Many investors focus on either individual companies or broad market factors like value and size. But our research reveals a powerful middle layer: the network of connections between stocks.

Here's the pattern: when a group of "leader" stocks moves today, related "laggard" stocks tend to follow tomorrow. This relationship is strong enough to generate meaningful returns beyond what traditional investment strategies capture.

We use a data-driven approach by treating the network of stock connections like an image and analyzing it with computer vision techniques. This reveals patterns that standard factors like market, value, profitability, and even momentum cannot explain. We also show that much of what was previously called "factor momentum" actually stems from these stock-to-stock lead-lag effects. And the most important connections vary over time rather than staying fixed to categories like industry or geography.

Management insight: Investors and firms should view the stock market as a dynamic network, not a list of isolated tickers. Systematically identifying which stocks lead and which follow creates an additional source of returns. These cross-stock connections can also signal where risks may propagate (e.g., supply-chain stress, liquidity crunches), valuable for early-warning systems and stress testing. Firms can apply existing AI and image-processing infrastructure to better understand these complex financial and supply-chain networks.