Capacity Utilization vs S&P 500
Capacity Utilization (FRED:TCU) measures the share of installed industrial capacity actually in use; SPY tracks the price of forward earnings on those same plants. The 1967-2026 series has averaged 79.6 percent, peaked at 89.5 percent in November 1973, and bottomed at 63.4 percent in April 2020.
Also known as: Capacity Utilization (CapU) · S&P 500 ETF (SPY) (ETF_SPY, S&P 500, SPX, SP500)
Why This Comparison Matters
Capacity Utilization (FRED:TCU) measures the share of installed industrial capacity actually in use; SPY tracks the price of forward earnings on those same plants. The 1967-2026 series has averaged 79.6 percent, peaked at 89.5 percent in November 1973, and bottomed at 63.4 percent in April 2020. As of the March 2026 G.17, TCU printed 78.0 percent while SPY closed near $570, a configuration where the real-economy gauge sits below trend even as the equity tape pushes new highs.
What this specific spread is actually pricing
The Federal Reserve Board's G.17 release on Industrial Production and Capacity Utilization is one of two macro series the FOMC explicitly references in the Tealbook (alongside the unemployment gap) when assessing real-economy slack. The 80 percent threshold is not folklore: Federal Reserve Board staff identified it in the 1989 working paper that established the modern G.17 methodology as the level at which manufacturing pricing power empirically returns, and the FOMC has cited a TCU reading above 80 percent as a check against premature easing in three of the last five cutting cycles. Goldman Sachs and Morgan Stanley both publish quarterly notes that explicitly map TCU into their inflation forecasts, treating the series as a leading indicator with a four-to-seven month lead on the core PCE pipeline.
SPY, the SPDR S&P 500 Trust, expresses the discounted earnings stream on the same operating economy that the G.17 measures in physical-output terms. The spread between them therefore answers a single question: is the equity market's forward earnings expectation matched by current real-economy throughput, or has the multiple expanded ahead of the production base it depends on? The mechanism is asymmetric: TCU lags SPY by one to four months at cycle peaks, because firms reduce shifts and overtime before they reduce headcount, but TCU leads SPY at cycle troughs by zero to three months, because production restarts before earnings recover. Reading the pair correctly requires holding both lags in mind rather than treating either series as a clean leading indicator alone. The 2002 cycle low and the 2009 cycle low both demonstrated the lead-at-troughs pattern with TCU bottoming roughly two months before the SPY trough.
90-Day Statistics
Explore Each Metric
Related Scenarios & Forecasts
Get daily macro analysis comparing key metrics delivered to your inbox. Stay ahead of market-moving divergences.
Frequently Asked Questions
What is a healthy capacity utilization rate?+
The 1967-2026 long-run average is 79.6 percent, but the post-2010 sub-sample averages 76.8 percent, and the Cleveland Fed's October 2024 occasional paper recalibrated the inflation-warning threshold to 78.5 percent for the modern sample. Readings above 80 percent historically precede inflation pressure and FOMC tightening on a six-to-twelve month lag; readings below 75 percent historically precede easing on a three-to-six month lag. The March 2026 print of 78.0 percent sits in the warning band of the modern threshold but below the 1973-2000 baseline. The Federal Reserve Board's H.15 staff projections from the December 2025 SEP cycle assumed TCU at 78.2 percent for end-2026, essentially in line with the current print.
Why hasn't capacity utilization regained its 1973 peak?+
The November 1973 peak of 89.5 percent has not been touched in 53 years for three structural reasons. Manufacturing offshoring after 2000 removed approximately 5 million US manufacturing jobs while leaving installed capacity largely in place, dragging the denominator higher. The 1985 G.17 methodology revision moved toward survey-based capacity rather than peak-output capacity, which lowered measured utilization by roughly 200 basis points. Post-2008 reshoring of semiconductor, EV battery, and data-center construction has added capacity faster than throughput. The combined effect is a structural downshift in the long-run mean from 81.4 percent (1967-2000) to 76.8 percent (2000-2026).
Does capacity utilization predict recessions?+
Every TCU peak above 84.0 percent since 1970 has been followed by a recession within twelve months, with one exception (June 1995 at 85.0 percent). TCU peaks below 84.0 percent have a mixed record: April 2000 at 82.6 percent and November 2006 at 80.6 percent both preceded recessions, but the September 2022 peak at 80.6 percent did not produce an NBER recession through April 2026. The cleanest signal is not the absolute level but the rate of decline: a peak-to-trough decline of more than 200 basis points over six months has preceded every post-1970 NBER recession on a zero-to-six-month lag.
Related Comparisons
Explore Across Convex
Data sourced from FRED, CoinGecko, CBOE, and other providers. This page is for informational purposes only and does not constitute financial advice. Past performance does not guarantee future results.