The Bond Market Prices Oil, Not Oil Fear

While digging into how this year’s oil shock has fed into the rates landscape, I kept running into the same assumption, and frankly wanted to know whether the data actually supported it. The underlying question was also discussed briefly in a recent interview, a conversation about the long end: if oil is driving treasury yields, is it the price of oil or the uncertainty around it. The timing and stakes are certainly not abstract, the Gulf conflict which began in late February pushed Brent spot to $138.21 a barrel in April and $135.51 as recently as October 2. Furthermore, In early October, both standard estimates of the 10-year term premium hit their highest levels in more than a decade: Kim-Wright 1.10% on October 5, highest since 2010; ACM 0.99% on October 7, highest since 2014. The prevailing read is that the long end is pricing the uncertainty of the energy shock moving forward, rather than the price of oil itself. I assumed the same, in fact it actually seemed obvious that a more uncertain oil market should make long bonds riskier to hold. Moreover, it’s not a naive view per se, it follows from some of the best-established research on the term premium. Now of course, oil is not simply the only force on the long end, such has been especially evident across this fall. All inclusive: fiscal supply, the Fed and the AI investment boom all matter, my question is narrower: when oil does move the long end, which part of oil does the moving? So I tested it. 

The term premium, or extra yield demand in order to hold longer maturity debt such as a 10 year Treasury rather than rolling over short-term bills is direct compensation for the risk that rates or inflation move adversely. Jonathan H. Wright details in a 2011 piece for the American Economic Review that across ten industrialized countries and nearly two decades, there was a strong positive relationship between term premia and inflation uncertainty. Where in term premia fell most in countries that cut inflation uncertainty by reforming their monetary policy. Michael D. Bauer, Glenn D. Rudebusch, and Jing Cynthia Wu (2014) furthermore expanded on his piece, correcting certain estimates for statistical bias and continued to find the relationship with inflation uncertainty remained intact. Now Wright’s evidence was about inflation uncertainty generally, not oil specifically, but headline CPI swings can mostly be attributed to energy, about three quarters in fact from 2007 - 2026, while energy’s swings itself is largely attributed to oil, 46% of monthly variance and 59% with a month’s lag. The most volatile parts of inflation are the hardest to predict. So uncertainty about inflation is, in large part, uncertainty about oil, which is why the belief that oil uncertainty raises the term premium is reasonable.

How energy’s share of inflation swings was measured: Using BLS consumer price data (headline, energy and core, seasonally adjusted), calculated the month-over-month change in each from June 2007 to September 2026, leaving out October and November 2025, when the government shutdown interrupted the data. Then estimated how much weight energy and core each carry in headline inflation, and split headline’s monthly swings between them. Energy accounts for roughly 73%, core for 23%, and the remaining 4% is mostly food.

How oil’s share of energy price swings was measured: EIA’s daily Brent spot price for each month was averaged and compared to its monthly changes with the monthly changes in energy CPI over the same period. On its own, the same month’s oil move explains 46% of the variation in energy CPI across 230 months. Adding the previous month’s oil move, to allow for the lag before crude prices reach the pump, raises that to 59%.

Now there is real academic groundwork on this question. It comes from a 2020 study from Balcilar, Gupta, Wang and Wohar, “Oil Price Uncertainty and Movements in the US Government Bond Risk Premia,” however there are a few distinctions to make. The authors tested realized excess returns of 2, 3, 4 and 5yr bonds over the 1-year. They measured oil uncertainty with a statistical model of WTI’s price swings (GARCH) with OVX as a robustness check, using monthly data (1953 - 2016) to analyze if it predicted anywhere in the distribution, especially the tails. They find that oil uncertainty helps predict excess returns on these bonds and, more strongly, their volatility. However, their standard test of the average relationship found no significant link; the effects appear only in the tails of the distribution. This article asks a narrower question: does oil uncertainty, measured by OVX, raise the 10-year term premium on average, week to week, through 2026? The 10-year is where long-run inflation uncertainty should matter most, and controlling for the Brent price and the VIX separates oil uncertainty from oil’s price and from fear in general, which their design does not do. 

Put as a test, if oil uncertainty raises the term premium then weeks when oil uncertainty jumps should be weeks when the term premium rises. Oil uncertainty being Cboe’s crude oil volatility, OVX, which uses oil option prices to show how much traders expect oil to swing over the next 30 days, it's the market’s own forward facing gauge into oil uncertainty. For term premium two estimates were used, the New York Fed’s ACM (built from Treasury yields alone) and FRB’s Kim-Wright (which uses surveys of rate forecasts). Together they hold a weekly correlation of .74. For controls, the VIX was used to gauge discrepancies between general market fear and OVX, Brent’s spot price was included so that OVX couldn’t take credit for the effect of oil’s price moves. 

The frequency tested was weekly changes (Friday - Friday), May 2007 to October 2026, or around 1,000 weeks, with May 2007 being when OVX began. However March to June 2020, the Covid crash where oil prices went negative and broke the oil volatility index, was removed to not distort the data. Each series was then converted into weekly changes: the term premium in basis points, OVX and the VIX in index points, and Brent as a percentage change. The weekly change in each term premium was then regressed on the weekly changes in OVX, VIX, and Brent together, therefore each variable's effect is measured while holding the other two fixed. 

Specifically, oil uncertainty is being measured to explain the term premium beyond just market fear and price, and if it cannot, then the belief fails. The results yielded that once fear (VIX) and oil price (Brent spot) are accounted for, oil uncertainty (OVX) has no measurable effect on the term premium, for either estimate. The effect is statistically indistinguishable from zero. This holds true across each era, 2007 - 14 (the financial crisis, zero rates, and quantitative easing), 2015 - 19 (Fed hike, oil crash), and 2021 - 26 (high inflation, hiking, 2026 oil shock), therefore the results are not driven by one period’s uniqueness. Market data obviously isn’t independent from one week to the next, with surprises carrying over and volatility coming in clusters, therefore standard errors were adjusted using the Newey-West method, allowing each week to be correlated with the four weeks before it. Re-running the test with eight- and twelve-week windows leaves the results unchanged. The results yielded can be visualized in the results table. Row one of the Main Results shows that per 1 point rise in OVX, with controls, yields no statistically significant effect, near zero with no detectable link. Coefficients of +0.004 for Kim-Wright and -0.035 for ACM, with respectively t values of 0.08 and -0.39. Therefore a 10 point jump in OVX, a big week of oil fear, implies a term premium change of under half a basis point on either estimate. In reality the test points instead that the bond market rather prices oil’s level not volatility / uncertainty. The Brent result was a +0.24bp on Kim-Wright (t = 5.46) and +.39bp on ACM (t = 3.28) per 1% change. Wherein a 10% rise, goes with roughly a 2.4 to 3.9bp change on term premium that week. The findings showed that the effect is most clearly significant since 2021 while weaker beforehand, it has strengthened over the past few years. 

Now our test is niche, with strict controls to block out general volatility and large oil price swing noise from general oil uncertainty. Perhaps the VIX and Brent controls are hiding too much of oil uncertainty’s effect. Therefore, they were dropped and gave OVX every advantage (Main Results, Row two). As a result, the effect turns negative and significant (-0.15 per OVX point, t = -2.9 on Kim Wright) and (-.23bp, t = -2.4 on ACM), higher oil fear goes with a lower term premium. Even at the extremes, with controls in place, in the 50 weeks of the sharpest OVX spikes, there’s no positive effect and when OVX was at its highest levels the term premium tended to fall. This is happening due to OVX spikes tending to occur when there are broad market crises, the crises themselves lower the term premium, therefore once VIX is controlled nothing is left for OVX. In oil volatility’s largest spikes 2008, 2015 - 16, and 2020 it came during demand collapses because the economy was weakening or a supply gut, investors expected rate cuts and falling inflation, long Treasuries gained value. The bonds act as insurance, investors accept a lower premium to hold. While oil price tells another story, higher oil lifts inflation the bond market expects, and an investor holding that 10yr needs more compensation for said risk. This occurs whether the economy is slowing down or in this year, holding up. What the bond market appears to price is where oil is and potentially going, not the uncertainty around it.

For this analysis, some limits to consider, the term premium is what remains of the 10-year yield after stripping out the expected path of short-term rates, so a different model gives a different number as seen with ACM and Kim-Wright’s own separate measurements. Yet the data still yielded the same results for our testing purpose. Furthermore, at the six month lag window the data for Kim-Wright shows that sustained oil volatility could seep into inflation worries the way it doesn’t in a single week, but across 38 observations the data is too thin to confirm. Our data is also limited in the regard we are only measuring OVX, VIX, and Brent together and while it was found that the three explain about 5 - 9% of week-to-week variation in term premium  (Kim-Wright R²= 0.093 and ACM R²= 0.051), there is a myriad of other unmeasured factors; Fed expectations, Treasury Supply, global demand for safe assets, inflation data, AI investment boom’s pressure on real rates and long-dated corporate supply to name a few. In regards to the data found by Balcilar and co-authors, our test does not rule out effects in the tails which prior research suggests may exist.

Moving forward, the current landscape can be analyzed in regards to this article's output. Right now from October 2nd to November 3rd Brent spot fell 21% from $135.51 to $107.03, while oil uncertainty rose to around 62 on OVX, its highest reading of the period. The response from the term premium, Kim-Wright fell 18bp and ACM 24 bp fell from October 2nd to November 3rd with oil’s price, while break-evens eased to 2.30%. Applying the full-sample estimates to this month, oil’s price and general market fear account for roughly half of that decline (43% on Kim-Wright, 50% on ACM), and OVX contributes essentially nothing. Therefore it is paramount to read the OVX spike in context before acting on it, if volatility jumps, check which way Brent is moving and in regard to VIX movement. A spike in conjugation with falling oil and rising fear has historically come with a lower term premium, therefore not necessarily a signal the long end will sell off. The 10% Brent move outlined in the sample has gone with a 2 to 4bp on term premium in the same week, stronger since 2021 when inflation was the markets central concern, were it to fade, expect the link to weaken. Investors worried about oil pushing long yields higher are better served by the exposure to the direct oil price itself, than by volatility. Using oil volatility as a term premium hedge pays for a risk which the long end hasn't historically priced.

Ciations MLA

Adrian, Tobias, et al. “Pricing the Term Structure with Linear Regressions.” Journal of Financial Economics, vol. 110, no. 1, 2013, pp. 110–38.

Adrian, Tobias, et al. “Treasury Term Premia.” Federal Reserve Bank of New York, www.newyorkfed.org/research/data_indicators/term-premia-tabs. Accessed 11 Nov. 2026.

Balcilar, Mehmet, et al. “Oil Price Uncertainty and Movements in the US Government Bond Risk Premia.” The North American Journal of Economics and Finance, vol. 52, 2020, article 101147, https://doi.org/10.1016/j.najef.2020.101147.

Bauer, Michael D., et al. “Term Premia and Inflation Uncertainty: Empirical Evidence from an International Panel Dataset: Comment.” American Economic Review, vol. 104, no. 1, 2014, pp. 323–37.

Board of Governors of the Federal Reserve System. “Term Premium on a 10 Year Zero Coupon Bond [THREEFYTP10].” FRED, Federal Reserve Bank of St. Louis, fred.stlouisfed.org/series/THREEFYTP10. Accessed 10 Nov. 2026.

Cboe Global Markets. “CBOE Crude Oil ETF Volatility Index [OVXCLS].” FRED, Federal Reserve Bank of St. Louis, fred.stlouisfed.org/series/OVXCLS. Accessed 11 Nov. 2026.

Cboe Global Markets. “CBOE Volatility Index: VIX [VIXCLS].” FRED, Federal Reserve Bank of St. Louis, fred.stlouisfed.org/series/VIXCLS. Accessed 11 Nov. 2026.

Federal Reserve Bank of St. Louis. “10-Year Breakeven Inflation Rate [T10YIE].” FRED, fred.stlouisfed.org/series/T10YIE. Accessed 6 Nov. 2026.

Kim, Don H., and Jonathan H. Wright. “An Arbitrage-Free Three-Factor Term Structure Model and the Recent Behavior of Long-Term Yields and Distant-Horizon Forward Rates.” Finance and Economics Discussion Series, no. 2005-33, Board of Governors of the Federal Reserve System, 2005.

Newey, Whitney K., and Kenneth D. West. “A Simple, Positive Semi-Definite, Heteroskedasticity and Autocorrelation Consistent Covariance Matrix.” Econometrica, vol. 55, no. 3, 1987, pp. 703–08.

U.S. Bureau of Labor Statistics. “Consumer Price Index for All Urban Consumers: All Items in U.S. City Average [CPIAUCSL].” FRED, Federal Reserve Bank of St. Louis, fred.stlouisfed.org/series/CPIAUCSL. Accessed 9 Nov. 2026.

U.S. Bureau of Labor Statistics. “Consumer Price Index for All Urban Consumers: All Items Less Food and Energy in U.S. City Average [CPILFESL].” FRED, Federal Reserve Bank of St. Louis, fred.stlouisfed.org/series/CPILFESL. Accessed 9 Nov. 2026.

U.S. Bureau of Labor Statistics. “Consumer Price Index for All Urban Consumers: Energy in U.S. City Average [CPIENGSL].” FRED, Federal Reserve Bank of St. Louis, fred.stlouisfed.org/series/CPIENGSL. Accessed 9 Nov. 2026.

U.S. Energy Information Administration. “Crude Oil Prices: Brent - Europe [DCOILBRENTEU].” FRED, Federal Reserve Bank of St. Louis, fred.stlouisfed.org/series/DCOILBRENTEU. Accessed 9 Nov. 2026.

Wright, Jonathan H. “Term Premia and Inflation Uncertainty: Empirical Evidence from an International Panel Dataset.” American Economic Review, vol. 101, no. 4, 2011, pp. 1514–34.