Understand the reading

Yield curves and market fragility

A yield tells you about a borrowing rate. A spread compares two yields. Follow that distinction from a simple example to historical Treasury rates and ForeGlass’s market-conditions index.

01 / The distinction

Read the level, then the difference

4.00%

A yield level

A bond yield is an annualized rate associated with its price and promised payments. It is not the same as the bond’s coupon or a guaranteed investment return.

10 years − 2 years

A term spread

A yield curve compares yields across maturities on one date. The 10-year minus 2-year spread summarizes two points on that curve; it does not describe every maturity.

1.00 pp = 100 bp

The units of a difference

If one yield is 4% and another is 3%, the gap is one percentage point, or 100 basis points. It is not a 1% relative increase.

A long-run chart of several rates shows their movement through time; a yield curve instead has maturity on its horizontal axis.

ForeGlass uses T10Y2Y: the 10-year Treasury yield minus the 2-year Treasury yield. A negative value means the 10-year yield is below the 2-year yield: an inversion between those two maturities.

02 / A small experiment

Change the yields. Watch the spread.

Synthetic teaching example · not current market data

Move either yield or choose a preset. “Higher rates, same spread” lifts both yields by the same amount: the borrowing-rate levels change while the difference stays the same.

Illustrative annualized Treasury yields

2-year yield 3.0%

10-year yield 4.0%

Use the arrow keys to change a focused slider. These controls change only this example; they do not recalculate the ForeGlass index.

The default example is shown below: 3% at two years and 4% at ten years. Interactive controls require JavaScript; all explanations and the historical table remain readable.

Two selected maturities, joined for illustration. This is not a fitted full yield curve.

Upward between these maturities

An upward slope is not an inversion signal. It does not prove that borrowers are resilient: debt burdens, refinancing needs and remembered earlier inversions can still matter.

4.0% − 3.0% = +1.00 pp
The 10-year yield is 100 basis points above the 2-year yield.

03 / The historical comparison

1997 Q1–2026 Q2· Fixed historical snapshot

The same quarters, three different quantities

First compare the two Treasury yield levels. Next look at their difference. Then compare that spread with the combined ForeGlass market-conditions index. Each panel keeps its own unit and scale.

How to read this history: Treasury points are quarterly averages over paired available daily observations. The index is a historical model replay, not a record of forecasts published in those quarters. The comparison can illustrate relationships; it does not establish causation or prove advance prediction.

Historical context: grey areas mark US recession peak-to-trough months; violet areas and dashed markers identify selected international, banking and policy events. Names appear in lanes above the plots. These annotations do not enter the index or establish what caused its movement.

Read the crisis dates, definitions and sources

Selected context, not an exhaustive crisis catalogue. Monthly recession shading includes the peak month through the trough month; NBER’s recession-duration count begins in the month after the peak. Boundaries are month precision, not official day-level dates. Recessions were dated retrospectively. A US interval does not classify every country as being in recession.

Label	Dates	Type	Interpretation / source
Asian financial crisis	Jul 1997–Dec 1998	Selected stress window	Selected international crisis; not a US recession band. Source
Dot-com / 2001 downturn	Mar–Nov 2001	US recession context	US peak-to-trough months. The technology-stock decline began before this interval. Source
Global financial crisis	Dec 2007–Jun 2009	US recession context	US peak-to-trough months; the wider financial crisis has different boundaries. Source
COVID-19 downturn	Feb–Apr 2020	US recession context	US peak-to-trough months, dated retrospectively. Source
Ukraine invasion	24 Feb 2022	Dated event	Start of Russia’s full-scale invasion; a dated event, not a crisis-duration estimate. Source
SVB failure	10 Mar 2023	Dated event	A bank failure marker; not the duration of all 2023 banking stress. Source
2025 tariff shock	2–11 Apr 2025	Selected stress window	Selected trade-policy market-stress window discussed in the Fed’s April report; not a designated US recession. Source

Quarterly means are positioned at quarter end on a calendar scale. Crisis boundaries can fall within a quarter; a quarterly point does not resolve the day on which a change occurred.

The complete dated values are available in the table below. Interactive charts appear when JavaScript is enabled and this snapshot can be read.

Read the historical data table · all quarters

Treasury yields are in percent. Their difference is in percentage points; multiply by 100 for basis points. The aggregate index is unitless, on a 0–1 scale. Missing values, if any, are shown as unavailable. Values are calculated before display rounding. The spread and the two yields are rounded independently, so subtraction of the printed yield values can differ slightly from the printed spread.

Download this historical snapshot as JSON.

Quarterly observations and historical index replay, 1997 Q1–2026 Q2
Quarter	2-year yield (%)	10-year yield (%)	Spread (pp)	Index (0–1)	Paired days
1997Q1	6.0447	6.5670	0.5223	0.099499	60
1997Q2	6.2759	6.6997	0.4238	0.112853	64
1997Q3	5.9037	6.2422	0.3384	0.142042	64
1997Q4	5.7345	5.9063	0.1718	0.199733	62
1998Q1	5.4497	5.5910	0.1413	0.242168	61
1998Q2	5.5581	5.5930	0.0349	0.282635	63
1998Q3	5.1361	5.2073	0.0712	0.334709	64
1998Q4	4.3763	4.6619	0.2856	0.399936	62
1999Q1	4.8616	5.0008	0.1392	0.454797	61
1999Q2	5.2842	5.5412	0.2570	0.460858	64
1999Q3	5.6323	5.8830	0.2506	0.461947	64
1999Q4	5.9482	6.1448	0.1966	0.446497	62
2000Q1	6.5263	6.4684	-0.0579	0.449076	63
2000Q2	6.5727	6.1849	-0.3878	0.546827	63
2000Q3	6.2170	5.8898	-0.3271	0.578677	63
2000Q4	5.7190	5.5706	-0.1484	0.628568	62
2001Q1	4.5802	5.0442	0.4640	0.610196	62
2001Q2	4.1917	5.2762	1.0844	0.591077	63
2001Q3	3.6762	4.9957	1.3195	0.596832	61
2001Q4	2.8671	4.7621	1.8950	0.609893	62
2002Q1	3.2005	5.0792	1.8787	0.527047	60
2002Q2	3.2339	5.1061	1.8722	0.381935	64
2002Q3	2.2388	4.2723	2.0336	0.369773	64
2002Q4	1.8897	4.0048	2.1152	0.364862	62
2003Q1	1.6490	3.9200	2.2710	0.316487	61
2003Q2	1.4216	3.6206	2.1990	0.246029	63
2003Q3	1.6783	4.2277	2.5494	0.187505	64
2003Q4	1.8563	4.2852	2.4289	0.140678	62
2004Q1	1.6866	4.0100	2.3234	0.132136	62
2004Q2	2.4547	4.5971	2.1424	0.115034	62
2004Q3	2.5566	4.3014	1.7448	0.109300	64
2004Q4	2.8224	4.1758	1.3534	0.096713	62
2005Q1	3.4549	4.3036	0.8487	0.073912	61
2005Q2	3.6459	4.1586	0.5127	0.063159	64
2005Q3	3.9580	4.2152	0.2572	0.099625	64
2005Q4	4.3646	4.4882	0.1236	0.144112	61
2006Q1	4.6047	4.5771	-0.0276	0.189942	62
2006Q2	4.9979	5.0727	0.0748	0.212856	63
2006Q3	4.9290	4.8935	-0.0356	0.225714	63
2006Q4	4.7376	4.6306	-0.1069	0.244620	62
2007Q1	4.7605	4.6790	-0.0815	0.257406	62
2007Q2	4.8037	4.8461	0.0423	0.263800	64
2007Q3	4.3908	4.7384	0.3476	0.292506	63
2007Q4	3.4897	4.2666	0.7769	0.328791	62
2008Q1	2.0293	3.6652	1.6359	0.380979	61
2008Q2	2.4164	3.8816	1.4652	0.387757	64
2008Q3	2.3600	3.8622	1.5022	0.396824	64
2008Q4	1.2155	3.2347	2.0192	0.450082	62
2009Q1	0.9043	2.7362	1.8320	0.450041	61
2009Q2	1.0176	3.3208	2.3032	0.429068	63
2009Q3	1.0300	3.5178	2.4878	0.326860	64
2009Q4	0.8779	3.4640	2.5861	0.251438	62
2010Q1	0.9177	3.7179	2.8002	0.187326	61
2010Q2	0.8731	3.4925	2.6194	0.181249	64
2010Q3	0.5387	2.7845	2.2458	0.189929	64
2010Q4	0.4871	2.8784	2.3913	0.182224	62
2011Q1	0.6937	3.4574	2.7637	0.152264	62
2011Q2	0.5616	3.2014	2.6398	0.134907	63
2011Q3	0.2792	2.4142	2.1350	0.189274	64
2011Q4	0.2636	2.0467	1.7831	0.224407	61
2012Q1	0.2895	2.0400	1.7505	0.203553	62
2012Q2	0.2895	1.8258	1.5362	0.204970	64
2012Q3	0.2578	1.6413	1.3835	0.192127	63
2012Q4	0.2675	1.7070	1.4395	0.165354	61
2013Q1	0.2628	1.9510	1.6882	0.153995	60
2013Q2	0.2702	1.9863	1.7161	0.147073	64
2013Q3	0.3659	2.7066	2.3406	0.135062	64
2013Q4	0.3273	2.7442	2.4169	0.124083	62
2014Q1	0.3748	2.7654	2.3907	0.116954	61
2014Q2	0.4195	2.6210	2.2014	0.104204	63
2014Q3	0.5150	2.4995	1.9845	0.100337	64
2014Q4	0.5366	2.2760	1.7394	0.114514	62
2015Q1	0.6046	1.9689	1.3643	0.129235	61
2015Q2	0.6127	2.1644	1.5517	0.139947	64
2015Q3	0.6925	2.2231	1.5306	0.179417	64
2015Q4	0.8382	2.1905	1.3523	0.204756	62
2016Q1	0.8362	1.9143	1.0780	0.236678	61
2016Q2	0.7712	1.7502	0.9789	0.209745	64
2016Q3	0.7281	1.5644	0.8363	0.181200	64
2016Q4	1.0095	2.1387	1.1292	0.155474	61
2017Q1	1.2444	2.4466	1.2023	0.144907	62
2017Q2	1.2976	2.2610	0.9633	0.140590	63
2017Q3	1.3627	2.2414	0.8787	0.130019	63
2017Q4	1.6944	2.3715	0.6771	0.113136	62
2018Q1	2.1605	2.7585	0.5980	0.122521	61
2018Q2	2.4770	2.9206	0.4436	0.134428	64
2018Q3	2.6659	2.9238	0.2579	0.173642	63
2018Q4	2.8028	3.0411	0.2384	0.227501	61
2019Q1	2.4851	2.6530	0.1679	0.245535	61
2019Q2	2.1279	2.3389	0.2110	0.247434	63
2019Q3	1.6859	1.7980	0.1120	0.258499	64
2019Q4	1.5911	1.7919	0.2008	0.254821	62
2020Q1	1.0815	1.3650	0.2835	0.337876	62
2020Q2	0.1937	0.6876	0.4940	0.429670	63
2020Q3	0.1409	0.6506	0.5097	0.416602	64
2020Q4	0.1497	0.8645	0.7148	0.395503	62
2021Q1	0.1321	1.3359	1.2038	0.365229	61
2021Q2	0.1742	1.5908	1.4166	0.308823	64
2021Q3	0.2258	1.3248	1.0991	0.241566	64
2021Q4	0.5318	1.5334	1.0016	0.203053	62
2022Q1	1.4645	1.9511	0.4866	0.230620	62
2022Q2	2.7205	2.9326	0.2121	0.302921	62
2022Q3	3.3822	3.1011	-0.2811	0.427195	64
2022Q4	4.3874	3.8267	-0.5607	0.521801	61
2023Q1	4.3427	3.6463	-0.6965	0.497715	62
2023Q2	4.2643	3.5957	-0.6686	0.480689	63
2023Q3	4.9167	4.1497	-0.7670	0.451830	63
2023Q4	4.8079	4.4471	-0.3608	0.436774	62
2024Q1	4.4823	4.1564	-0.3259	0.448943	61
2024Q2	4.8262	4.4487	-0.3775	0.455027	63
2024Q3	4.0406	3.9547	-0.0859	0.464269	64
2024Q4	4.1455	4.2755	0.1300	0.469780	62
2025Q1	4.1487	4.4536	0.3049	0.482905	61
2025Q2	3.8618	4.3618	0.5000	0.507300	62
2025Q3	3.7208	4.2611	0.5403	0.494050	64
2025Q4	3.5224	4.1000	0.5776	0.441682	62
2026Q1	3.5807	4.1979	0.6172	0.400211	61
2026Q2	3.9659	4.4219	0.4560	0.342518	63

Data definitions and snapshot dates

The Treasury series are DGS2, 2-year constant maturity, and DGS10, 10-year constant maturity. The spread is 10-year minus 2-year. Averaging the difference over the same paired dates equals the difference between those two averages.

Pairing matters: subtracting averages built from different sets of dates need not give the same result. These retrospective source vintages are not a reconstruction of everything a reader could have known at the time.

Treasury observations: Federal Reserve Board H.15, retrieved through FRED. Quarterly aggregation by Comuvia. These two Treasury series are public domain, with citation requested; source-specific terms apply to other datasets.

History shown: 1997 Q1 to 2026 Q2
Treasury data vintage: 2026-08-22
Index snapshot date: 2026-09-20
Index snapshot SHA-256: Included in the downloadable data record.

This page is a fixed educational snapshot, not today’s reading. For the latest published index and its source status, visit Daily Fragility.

04 / From a spread to an index

Why the index does not trace the spread

The curve is one input. The index also uses a corporate credit spread, equity volatility and equity valuation. A change in one does not describe all four.
The inputs are summarized by quarter. Daily market observations enter quarterly averages; a still-open quarter uses the available days so far. The valuation input updates more slowly.
Recent curve history is retained. An earlier flat or inverted curve can continue to affect the index after the current spread turns positive.
The combined reading is smoothed. It need not move as quickly as any one market observation. Daily publication does not mean all underlying inputs changed that day.

A credit spread is a different comparison

The term spread compares Treasury maturities. A corporate credit spread compares corporate bond yields with a Treasury benchmark.

Illustration: 5.5% − 4.0% = 1.5 pp = 150 bp

These are synthetic values. ForeGlass’s actual credit-spread series is BAA10Y: Moody’s seasoned Baa corporate yield minus the 10-year Treasury yield. That market difference also reflects maturity and liquidity differences; it is not a pure credit-risk premium, a default probability or an amount of debt.

Keep the claims separate. ForeGlass reports experimental market conditions. A reading of 0.30 does not mean a 30% recession probability. The index does not directly measure household payment capacity, corporate maturity walls, bank losses or all five forms of economic vulnerability.

Read the index explanation or explore the five vulnerability questions.

05 / What a curve can tell you

Expectations and a premium for time

Longer-term Treasury yields reflect expectations about future short-term interest rates and a term premium. The premium is model-estimated and can vary over time, including becoming negative. A change in the curve therefore has more than one possible explanation. Federal Reserve yield-curve models distinguish these components.

A flat or inverted spread is useful context for investigating financing conditions and the economic outlook. It is not a mechanical cause, a countdown to recession or a complete account of financial vulnerability. Different maturity pairs and forecasting models answer different questions.

Do not borrow another model’s probabilities. The cited New York Fed recession model uses a monthly 10-year minus 3-month Treasury spread to estimate recession probability twelve months ahead. ForeGlass uses 10-year minus 2-year as one input to a different index. Their values and interpretations are not interchangeable.

06 / Put the other numbers in context

“Interest rates” can mean several things

These measures describe different markets or decisions. The federal funds rate provides policy context; it is not the short-maturity yield subtracted in T10Y2Y.

Keep the instrument, maturity and unit with the number
Measure	What it describes	Relation to this index
Effective federal funds rate	The volume-weighted median rate on actual overnight federal funds transactions.	Policy and money-market context. It is not the 2-year Treasury yield and is not one of the four direct inputs.
Federal funds target range	The policy range set by the FOMC, rather than a single observed transaction rate.	Context for monetary policy; distinct from the effective rate and longer-maturity Treasury yields.
2-year and 10-year Treasury yields	Market benchmark yields at two stated constant maturities.	Their difference is the term-spread input. They are not rates directly set by the FOMC.
Mortgage rate	A borrowing rate for a specified mortgage product. A new-loan quote differs from the rate on an existing loan.	Household financing context, not a direct input. Mortgage terms and borrower characteristics matter.
CPI and CPI inflation	CPI measures a consumer-price level; its change over a named interval measures inflation, not a bond yield.	Price and purchasing-power context, not a direct input. A nominal yield and inflation rate describe different quantities.

For the effective federal funds rate and its distinction from the policy target, see the New York Fed’s rate definition.

Definitions you can check

FRED: 10-year minus 2-year Treasury spread, with the definitions of DGS2 and DGS10.
FRED: Baa corporate yield relative to the 10-year Treasury yield.
Federal Reserve: yield curves and term-structure models, including expected short rates and the term premium.
New York Fed: The Yield Curve as a Leading Indicator, 2006 (PDF), a distinct model using a different maturity pair.

These sources define the underlying concepts. They do not endorse ForeGlass or validate its index as a forecast. The interactive example is original and synthetic.