Why slopes carry meaning
A normal upward-sloping curve means investors demand a term premium to hold long bonds — they want compensation for inflation, default risk, and the opportunity cost of locking up capital. When the slope flattens or inverts, the bond market is making a specific bet: that short rates will be lower in the future than they are today. Since the Fed sets short rates and only cuts during downturns, an inversion is shorthand for 'recession expected.'
The 2s10s and 3m10y benchmarks
Two slope measures get the most attention. The 2s10s spread (10-year minus 2-year) is the trader's favourite — it reacts quickly to Fed expectations. The 3m10y spread (10-year minus 3-month T-bill) is the academic's preference because it's the one the New York Fed uses in its recession-probability model. Both inverted before every US recession of the past 70 years, with an average lead time of 12-18 months from inversion to recession start.
- →2s10s inversion: Fast, noisier, more useful for trading rates and curve trades.
- →3m10y inversion: Slower, cleaner, the more reliable recession signal historically.
- →Both inverted simultaneously: Almost universal precondition for past US recessions.
The three movements that matter
The headline shape gets the press, but slope dynamics drive the trades. Bull steepening (long yields fall faster than short) signals growth worry and often coincides with risk-off equity moves. Bear steepening (long yields rise faster) signals inflation or supply concern — common during fiscal expansions. Bull flattening (short yields fall faster) is the early-cutting-cycle pattern. Bear flattening (short yields rise faster) is the late-hiking-cycle pattern that often precedes inversion.
Inversion vs un-inversion: which is the real signal?
The widely-cited recession lead time of 12-18 months actually measures from un-inversion — when the curve starts re-steepening after a period of inversion — not from the initial inversion itself. Most recessions in the past 50 years have begun shortly after the curve normalised. The market reads the un-inversion as confirmation that the Fed will be forced to cut, which it does because the economy is rolling over. Watch the steepening as carefully as the inversion.
What can break the signal
Quantitative easing distorts long-end yields by suppressing them artificially, which can keep the 10-year low even when growth expectations would normally lift it. Foreign demand for Treasuries — particularly Asian central banks recycling dollar reserves — does the same thing. Both effects mean the curve can stay flat or inverted for longer than usual without triggering a recession on the historical schedule. The signal still works, but the lead time becomes less reliable.