Why they lead in the first place
Economists sort data by when it moves. Leading indicators turn before the economy does, coincident ones move alongside it, and lagging ones just confirm what already happened. That split isn't arbitrary. It falls out of how activity actually flows. Orders come before production. Production comes before hiring. Hiring feeds income, income feeds spending, and spending loops back into more orders.
Leading indicators work because they sit at the front of that chain. When factories start booking fewer new orders, it takes weeks or months before that becomes lower output, longer still before it becomes layoffs, and longer again before it shows up in what people are buying. Watch the early links and you get a look at where things are headed before they get there.
The Conference Board's ten pieces
The Leading Economic Index from the Conference Board is the composite everyone quotes for the US. It rolls up ten components, each chosen because decades of data show it tends to turn ahead of the business cycle rather than with it or after it.
A couple come from the labor market. Average weekly manufacturing hours matter because employers cut hours before they cut heads. Initial jobless claims matter because new filings climb before layoffs go broad. A couple come from the factory floor: new orders for consumer goods and materials, and new orders for non-defense capital goods excluding aircraft. Those catch demand shifts early in the pipeline, roughly what the ISM new orders index is picking up.
Finance contributes three. The S&P 500, since stock prices are a bet on the future. The spread between the 10-year Treasury and the fed funds rate, which is the yield-curve signal. And the Leading Credit Index, a read on how tight financial conditions are. The last two, housing and the consumer, come from building permits for new private housing and the University of Michigan survey of consumer expectations. Permits lead because someone decides to build months before anyone pours concrete, and sentiment shifts before wallets do.
Why some lead by more than others
The ten don't share a lead time. Building permits run out ahead by roughly 9 to 12 months, because the gap between filing a permit and the actual construction, furnishing, and related spending is long. Initial claims lead by more like 3 to 6 months, since they track today's labor market pretty directly.
Stocks are the odd one. They've been in the LEI since the start and they do lead on average, but they're noisy about it. The market, as Paul Samuelson put it, has predicted nine of the last five recessions. Equities sell off hard sometimes and no recession follows, which is exactly why the stock component earns its keep inside a composite rather than on its own.
The credit and rate pieces have grown more important lately, as financial conditions do more of the work of pushing monetary policy into the real economy. When credit tightens, it bleeds into business investment, consumer borrowing, and housing, and you see it in GDP many months down the road.
Reading the composite
The LEI usually gets reported as a month-over-month percent change plus a six- or twelve-month change to smooth the noise. The Conference Board's own rule of thumb is that three straight monthly declines alongside a negative six-month change flag recession risk. Historically that combo has led recessions by anywhere from 7 to 20 months, averaging around 10.
Size matters too. A six-month annualized drop of negative 4% or worse has come before every recession in the modern era. Shallower dips sometimes clear on their own, especially when one or two noisy components are doing the damage, like the stock piece dragging the index down during a correction that never turns into a real contraction.
One practical note. The LEI ships with a one-month lag, so January's number lands in late February. If you want faster reads, most of the underlying components are available sooner. Claims are weekly, stock prices are live, building permits are monthly and barely get revised.
Where it falls short, and what else is out there
The mix has been tweaked several times over the years, and the fair criticism is that some components have aged as the economy tilted from making things to providing services. The manufacturing-heavy pieces, new orders and factory hours, can understate a services-driven economy and throw off the signal when factories shrink while services keep growing.
There are other composites if you want them. The OECD builds leading indicators for member countries on similar logic. ECRI runs its own proprietary index with a longer track record in some respects. And plenty of analysts just wire up their own dashboards from whatever components fit their horizon.
The point isn't that one composite is the right one. It's that stacking several forward-looking series gives you a cleaner forecast than staring at any single one. The same reason diversification helps a portfolio helps here too. The errors in the individual pieces partly cancel, and what's left is a steadier read on which way the economy is leaning. When I'm sizing up macro risk, I treat the LEI as one input among several rather than a verdict, and I lean on the faster components between releases.