BAA vs AAA Corporate Spread

Moody's Investors Service publishes the Seasoned Baa and Seasoned Aaa indices daily on FRED as DBAA and DAAA respectively, with the monthly averages also available as BAA and AAA. The 'seasoned' qualifier means the indices include only bonds with at least 20 years to maturity that have been outstanding for at least one month, which removes new-issue concessions and short-end rate noise. Both legs have approximately matched durations of 13-15 years and both are constructed from US-domiciled corporate issuers. Subtracting one from the other strips the risk-free rate and leaves the pure quality-tier credit premium. The same risk-free-rate cancellation is the reason the BAA-AAA spread is preferred over the BAA-10Y or AAA-10Y series for measuring quality-tier stress, since the latter two embed Treasury yield moves that have nothing to do with credit quality.

This spread has continuous monthly data back to January 1919, longer than any other commonly cited US credit-spread series. The Bloomberg US Investment Grade OAS (LUACOAS) starts in 1989. The ICE BofA US Corporate Index OAS starts in 1996. Only the Moody's-based BAA-AAA spread covers the 1929 stress, the 1932 trough, the 1973-1975 stagflation, the 1979-1982 Volcker tightening, the 2008 GFC, and the 2020 COVID episode in a single internally consistent series. That continuity is the principal reason the BAA-AAA spread remains the academic-grade benchmark for IG credit-quality stress despite the wider availability of OAS-based alternatives. NBER recession-dating committees have referenced the spread in retrospective dating papers since the 1960s, and the Federal Reserve Board's Financial Stability Report cites it as one of four core credit-stress indicators.

The series peaked at 553 basis points in May 1932 during the Great Depression default cycle, when Baa-rated corporate yields averaged 11.0 percent against Aaa yields of 5.5 percent. The June 1934 reading of 437 bps marked the second-highest print and reflected the residual stress as the New Deal recovery began. The next major peak was 280 bps in December 2008 during the GFC, against an Aaa yield of 5.05 percent and a Baa yield of 7.85 percent. The COVID episode produced a 350 basis point spike between Baa and Aaa over a six-week window in March-April 2020, with the BAA-AAA monthly average peaking at 165 bps in May 2020 before the Fed's Primary and Secondary Market Corporate Credit Facilities (PMCCF and SMCCF, announced March 23 and April 9 respectively) compressed the spread back to 110 bps by August 2020. The 1981 stagflation peak of 235 bps in August 1981, during the Volcker tightening cycle, marks the only modern episode where the spread broke 200 bps without a recession-coincident default cycle.

Beyond the named crises, four other episodes broke 110 bps in the post-1990 record: the 2002 telecom-debt crisis (peak 162 bps in October 2002, driven by WorldCom and Tyco defaults), the 2011 European sovereign stress (peak 122 bps in October 2011), the 2016 energy-sector default cycle (peak 138 bps in February 2016, driven by oil-and-gas high-yield migration into stressed IG), and the 2022 rate shock (peak 107 bps in October 2022). The April 2026 reading of approximately 75 bps sits well below the post-1990 stress threshold of 110 bps and roughly 21 bps below the 30-year average. The all-time low of 32 bps in February 1966 occurred during the late-1960s credit boom that preceded the 1969-1970 mild recession, and the post-1990 low of 47 bps was reached in May 2007, two months before the August 2007 onset of the GFC stress phase.

The structural baseline reflects the embedded difference in default-adjusted recovery between the two rating tiers. Moody's historical default study places one-year Baa default rates at 0.18 percent against Aaa default rates of 0.00 percent, with five-year cumulative rates of 1.79 percent versus 0.10 percent. The implied default-adjusted yield differential, holding loss-given-default at 60 percent, accounts for approximately 35-50 bps of the average spread. The remaining 50-70 bps reflects liquidity differentials (the Aaa universe is now extremely small, with fewer than 10 US issuers carrying the Aaa rating in 2026, against more than 200 Baa-rated issuers), pension-and-insurer demand for the highest tier, and ratings-cliff insurance value (the cost of avoiding the IG-to-HY transition). The 2003 Federal Reserve Board working paper on credit-spread decomposition estimated this liquidity-and-segmentation premium at approximately 60 bps in the 1990-2002 sample, which has held remarkably stable in subsequent updates.

The cyclical driver is the change in expected default rates rather than the level. The spread widens on rising default expectations and tightens on falling default expectations, with a one-quarter lag to the Moody's expected default frequency for the Baa universe. The 2002, 2008, 2016, 2020, and 2022 stress episodes all featured EDF spikes that preceded the BAA-AAA spread widening by six to twelve weeks, and disciplined credit overlays use the EDF lead to position before the spread move rather than after. The cleanest test of this lead-lag relationship was the late-2007 episode, where the Baa-issuer EDF rose from 0.4 percent in June 2007 to 1.1 percent in November 2007, ten weeks before the BAA-AAA spread broke through 110 bps in late January 2008.

The Aaa-rated US corporate universe has shrunk dramatically. In 1980, 65 US corporate issuers carried the Moody's Aaa rating. By 2008, the count had fallen to 6 (Berkshire Hathaway, ExxonMobil, Microsoft, Johnson & Johnson, Pfizer, and General Electric, with GE then losing the rating in March 2009). By 2026, only Microsoft and Johnson & Johnson retain the Aaa, and Microsoft has carried it as the dominant Aaa issuer since ExxonMobil's downgrade in March 2016. Pfizer lost its Aaa in 2009 amid the Wyeth acquisition, and Berkshire Hathaway's Aaa was downgraded by Moody's to Aa1 in April 2010 before being restored. The Aaa index that Moody's publishes has therefore been progressively diluted with seasoned issues from previously-Aaa firms whose ratings have since fallen.

The practical implication is that the BAA-AAA spread in 2026 captures less idiosyncratic Aaa-firm risk and more pure-rating-tier premium than the same spread in 1980 did. The structural break is most visible in the post-2008 period: the spread averaged 91 bps from 1990-2007, but only 96 bps from 2008-2026, a structurally similar level despite the dramatically different composition. The implication is that the historical thresholds (110 bps as warning, 150 bps as recession-confirming, 200 bps as crisis) remain useful even with the composition shift, because the seasoned-issuer methodology has preserved the intended factor exposure. The S&P-rated AAA universe is similarly small (only Microsoft and Johnson & Johnson among non-financials), which is why the two ratings agencies' AAA universes track each other closely despite their different methodologies.

The Convex Risk Appetite Index (CRAI) treats the BAA-AAA spread as one of three core credit-tier inputs alongside the HY-IG spread and the IG OAS percentile. The current 75 bps reading places the spread in the 18th percentile of post-1990 readings, the second-tightest decile and historically associated with periods of credit complacency. The complacency band extends through the spread tightening to 65 bps or below; the warning band begins at 95 bps; the stress band begins at 110 bps; the recession-confirming band begins at 150 bps. CRAI uses the 30-day rate of change rather than the absolute level as its primary positioning input, because the level alone has too long a half-life to drive tactical allocation.

The 18th-percentile reading is associated with three forward configurations on the historical record. First, continued tight spreads through the next twelve months, last observed in 2017-2018 (spread averaged 78 bps for sixteen months, until the December 2018 widening to 105 bps). Second, a benign re-widening to the 90-100 bps band as default expectations normalize, last observed in 2003-2004 (spread re-widened from 75 to 95 bps over fourteen months without recession). Third, a sharp widening on a discrete shock, last observed in late 2007 (spread widened from 80 bps in June 2007 to 280 bps in December 2008, with Bear Stearns in March 2008 and Lehman in September 2008 marking the inflections). CRAI does not currently flag the third scenario but does flag the spread as positioned for asymmetric upside on any credit shock, with the implied probability-weighted distribution skewed toward widening rather than further compression.

The pair is best used as a regime gate for IG credit allocation rather than as a directional signal. Three specific configurations have produced consistent outperformance in the post-1990 record. First, the spread above 150 bps with falling EDFs has produced a 12-month forward IG total return of +14.2 percent on average (n=4 episodes: 2009 H1, 2011 Q4, 2016 Q1, 2020 Q3), against an unconditional average of +6.8 percent. Second, the spread below 70 bps with rising EDFs has produced a 12-month forward IG total return of +1.1 percent on average (n=6 episodes), against the same unconditional baseline. Third, the spread changing direction by more than 30 bps over a four-week window, in either direction, has consistently led the broader IG OAS by two to five weeks. The lead-lag relationship reflects the slower update cadence of the seasoned-issuer methodology relative to the faster-marking OAS-based indices.

The spread does not tell you anything useful about HY credit, equity-market direction, or recession timing in isolation. It is specifically a quality-tier indicator within IG. Reading it as a recession indicator is a common error: the spread widened to 138 bps in February 2016 without an NBER recession, and it failed to widen above 110 bps before the 2001 NBER recession start. Reading it alongside the HY-IG spread, the unemployment claims series, and the yield curve inversion provides cross-confirmation that the BAA-AAA spread alone cannot. The Federal Reserve Bank of Cleveland publishes a composite credit-stress indicator that combines the BAA-AAA spread with four other inputs precisely because the single-spread signal has historically generated too many false positives and false negatives to drive a defensible cycle call.