Credit Spreads Explained: The Stress Gauge That Often Moves First

On October 19, 2026, US high-yield corporate bonds paid 3.29 percentage points a year more than Treasuries, and investment-grade bonds 0.87 points more. Those two ICE BofA spreads are the bond market's running price for corporate default risk plus the premium for carrying it. Both are tight by historical standards: the high-yield spread was narrower than on four of every five trading days since the series began in late 1996, the investment-grade spread narrower than on 87% of them.

Spreads have a reputation as the gauge that moves before stocks. The record is mixed: they led in 2007 and 2015, moved in step with equities in 2020, understated the 2022 bear market, and since 2020 have carried the imprint of central-bank backstops. There is a second complication. CrisisBoard does not read spread indexes; it reads the prices of credit ETFs such as HYG and LQD, which measure something related but different.

What a credit spread measures

A corporate bond's yield has two layers: the yield on a Treasury of similar maturity, and an extra margin for the risk that the company does not pay in full and on time. The spread is that margin. Quoting it separately removes the level of interest rates, so 2008, when Treasury yields fell, can be compared with 2022, when they rose fast. One percentage point is 100 basis points.

Many corporate bonds are callable, which inflates the raw yield gap. An option-adjusted spread (OAS) uses a rate model to strip that call option out. ICE BofA computes an OAS for each bond against a spot Treasury curve and weights the results by market value. Two of its series carry most US credit analysis:

• US Corporate Index OAS (FRED code BAMLC0A0CM): investment grade, rated BBB or better. It usually moves in tenths of a point.
• US High Yield Index OAS (FRED code BAMLH0A0HYM2): rated BB or below. In crises it moves in whole points, because a recession changes the default arithmetic for leveraged issuers far more than for investment-grade ones.

Since April 2026, FRED carries only the latest three years of these ICE series; the full history is licensed by ICE Data Indices. The long-run figures below come from earlier full-history downloads of the same FRED series.

What moves spreads

A spread pays for four things that rarely move in lockstep.

1. Expected default losses. Roughly the default probability times the share not recovered. As an illustration, if 4% of high-yield issuers default in a year and holders recover 40 cents on the dollar, losses run at 2.4 points a year and a 3% spread barely covers them. That is why high yield carries most of the recession signal.
2. Recovery. Recoveries tend to be weakest in default waves, when many distressed companies are restructured at once, so loss expectations can rise faster than default forecasts.
3. Liquidity. Corporate bonds trade over the counter through dealers. When dealers stop absorbing inventory, spreads widen for reasons unrelated to any issuer. The Fed's May 2020 Financial Stability Report described this in March 2020: corporate bid-ask spreads widened considerably while bond funds and ETFs faced large outflows.
4. Risk appetite. Simon Gilchrist and Egon Zakrajšek split spreads into a part explained by expected default and an excess bond premium that tracks investor sentiment; Fed economists have used that premium to estimate twelve-month recession odds. In May 2020 the Fed reported the premium had risen well above its historical median.

A spread widening on default forecasts says something about the economy. One widening because dealers are full says something about market plumbing, and can reverse within days once a central bank steps in.

Three decades of US spreads in one table

Three patterns stand out. Peaks are not proportional to economic damage: 2011 and 2016 pushed high yield to about 9% without a US recession, while the 2020 recession peaked at 10.87%, barely higher. Investment-grade spreads move less in points but often more in proportion; in just over a month in 2020 they nearly quadrupled while high yield tripled. And the recent lows are extreme: 2.59% in high yield was the tightest since June 2007, and 0.73% in investment grade the tightest since mid-1998.

When spreads led, lagged, or cried wolf

2007: the textbook lead

The high-yield spread bottomed at 2.41% on June 1, 2007, and reached 4.59% by August 16. When the S&P 500 set its record close on October 9, it had eased to 3.95%, still more than 1.5 points above its low. The National Bureau of Economic Research dates the recession from December 2007. A spread watcher had about four months of warning before the equity peak and six before the recession.

2014–16: a lead without a recession

The spread made a cycle low of 3.35% on June 23, 2014, then climbed as energy bonds repriced for the collapse in oil. It was 4.51% at the S&P 500's record on May 21, 2015, 6.14% in the August 24, 2015 selloff, and 8.87% on February 11, 2016, the day of that selloff's closing low in stocks. Spreads led by roughly a year, yet the NBER records no recession between June 2009 and February 2020. The 2011 peak of 9.10% has the same profile: genuine credit events for particular sectors and regions, false alarms as forecasts of a US downturn.

2020: no lead at all

On February 20, 2020, the day after the S&P 500's record close, high yield stood at 3.62%, barely above its January low of 3.38%. Spreads and stocks broke together, and the spread peaked at 10.87% on March 23, the day of the S&P 500's closing low. A pandemic gave credit nothing to anticipate.

2022: understating an equity bear market

The S&P 500 fell about 25% from its January 2022 peak to its October low. High yield rose from about 3.1% at the end of 2021 to 5.99% in July, 1.5 points above its long-run median, and investment grade never passed 1.71%. The selloff was about inflation, not defaults: the 10-year Treasury yield went from 1.63% on January 3 to 4.25% on October 24. Spreads correctly said companies could keep paying, and little about what rates were doing to asset prices.

When the central bank buys the gauge

On March 23, 2020, the Federal Reserve established two corporate credit facilities: one to lend to investment-grade companies, and the Secondary Market Corporate Credit Facility (SMCCF) to buy their existing bonds and US-listed ETFs tracking investment-grade bonds. Revised terms on April 9 set a combined size of up to $750 billion, admitted "fallen angels" rated at least BBB-/Baa3 on March 22, 2020 and no lower than BB-/Ba3 when bought, and let a minority of its ETF holdings be high-yield funds.

High yield peaked on the day of the first announcement and was 7.96% by April 9. The SMCCF bought its first ETF on May 12, with the spread already near 7.4%, and its first bond on June 16. When purchases stopped at the end of 2020 it held about $14.2 billion, under 2% of the combined capacity. More than three points of tightening came before it bought anything. New York Fed researchers found that the announcements improved primary-market conditions and that purchases also mattered, direct bond buying more than ETF buying.

That leaves a lasting interpretation problem. Investors have seen the Fed act as buyer of last resort for investment-grade credit and can price the chance it does so again, so a narrow spread can mean low default risk or confidence in a backstop. The ECB has been there longer: it added investment-grade euro corporate bonds to its purchases in March 2016, and on March 18, 2020 launched a €750 billion pandemic programme covering every asset class its existing purchases allowed. CrisisBoard tracks the policy side on its central banks page.

Europe and emerging markets

European high yield now trades at US levels. ICE BofA's euro high-yield OAS was 3.31% on October 19, 2026, against 3.29% in the US, and peaked at 4.29% on April 9, 2025, below the US tariff-shock peak of 4.61%. Since ECB corporate purchases target investment-grade issuers, euro high yield is the less managed of the two European gauges.

Emerging-market dollar bonds add sovereign risk and the dollar itself. CrisisBoard reads EM credit through EMB, a fund of dollar-denominated sovereign and quasi-sovereign bonds whose price blends Treasury yields, sovereign spreads and fund flows. In the five sessions to March 18, 2020, EMB fell 16% and the high-yield EM fund EMHY 19%, against 9.7% for HYG. The currency side is covered in our guide to emerging-market currency crises.

Spread indexes versus ETF prices

CrisisBoard's rules run on market prices, so its credit inputs are funds: HYG and SRLN in the US, London-listed euro high-yield and investment-grade funds for Europe, LQD, EMB and others elsewhere. A fund price differs from an OAS index in three ways.

• Duration. Prices respond to Treasury yields as well as spreads. On October 19, 2026, iShares reported effective durations of 7.54 years for LQD and 3.21 for HYG: a one-point rise in yields costs LQD roughly 7.5%, a one-point spread widening costs HYG roughly 3%. In a flight to safety, falling Treasury yields offset part of the credit move.
• Income. HYG and LQD pay monthly, and CrisisBoard reads unadjusted closes, so ex-dividend dates show up as small price drops, far too small to trip a rule alone.
• Premiums and discounts. The ETF trades continuously on an exchange; its bonds trade over the counter and are often valued from dealer quotes. Under stress the two separate.

The duration arithmetic sets a high bar. With HYG's duration near 3, a 5% fall in five sessions needs roughly 1.5 points of spread widening in a week with Treasuries flat, and an 8% fall about 2.5. In April 2025 the high-yield OAS jumped 119 basis points in three sessions, from 3.42% to 4.61%; HYG's worst five-day decline was 3.7%.

March 2020 shows the discount problem. BIS research found that bond ETFs fell to steep discounts to net asset value that turned into large premiums within days, and noted that NAVs can rest on stale quotes. Between the March 20 and March 23 closes, the 10-year Treasury yield fell 16 basis points and the investment-grade OAS widened 14, leaving the underlying yield roughly unchanged. Yet LQD rose 7.4% on March 23, the day the Fed said it would buy investment-grade ETFs. The arithmetic suggests most of that gain was the fund's discount closing, not a recovery in credit.

How CrisisBoard's rules use credit

The methodology explains how signals add up. The credit rules are price-change triggers on ETFs, scored in points toward a regional total. The US region reaches MONITOR at 8 points, PHASE 1 at 10 and PHASE 2 at 12; Japan, the UK and the EM basket use 5, 7 and 9.

Two quirks matter. The US region has no investment-grade fund, so the HY-versus-IG comparison, the nearest thing on the board to a true spread, runs only in Europe. And Japan's credit slot uses LQD, a US fund, so it is effectively a second reading of US corporate credit.

Put plainly, CrisisBoard's credit rules confirm acute dislocations; they are not built to call the turn in a slow credit cycle. The warning spreads gave in 2007 and 2015 came from levels creeping up over months, which no five-day rule sees, so watch the OAS series alongside the board. For how volatility interacts with credit, see our VIX guide; for an episode where credit stayed comparatively calm while bank stocks broke, see the SVB guide.