Skip to content
LB Research
Black-and-white view over rail tracks towards the towers of Hudson Yards beneath a broad clouded sky.
Back to insights

Research insight · Issue 003

When shock transmission rotates

Aggregate connectedness can hide a changing transmission structure.

Published
Data through
7 August 2026
Download canonical PDF
ContentsExecutive summary

Executive summary

Aggregate connectedness can hide a changing transmission structure.

  1. 01Aggregate connectedness looks similar across quantiles.
  2. 02Sector transmission roles change as drawdowns deepen.
  3. 03Industrials strengthen while Technology weakens in deep drawdowns.
Drawdown regimes
4
Deep drawdowns examined
4
Industrials calm tail shift
−3.57
Industrials >20% tail shift
+1.70
Technology calm tail shift
+1.22
Technology >20% tail shift
−1.63
Largest deep-drawdown event shift
Energy
+15.69 in 2020

Why it matters

Aggregate connectedness can hide important changes underneath. The level of transmission matters. So does where it comes from.

Analysis 01

Aggregate connectedness hides what changes underneath

Lower-tail, median and upper-tail total connectedness follow remarkably similar paths through time. Major increases and declines in aggregate connectedness generally occur across all three states together, with relatively modest separation between the series. At the system level, this can make downside transmission appear similar to ordinary market transmission. That interpretation is incomplete. The aggregate index measures the overall intensity of cross-sector transmission. It does not show whether the same sectors are transmitting and receiving those shocks in each state. A relatively stable aggregate relationship can therefore coexist with substantial changes in the network underneath. This distinction motivates the rest of the analysis: rather than asking only whether connectedness changes during drawdowns, the relevant question becomes who carries the shock when market conditions deteriorate?

Figure 1. Total connectedness across lower-tail, median and upper-tail quantiles.

Lower-tail, median and upper-tail total connectedness rise and fall together through most of the sample, generally remaining close even during major increases and declines in aggregate transmission.

Figure 1

Aggregate connectedness hides what changes underneath

100%

Figure 1. Total connectedness across lower-tail, median and upper-tail quantiles.

Lower-tail, median and upper-tail total connectedness rise and fall together through most of the sample, generally remaining close even during major increases and declines in aggregate transmission.

Download PNG

Analysis 02

Sector transmission changes as drawdowns deepen

Sector roles change progressively across drawdown regimes. The difference between lower-tail and median transmission is not uniform across sectors, and the direction of that difference can reverse as losses deepen. Industrials provide the clearest example. Their median tail shift moves from −3.57 in 0–5% drawdowns to +1.70 when drawdowns exceed 20%. Technology moves in the opposite direction. Its tail shift changes from +1.22 in calm conditions to −1.63 in the deepest regime. Other sectors show smaller or less monotonic changes. Financials remain negative across all four regimes, while Healthcare stays mostly positive. Consumer Staples shift towards stronger downside transmission in the 10–20% regime before weakening again beyond 20%. The result is not simple convergence. Drawdowns reorganise the relative transmission roles inside the network.

Figure 2. Median lower-tail transmission shift across S&P 500 drawdown regimes.

Industrials move from a median tail shift of −3.57 in 0–5% drawdowns to +1.70 above 20%, while Technology moves from +1.22 to −1.63. Other sectors show smaller or less monotonic changes.

Figure 2

Sector transmission changes as drawdowns deepen

100%

Figure 2. Median lower-tail transmission shift across S&P 500 drawdown regimes.

Industrials move from a median tail shift of −3.57 in 0–5% drawdowns to +1.70 above 20%, while Technology moves from +1.22 to −1.63. Other sectors show smaller or less monotonic changes.

Download PNG

Analysis 03

Industrials and Technology shift in opposite directions

The opposing movement in Industrials and Technology is visible not only in their median values but in the underlying distributions. During 0–5% drawdowns, Industrials have a median tail shift of −3.57, indicating a weaker lower-tail transmission role relative to their median-state behaviour. During >20% drawdowns, the median moves to +1.70. Technology follows the reverse pattern. Its median moves from +1.22 in calm conditions to −1.63 once the drawdown exceeds 20%. The distributions remain wide, particularly in calm conditions, so neither sector occupies a fixed role at every point in time. But their central tendencies move in opposite directions as market losses become severe. This is important because stress does not simply make every sector a stronger transmitter. It changes which sectors become relatively more important to the transmission process.

Figure 3. Distribution of lower-tail transmission shifts for Industrials and Technology in calm and >20% drawdowns.

Industrials shift from a negative median in calm drawdowns to a positive median in deep drawdowns. Technology shifts from a positive median to a negative median, while both sectors retain wide distributions.

Figure 3

Industrials and Technology shift in opposite directions

100%

Figure 3. Distribution of lower-tail transmission shifts for Industrials and Technology in calm and >20% drawdowns.

Industrials shift from a negative median in calm drawdowns to a positive median in deep drawdowns. Technology shifts from a positive median to a negative median, while both sectors retain wide distributions.

Download PNG

Analysis 04

Shock transmission rotates in deep drawdowns

The cross-sector comparison makes the rotation more explicit. Industrials show the largest positive change between calm and severe drawdowns, increasing by approximately 5.27 points. Consumer Discretionary also moves towards stronger relative downside transmission. Technology shows the largest negative change, falling by approximately 2.85 points, followed by Utilities. Several other sectors remain closer to zero, suggesting comparatively smaller changes in their relative transmission role. The important feature is the dispersion across sectors. A deep drawdown does not push every part of the network in the same direction. Some sectors become more important transmission channels while others recede. System-wide stress therefore changes the composition of the network, not merely its overall intensity.

Figure 4. Change in median tail transmission between 0–5% and >20% S&P 500 drawdowns.

Industrials record the largest positive calm-to-deep-drawdown change at approximately +5.27 points. Technology records the largest negative change at approximately −2.85 points, with other sectors distributed between them.

Figure 4

Shock transmission rotates in deep drawdowns

100%

Figure 4. Change in median tail transmission between 0–5% and >20% S&P 500 drawdowns.

Industrials record the largest positive calm-to-deep-drawdown change at approximately +5.27 points. Technology records the largest negative change at approximately −2.85 points, with other sectors distributed between them.

Download PNG

Analysis 05

Each deep drawdown has a different transmission structure

The composition of downside transmission also changes substantially from one major drawdown to another. In the 2002 decline, Utilities show the strongest positive tail shift at +8.51, while most other sectors remain negative. The 2009 drawdown has a different structure, with Financials moving to +4.37 and Industrials to +2.24. During the 2020 drawdown, Energy becomes the dominant positive shift at +15.69, while Utilities, Technology and Healthcare move in the opposite direction. The 2022 decline rotates again. Industrials rise sharply to +10.83, Consumer Discretionary reaches +5.32, while Energy falls to −6.38 and Technology to −7.83. The largest drawdowns therefore do not share a fixed sector transmission hierarchy. Stress repeats. Its transmission structure does not.

Figure 5. Sector tail shifts across major peak-to-trough drawdowns.

Utilities lead the positive shift in 2002 at +8.51, Financials in 2009 at +4.37, Energy in 2020 at +15.69, and Industrials in 2022 at +10.83. Technology is negative in all four events and reaches −7.83 in 2022.

Figure 5

Each deep drawdown has a different transmission structure

100%

Figure 5. Sector tail shifts across major peak-to-trough drawdowns.

Utilities lead the positive shift in 2002 at +8.51, Financials in 2009 at +4.37, Energy in 2020 at +15.69, and Industrials in 2022 at +10.83. Technology is negative in all four events and reaches −7.83 in 2022.

Download PNG

Analysis 06

Technology's relative downside transmission role falls

Technology records a negative tail shift in each of the four major drawdowns examined. The magnitude becomes progressively larger across the events shown, moving from approximately −0.99 in 2002 to −2.44 in 2009, −4.14 in 2020 and −7.83 in 2022. This should not be interpreted as Technology becoming unimportant during market stress. The measure is relative: it compares the sector's lower-tail network role with its role under the median state. The finding therefore suggests that Technology's relative contribution to downside transmission weakens, even when the sector itself may remain highly exposed to the broader market shock. That distinction is important. Large losses and systemic transmission are related concepts, but they are not the same thing.

Figure 6. Technology tail shift across major >20% peak-to-trough drawdowns.

Technology's tail shift is −0.99 in 2002, −2.44 in 2009, −4.14 in 2020 and −7.83 in 2022, becoming progressively more negative across the four events shown.

Figure 6

Technology's relative downside transmission role falls

100%

Figure 6. Technology tail shift across major >20% peak-to-trough drawdowns.

Technology's tail shift is −0.99 in 2002, −2.44 in 2009, −4.14 in 2020 and −7.83 in 2022, becoming progressively more negative across the four events shown.

Download PNG

Methodology

How the analysis is constructed

Sample period
2000–7 August 2026
Drawdown benchmark
SPY, used for S&P 500 drawdown measurement
Data source
Yahoo Finance via yfinance. Daily auto-adjusted closing prices.
Data through
7 August 2026
Sector universe
  • XLE Energy
  • XLF Financials
  • XLI Industrials
  • XLK Technology
  • XLP Consumer Staples
  • XLU Utilities
  • XLV Healthcare
  • XLY Consumer Discretionary

Data

Yahoo Finance via yfinance. Daily auto-adjusted closing prices.

Measures

Sector return dynamics are estimated using quantile vector autoregressions, allowing transmission structure to vary across different parts of the conditional return distribution rather than imposing a single average relationship.

Quantile VAR

Sector return dynamics are estimated using quantile vector autoregressions at Q05, lower-tail state; Q50, median state; and Q95, upper-tail state.

This allows the transmission structure to vary across different parts of the conditional return distribution rather than imposing a single average relationship.

Total connectedness index

Forecast-error variance decomposition is used to measure the share of forecast uncertainty attributable to shocks originating in other sectors. The total connectedness index summarises the aggregate level of cross-sector transmission.

Net transmission

For each sector:

NET = shocks transmitted TO others − shocks received FROM others

  • Positive values indicate a net transmitter.
  • Negative values indicate a net receiver.

Tail shift

The central measure for this report is:

Q05 NET − Q50 NET

  • Positive values indicate that a sector becomes relatively more important as a transmitter in the downside tail.
  • Negative values indicate that its relative transmission role weakens.

Drawdown design

S&P 500 drawdown is measured relative to the preceding market peak. The four regimes are 0–5%, 5–10%, 10–20% and >20%.

For the event analysis, major episodes are defined using true peak-to-trough drawdown cycles rather than individual observations within an ongoing decline. The four >20% events retained with sufficient QVAR observations are 2002 at 47.5%, 2009 at 55.2%, 2020 at 33.7% and 2022 at 24.5%.

Research workflow

Data
Daily SPY and sector ETF prices
Returns
Daily sector returns
State model
Quantile VAR at Q05, Q50 and Q95
Network measure
Generalised FEVD connectedness
Sector measure
Net directional connectedness
Tail measure
Q05 NET − Q50 NET
Drawdown measure
S&P 500 peak-to-trough decline
Regimes
0–5%, 5–10%, 10–20%, >20%
Event test
Major >20% peak-to-trough cycles
Interpretation
Change in sector transmission relative to ordinary market conditions

Aggregate connectedness captures the overall level of transmission. Sector tail shifts show where that connectedness comes from as market conditions deteriorate.

Research implications

What the evidence shows

Connectedness alone does not describe the full network.

  • 01Aggregate connectedness across Q05, Q50 and Q95 follows broadly similar long-run dynamics.
  • 02Sector-level tail transmission varies materially across drawdown regimes.
  • 03Some sectors change from relatively weaker to stronger downside transmitters as losses deepen.
  • 04Industrials and Technology show particularly clear opposing changes.
  • 05Major historical drawdowns exhibit different sector transmission structures.
  • 06No single sector hierarchy characterises all severe market declines.
  • 07Technology's relative downside transmission role is negative across all four >20% events examined.

Together, these results suggest that monitoring only the total level of connectedness can conceal important changes beneath the surface. Two periods may display similar system-wide connectedness while presenting quite different sources and pathways of shock transmission. For portfolio and systemic-risk analysis, the relevant question is therefore not simply: How connected is the market? It is also: Where is that connectedness coming from?

Publication record

Sources and disclosures

Data

  • Yahoo Finance, accessed through yfinance.
  • SPDR sector ETFs: XLE, XLF, XLI, XLK, XLP, XLU, XLV and XLY.
  • SPDR S&P 500 ETF Trust, SPY.
  • Data through 7 August 2026.

Methodology reference

  • Koenker, R. and Bassett, G. (1978). “Regression Quantiles.” Econometrica, 46(1), 33–50.
  • White, H., Kim, T.-H. and Manganelli, S. (2015). “VAR for VaR: Measuring Tail Dependence Using Multivariate Regression Quantiles.” Journal of Econometrics, 187(1), 169–188.
  • Diebold, F. X. and Yilmaz, K. (2012). “Better to Give than to Receive: Predictive Directional Measurement of Volatility Spillovers.” International Journal of Forecasting, 28(1), 57–66.

Software

Python, pandas, NumPy, statsmodels, SciPy, matplotlib and yfinance. The yfinance package provides programmatic access to Yahoo Finance market data.

This publication has been prepared by LB Research for informational and educational purposes only. It does not constitute investment advice, investment research as defined under applicable regulation, a recommendation, or an offer to buy or sell any financial instrument.

The analysis is based on publicly available market data and the methodologies described within this publication. While reasonable care has been taken in preparing this report, no representation or warranty is made regarding the accuracy, completeness or timeliness of the information presented.

Any opinions expressed reflect the author's judgement at the publication date and may change without notice. Past performance is not indicative of future results.

Readers remain solely responsible for their own investment decisions.

© 2026 LB Research. All rights reserved. Questions or feedback: Contact LB Research

Download canonical PDF