RateWalk: Forecasting Fed Decisions
Naive inflation-regime conditioning overfits; empirical-Bayes shrinkage fixes it in the US, UK and Germany, and a blend with a market proxy reaches log-loss 1.056 and a 72.7% hit rate.
Research / Too Big to Co-Move
Too Big to Co-Move: Concentration and Sector Connectedness Around the World
The ten largest firms rose from 24% to 41% of S&P 500 market capitalization between 2018 and 2025, and by August 2026 connectedness among the eleven sectors had reached its lowest level in the sample. Does concentration explain the disconnection?
Key results
-0.75 pts
monthly R² against the other ten sectors per one-point rise in top-three share, uncapped U.S. returns (permutation p = 0.04)
78%
of 933 sector-years, across eleven markets, in which a sector's three largest firms co-move less with the market than the rest of the sector
3 of 10
other markets where the concentration link reappears: Japan, Canada and, more weakly, the United Kingdom
< 1 pt
change in U.S. average squared correlation from holding top-three weights at 2018 levels, against a 28-point fall since 2018
88.9%
peak total connectedness (16 March 2020); sample low 53.7% (August 2026)
0 of 110
Granger lead-lag links surviving heteroskedasticity-robust tests; classical tests appear to find 90
Abstract
The ten largest firms rose from 24% to 41% of S&P 500 market capitalization between 2018 and 2025, and by August 2026 connectedness among the eleven GICS sectors had reached its lowest level in the sample. This paper asks whether concentration explains the disconnection, using daily returns of the Select Sector SPDR funds from October 2018 to August 2026, an uncapped constituent panel with point-in-time membership, a filtered TVP-VAR, and constituent panels for ten other major markets. Concentrated sectors co-move less with the rest of the market. On uncapped U.S. returns, a one-point rise in a sector's top-three share lowers its monthly R² against the other ten sectors by 0.75 points, and the result holds when concentration paths are randomly reassigned across sectors (p = 0.04). The mechanism is granular: in every market studied, a sector's three largest firms co-move less with the market than the rest of the sector, in 78% of 933 sector-years. The link between concentration and co-movement reappears in Japan, Canada and, more weakly, the United Kingdom, but not in the other seven markets. Everywhere, the shift in weight toward the largest firms is small in aggregate, and sector co-movement fell after 2025 in eight of eleven markets, including several where concentration fell. Net directional connectedness, the usual way to read such results, depends on how the variance decomposition is normalized and overstates the link. Concentration explains which sectors are least connected, not why markets disconnected.
Methods
- Filtered TVP-VAR connectedness with forgetting factors (Antonakakis, Chatziantoniou and Gabauer 2020), so every date uses only data available on that date
- Uncapped constituent returns on point-in-time S&P 500 membership: historical share counts, the 2023 GICS reclassification, dual-class deduplication, 78 since-removed firms recovered; the same design for ten other markets
- Panel regressions with sector and month fixed effects, Driscoll-Kraay errors, a wild cluster bootstrap and a permutation test that reassigns concentration paths across sectors
- Firm-level decomposition of every sector into its three largest firms and the rest, plus a counterfactual that holds top-three weights at their 2018 levels
What it does not show
- The cross-market link is weak: it appears in Japan, Canada and, more weakly, the United Kingdom, not in the other seven markets, and other-market panels use current index membership.
- Concentration explains which sectors are least connected, not why markets disconnected: holding top-three weights at 2018 levels moves the U.S. result by less than one point.
- The usual net-directional reading depends on how the variance decomposition is normalized and overstates the link: under the scalar normalization the concentration effect on net connectedness disappears.
JEL: C12, C23, C32, G12, L11
Cite
BibTeX
@article{yadav2026toobig,
title = {Too Big to Co-Move: Concentration and Sector Connectedness Around the World},
author = {Yadav, Tanishk},
year = {2026},
journal = {SSRN Electronic Journal},
note = {Preprint, abstract 6475898}
}
Tanishk Yadav
MS Financial Engineering at New York University, graduating May 2027, and looking for full-time quantitative research roles. My master's thesis measures, in real time, when US debt dynamics turn self-reinforcing.
More research
Other papers
Calm Months, Crash Months
Built with only past data, capped leverage and trading costs, volatility management lifts momentum's Sharpe ratio from 0.43 to 0.94 and helps none of the other five factors. A closed-form formula for momentum's market beta explains why.
VolEdge: The Variance Risk Premium
A full-stack platform that extracts risk-neutral moments from live option chains without a pricing model; a GARCH vol-targeted sleeve beat SPY on risk (Sharpe 0.85 vs 0.68) while naive short-variance harvesting lost money.
Multi-Scale Regime Detection
Four independent stress signals that must agree before de-risking: maximum drawdown -6.2% vs -41.7% for SPY, and the same calibration transfers to factor and international portfolios.