Auto Sales vs S&P 500

Auto purchases sit at the intersection of household credit, labor-income confidence, and durable-goods replacement, and the SAAR is the highest-frequency clean read on the consumer balance sheet that the BEA publishes. The Conference Board includes vehicle sales in the supplementary panel that supports its Leading Economic Index, and Federal Reserve research notes that auto sales per capita roll over an average of nine months before post-war business-cycle peaks. The Richmond Fed's May 16, 2023 macro-minute brief specifically called out the indicator's leading properties through the past three cycles. Pairing the series against SPY is the analytic move: SPY is the discount-rate-sensitive long-duration claim on aggregate corporate earnings, while TOTALSA is a real-economy credit-sensitive flow that responds to financing rates with roughly a two-quarter lag.

The pair shows up specifically on Fed economic-data dashboards and on Yardeni Research's consumer-indicator panel because it answers a single question: is the equity rally being underwritten by household demand, or by margin expansion in spite of softening household demand. April 2026's 16.0 million SAAR is a 7.3 percent decline from the tariff-pull-forward distorted April 2025 print of 17.2 million, and SPY is 5.16 percent higher year-to-date. That divergence, sustained across the past 24 months while SPY compounded roughly 60 percent off the October 12, 2022 low, is exactly the configuration the pair is built to surface. The 1998-1999 analog is the closest historical precedent for this divergence pattern, and that episode resolved through SPY catching down to the labor and credit signal between March 24, 2000 and October 9, 2002.

The historical track record gives the pair its predictive credibility. In the 2007-2009 downturn, SAAR fell from 17.0 million in early 2007 to 9.3 million in Q1 2009, a 45 percent peak-to-trough decline that began two quarters before SPY topped on October 9, 2007. Annual sales fell from approximately 17.0 million in 2006 to 10.6 million in 2009, a roughly 40 percent collapse. The 2020 COVID episode compressed the lead time: SAAR fell from 16.8 million in February 2020 to 9.0 million in April 2020, the lowest reading since April 2010, while SPY drew down 34 percent over five weeks between February 19 and March 23, 2020 before recovering through August.

The 2001 recession showed a milder 12 percent SAAR decline while SPY fell roughly 49 percent peak-to-trough between March 24, 2000 and October 9, 2002. The 1990-1991 cycle saw SAAR decline by 17 percent peak-to-trough with SPY drawing down 20 percent into the October 11, 1990 low. The long-run record is that every post-1970 recession has been preceded by a SAAR decline of at least 8 percent from the prior 12-month peak, and the lead time has averaged nine months but ranged from one (2020) to eighteen (1990) months. The pair is therefore not a precise market-timing tool but a regime classifier: when SAAR is rolling and SPY is making new highs, the burden of proof shifts onto the equity bull case to explain why this cycle is different. Federal Reserve research dating to 1939 documents the manufacturing workweek leading business-cycle peaks by an average of 10 months, and the auto-sales series carries similar leading properties.

The post-2021 cycle has produced the largest sustained TOTALSA-SPY divergence in the FRED record. SPY has compounded roughly 60 percent off the October 12, 2022 low while SAAR has held in a narrow 15.4 to 16.5 million band, well below the 17.1 million 2019 peak. Three structural factors explain the wedge. First, vehicle prices rose 22 percent between 2020 and 2023 while financing rates climbed from 4 percent to over 7 percent, pulling annual affordability roughly 200 dollars per month higher even at flat unit volumes. Second, the equity rally has been concentrated in mega-cap technology, where AI-capex-driven free-cash-flow generation has decoupled from the household credit cycle. Microsoft's April 29, 2026 fiscal Q3 capex guide of 190 billion dollars for FY2026, up 61 percent year-over-year, exemplifies the earnings-driver decoupling.

Third, the labor market has not yet produced the income shock that historically converts a SAAR plateau into a SAAR decline. Average weekly hours have ticked down from 34.4 hours to 34.2 between mid-2024 and March 2026, but unemployment remains sub-4.5 percent. The pair is therefore in an unusual configuration: the leading credit-sensitive series has rolled into a plateau without confirming a recession, and the equity leg has been free to rerate on the AI-capex story. Watching for either a SAAR break below 15 million or a sub-34.0 hours print is the way to time when the historical relationship reasserts. The May 2026 employment situation report, scheduled for May 2, 2026, is the next confirmation point on the labor leg. The May 28, 2026 TOTALSA monthly release will confirm whether the April 16.0 million projection holds.

Roughly 86 percent of new vehicles and 55 percent of used vehicles are financed in the US, which makes TOTALSA mechanically sensitive to two specific rate benchmarks: the 60-month new-car loan rate (FRED:RIFLPBCIANM60NM) and the auto-ABS spread to Treasuries. Both channels transmit Fed policy into household demand within roughly two quarters, faster than the housing channel and slower than the equity-wealth channel. The Q1 2026 average new-car loan rate held near 7.3 percent against a federal funds upper bound of 4.50 percent, leaving a roughly 280 basis points spread that is wide by post-2010 standards. The auto-ABS spread to Treasuries has held above 110 basis points through 2025-2026, reflecting subprime-default trends that lenders are pricing into the marginal-borrower curve.

SPY responds to the same rate path with a different transmission: long-duration discount-rate compression on the multiple, plus earnings-impact through cyclical sectors. The asymmetry is what makes the pair informative. When SPY rerates higher because the market prices rate cuts, but TOTALSA fails to respond because the auto-ABS spread keeps financing costs sticky, the equity rally is leaning on a discount-rate move that the household sector is not actually receiving. That setup has historically resolved through SPY catching down to TOTALSA rather than the reverse. The 2007 cycle is the canonical example: subprime auto-ABS spreads widened by 200 basis points between Q1 2007 and Q3 2007 even as SPY made new highs into October 2007, and the resolution came through SPY rather than through ABS spread compression. The 2025-2026 setup carries similar credit-spread dynamics with markedly tighter overall spreads.

The Convex Net Liquidity Impulse decomposes the macro liquidity backdrop into outright Fed assets, the Treasury General Account, and reverse-repo balances. For the TOTALSA-SPY pair, CNLI captures the marginal driver of SPY but not of TOTALSA: the equity leg is liquidity-sensitive on a 30-to-60-day horizon, while the auto leg is liquidity-sensitive on a two-to-four-quarter lag through the credit channel. That mismatch is why SPY can rerate ahead of any improvement in vehicle sales, and why a CNLI reading alongside the pair is the sharper diagnostic than either series alone. The April 2026 CNLI reading is roughly flat following the FOMC's December 17-18, 2024 meeting that announced the end of quantitative tightening, finalized in December 2025.

The Convex Recession Probability Index uses TOTALSA as one of its consumer-cycle inputs, weighted alongside the Sahm rule unemployment trigger, AWHAETP weekly hours, building permits, and the 10y-2y yield curve inversion signal. In April 2026, TOTALSA contributes a moderate-not-acute reading to CVRP because the SAAR plateau is not yet a SAAR break, while the labor-hours input is the more pressing component. Reading the SPY-versus-TOTALSA spread alongside the CVRP composite gives the user the signal level (probability) and the marginal contributor (which input drives the move) in one workflow. The composite framework also distinguishes between cyclical SAAR weakness and structural shifts (EV-transition demand-substitution, work-from-home commuting reduction, tariff pull-forward distortions) that may be inflating the noise in the headline reading.

The actionable read is that the pair is in a middle-information configuration: TOTALSA has plateaued well below the 2019 peak, SPY is up 5.16 percent year-to-date with a record-high April 27, 2026 print, and the divergence is among the largest in the FRED record without a confirming labor-market break. The historical analog is the 1998-1999 setup, where SAAR plateaued near 16 million for six quarters while SPY rerated 50 percent before the relationship reasserted between March 24, 2000 and October 9, 2002. The base-rate odds favor mean reversion through SPY catching down rather than TOTALSA catching up, but the timing has historically required a labor-market trigger.

The horizon for the signal is two to four quarters. The two specific watches are a SAAR break below 15.0 million, which would confirm the credit-channel transmission has tipped over, and an AWHAETP print below 34.0 hours, which would historically convert a TOTALSA plateau into a TOTALSA decline. Either trigger reasserts the textbook reading; absent either, the divergence can persist as it did through 1998-1999. The May 2 employment situation report and the May 28 TOTALSA monthly release are the next two data points on the calendar. Sizing decisions should treat the pair as a regime probability input, not a near-term timing signal: position sizing benefits from the fact that the lead time from threshold break to confirmed equity drawdown has historically run three to nine months, providing a meaningful action window for hedge construction or beta-reduction.