Rather than assuming leadership a priori, we let the data reveal which companies tend to move first.
We conclude by revisiting our main research questions and summarizing the empirical insights provided by the leader–follower analysis.
How do we define a “leader” and a “follower” in stock movements?
A leader is a stock whose past daily returns improve the prediction of another stock’s future returns. A follower is a stock that reacts with a short delay to movements in the leader.
In practice, leadership is assigned only when two conditions are met:
- a statistically significant lead–lag correlation at a positive time shift, and
- a Granger causality test confirming predictive precedence in one direction.
How can directional influence be detected within sectors?
Directional influence is detected through a two-step statistical pipeline. First, cross-correlation analysis identifies candidate lead–lag relationships. Second, Granger causality tests verify whether past returns of one stock improve prediction of another beyond its own history.
Only relationships supported by both steps are retained, ensuring that detected links reflect directional predictability rather than simple co-movement.
Can daily return time series reveal short-term influence?
Yes. Daily return series are sufficiently granular to reveal lead–lag effects over horizons of a few trading days. While these effects are moderate in magnitude, they are statistically meaningful and consistently detected across multiple sectors.
Are leader–follower dynamics consistent across sectors?
No. Leadership patterns are strongly sector-dependent. Some sectors exhibit structured hierarchies with multiple leaders, others display only a few isolated leader–follower pairs, and several sectors show minimal detectable leadership.
Sectors such as Health Care, Energy, and Transportation tend to form denser and more structured leadership networks, suggesting faster or more coordinated information transmission within these industries.
Detecting Leadership in Stock Movements
Financial markets rarely move in isolation. Price changes in one firm are often followed—sometimes within days—by reactions in others. We search for short-term leader → follower relationships within sectors and visualize how information appears to flow across companies.
What “leadership” means here
An arrow from A to B means that A tends to move first and B tends to react afterward over a short delay. This is a statistical notion of predictive precedence, not a claim of true economic causality.
How a link is selected
We first screen for short lead–lag alignments between two return series across small delays (up to 7 trading days). Then we keep only pairs that pass a directional predictability test (Granger causality) in one direction but not the reverse.
Mathematical details
Lead–Lag Cross-Correlation
For each ordered pair of stocks \( (i,j) \) within a sector, we test whether movements in \( i \) tend to precede movements in \( j \) by computing:
\[ \rho_{ij}(k) = \mathrm{Corr}\!\big(r_{i,t},\, r_{j,t+k}\big), \qquad k = 1,\dots,7 \]Granger Causality Test
To establish directionality, we compare a baseline autoregressive model to an augmented model that includes lagged returns of the leader:
\[ r_{j,t} = \alpha + \sum_{\ell=1}^{p} \beta_\ell r_{j,t-\ell} + \sum_{\ell=1}^{p} \gamma_\ell r_{i,t-\ell} + \varepsilon_t, \qquad p = 3 \]Interactive Sector Leadership Heatmap
This heatmap shows, for each sector, how strongly each stock (rows) appears to lead others (columns) based on our cross-correlation.
Leader–Follower Network Graph (per sector)
Each node is a stock. Arrows point from leader → follower. Edge thickness reflects the strength of the relationship (|correlation|).
• Nodes represent companies in the selected sector.
• Arrows point from the Leader → Follower stock.
• Thickness of arrows represents the strength of the statistical link (|correlation|).
• Hover over a node to see the ticker.
• Drag nodes to explore the structure.
Sector-by-Sector Leadership Insights
The analysis identifies sparse but structured leader–follower relationships within sectors. Rather than dense interactions, leadership effects concentrate around a limited number of firms and propagate in clearly defined directions.
Basic Industries
Leadership is shared among several industrial firms, with Steel Dynamics (STLD), WD-40 Company (WDFC), and Matrix Service (MTRX) acting as upstream movers. Their influence propagates toward firms such as Stericycle (SRCL) and Codexis (CDXS), forming a layered but non-centralized structure.
Capital Goods
A small group of technology-oriented firms drives leadership. II-VI (IIVI) and Mercury Systems (MRCY) initiate movements that are followed by Nordson (MKSI) and FLIR Systems (FLIR), producing clear and directional influence paths.
Consumer Durables
Leadership is fragmented across independent channels. American Superconductor (AMSC), Central Garden & Pet (CENT), and iRobot (IRBT) each lead specific followers, resulting in localized influence rather than a single sector-wide hierarchy.
Consumer Non-Durables
The sector is characterized by a few strong but isolated links. Fossil Group (FOSL) leads Crocs (CROX), while Columbia Sportswear (COLM) influences Sanderson Farms (SAFM), with limited interaction beyond these pairs.
Consumer Services
Leadership is highly concentrated. DISH Network (DISH) emerges as the sole leader, with its movements preceding those of Starbucks (SBUX), indicating a narrow but detectable information channel.
Energy
Energy exhibits one of the strongest leadership structures. Firms such as TUSK Energy (TUSK), Diamondback Energy (FANG), and Viper Energy (VNOM) influence multiple downstream companies, including Centennial Resource Development (CDEV), reflecting strong and widespread propagation effects.
Finance
Leadership is distributed among several financial institutions. Carlyle Group (CG), SEI Investments (SEIC), and Principal Financial (PFG) influence firms such as Ameritrade (AMTD) and T. Rowe Price (TROW), forming a sparse but multi-path network.
Health Care
Health Care displays rich internal structure. Biomarín (BMRN), Intuitive Surgical (ISRG), and Gilead Sciences (GILD) act as prominent leaders, influencing firms such as Vertex Pharmaceuticals (VRTX) and Hologic (HOLX) through both positive and negative effects.
Miscellaneous
Leadership is spread across platform-oriented firms. HealthEquity (HQY), CoStar Group (CSGP), and MercadoLibre (MELI) influence several followers, including Zillow Group (ZG) and GoPro (GPRO), producing moderate but consistent directional links.
Public Utilities
The sector shows multiple localized leadership relationships rather than a dominant hub. Firms such as NextEra Energy (NEXT), ADTRAN (ADTN), and Spark Energy (SPKE) influence utilities like Clean Energy Fuels (CLNE) and Casella Waste Systems (CWST).
Technology
Leadership is driven by large platform and semiconductor firms. Broadcom (AVGO) leads both Apple (AAPL) and Microsoft (MSFT), while Meta Platforms (FB) and Adobe (ADBE) form an opposing influence channel, highlighting delayed information transmission.
Transportation
Transportation exhibits multi-channel leadership across logistics and airlines. Old Dominion Freight Line (ODFL), Hub Group (HUBG), and JetBlue (JBLU) influence firms such as Ryder (RYAAY) and Werner Enterprises (WERN), forming a well-connected but non-centralized network.
Takeaway
Across sectors, leadership effects are directional, heterogeneous, and concentrated. A limited set of firms consistently act as short-term influencers, while most stocks respond passively or exhibit no detectable leadership dynamics.