Market Structure Weekly -- Week of May 4, 2026

The VIX closed the week at 16.2, a fifth consecutive week of compression and the lowest weekly close in the trailing 90 sessions. This level places implied volatility in the lower half of the 15-18 band that historically characterizes a fully settled low-volatility regime. The intraday VIX range has compressed further, averaging 0.4 points across the five sessions with one session printing a sub-0.3-point range, a configuration that is structurally typical of low-variance environments where macro-driven repricing flow is essentially absent. Realized volatility on a 10-day trailing basis declined to approximately 12.1% annualized, and the 5-day measure is now firmly below 12% across every session. The alignment between spot VIX, short-trailing realized, and option-implied paths is now within ordinary cross-section noise bands. There is no remaining signature in the input data that would suggest residual stress at any horizon, and the configuration this week is operationally indistinguishable from the pre-disruption baseline volatility regime.

Term structure has held its constructive contango configuration at the level established last week, with the VX1-VX2 spread holding near +0.78 points across all sessions. The full VX curve from the first through sixth month exhibits a stable monotonically increasing slope with coefficients near the median of the long-run normal-regime distribution. SPX skew metrics have converged to within ordinary noise of their long-run baselines: the 25-delta put-call skew is now within 0.1 volatility points of its trailing 90-day median, and the 90-day put-call skew is within 0.2 points. Variance risk premium on a 30-day basis has stabilized at +2.8 points, marginally above its long-run median, consistent with an option market operating under standard pricing conventions. The convergence of these distinct measures within the noise band of their respective normal ranges represents the cleanest week-over-week stability of the entire post-disruption arc and supports the analytical inference that the volatility regime is now stable rather than still resolving.

Regime classification models that incorporate level, term structure, skew, and realized-implied spread inputs now assign approximately 78% probability to a low-volatility state and 22% to a transitional state, the highest low-volatility weight observed in the trailing 60 sessions. The transitional probability has fallen below one-quarter for the first time since early March, and the high-volatility regime probability remains at zero. GARCH(1,1) conditional variance estimates have settled at levels statistically indistinguishable from the pre-disruption baseline. The transitional probability that remains is now structurally inseparable from ordinary classifier uncertainty under low-volatility regimes and does not reflect any identifiable residual artifact of the prior dislocation. The volatility component of the framework is now operating under standard regime assumptions, and lookback windows for variance estimation can be returned to their normal-regime defaults without incorporating dislocation contamination.

The percentage of S&P 500 constituents trading above their 50-day moving average extended further this week, closing Friday at approximately 56%, up from 51% the prior week and continuing the trajectory established since crossing the 50% threshold. The 56% reading is now within the 55-65% range that characterizes a fully developed positive participation regime, and the metric has held above 50% on every session this week. The sectoral contribution pattern continued to broaden: ten of eleven major sectors posted week-over-week breadth improvement, and the lone exception (Real Estate) declined only marginally and remained well above its 50% threshold. The cumulative advance-decline divergence that had been the dominant breadth feature of the disruption phase has now closed completely, with the equal-weight and cap-weight S&P 500 indices moving in essentially parallel paths this week.

Advance-decline data showed net positive readings on all five sessions for the third consecutive week. New 52-week high counts on the NYSE have continued to expand, with the weekly aggregate now approaching the levels typical of established positive participation regimes. The new-highs-to-new-lows ratio settled near 4.2 on a trailing five-day basis, building on last week's 3.4 reading. The McClellan Oscillator closed near +24, holding within the constructive +20 to +45 range, and the McClellan Summation Index extended its uptrend with the steepest weekly slope observed in the trailing 12 weeks. Volume-weighted breadth confirmed the participation broadening: advancing volume exceeded declining volume by margins consistent with established positive regimes, and the compositional shift toward small and mid-capitalization names that began two weeks ago has continued, with the Russell 2000 outperforming the S&P 500 by approximately 50 basis points this week.

The structural configuration that had defined the disruption and recovery phases over the prior eight weeks is now substantively absent. Breadth is no longer the analytical question of the week. The relevant breadth question has shifted from whether participation can recover to whether the established participation regime will be durable, which is a question that requires a different analytical horizon than the weekly framework can address. For systematic frameworks that incorporate breadth state as an input, the configuration is now resolved into a clear positive-participation reading, and breadth-conditional models can operate under their standard regime assumptions without dislocation-adjusted modifications. The post-disruption breadth normalization that began nine weeks ago has fully completed, and the framework now treats breadth dynamics as ordinary regime-state inputs rather than as recovery-phase signals.

Sector rotation dynamics this week extended the fully distributed leadership configuration noted last week, with no single sector establishing meaningful leadership separation and ten of eleven major sectors posting positive relative performance within a tight 70-basis-point band. The rank changes among sectors averaged fewer than one position per sector for the second consecutive week, the lowest level of week-over-week rank instability in the trailing 12 weeks. Industrials, Financials, and Materials retained their modest leading positions but no longer displayed the separation patterns that would identify them as a coherent leadership cohort. Technology posted balanced performance across sub-sectors. Healthcare, Communication Services, Consumer Staples, and Real Estate all participated in the broad-based positive movement. Energy faded further from its earlier dominant position and is now slightly trailing the broad index. The cross-sector dispersion of weekly returns reached its lowest level in 11 weeks, consistent with the fully settled rotation environment characteristic of stable regimes.

The industry-group correction count improved further to 3 of the 25 tracked groups remaining in technical correction territory at week close, down from 4 last week and 16 five weeks ago. The 3 remaining groups are now distributed across uncorrelated drivers: one rate-sensitive (commercial REITs in a specific subsegment), one regulatory-pressured (managed care, unchanged from last week), and one idiosyncratic (a defense-policy-exposed transportation subsegment, unchanged). The composition is structurally indistinguishable from a normal-regime distribution of correction-state groups, where the count typically sits in the 2-5 range and the membership is dominated by group-specific dynamics rather than systemic factors. The macro-clustered correction structure that defined the prior six-week period is now fully absent, and the framework's correction-state analytics have returned to their normal-regime calibration.

Rotation analytics within the quantitative framework now indicate that the sector leadership configuration is operating fully within the parameters that characterize stable normal-regime rotation. Factor-based decomposition of sector returns indicates that idiosyncratic sector-level dynamics now explain approximately 74% of weekly sector return variance, up from 71% last week and well above the typical 65-70% range that characterizes baseline regimes. The sector-to-market variance ratio is at its highest reading in 13 weeks, supporting the inference that sector selection rather than directional beta is the primary driver of cross-sectional outcomes. For systematic frameworks that incorporate sector rotation as an input, the configuration is now fully consistent with normal-regime rotation dynamics, and the analytical framework no longer applies any post-dislocation adjustments to its sector-rotation models.

Pairwise correlation across S&P 500 constituents settled at approximately 0.27 by Friday close, down from 0.28 the prior week and now precisely at the pre-disruption baseline. The trajectory has fully completed its normalization arc, and the remaining week-over-week variation in the correlation metric is now within the standard noise band of the long-run estimate rather than reflecting any structural movement. The first principal component of cross-sectional return variance now explains approximately 28% of total variance, statistically indistinguishable from the long-run pre-disruption average. The second through fifth principal components have reasserted their typical share of explained variance, and the cross-sectional variance is now distributed across factors in a configuration that matches normal-regime patterns precisely. The multi-factor structure that defines stable equity market regimes is now fully re-established.

Intra-sector correlation within Technology settled at approximately 0.38, precisely at its pre-disruption baseline. Sub-sector dispersion within Technology continues to widen modestly, with the trailing 5-day intra-sector correlation moving within the standard variation range of normal regimes. Intra-sector correlations across other sectors are similarly within their normal regime ranges: Financials at 0.35, Consumer Discretionary at 0.32, Industrials at 0.34. The within-sector dispersion patterns across the broad index match the multi-factor configuration that characterized the pre-disruption baseline within ordinary noise tolerances. The within-sector correlation environment is now operating under standard regime assumptions, and sector-conditional models that had been applying dislocation-adjusted correlation inputs can now revert to their normal-regime calibrations.

Cross-asset correlation dynamics have completed their normalization. The equity-bond correlation on a 10-day rolling basis settled near -0.18, fully within the -0.15 to -0.20 range that characterized the pre-disruption regime. Credit spreads tightened by a further 2 basis points this week, and the equity-credit correlation is now within ordinary noise of its pre-disruption baseline. The equity-commodity correlation with energy prices has fully normalized, the equity-dollar correlation has stabilized at its pre-disruption value, and the relationship between cross-asset volatility surfaces matches the configuration typical of stable regimes. The aggregate cross-asset configuration is now operating under standard normal-regime diversification dynamics, and quantitative frameworks that depend on cross-asset correlation estimates have reintroduced standard longer-window estimation without incorporating dislocation contamination. The cross-asset correlation environment is structurally indistinguishable from the pre-disruption baseline within ordinary noise bands.

Scanner flag counts at the standard z-score 2.8 threshold settled at 3 names this week, at the lower end of the 3-6 flag range that characterizes a settled low-volatility regime and consistent with the long-run median count of approximately 3.2. The 3 flags are distributed across unrelated names with idiosyncratic drivers: one earnings-related move in a small-cap consumer discretionary name, one merger-arbitrage convergence in a mid-cap industrial, and one specialty-pharmaceutical name reflecting a clinical-trial outcome. None of the flags share factor exposure or sector concentration. The compositional pattern is fully consistent with the normal-regime distribution of statistical extremes, and the scanner is now operating under its standard low-vol-regime calibration without any dislocation-adjusted modifications.

The momentum factor cross-sectional dispersion expansion noted in last week's report has continued and broadened. The 90-day momentum factor's intra-cohort dispersion has risen from the 67th percentile last week to the 75th percentile this week, and the size factor's intra-cohort dispersion has begun a parallel expansion, moving from the 51st to the 64th percentile over the same window. This dual expansion across two structurally distinct factors is characteristic of post-recovery regimes where the broad-market normalization has completed and individual-name factor dynamics begin to differentiate as the multi-factor structure fully re-establishes. The rate of dispersion expansion remains within historical norms for the post-recovery phase, but the breadth of the expansion across two factors simultaneously represents a structurally meaningful pattern that the framework continues to monitor as a potential early-stage signal of the next regime evolution. The pattern does not yet meet the threshold for a distinct regime-state characterization, and the framework continues to apply standard cautions regarding the inference of structural change from dispersion patterns alone.

The mean-reversion scanner's calibration files now treat the late-March dislocation episode as a closed reference event with finalized retrospective metrics: 78% reverted at 1.0 standard deviation, 1.5 standard deviation median reversion magnitude, 2.1 risk-adjusted dispersion. These metrics are stored alongside the framework's library of prior dislocation-episode references and will be used as comparison points for future regime-change diagnostics. The scanner output continues to describe statistical patterns observed in the data and is not a source of directional recommendations. The regime characterization presented here is a probabilistic inference drawn from available data rather than a forecast of subsequent market behavior, and the framework's standard cautions regarding the limits of regime-state inference apply unchanged.