Home » Research Blog » Commodity trading strategies and convenience yields

Commodity trading strategies and convenience yields

Convenience yield can be interpreted as a leasing rate for physical commodities. Returns on convenience claims are premia earned by investment strategies for providing this leasing service. An empirical analysis suggests that they depend on risk factors related to other asset classes, however. The inertia in these risk factors seems to help predicting returns on convenience claims.

Bollinger, Thomas and Axel Kind (2015), “Risk Premiums in the Cross-Section of Commodity Convenience Yields”, University of Konstanz, Department of Economics, Working Paper Series, 2015-17

The below are excerpts from the paper. Headings and cursive text have been added.

Understanding convenience yields

“Commodity processors, such as oil refineries, metallurgical plants, and food manufacturers, require physical availability of raw materials to sustain their business activity. To avoid a costly disruption of the production process, they may be forced to purchase a commodity at a spot price much higher than the corresponding futures price…The theory of storage, developed by Kaldor (1939) and Working (1949), introduces the concept of convenience yield, i.e. the benefit that accrues to the owner of a physical inventory, to explain the observed discrepancy between spot and futures prices and thus the shape of the commodity term structure.”

“A convenience yield can be interpreted as a return on a particular asset, termed ‘convenience claim’…This asset consists of a short position in a commodity futures combined with a spot purchase of the related underlying. Convenience claims thus correspond to a short-term leasing contract of one unit of inventory, or, in other words, a calendar spread. Hence, investments in convenience claims allow one to receive convenience yield as compensation for the temporary physical provision of an underlying commodity. More precisely, since convenience yields are known in advance, the returns of the convenience claim correspond to the negative convenience yield in the previous month.”

N.B.: The value of physical inventories typically increases when scarcities arise. As a consequence, convenience yields help predicting future demand and price changes (view post here).

An empirical analysis

“The empirical analysis is conducted on a sample of 22 commodity futures in the period from January 1991 to December 2011 and includes a total of 36, 319 monthly futures prices…Energy products, metals, and agricultural products are included.”

“We choose to implement the Schwartz three-factor model because it is widely used…The state variables…are the spot price, the instantaneous convenience yield, and the risk-free interest rate…[This method] allows one to extract and separate the convenience yield from other determinants affecting futures prices (i.e., commodity spot prices and interest rates)…By interpreting convenience yields as returns of convenience claims we can estimate multi-factor asset pricing models…and can investigate the risk exposures and the existence of risk premiums in the cross-section of convenience yields. “

“Five economic variables…represent plausible economy-wide risk factors that are commonly used in the empirical asset pricing literature…returns of the S&P 500 index, the Citigroup world government bond index…the Goldman Sachs commodity index…the growth rates of the U.S. industrial production and…unexpected inflation.”

Instrumental variables are used to predict changes of risk premiums associated to each risk factor. The four instrumental variables used in this paper are…[i] the lagged dividend yield of the S&P 500 index…[ii] the average lagged spread of yields between Baa-rated and Aaa-rated corporate bonds…[iii] the capacity utilization rate of all industries [in the U.S.] and the level of new orders in the [U.S.] economy.”

What drives convenience yields?

“For all commodities except some precious metals, clear evidence of stochastic and mean-reverting convenience yields is found.”

“[Our] analysis…relates returns of convenience claim investments (which are directly affected by the term structure of commodity futures) to risk factors affecting stock and bond returns.… It reveals the existence of significant premiums embedded in convenience yields for systematic risk factors typically related to other asset classes… Exposure toward some of the economy-wide risk factors is rewarded. More specifically, convenience claims that load on the bond and the commodity spot market earn a statistically on significant premium…The results prove to be robust with respect to the cross-sectional composition of the data sample and the specification of the asset pricing model.”

“Changes in conditional betas are found to forecast variations in convenience yields… Since variations in the convenience yield are driven by systematic factors, the roll gains in commodity trading strategies cannot be seen as purely idiosyncratic return components of commodity investments.”

Forecasting returns

“The identified risk premiums are only marginally predictable by a set of commonly used instrumental variables…the majority of risk premiums in commodity markets do not seem to be predictable.”

“However, a more promising approach for predicting convenience-claim returns is to exploit the strong degrees of autocorrelation detected in their factor loadings. Selecting equally sized portfolios of commodities with high and low conditional betas with respect to the two significant risk factors (i.e., the bond and commodity spot market) results in significantly different average convenience-claim returns.”

“To test this simple relation, we construct for each of the five risk factors portfolio pairs consisting of the 11 commodities with the highest (lowest) conditional betas. Each portfolio is rebalanced at the beginning of each month according to the betas measured for the previous month.…Statistically significant differences between portfolio pairs are detected when sorting according to…[bond market] loadings and…[commodity index price] loadings. In both cases, the return differential exceeds four percentage points per annum and the high-beta portfolios returns are significantly larger than zero.”


Related articles