How tight or loose is the financial system right now

FinObservatory Financial Conditions Index

A PCA-weighted z-score of eight credit, curve, survey and market indicators, standardized to mean 0 and standard deviation 1, higher = tighter. The latest reading is -0.68 (Loose, Mar 31, 2026), at the 33rd percentile of its 1991–2026 history: looser than about 67% of all quarters. Its one extreme tight regime is the 2008–2011 global financial crisis. Shaded bands are NBER recessions.

Source: FinObservatory macro engine (absorbed from argus), driven unmodified | FRED, Federal Reserve Bank of St. Louis Quarterly, 8 of the engine's 12 component slots; PC1 explains 44.0% of common variance. Recession bands from the FRED USREC series. Methodology

The Treasury yield curve

The current curve of constant-maturity Treasury yields (as of Jul 16, 2026), and the 10-year-minus-2-year slope back to 1976. An inverted slope (below zero, shaded red) went negative at some point in the 2 years before 6 of the 6 NBER recessions in that sample (gray bands), which is why it is the market’s single most-watched recession signal. The slope is currently +0.37 pp (positive).

Credit: the Basel III gap and the credit impulse

The credit-to-GDP gap is the deviation of private non-financial credit from its long-run trend, the Basel III anchor for the countercyclical capital buffer. The credit impulse is the acceleration of that credit, which tends to lead GDP. The US gap is currently -11.5 pp (normal), far below trend after the post-2008 and post-2020 deleveraging.

The US credit cycle: loan demand, standards, and sectoral debt

Two views of the same cycle. The Fed’s quarterly Senior Loan Officer Opinion Survey (SLOOS) reads the bank lending channel from both sides: loan demand (the net percentage of banks reporting stronger demand for commercial and industrial loans) against lending standards (the net percentage tightening them). The Z.1 Financial Accounts track the debt stock that cycle leaves behind, by sector, back to 1945. As of 2026Q2, large and middle-market C&I demand reads +4.8 and standards +8.1 (both net % of banks); total US debt securities and loans outstanding stand at $115.6 trillion (2026Q1).

Money markets: SOFR, EFFR, and the FOMC target range

The overnight rates that anchor the short end of the curve. SOFR, the Secured Overnight Financing Rate, is the volume-weighted median of overnight Treasury repo and is the successor benchmark to USD LIBOR; EFFR, the Effective Federal Funds Rate, is the volume-weighted median of overnight unsecured interbank lending and is the rate the FOMC steers into its target range. Latest SOFR is 3.62% and EFFR 3.63% (Jul 16, 2026), both inside the FOMC target range of 3.50–3.75%; the 30-day compounded SOFR average is 3.62%. These NY Fed reference rates are the canonical short-rate source here, superseding any single SOFR series carried on the FRED spine. These are the rates; the corresponding repo and money-fund volumes have their own page.

SOFR vs EFFR vs the FOMC target range, 2016–present

Both overnight rates sit inside the shaded target band in normal times. Three departures stand out: the September 17, 2019 repo spike, when a collateral-and-reserves squeeze drove SOFR to 5.25% (its 99th percentile hit 9.00%) far above the band; the March 2020 cut to the zero lower bound (target 0–0.25%); and the March 2023 SVB week. SOFR is secured and starts at its 2018 inception, so its line begins mid-chart.

Source: Federal Reserve Bank of New York, Reference Rates (Markets Data API) Daily, NY Fed Markets Data API; SOFR/EFFR/SOFR-averages carried unaltered. FOMC target range published alongside EFFR. Chart plots each window's extreme daily prints: every point is a real observation, spikes preserved. Methodology

The natural rate of interest: r-star

Where does policy stand relative to neutral? The NY Fed's Holston-Laubach-Williams model puts the US natural rate of interest, the real short rate consistent with full strength and stable inflation, at 1.06% in 2026Q1 (the original Laubach-Williams model reads 1.70%). The comparable real policy rate, built here as the effective federal funds rate (FRED FEDFUNDS) averaged over the same quarter minus 12-month core PCE inflation (FRED PCEPILFE) averaged over that quarter, is 0.50% (3.64% nominal less 3.14% inflation): -0.55 pp below the HLW natural rate on this construction.

US r-star, two models, 1961–present

Six decades of the estimated natural rate: from around 5% in the 1960s to a post-GFC low just above half a percent in 2014, and only a partial recovery since. Trend growth (HLW g, not charted) is 2.42% in 2026Q1, so the low r-star is carried by the model's negative other-determinants component, not by growth.

Source: Federal Reserve Bank of New York, Measuring the Natural Rate of Interest | Holston, Laubach and Williams (2023), NY Fed Staff Reports no. 1063; Laubach and Williams (2003), Review of Economics and Statistics 85(4) Quarterly, one-sided (filtered) estimates, 261 quarters. The r* = c*g + z identity was re-derived on every row at build using the files' own parameters. The real-rate construction is FinObservatory's, stated above; HLW itself uses different inflation expectations. Methodology

What households expect inflation to do

The Survey of Consumer Expectations asks a rotating panel of household heads where inflation is going. In June 2026 the median respondent expected 3.7% inflation over the next year (from 3.5% the month before) and 3.3% at the three-year horizon. Disagreement is wide: the middle half of one-year answers spans 2.0% to 6.0%. The one-year median peaked at 6.8% in June 2022, the post-pandemic inflation shock.

Source: Federal Reserve Bank of New York, Survey of Consumer Expectations Monthly since 2013 (157 surveys), median point forecasts from the survey's density questions; the interactive-chart data download, ingested unaltered. Methodology

The US household balance sheet

Household debt by product and its delinquency, from the New York Fed’s Consumer Credit Panel (an anonymized 5% sample of Equifax credit files). Total household debt stands at $18.79 trillion (2026Q1), of which mortgages are $13.19 trillion (70.2% of the total). The stress is concentrated in unsecured revolving credit: credit-card balances 90+ days delinquent have reached 13.12%, against just 1.09% on mortgages.

The lender of last resort: the Federal Reserve discount window

Loan-level discount-window borrowing, aggregated by credit type. Primary credit is the standby facility for sound banks; secondary and seasonal credit are narrower programs. The single largest quarter is 2023Q1, when primary-credit borrowing summed to $3.14T across 701 distinct borrowers, in the week of the Silicon Valley Bank failure. These are loan originations summed over the quarter, a flow, not outstanding balances. Primary credit is dominated by overnight loans re-originated every business day, so the summed flow far exceeds the point-in-time stock: the Fed’s H.4.1 release shows primary credit outstanding peaking near $153 billion the week ending March 15, 2023, more than twenty times smaller than the $3.14T of originations that same quarter. Over the full record (2010Q3–2024Q2), all credit types sum to $4.20T of originations. Aggregates only: no borrower is named.

Quarterly borrowing by credit type, 2010Q3–2024Q2

Originations summed within each quarter, USD billions. The 2023Q1 primary-credit spike (mostly the two FDIC bridge banks after the March 2023 failures) dwarfs every other quarter on this flow basis, which is exactly why the originations-vs-outstanding distinction matters. Secondary and seasonal credit are small throughout; two quarters carry no secondary or seasonal loans and break those lines rather than reading zero.

Source: Board of Governors, Discount Window loan-level disclosures (Dodd-Frank section 1103) Aggregates only, never borrower names. Values are originations SUMMED over each quarter (a flow), not outstanding balances (a stock). Published on the ~2-year Dodd-Frank section 1103 lag, so coverage ends at 2024Q2. Methodology

Long-run valuation: the Shiller CAPE

The cyclically adjusted price-to-earnings ratio (CAPE, or P/E10): the real S&P Composite price divided by the ten-year average of real earnings, Robert Shiller’s standard gauge of how richly US equities are valued relative to their own long history. The latest available reading is 41.4 (July 2026), at the 99th percentile of the full 1881–2026 monthly history: richer than about 99% of all months since 1881. Its all-time high is 44.2 (December 1999, the dot-com peak); its all-time low is 4.8 (December 1920). This is a descriptive valuation measure, not a market-timing signal or investment advice.

Source: Robert J. Shiller, ie_data (shillerdata.com) Monthly, 1881–2026-07. No explicit open-data license: freely downloadable research data, displayed here with citation to Shiller. The shipped vintage ends July 2026, so "latest" is the most recent published observation, not today. Methodology

For the other side of the global monetary picture, see how the world holds its reserves: reserve-currency composition.

Global financial stress

Two official daily stress indices that both span the 2008 and 2020 crises. The OFR Financial Stress Index is the U.S. Treasury Office of Financial Research’s market-based gauge, built as the sum of stress contributions from the United States, other advanced economies and emerging markets, and centered so zero is its long-run average. The ECB CISS is the European Central Bank’s composite indicator of systemic stress, bounded in [0, 1] and constructed to rise when several market segments are stressed at once. Both are official indices carried here unaltered, and both currently read well below their crisis levels.

Official stress indices, side by side

The authoritative financial-stress gauges from four official bodies, each standardized so higher = more-stressed. These are the official indices from the Federal Reserve (NFCI, STLFSI4), the U.S. Treasury (OFR FSI) and the European Central Bank (CISS), distinct from the FinObservatory FCI above (our own transparent, long-history composite, shown for contrast). Every US and euro-area gauge here currently reads below its own average.

Source: FRED, Federal Reserve Bank of St. Louis | OFR Financial Stress Index, U.S. Treasury | ECB CISS, European Central Bank Data Portal NFCI and STLFSI4 (Federal Reserve), OFR FSI (U.S. Treasury) and CISS (ECB) are official indices on different scales; the FinObservatory FCI is our own composite, shown alongside for contrast, not as an equal authority. Values are each index's own units, not cross-comparable levels. Methodology

How does today compare to 2007 and 2020

Seven stress indicators (all oriented so higher = more stressed, all covering both crises) at today’s reading, at their most-stressed reading during the 2007–2009 global financial crisis, and during the 2020 COVID crash. Today sits far below both crisis peaks on every measure.

Source: FinObservatory macro engine (absorbed from argus), driven unmodified | FRED, Federal Reserve Bank of St. Louis | OFR Financial Stress Index, U.S. Treasury | ECB CISS, European Central Bank Data Portal Most-stressed reading within each crisis window (2007-01 to 2009-12; 2020). Each indicator is in its own units, not cross-comparable levels. The FinObservatory FCI is quarterly, so its brief-and-averaged-out March-2020 spike reads near zero, a documented limitation. Methodology