NFCI Compressed in 2024-2026: Post-GFC Anomaly or New Regime?

May 2026: NFCI at −0.50, a compression rarely observed after a monetary tightening phase. Fed-to-financial-conditions transmission appears to operate differently than in previous cycles.

Three competing readings circulate in strategist notes and recent academic literature. This article lays out the three without adjudicating.

The observation: compression sustained despite ZIRP end and QT

May 2026, weekly NFCI measurement: −0.50. The level has settled around this value over the past eight weeks, after a slow drift from the local high of October 7, 2022, at −0.10. One point of chronology is routinely misstated: the index has touched nothing positive since May 1, 2020. The 2022-2023 tightening brought it back to the edge of the neutral zone without ever crossing it, and it has traded below −0.35 since January 2024.

The monetary configuration in place makes this compression analytically surprising. The Federal Funds Rate peaks at 5.25-5.50% between July 2023 and September 2024 — a range held for 14 months. The cutting cycle that follows cumulates 175 basis points between September 2024 and December 2025, bringing the rate to 3.50-3.75%, the range still in force in May 2026. In parallel, the Fed balance sheet measured by WALCL has shrunk by 2.3 trillion dollars from its April 13, 2022 peak of 8.97 trillion, consequence of quantitative tightening engaged in June 2022 (initial cap 47.5 billion per month, raised to 95 billion in September 2022, slowed to 60 billion in June 2024, then 40 billion in April 2025). That balance sheet run-off is over: WALCL bottoms at 6.54 trillion on December 3, 2025 and has risen since.

Traditional monetary transmission logic, as observed over 1971-2010, would lead to expect NFCI near zero or weakly restrictive in this configuration. The intuition survives the count poorly: only two cycles since 1971 cumulate more than 500 basis points in two years, and the index holds above +0.3 for ten months after the July 1974 peak, nineteen months after the June 1981 one. The 2004-2006 cycle, examined below, ends on an NFCI already negative. The index does the opposite: it has fallen continuously and settled in markedly negative territory. This divergence between monetary policy and the aggregate index is the object of the ongoing analytical debate, and cannot be addressed without the reading of NFCI definition and analytical reading developed in the cluster MAJEUR.

Reading 1: internal decomposition, risk is what pulls the aggregate

The first reading, most mechanical, examines what is happening inside the aggregate via the sub-index decomposition. The finding is sharp, and it cuts against the common reading: the risk sub-index (NFCIRISK) is at −0.59 in May 2026. The credit sub-index (NFCICREDIT) is at −0.03, barely below neutrality, and its post-2010 minimum only reaches −0.44, so the −0.7 level sometimes attributed to it does not exist in the modern period. The leverage sub-index (NFCILEVERAGE) moves from +0.17 to +0.27 over the single month of May, squarely restrictive. The aggregate result is pulled by the risk component, not by credit.

This risk sub-index compression has measurable counterparts in the credit market. ICE BofA US Investment Grade Master OAS spreads average 76 basis points in May 2026, in a 73 to 81 band. HY OAS is at 277 bps, against 393 bps in September 2023. Bank lending conditions measured by the Senior Loan Officer Opinion Survey have loosened without ever crossing to the accommodative side: net tightening reported on large enterprise loans peaks at 50.8% of banks in the third quarter of 2023, falls back to 8.1% in the second quarter of 2026, then to zero in the third. The loosening is real; net easing never happened.

One methodological caution belongs here, because it is routinely skipped. The Chicago Fed publishes no weighting of the three sub-indices within the aggregate: the NFCI is estimated in a single pass over its 105 series by a dynamic factor model, and the sub-indices are built separately from the same series. No contribution calculation is therefore available, and the 40% or 50% weights that circulate in strategist notes come from no source at all. The decomposition is read by comparing levels, which is enough to settle the question: risk at −0.59 against credit at −0.03 and leverage at +0.27 locates the compression unambiguously.

This reading is descriptive: it explains the how without explaining the why. It validates that aggregate NFCI is consistent with its components, but says nothing about why the risk sub-index is itself compressed. It is a first-level reading, which sets up the second and third readings.

Reading 2: structurally weakened transmission, the lock-in effect

The second reading postulates that the post-2020 regime has modified monetary transmission itself, and that accommodative NFCI reflects this structural change rather than a cyclical anomaly. Two main mechanisms are mobilized.

First mechanism: lock-in of fixed-rate financing during 2020-2021. Freddie Mac’s weekly survey gives the measure of the lock. The conventional 30-year prints at 2.96% on average across 2021, with a floor at 2.65%, crosses 6.5% on September 29, 2022 and peaks at 7.79% on October 26, 2023; it still holds 6.95% on September 17, 2026. A household that signed in 2021 therefore carries a rate some 400 basis points below what it would pay today, and has no reason to refinance. Since the U.S. mortgage stock is overwhelmingly thirty-year fixed, that lock unwinds at the pace of house moves, not at the pace of the Fed. The corporate side ran the same sequence: the 2020-2021 wave of long-duration issuance means the average coupon paid on outstanding debt follows the marginal rate with a multi-year lag. These two structural features together imply that the marginal rate increase signaled by the Fed transmits to the average effective rate paid by the private sector with a multi-year lag.

Second mechanism: the ample reserves system installed since 2008 and confirmed since 2020, combined with the post-Dodd-Frank robustness of bank balance sheets. Common equity tier 1 capital ratios of large U.S. banks (G-SIBs) exceed 13% in 2026, against a regulatory requirement around 11% and pre-GFC levels of 6-7%. Eco3min’s analysis of the 2008 crisis traces this episode. This balance sheet robustness limits transmission of policy rate tightening to bank lending conditions — banks can absorb moderate tightening without passing it on to borrowers.

In this framework, accommodative NFCI is not an anomaly but the reflection of a structurally different regime where monetary transmission operates with reduced lag and amplitude. The structural reading is consistent with empirical lock-in and bank robustness observation, but stays partially non-testable: it will only be verifiable in case of a major shock to the system that would reveal the real elasticity of transmission. This reading fits within financial conditions as monetary transmission across cycle, whose very nature is in post-2020 mutation.

Reading 3: the 2006-2007 analogy, dangerous late-cycle complacency

The third reading is cyclical and pessimistic: current compression is analogous to 2006-2007, when NFCI settled between −0.3 and −0.7 despite a Fed hiking cycle complete since June 2006. This analogy frames May 2026 compression as dangerous complacency that will not survive the first systemic shock.

The chronological reminder of 2006-2007 is instructive. The 2004-2006 hiking cycle cumulated 425 basis points (from 1.00% to 5.25%), concluded in June 2006. NFCI settles between −0.29 and −0.68, averaging −0.59, between July 2006 and July 2007: a phase of apparent accommodative normality. On August 9, 2007, BNP Paribas suspended three money market funds with subprime exposure; the ABCP market froze within weeks. NFCI crossed zero on August 17, 2007, peaked at +0.22 in early September 2007, and would reach +3.08 on November 28, 2008. The transition from accommodative regime to systemic crisis took fifteen months, a window during which the composite index sent no signal of imminent rupture.

Defenders of this reading for 2024-2026 point to several parallels. The compression of HY spreads, at 277 bps in May 2026 against 393 bps in September 2023, in a structurally higher default environment than 2006. The vigor of leveraged loan and private credit issuance, feeding a demand for credit risk reminiscent of 2006-2007. Above all, the weakness of the equity risk premium. The calculation calls for precision here, because the metric is routinely misnamed: Shiller’s excess CAPE yield, the inverse of the adjusted PER less the real yield on long bonds, falls to 1.40% in May 2026 against a 1970-2024 average of 3.71%. The inverse of the adjusted PER on its own, sometimes presented as a risk premium when it is only an earnings yield, stands at 2.49% against a historical average of 5.75%.

The pessimistic reading nonetheless has its limits. The 2006-2007 analogy is partial: credit market structure has substantially changed (private credit outside the banking perimeter now represents 2.4 trillion against virtually nothing in 2006), banking regulation is significantly tighter (Dodd-Frank, Basel III), the Fed’s mandate now explicitly includes financial stability, and Treasury market depth is sufficiently maintained by SRF and FIMA facilities to limit risks of systemic dislocation seen in March 2020. A rigorous historical comparison refers to the major financial ruptures documented since 1971, which show that each cyclical analogy has its limits.

The three readings are not falsifiable in the short term

None of the three readings is falsifiable in the immediate horizon. The decomposition reading (reading 1) is validated by the data but does not explain why the risk sub-index is compressed — it is descriptive without being explanatory. The structural reading (reading 2) is consistent with the lock-in observation but will only be testable in case of a major shock revealing the real elasticity of monetary transmission. The pessimistic reading (reading 3) draws the 2006-2007 analogy but the analogy is partial and market conditions have materially changed.

The absence of immediate falsifiability is not an analytical defect: it is inherent to the nature of late-cycle phases. The test will come from the next quarters. Three indicators deserve surveillance in this perspective. First, the NFCI inter-sub-index dispersion: if it narrows (the three sub-indices converge toward a common value), the aggregate signal becomes readable again and reading 1 loses relevance. If it stays elevated, the aggregate continues to mask internal structure and the other readings retain value. A sub-index reading grid remains indispensable for tracking this dispersion. The underlying weekly data needed to monitor the three sub-indices is available via the live FRED NFCI weekly data.

Second, the evolution of the leverage sub-index: it moves from +0.17 to +0.27 over the single month of May 2026, after a 2024 spent between −0.35 and −0.01. A durable break above +0.3 would be consistent with a late-cycle scenario following the 2006-2007 pattern, where that same sub-index turns positive on June 15, 2007, two months ahead of the aggregate. Third, market functioning indicators (Treasury bid-ask, SOFR-OIS spread, primary dealer positioning): their gradual degradation would constitute an early warning signal, independent of aggregate NFCI.

The Eco3min analysis does not adjudicate among the three readings because adjudication would not be analysis but forecasting. The editorial position consists in laying out the three readings with their respective implications, signaling the surveillance indicators that will discriminate among them as quarters pass, and letting the reader form their own informed judgment.

Update, September 22, 2026. The compression has extended beyond the window described above: the NFCI fell to −0.56 on September 11, 2026, its low for the sequence. The internal structure, however, moved in the direction of reading 1 rather than reading 3: the leverage sub-index peaked at +0.27 on June 5, then fell back to +0.06, never reaching the +0.3 marker proposed above, while credit slipped to −0.07. The Fed has also interrupted its cutting cycle and raised its target to 3.75-4.00% on September 17, 2026.

Last updated — 8 October 2026