AAA Corporate Yield vs 10Y Treasury

FRED's AAA10Y daily series runs from January 1983 onward, giving a 43-year window across four full credit cycles (1989-1991, 2000-2002, 2008-2009, 2020). The series median is 102 basis points, the mean is 109 basis points, and 60 percent of daily readings sit between 70 and 130 bps. AAA is Moody's highest corporate credit rating, awarded to issuers with the lowest probability of default in their statistical models, so the spread captures default premium and liquidity premium more cleanly than BAA-10Y does. The AAA universe has shrunk dramatically: in 1983 there were over 60 US non-financial AAA issuers; by April 2026 only Microsoft and Johnson & Johnson hold AAA ratings from Moody's (S&P AAA list is similarly thin: Microsoft, Johnson & Johnson). The shrinking universe is itself a structural input to the spread, because the seasoned BAA index captures a broader base of corporate borrowers while seasoned AAA increasingly approximates the spread for the highest-quality global credit. The Fed's Treasury Borrowing Advisory Committee (TBAC) and the BIS Quarterly Review both reference AAA-10Y as the cleanest US investment-grade default-risk indicator, and the April 2026 reading of 95 bps places it in the 35th percentile, in the lower-middle of the distribution. The percentile classification matters more than the absolute level because of the structural shifts that have repriced the post-2014 era versus the pre-2008 baseline, and reading the spread without normalising to its post-2014 distribution will systematically misclassify the credit regime.

The December 2008 peak at 372 basis points is the modern series high. The October 2002 peak hit 211 basis points during the corporate-fraud cycle. The March 23, 2020 COVID peak reached 277 basis points before the Fed's PMCCF and SMCCF announcement compressed the spread by approximately 100 bps within a week. Each episode produced an asymmetric move: the AAA leg widened by 80-120 bps while the 10-year fell by 100-150 bps on flight-to-quality, producing the upper-tail spread readings. The structural break in the series occurred between 2008 and 2014: pre-2008 average AAA-10Y was 88 bps, post-2014 average is 116 bps. The 28 bp lift reflects three structural shifts: (1) the shrinking AAA universe pushed the seasoned AAA index toward longer-duration issuers, raising the average yield; (2) post-Dodd-Frank dealer balance-sheet constraints widened the bid-ask premium structurally; (3) the post-QE term-premium compression suppressed 10-year yields by an estimated 50-80 bps, mechanically lifting the spread. Reading post-2014 spreads against pre-2008 baselines therefore overstates the credit-risk component. The 1989-1991 S&L crisis pushed AAA-10Y to 184 bps in October 1990, the highest pre-2002 reading, and the 1998 LTCM episode produced a brief 145 bp spike before the Fed's October 15, 1998 emergency 25 bp cut compressed the spread within four trading days. Each historical peak is dated to within days of the central-bank policy response that compressed it, which is a feature of the AAA leg's shorter reaction time versus longer-rated credit.

The seasoned Moody's AAA index has an effective duration of approximately 11.5 years versus the 10-year Treasury duration of approximately 8.5 years, a 3-year duration gap. When real yields move, the AAA leg responds with a larger price change than the 10-year. This is why AAA-10Y carries a structural bias to widen during periods of falling real yields (the 10-year yield has more headline rate sensitivity, the AAA leg has more total-return sensitivity through duration). The liquidity premium component is particularly pronounced in AAA versus BAA, because the seasoned AAA universe is so concentrated that the bid-ask premium can widen sharply on single-issuer news. The Berkshire Hathaway downgrade rumour in April 2009, AT&T's loss of AAA in 1994, and General Electric's loss of AAA in 2009 each produced 10-25 bp spread spikes on idiosyncratic news. Policy distortion is the cleanest mechanism for the post-QE era: Fed asset purchases concentrated demand at Treasuries and agency MBS, pulling the 10-year leg down faster than AAA could follow, mechanically widening AAA-10Y. The 2022-2024 QT cycle reversed this, lifting the 10-year by an estimated 60 bps relative to AAA and compressing the spread. The duration mismatch also produces the asymmetric reaction to Fed easing announcements: the AAA leg responds faster to policy easing than to policy tightening, because high-quality issuers retain primary-market access continuously and the duration component prices in immediate forward rate expectations, whereas the BAA leg is more sensitive to credit-cycle conditions that take longer to reset.

AAA-10Y at 95 basis points (April 2026) sits in the 35th percentile of the 1983-2026 daily distribution. The 25th percentile is 80 bps, the 75th percentile is 142 bps, and the 90th percentile is 178 bps. The current configuration is one of the lowest readings of the post-2014 era and is consistent with the broader bottom-third credit-spread regime that BAA-10Y, ICE BofA IG OAS (52 bps in February 2026), and CDX IG (50 bps in April 2026) all read. The 10-year yield component closed at 4.31 percent on April 24, 2026, and the AAA component at approximately 5.26 percent. The Convex Net Liquidity Impulse (CNLI) reads AAA-10Y as one of three pure-IG credit-stress inputs (alongside ICE IG OAS and CDX IG), and the live CNLI cyclical-credit reading is consistent with the bottom-third spread classification. The AAA-BAA differential at approximately 77 bps is in the 25th percentile of the post-1986 distribution, confirming that the compressed regime reflects broad-based investment-grade tightness rather than just AAA-specific dynamics. The configuration most closely resembles the 2017-2019 setup, when AAA-10Y sat in the 25-40th percentile range for an extended period before the 2020 COVID stress produced a sharp but short-lived widening, and the 1995-1998 setup, when AAA-10Y sat below 95 bps for an extended pre-LTCM expansion phase. Both analogues are consistent with continuation of the late-cycle tight-credit regime barring a defined catalyst.

AAA-10Y and BAA-10Y move together but with different cyclical behaviour. In stable expansions, the two spreads move within tight bands of each other (correlation above 0.85 over rolling 6-month windows). In stress, BAA widens more sharply than AAA (the 2008 episode saw BAA-10Y reach 594 bps versus AAA-10Y at 372 bps, a 222 bp differential at the peak). In recoveries, AAA compresses faster than BAA because the highest-quality borrowers regain primary-market access first. The differential between the two spreads (BAA-10Y minus AAA-10Y, also called the rating-quality differential) is the most actionable composite read: it widens to 200+ bps in major credit events and compresses to under 70 bps during late-cycle complacency phases. The April 2026 differential of approximately 77 bps sits in the bottom quartile of the post-1986 sample, in line with the late-1990s, mid-2000s and 2017-2019 periods that all preceded eventual credit-cycle inflection points. The differential's role inside CNLI is as a late-cycle warning: when it sits below 80 bps for sustained periods, the historical base rate is a credit-spread widening cycle within 18-24 months. The cleanest historical false-positive is the 1995-1998 episode, when the differential sat below 80 bps for almost three years before the LTCM-era widening, demonstrating that the warning has a long lead time and requires patience to act on. The differential's data cadence is daily for both legs from FRED, which makes it operationally easier to track than the credit-default-swap-derived alternatives that some desks substitute for it.

AAA-10Y at 95 bps and the AAA-BAA differential at 77 bps both sit in the bottom quartile of their distributions. The actionable read: investment-grade credit is priced for a benign cyclical environment, with little buffer for default-rate surprises or for a Fed reaction-function shift. The historical base rate for forward 12-month investment-grade total returns conditional on AAA-10Y in the bottom quartile is 5.4 percent, against 7.1 percent for the second quartile, 8.6 percent for the third, and 11.4 percent for the top quartile (1983-2025 sample, IG total return as proxied by the Bloomberg US Corporate Total Return Index). The forward equity-return implication is similar: bottom-quartile AAA-10Y has historically been associated with subsequent S&P 500 12-month returns of 6.8 percent versus 14.7 percent for top-quartile AAA-10Y. The horizon for the signal is 6-18 months. The watches that change the read are AAA-10Y sustained above 142 bps (the 75th percentile, indicating a transition into the upper half of the distribution) and the BAA-AAA differential widening above 120 bps (indicating credit-quality differentiation accelerating, a late-cycle warning that has historically led the broader credit-spread cycle by 3-6 months). Convex publishes the AAA-10Y percentile rank daily alongside the BAA-AAA differential, the ICE IG OAS cross-check and the CNLI cyclical-credit composite, providing the multi-input frame that any single-spread reading needs to be interpreted within. The April 2026 cross-check confirms the bottom-quartile classification holds across all four indicators, which is the configuration with the lowest historical forward-return distribution in the post-1986 sample.