Home › Guides › Corporate Bond Yields by Credit Rating, Explained
What companies pay to borrow - and how much more the risky ones pay than the safe ones.
When a company borrows money by selling bonds, the interest rate it pays depends on how likely investors think it is to pay the money back. Rating agencies (S&P, Moody's, Fitch) grade that likelihood on a letter scale: AAA is the safest, then AA, A, BBB, and on down through BB, B, and CCC for the riskiest borrowers.
What companies pay to borrow - and how much more the risky ones pay than the safe ones.
When a company borrows money by selling bonds, the interest rate it pays depends on how likely investors think it is to pay the money back. Rating agencies (S&P, Moody's, Fitch) grade that likelihood on a letter scale: AAA is the safest, then AA, A, BBB, and on down through BB, B, and CCC for the riskiest borrowers.
This page shows the average yield for each grade, updated every business day. Two things jump out. First, the ladder: at any moment, each step down in rating costs more - a CCC-rated company might pay two or three times the rate of an AAA one. Second, the ladder stretches and compresses: in calm markets the gap between safe and risky shrinks, and in scary markets (2008, March 2020) it blows wide open as investors flee risk.
The big dividing line is between BBB and BB. Everything BBB and above is called "investment grade" - safe enough that pension funds and insurers can hold it freely. BB and below is "high yield," or less politely, "junk." Crossing that line - getting downgraded from BBB to BB - is a big, expensive deal for a company.
ICE BofA index effective yields by bucket; OAS isolates the credit premium from the rates component.
Each line is an index: the market-weighted effective yield of all the US corporate bonds in one rating bucket (ICE BofA index family, redistributed daily by FRED - which carries roughly a trailing three-year window of these series). A corporate yield has two components - the underlying Treasury rate for that maturity, plus a credit spread for default and liquidity risk. The yield view shows the sum; the OAS (option-adjusted spread) view shows just the spread, in percentage points over maturity-matched Treasuries, with any embedded call options priced out - hence 'option-adjusted.'
OAS is the cleaner risk signal. Yields can rise simply because the Fed hiked; OAS rises only when the market demands more compensation for credit risk. Typical ranges: AAA OAS runs a few tenths of a percent; BBB nearer 1-2pp; CCC anywhere from ~4pp in booms to 20pp+ in panics. Watching BBB (the largest bucket, and the cliff-edge of investment grade) and CCC (the distress bellwether) covers most of the information.
Moody's seasoned Aaa and Baa series carry the long yield history back to 1983 on this page (and to 1919 monthly on FRED). They're long-maturity averages, roughly comparable to AAA and BBB, and they're the series used in decades of academic work - the 'Baa-Aaa spread' is a classic recession indicator.
Index construction caveats, the BBB cliff and fallen angels, spreads as a macro signal vs default compensation.
Construction caveats worth knowing before leaning on these series: bucket composition drifts (the BBB share of the investment-grade index roughly doubled after 2008 as companies levered up to the cheapest rating that keeps IG status), duration differs across buckets (high-yield indices are shorter), and effective yield is a portfolio average - it conflates maturity mix with credit quality. Comparing a bucket's yield across decades partly compares different bond populations.
The BBB/BB boundary is an institutional cliff, not just a label: many mandates cap or forbid sub-IG holdings, so a downgrade to BB (a 'fallen angel') triggers forced selling. That structural break is why the BB-BBB yield gap persists beyond what default statistics alone justify, and why fallen-angel bonds have historically been a systematically cheap corner of the market.
As a macro input, credit spreads are among the best market-based recession signals - arguably better than the yield curve at short horizons, because they price realized stress rather than expected policy. The Gilchrist-Zakrajšek 'excess bond premium' literature formalizes this: the component of spreads NOT explained by measured default risk is what forecasts activity. Rule of thumb for reading this chart: OAS widening led or coincided with every downturn since 1996, but it also false-alarms on pure market events (2011, 2015-16 energy) - condition it on the funding-stress and labor indicators elsewhere on this site.
Spread decomposition, the CDS-cash basis, and what index OAS can and cannot support analytically.
Decomposition: raw OAS ≈ expected default loss + risk premium + liquidity premium + technical/structural terms. Expected-loss compensation is a minority of IG spreads (for AAA/AA, historically well under a quarter of OAS - default rates at those grades are near zero), so IG OAS is mostly risk and liquidity premia; at CCC the expected-loss share dominates. This is why IG OAS is a sentiment gauge while CCC OAS is closer to a default forecast.
The CDS-cash basis (CDS spread minus cash-bond spread) isolates funding and liquidity effects: it went deeply negative in 2008 when balance-sheet-constrained arbitrageurs couldn't fund the cash-bond leg - a clean demonstration that cash-bond spreads embed a funding premium beyond credit risk. Related: the box-spread section on this site's Stock Market page measures the analogous premium in the options market.
Analytical limits of what's shown here: index-level OAS is fine for regime work, spread-momentum, and macro forecasting, but it cannot support issuer-level work (composition drift), rating-migration studies (buckets re-populate on migration - survivorship in place), or precise curve-relative trades (the Treasury-matching inside OAS is the index provider's model). For those you need constituent-level data (ICE, Bloomberg) or TRACE prints. Insurance-relevant note: NAIC designations map onto these buckets (NAIC 1-2 ≈ IG, 3-6 ≈ HY) with RBC charges that jump at the same BBB/BB cliff, so the fallen-angel dynamic has a direct statutory-capital analog.