Leading and Lagging Indicators
Some economic data moves before the economy turns, some moves after — knowing which is which is the difference between forecasting a recession and confirming one that already happened.
Prerequisites: GDP and the National Accounts
The unemployment rate rises after a recession has already begun — it is one of the last things to move, because businesses cut hours and spending before they lay off staff. New building permits, by contrast, tend to fall months before a downturn shows up anywhere else, because builders sense weakening demand before it hits GDP. The distinction between data that moves ahead of the cycle and data that confirms it after the fact is what separates a leading indicator from a lagging one.
Leading indicators turn before the economy does and are useful for forecasting; lagging indicators turn after and are useful for confirming a turn already happened. Coincident indicators move roughly in step with the cycle itself. Mixing them up means reacting to news that's already stale.
What sits in each bucket
Common leading indicators: building permits, new orders for manufactured goods, the yield curve slope, stock prices, and consumer expectations surveys. Each reflects a decision made before its economic effect shows up — a builder applies for a permit months before a house is finished and sold, a manufacturer books an order before production ramps.
Common coincident indicators: industrial production, personal income, and payroll employment — these move roughly alongside the cycle itself, useful for describing where the economy is right now.
Common lagging indicators: the unemployment rate, average duration of unemployment, and commercial lending rates — these confirm a cycle turn only after it's underway, because businesses adjust staffing and banks adjust credit terms in response to conditions that already changed.
Worked example
Suppose building permits (leading) have fallen 8% over six months, the yield curve (leading) has been inverted for four months, industrial production (coincident) is still flat, and the unemployment rate (lagging) hasn't moved at all. Read individually, this looks contradictory — "the economy is fine, unemployment is low." Read as a set, it's coherent: leading indicators are warning of a slowdown that hasn't yet reached coincident data, and won't show up in the lagging unemployment rate until well after the slowdown is already underway, if it materializes. A forecaster weighting only the unemployment rate would miss the warning entirely; one weighting only permits might call a recession that some other force (like a rate cut) prevents from arriving.
What this means in practice
Composite indices like the Conference Board's Leading Economic Index bundle several leading series into one number specifically to smooth out the noise any single series carries. Traders use the distinction to decide which data to react to hardest: a rates desk pricing in cuts cares far more about a leading indicator rolling over than about the lagging unemployment rate finally confirming what everyone already priced in months earlier.
Leading indicators are early, not infallible — they have flagged recessions that didn't arrive, and a genuinely leading series can occasionally lag in a specific cycle because of a one-off distortion. Treat them as probabilistic signals to weigh together, not a single switch that flips.
Related concepts
Practice in interviews
Further reading
- Conference Board, 'The Leading Economic Index'