly nominal Food Index calculated by the IMF (2005=100) (blue line); the red line is one possible representation of the trend (calculated using the Hodrick-Prescott filter); and the green line (in the left axis) is the absolute value of the percentage deviation of the Food Index from the HP trend (a variation of the usual coefficient of volatility which is the standard deviation of a variable divided by the average of that variable over a certain period). It starts in January 1960 and ends in January 2011.
There are at least two stories in the 2000s: the fact that the trend (red line) has been moving upwards since hitting a nominal bottom in late 1990s and early 2000s and the about 30% spike in 2007/2008 above trend (green line). As I discussed in other places (“Globalisation of Agriculture and Food Crises: Then and Now” in “Food Crises and the WTO.” B. Karapinar and C. Häberli (eds.), Cambridge University Press. 2010; http://www.cambridge.org/aus/catalogue/catalogue.asp?isbn=9780521191067), it is clear that the 2008 spike was smaller and took place over a more extended period than the one in the early 1970s: in the 1970s there was an almost 200% increase in about 5 years, while in the 2000s the increase was less than 140% over almost 9 years. Also the timing and movement of the prices of other commodities were different in the 1970s and in the 2000s.
In the 1970s the breakdown of the Bretton Woods arrangement for exchange rates and a series of supply and demand shocks led to a step adjustment in nominal food prices (and other commodities). As can be seen in the graph, food nominal prices had been oscillating around that new plateau until recently. The bottom in nominal prices in the late 1990s and early 2000s was deeper and more extended than the previous ones (the result of a series of financial crises -starting with the 1997 Asian crisis and ending with the 2002 Argentine crisis- which reduced demand and/or increased supply of food (and commodity) products, the global growth decline in early 2000s and the peaking of the last cycle of appreciation of the US dollar; I discuss the macro factors in both the 1970s and the 2000s in the paper mentioned above). But up until mid 2007, the nominal increase was in line with previous nominal cycles. The question is why it has been going up further since then and whether this trend may continue or not.
I do not believe we can discuss those questions properly focusing only on “volatility” (however defined). Rather, we need to discuss two aspects: the variability of the trend and the variability around the trend. In macroeconomics, the issue of variable trends has been emphasized long ago (see, for instance, Stock and Watson (1988), “Variable trends in economic time series.” Journal of Economic Perspectives, 2(3), 147–174). Also the factors and policies that affect the trend and those that affect the variability around it are usually different (although there may be cross effects).
If we accept the argument that we need to look at both the variability of the trend and the variability around the trend, then the issues to be defined in the analysis expand significantly: the analysis needs to clarify how the trend is defined and measured; whether it is expected or unexpected variability, and in the case of the latter whether those shocks may fall outside a “normal range” (which also requires a definition of what is “normal”); whether volatility is considered in world or in domestic prices; if the focus is on world prices, it is necessary to define the currency in which prices are quoted (such as US dollars, Euros, Special Drawing Rights, and so on); if the analysis centers on domestic prices, we need to define the relevant markets for price formation and measurement along the production, processing and distribution chain that links primary producers to final consumers; it is also important to clarify whether volatility is analyzed for nominal prices or for real prices (and in the latter case, an appropriate deflator must be identified (such as the Export Unit Value (EUV) Index for Advanced Economies, calculated by the IMF, the US Consumer Price Index (CPI) or the US Producer Price Index); whether the analysis focuses on specific commodities or on broader aggregates of commodities; it is also important to make explicit the relevant time horizon for the volatility calculations: is it an annual, seasonal, monthly or even daily time window? (the time horizon selected depends on the purpose of the analysis: for instance if the focus is on consumers, perhaps a shorter horizon (monthly) may be needed than in the case of producers who make decisions on longer time frames (at least yearly for planting decisions of many crops, and even longer for investment decisions) (see a somewhat more detailed discussion in http://www.acp-eu-trade.org/library/files/%20Diaz-Bonilla_Ron_En_01122010_ICTSD_Food%20Security%20Price%20Volatility%20and%20Trade.pdf)
Just to illustrate the differences I add two charts: one is the same Food Index as before but in real terms (deflated by the EUV); and the second one is with prices measured in SDRs.
Again, we have two different stories in the 2000s: after a long period of decline or at least stagnation, real prices started to move upwards (change in the trend) and, at the same time, there has been some more volatility than in later decades (change in variability around the trend), but without matching the variability of the 1970s.
The index in nominal SDRs (the red line) does not show the same strong upward trend as the index in US dollars (blue line), suggesting that part of the recent price increases (both trend and spike) may be just consequence of the devaluation of the dollar against other currencies since the peak in the early 2000s. An implication is that if developing countries have been revaluing their currencies against the dollar (as it seems to have been the case for many of them), the impact on domestic price would be lower (and this effect is separate from the issue of transmission related to infrastructure, input-output coefficients, production and marketing margins, market structures, and polices other than the exchange rate).
Still all this does not say much about the impact on countries, producers and consumers: that analysis requires a better definition not only of the notion of “volatility” being utilized but also a proper scaling of the shocks by some macroeconomic variable such as GDP, exports or fiscal accounts (at the country level), and household income or consumption (at the level of producers and consumers). An example of such scaling at the country level is the series of studies conducted by Bela Balassa in the early 1980s to analyze the shocks of the 1970s (see for instance, Balassa’s “Policy Responses to Exogenous Shocks in Developing Countries.” The American Economic Review, Vol. 76, No. 2, Papers and Proceedings of the Ninety-Eighth Annual Meeting of the American Economic Association (May, 1986), pp. 75-78.)
For instance, a possible country indicator may be food imports over total exports (i.e. how much of the income from all exports a country needs to pay for the food import bill). I present this indicator for 1970 to 2008 (the last year with data from FAO) for two groups of countries: net food importing countries (a WTO definition with some implications for trade negotiations) and low income food deficit countries (a category utilized for FAO). Of course, this indicator should be calculated at the country level, and not for groups of countries; also, the full impact may require 2009 data to be perceived in the series. But still the chart gives some idea of the size of the shocks in the aggregate.
The shock of the 1970s is clearly more visible for those two groups of countries than the last one (at least until 2008).
Going back to the G-20 paper, my comments do not detract from the value of the document mentioned, which covers a lot of ground and is very much worth reading (I read it on a sleepless night, flying from Washington DC to Geneva). At the same time, I am pretty sure that those that have been working on it, all intelligent people writing under the pressure of having to finish something for the next meeting of the principals, would agree with me that with more time it would be useful to conduct an analysis that a) distinguishes trends and volatility (and defines the latter according to the focus of analysis, i.e. country, consumer, producer); b) takes a systemic view of trends, cycles, shocks and crises, considering full macroeconomic cross effects (fiscal, monetary, inflation, exchange rates) of increases in all commodity prices (not only food) and other world variables (such as global interest rates) (such as the Balassa’s work in the 1980s); and c) differentiate impacts and situations of vulnerability (at the country and household levels) using adequately scaled metrics of the shocks. But certainly that is a research project that needs more time than what is usually available to finish a position paper for a rapidly approaching international meeting.