Author here. Built a 1970–2026 dataset of US oil shocks to test the Hamilton-style “oil shock → recession” pattern using oil burden, not nominal price. Burden = (WTI × US petroleum consumption × 365) / nominal GDP. Sources: FRED (WTISPLC, CPIAUCSL, GDP), EIA Monthly Energy Review Table 3.1 (PATCPUS), NBER BCDC. Reproducible from public endpoints, no API keys.
Episodes by peak burden:
1979 Iran Revolution — 8.65% — recession
2008 Peak — 6.47% — recession
2011 Arab Spring — 4.80% — no recession (only false positive)
1973 OPEC — 4.34% — recession
1990 Gulf War — 3.70% — recession
2022 Ukraine — 3.31% — no recession
2000 Y2K/OPEC — 2.33% — no recession
2026 Iran War — 2.57% — ongoing
Every completed episode at or above 4% of GDP was followed by an NBER recession within 18 months from shock onset; sub-threshold episodes were not, except 2011. The 2026 episode is still open. Caveats discussed in the writeup, not glossed: small sample (n=7 completed), the 2000 episode is sensitive to onset-based vs peak-based timing, and 2008 was partly endogenous to the financial conditions of that fall. CSVs, chart, and build script are in the writeup. Happy to discuss.
RealisticForYou on
In reference to vehicle gas consumption the typical U.S. consumer pays a much smaller percentage of their earnings towards gas….in comparison to all of your prior data. With better gas mileage and with the boatload of EV’s and hybrids, it could be a ‘gas” recession may not occur.
And how foreign oil dependent was the U.S. in those prior years? Today, U.S. produces nearly all of its oil.
We are fortunate this conflict is occurring during summer months without the need for heating oil.
2 Comments
Author here. Built a 1970–2026 dataset of US oil shocks to test the Hamilton-style “oil shock → recession” pattern using oil burden, not nominal price. Burden = (WTI × US petroleum consumption × 365) / nominal GDP. Sources: FRED (WTISPLC, CPIAUCSL, GDP), EIA Monthly Energy Review Table 3.1 (PATCPUS), NBER BCDC. Reproducible from public endpoints, no API keys.
Episodes by peak burden:
1979 Iran Revolution — 8.65% — recession
2008 Peak — 6.47% — recession
2011 Arab Spring — 4.80% — no recession (only false positive)
1973 OPEC — 4.34% — recession
1990 Gulf War — 3.70% — recession
2022 Ukraine — 3.31% — no recession
2000 Y2K/OPEC — 2.33% — no recession
2026 Iran War — 2.57% — ongoing
Every completed episode at or above 4% of GDP was followed by an NBER recession within 18 months from shock onset; sub-threshold episodes were not, except 2011. The 2026 episode is still open. Caveats discussed in the writeup, not glossed: small sample (n=7 completed), the 2000 episode is sensitive to onset-based vs peak-based timing, and 2008 was partly endogenous to the financial conditions of that fall. CSVs, chart, and build script are in the writeup. Happy to discuss.
In reference to vehicle gas consumption the typical U.S. consumer pays a much smaller percentage of their earnings towards gas….in comparison to all of your prior data. With better gas mileage and with the boatload of EV’s and hybrids, it could be a ‘gas” recession may not occur.
And how foreign oil dependent was the U.S. in those prior years? Today, U.S. produces nearly all of its oil.
We are fortunate this conflict is occurring during summer months without the need for heating oil.