“Economists have recognized that not all oil price shocks are alike.… The responses of industrial production, unemployment, core inflation, and the Fed Funds rate depend critically upon whether the oil shock was a demand shock or a supply shock.”
– Daniel P. Ahn, Columbia University, and Leonid Kogan, MIT Sloan School of Management, “Crude or Refined: Identifying Oil Price Dynamics through the Crack Spread,” Working Paper, 2011
Many sophisticated investors view the world in terms of risk factors, as opposed to asset classes. In our opinion, diversifying across risk factors can be more effective than diversifying across asset classes, because asset classes are simply carriers of risk factors. While correlations among most risk factors have historically been, on average, low and stable over different regimes (we view this as a good thing), some risk factors have had widely varying correlations depending on underlying economic conditions. Oil and its correlation with equities is an interesting case in point.
The link between oil price dynamics and economic growth generally depends on whether oil shocks are driven by changes in supply or demand. Consider the case of rising oil prices:
- In a demand regime, oil prices may rise due to a booming economy, and say, demand from an increased number of people buying cars in China. In this case, the link between economic growth and oil price dynamics is strong.
- In a supply regime, oil prices may rise due to political turmoil in oil-producing countries, which leads to fears of shortages. In this case, oil price dynamics have little to do with economic growth.
Interestingly, oil price regimes are usually driven by both demand and supply, simultaneously, according to the recent working paper, “Crude or Refined: Identifying Oil Price Dynamics through the Crack Spread,” by Daniel Ahn of Columbia University and Leonid Kogan of MIT. Sometimes demand explains most of the variations in oil prices, and sometimes supply dominates, but rarely are oil price regimes exclusively demand or supply driven.
While academic research has so far focused on policymaking implications (what should the Federal Reserve do in response to an oil shock?), it has important investment implications as well. If oil exposure appears to be increasing portfolio equity beta, everyone involved in asset allocation needs to know.
Using the Crack Spread to Determine Supply-Demand Dynamics
The crack spread – the price difference between crude oil and refined products – helps disentangle demand and supply shocks in oil prices. If prices for refined products increase faster than the price of crude oil, then the crack spread widens and we suspect a demand shock. The logic is simple: Consumers are buying more gas at the pump, which drives gasoline prices higher. This increase is only partly transferred into higher crude oil prices, as refineries use their inventory cushions.
Similarly, if crude oil prices rise faster than refined products, then the crack spread narrows and we suspect a supply shortage. The onset of the recent turmoil in the Arab Gulf region provided a good example. Crude oil prices rose, but gasoline prices did not rise proportionately. In other words, the crack spread narrowed because of an incomplete pass-through of the oil price increase.
Obviously these examples are simplistic, as more factors are at play, but they provide insight into the “incomplete pass-through” principle.
The chart below shows how to attribute crude oil price changes to demand or supply shocks based on changes in the crack spread. This regime-specific analysis of oil price changes reveals opportunities to trade oil as an equity surrogate when two conditions occur:
1. Demand appears to be driving oil prices, and
2. Turmoil in oil-producing countries appears likely to jeopardize supply.
Under these conditions, trading oil has the potential to provide better equity beta exposure than investing directly in equities, because unlike most equities, oil is also likely to benefit from a supply shortage – a potential win-win proposition. Investors who focus on risk factor allocation instead of asset allocation often look for such opportunities to trade risk exposures that cut across asset classes. In this case, the investor wants to be long the growth factor, but also wants to limit downside.
Based on data collected from the S&P 500 and oil futures, the beta between oil and equities was near zero from January 1971 to December 2010, as the following chart shows. However, betas in the supply and demand regimes differed significantly:
- When oil prices rose due to increased demand, oil’s equity beta was 0.19, compared to -0.15 when oil prices rose due to a supply shortage.
- Similarly, when oil prices fell due to decreased demand, oil’s equity beta was 0.11, compared to -0.16 when supply increased.
A statistical test called the “t-test” reveals the differences between demand and supply betas were statistically significant at 85% for oil price decreases and 93% for oil price increases. (A t-test is a test designed to determine whether the average of two groups are statistically different. Here, each beta is estimated with a level of uncertainty, hence the “groups” represent distributions or ranges of possible betas.)
When demand pushes oil prices higher, oil’s average equity beta has not historically risen all the way to 1.0 over the regime because the relationship depends on a number of variables. For example, oil prices generally increase in growth environments (more people buying cars), but when oil prices rise too fast, growth might suffer from a decrease in consumption (less disposable income after buying gas for the car). This circular reference partly explains the imperfect correlation. Indeed, when oil reaches a peak, a recession generally follows, as James Hamilton shows in “Nonlinearities and the Macroeconomic Effects of Oil Prices” (to be published in the journal, “Macroeconomics Dynamics”).
Initial conditions in the equity market also matter. When valuations are very high or very low, the impact of rising growth on equity returns tends to differ from what might be expected under fair valuation, and the oil-equity-growth connections loosen. And oil price dynamics are driven by both demand and supply simultaneously, as Ahn and Kogan have shown. Hence there’s no such thing as a “pure” demand regime. This combined effect of demand and supply is what we believe has the potential to make the oil-as-equity-surrogate trade attractive.
So overall, while the average equity beta of oil does not typically reach the +1.0 and -1.0 levels over a regime – although it can in sub-periods of the regime when the absolute value of the correlation between equity and oil is very high – the demand and supply equity betas for rising oil prices suggest that in markets subject to supply shocks during a period of secular demand growth, oil may provide growth exposure with an asymmetric payoff that may be tactically superior to equities.
Tilting Equity Beta Through Oil Exposure
Historically, the correlation between oil prices and the equity market has been highly unstable. To disentangle oil price dynamics into demand and supply regimes helps identify when we can expect correlations to be high or low. This in turn helps inform broadly diversified investors when their oil exposure is tilting their portfolio toward a higher or lower equity beta. And we believe making the right call on equity factor exposure in a portfolio is one of the most important decisions an investor can make.
PIMCO uses the risk factor approach to help clients understand their portfolio positioning in the context of risk and the firm’s tactical, cyclical and secular macroeconomic views. For more on risk factor allocation, please see PIMCO Viewpoints, “The Myth of Diversification: Risk Factors vs. Asset Classes,” September 2010.
We thank the following colleagues and industry compatriots for their insights and contributions to this paper: Daniel Ahn, Vineer Bhansali, Helen Guo, Fei He, Leonid Kogan, Ravi Mattu and Curtis Mewbourne.