Lead-Lag Relationships Across Assets

A large index future can react to macro news within seconds, while the individual stocks that make up that index — especially the more thinly traded ones — take longer to fully reprice. For a short window, the future is telling you information the laggard stock hasn't absorbed yet. A lead-lag strategy trades that gap: watch the leader move, then trade the follower before it catches up.

Where the lag comes from

The mechanism is almost always about speed of price discovery, not about one asset literally causing the other to move. A liquid, widely-watched instrument (an index future, an ETF, a large-cap bellwether) absorbs new information into its price almost immediately because it trades constantly and is watched by everyone. A less liquid or less-covered instrument exposed to the same underlying news — a smaller stock in the same sector, a related but thinner futures contract — reprices more slowly, because fewer participants are watching it closely enough to trade on the news the instant it lands. The gap between those two speeds is what a lead-lag strategy captures.

Worked example

A statistical test (Granger causality on 1-minute bars) finds that a sector ETF's return in minute $t$ has meaningful predictive power for a thinly-traded stock's return in minute $t+3$, but not the reverse. On a day the ETF jumps 0.8% in one minute on a sector-wide catalyst, and the stock has only moved 0.2% so far, a lead-lag desk buys the stock expecting it to converge toward roughly the same 0.8% move over the next few minutes as slower participants catch up and trade it — capturing the remaining 0.6 points of expected catch-up, adjusted for the stock's usual beta to the sector.

What this means in practice

Lead-lag effects live on very short horizons — minutes, sometimes seconds — because that's how long the information gap between a fast and slow instrument typically persists before arbitrageurs and slower traders close it themselves. That makes the strategy dependent on low-latency execution and cheap trading costs; the same signal on daily bars is usually too small, and too eroded by fees, to trade profitably.