Working Paper

Informing SPR Policy through Oil Futures and Inventory Dynamics


New analysis provides guidance and insight on the impacts of using the Strategic Petroleum Reserve (SPR).

Key Findings

  • Examining the relationship between spot and future oil prices can better guide decisions about how and when to use the Strategic Petroleum Reserve (SPR) than the traditional “counting barrels” approach.
  • We statistically estimate the relationship among oil prices, price spreads, and oil inventories to analyze the potential impact of SPR releases on oil markets.
  • An SPR release of 10 million barrels could temporarily reduce oil prices by 2–3 percent and mitigate backwardation (when the price of oil delivered in the future is lower than the price delivered today) by 0.8 percentage points.
  • Historical simulations suggest that SPR releases in 1991 and 2011 led to oil prices about 15–20 percent lower for a time and avoided about 5 percentage points of backwardation, compared to a no-release counterfactual.


We examine how information on the time pattern of expected future prices for crude oil, based on the term structure of futures contracts, can be used in informing whether to draw down, or contribute to the Strategic Petroleum Reserve (SPR). Such price information provides insight on expected changes in the supply-demand balance in the market and can also facilitate cost-effective transitions for SPR holdings. Backwardation in futures curves suggests that market participants expect shocks to be transitory, creating a stronger case for SPR releases. We use vector autoregression to analyze the relationship between the term structure of futures contracts, the management of private oil inventories, and other variables of interest. This relationship is used to estimate the magnitude of the impacts of SPR releases into the much larger global inventories system. Under the assumption that strategic releases can be modeled as surprise inventory additions, impulse response functions suggest that a strategic release of 10 million barrels would temporarily reduce spot prices by about 2% to 3% and mitigate backwardation by approximately 0.8 percentage points. Historical simulations suggest that past releases reduced spot prices by 15% to 20% and avoided about 5 percentage points of backwardation in futures curves, relative to a no-release counterfactual. This research can help policymakers determine when to release SPR reserves based on economic principles informed by market prices. It also provides an econometric model that can help inform the amount of SPR releases necessary to achieve given policy goals, such as reductions in prices or spreads.