Market Structure Weekly -- Week of April 13, 2026

The VIX closed the week at 19.2, completing a second consecutive week of compression from the 26.78 intraday spike observed two weeks prior. This level places implied volatility within the upper boundary of the 18-20 range that prevailed before the late-March disruption, though the path back has been orderly rather than abrupt. The current reading represents a roughly 28% decline from the prior week's 21.4 close, a rate of mean-reversion consistent with historical decay patterns following single-spike events that lack a secondary catalyst. The fact that the decline has proceeded without a meaningful intraday reversal suggests that the hedging demand responsible for the initial spike has been substantially unwound. Intraday VIX ranges narrowed progressively through the week, with the Monday-to-Friday high-low band compressing from 2.1 points to 0.9 points, a pattern consistent with declining options market maker re-hedging activity. Open interest in near-term VIX calls also declined, reinforcing the interpretation that the protective positioning cycle associated with the prior spike has largely run its course. The current level does not, however, represent a return to the compressed sub-16 regime observed in January and February.

Term structure has re-established a modest contango slope in the front two months, with the VX1-VX2 spread moving to approximately +0.35 points. This is a meaningful shift from the flat-to-mildly-backwardated configuration observed last week and indicates that the market's pricing of near-term uncertainty has declined relative to medium-term expectations. However, the contango slope remains shallower than the +0.70 to +0.90 range typical of settled low-volatility environments, which suggests that structural comfort with forward risk has not fully returned. Skew metrics on the SPX options surface have also moderated, with the 25-delta put-call skew compressing by approximately 1.5 volatility points on the week. The term structure configuration across the full VX curve, from the first through sixth month, now shows a monotonically increasing pattern for the first time since the disruption began, though the slope coefficients remain below their 90-day trailing averages. This partial normalization of the term structure is an intermediate-stage observation, neither indicating all-clear nor suggesting renewed stress.

From a quantitative modeling perspective, GARCH(1,1) conditional variance estimates have continued their decay trajectory and are now approaching levels consistent with a realized volatility regime in the 14-16% annualized range. The half-life of the March shock, as estimated from the observed decay path, appears to be approximately 8-10 trading days, which is within the normal range for equity index volatility shocks that do not involve a fundamental earnings revision or macro policy shift. The transitional regime characterization from last week remains appropriate: volatility has normalized meaningfully but has not yet established the kind of compressed, low-variance baseline that would indicate a clean regime shift back to quiescence. Realized-to-implied volatility ratios have moved below 1.0 on a 5-day trailing basis, which historically has been associated with the late stage of a volatility normalization cycle where implied volatility is the last metric to fully adjust. The persistence of a modest premium of implied over realized suggests residual caution in the options market, even as spot VIX has largely mean-reverted. Regime classification models that incorporate both level and term structure inputs currently assign approximately 60% probability to a transitional state and 40% to low-volatility, with the high-volatility regime probability having declined to near zero.

The percentage of S&P 500 constituents trading above their 50-day moving average improved to approximately 35% by Friday close, up from 29% the prior week and continuing the recovery from the 19% trough observed two weeks ago. This represents a steady but measured improvement -- the metric has recovered roughly half of the ground lost during the late-March breadth deterioration, when it declined from 52% to 19% over a two-week span. The recovery trajectory is consistent with a gradual re-engagement pattern rather than the kind of broad-based thrust that would indicate decisive participation restoration. For the recovery to register as analytically significant within this framework, the metric would need to re-establish above 40% on a sustained multi-session basis. On a sector-decomposed basis, the improvement was concentrated in Industrials and select Financials, while Technology breadth showed only marginal improvement and Consumer Discretionary remained flat. This uneven sectoral contribution to aggregate breadth improvement limits the strength of the recovery interpretation, as broad participation recoveries historically require contribution from at least four of the six largest sectors to sustain beyond the initial rebound phase.

Advance-decline data showed modest net positive readings on four of five sessions this week, an improvement from the three-of-five pattern last week. New 52-week low counts on the NYSE continued to decline, with the trailing five-day average falling below the 20-day average for the first time since the late-March disruption began. The new-highs-to-new-lows ratio moved fractionally above 1.0 on a trailing five-day basis, marking the first time this threshold has been breached since early March. While the directionality is constructive, the absolute level of new highs remains subdued, suggesting that the improvement is driven more by the contraction of new lows than by an expansion of leadership. The McClellan Oscillator turned positive mid-week and closed near +15, which is above its trigger line but well below the +40 to +60 range typically associated with breadth thrusts that have historically preceded sustained participation recoveries. Volume-weighted breadth metrics tell a similar story: advancing volume has exceeded declining volume on most sessions, but the margin has been narrow and concentrated in larger-capitalization names rather than distributed across the full market-cap spectrum.

The cumulative advance-decline divergence that developed over the prior six weeks has narrowed but remains unresolved. Large-cap index prices have held relatively close to their pre-disruption levels, while the underlying breadth metrics tell a story of selective participation rather than broad confirmation. This type of divergence has historically resolved in one of two ways: either breadth catches up to price through sustained improvement, or price eventually adjusts to reflect the weaker internal condition. The current data does not provide sufficient evidence to favor either resolution path, which is consistent with the transitional characterization applied to this market regime. It is worth noting that the equal-weight S&P 500 has underperformed the cap-weight index by approximately 180 basis points since the disruption began, a spread that has widened slightly this week even as the above-50-day-MA metric improved. This divergence between equal-weight relative performance and improving breadth counts is unusual and suggests that the median constituent is participating in the recovery but with smaller magnitude moves than the largest names. The analytical implication is that headline breadth improvement may overstate the depth of the recovery when measured on a dollar-weighted or return-magnitude basis.

The broadening of sector participation that began tentatively last week continued in modest fashion. Energy retained its relative-strength leadership position but the margin of outperformance continued to narrow as Industrials and Materials posted their second consecutive week of stabilization. Technology, which had underperformed during the late-March disruption, showed mixed results at the sub-sector level: semiconductor names recovered more ground than software, creating an internal dispersion pattern within the sector that has not been present in recent months. Healthcare posted a quiet week of slight relative outperformance, consistent with its typical behavior during transitional periods when defensive positioning partially unwinds. Communication Services was essentially flat on a relative basis, which represents an improvement from the prior two weeks of relative weakness. The cross-sector dispersion of weekly returns, measured as the standard deviation of GICS sector returns, declined to its lowest level in four weeks, a development that is arithmetically consistent with the correlation normalization discussed in the following section and qualitatively consistent with a market environment where sector-specific dynamics are reasserting themselves after a period dominated by macro-driven co-movement.

The industry-group correction count improved further, with 10 of the 25 tracked groups remaining in technical correction territory at week close, down from 13 last week and 16 the week before that. The groups exiting correction this week were concentrated in domestically-oriented industrials and select consumer discretionary sub-industries, specifically machinery, building products, and specialty retail names with lower rate sensitivity. Rate-sensitive groups -- including homebuilders, REITs, and select financials -- remained among those still in correction, reflecting the ongoing sensitivity to interest rate expectations that has characterized this cycle. Utilities and Consumer Staples continued to retrace from their volatility-event outperformance, consistent with the normalization of defensive positioning as the acute uncertainty phase recedes. The pattern of correction resolution is notably sequential rather than simultaneous: domestically-oriented cyclicals are recovering first, followed by industrials, with rate-sensitive and export-sensitive groups lagging. This sequencing is consistent with historical post-disruption recovery patterns where the most fundamentally constrained groups are the last to normalize, as their recovery depends on the resolution of specific macro conditions rather than simply the abatement of broad risk aversion.

Rotation analytics within the quantitative framework indicate that the sector leadership configuration is becoming incrementally more stable than it was last week, though it has not yet reached the level of persistence that would support high-confidence regime characterization. The relative-strength rankings across sectors have shown less week-over-week rank change than in the prior two weeks, which is a necessary but not sufficient condition for rotation stability. Factor-based decomposition suggests that the sector movements this week were driven more by idiosyncratic sector-level dynamics than by macro factor rotation, which is consistent with the correlation normalization discussed below. The ratio of sector-level variance to market-level variance increased this week, indicating that a greater proportion of total market variance is now attributable to sector selection rather than directional beta -- a structural feature that tends to emerge during the middle-to-late stages of post-dislocation normalization. For systematic frameworks that incorporate sector rotation as an input, the current environment remains analytically ambiguous: the rotation pattern is becoming more readable but has not yet stabilized sufficiently to support high-confidence modeling of forward sector dynamics.

Pairwise correlation across S&P 500 constituents continued its decline, settling near 0.34 by Friday close, down from 0.41 the prior week. The trajectory toward the pre-disruption baseline of 0.27 is proceeding at a pace consistent with historical normalization patterns following correlation spikes that are not accompanied by a sustained fundamental deterioration. At 0.34, the current level sits roughly midway between the spike peak of 0.52 and the baseline, suggesting that the normalization process is approximately half complete on a magnitude basis. The rate of decline has slowed modestly compared to the prior week, which is typical of exponential decay dynamics in correlation structures where the initial rapid compression gives way to a more gradual convergence toward equilibrium. Decomposing the aggregate pairwise correlation into its principal component structure, the first principal component now explains approximately 38% of cross-sectional return variance, down from 44% last week and 51% at the spike peak. This compression indicates that market returns are becoming less dominated by a single systematic factor and more reflective of multiple independent sources of variation -- a structural prerequisite for the kind of differentiated stock-level behavior that characterizes normal-regime market conditions.

Intra-sector correlation within Technology compressed further to approximately 0.48, down from 0.58 last week and 0.71 at the spike peak. This represents meaningful progress toward stock-level differentiation within the sector, and the pace of normalization has been somewhat faster than the broad-market correlation decline, which is consistent with the observation that intra-sector correlations tend to mean-revert more quickly than cross-sector correlations following systemic events. The divergence between semiconductor and software sub-sector behavior noted in the rotation section above is reflected in this correlation metric: as sub-sector differentiation increases, the aggregate intra-sector correlation mechanically declines. Other sectors showed similar but less dramatic intra-sector correlation normalization, with Financials moving from 0.49 to 0.42 and Consumer Discretionary from 0.44 to 0.38. The rate-sensitive sectors that remain in technical correction show the slowest intra-sector correlation normalization, which is analytically consistent with the interpretation that these groups continue to respond primarily to a common macro factor -- interest rate expectations -- rather than to stock-specific fundamentals. This factor-driven co-movement within lagging sectors is a feature that our correlation monitoring framework tracks as an indicator of sector-specific stress persistence.

Cross-asset correlation dynamics showed incremental improvement. The equity-bond correlation on a 10-day rolling basis moved from positive territory back toward zero, closing the week at approximately +0.05, though it has not yet re-established the negative correlation of -0.15 to -0.20 that characterized the pre-disruption regime. Credit spreads, as measured by investment-grade indices, tightened by approximately 5 basis points on the week, reducing the elevated equity-credit correlation that had persisted since the late-March volatility event. The equity-commodity correlation, particularly with energy prices, remained modestly positive, which is consistent with Energy's continued relative-strength leadership and the sector's ongoing influence on broad index returns. The normalization of cross-asset relationships is a lagging indicator relative to intra-equity correlation, and the current data is consistent with early-stage but incomplete normalization of diversification dynamics across asset classes. For quantitative frameworks that rely on cross-asset correlation estimates for portfolio construction or risk modeling, the current regime warrants continued use of shorter lookback windows for correlation estimation, as longer-window estimates still incorporate the disruption period and may understate the degree of normalization that has already occurred in more recent data.

Scanner flag counts at the standard z-score 2.8 threshold declined to 8 names this week, continuing the reduction from 14 last week and 41 at the peak of the dislocation event. This progression is consistent with the statistical mean-reversion cycle that typically follows broad dislocation: the initial burst of extreme z-scores resolves over a 10-15 session window as prices revert toward distributional norms. The 8 remaining flags are concentrated in a narrower set of rate-sensitive consumer sectors -- specifically, select specialty retail and consumer finance names -- where z-score persistence has now exceeded two standard reversion cycles as measured by the model's historical calibration. This extended duration is an empirical observation that is tracked within the framework for duration-of-dislocation analysis. The concentration of persistent flags in rate-sensitive names is consistent with the sector rotation and correlation observations above, suggesting that the residual statistical dislocation has a identifiable fundamental driver rather than being a random artifact of the reversion process. Historical analogues suggest that z-score persistence of this duration in a defined sector cohort typically resolves over an additional 5-10 sessions once the macro variable driving the co-movement stabilizes, though this observation is drawn from a limited sample of comparable episodes.

The mean-reversion scanner's retrospective tracking of the original 41-name flag list from two weeks ago shows that approximately 68% of flagged names have reverted by at least one standard deviation in the direction implied by their initial z-score extreme. This hit rate is within the historical range of 60-75% observed in prior dislocation episodes of comparable magnitude, though the median reversion magnitude has been somewhat smaller than the historical median, consistent with a market environment where the underlying volatility regime has not fully settled. The remaining unreversed names are disproportionately drawn from the rate-sensitive sub-sectors noted above, reinforcing the observation that the current dislocation has a sector-specific component that is resolving more slowly than the broad statistical pattern. On a risk-adjusted basis, the reversion dynamics of the flagged cohort have generated a Sharpe-equivalent dispersion metric of 1.8, which is within the range of prior episodes but below the 2.2 median, further supporting the characterization that this reversion cycle is proceeding at a slower-than-typical pace. The retrospective tracking exercise is conducted for model calibration and performance monitoring purposes within the research framework, not as a measure of exploitable opportunity.

Cross-sectional factor dispersion metrics from secondary scanner output show that size-factor and momentum-factor dispersion have both declined from last week's elevated readings, though both remain above their trailing 60-day averages. The size-factor dispersion, measured as the return spread between the top and bottom quintile by market capitalization, compressed from the 95th percentile reading observed during the dislocation to approximately the 72nd percentile, indicating meaningful but incomplete normalization. Momentum factor dispersion followed a similar trajectory, declining from extreme levels to the 68th percentile of its trailing distribution. The compression of factor dispersion is directionally consistent with the broader correlation normalization: as pairwise correlations decline from spike levels, the initial effect is often increased factor dispersion, which then compresses as the correlation structure settles into a new equilibrium. The current readings suggest that the factor environment is in the later stages of post-dislocation adjustment but has not yet fully stabilized. The value-growth spread, which was less affected by the initial dislocation than the size or momentum factors, has remained relatively stable throughout the period and is currently near the 55th percentile of its trailing distribution. Scanner output continues to describe statistical patterns observed in the data, not directional opportunities or actionable recommendations.