Yield Curve Shape and Bank Net Interest Margin

The default bank business model is 'borrow short, lend long.' Deposits are typically short-duration (checking accounts can be withdrawn instantly; CDs lock in for a few years at most). Loans are typically longer-duration (mortgages run 15-30 years; commercial real estate often 5-10). When the yield curve slopes upward, that mismatch is profitable — the bank earns the long-rate, pays the short-rate, and keeps the spread.

Deposit beta is the bank-by-bank variable that matters most. Deposit beta is the fraction of a Fed rate move that the bank passes through to depositors. A bank with deposit beta of 20 percent passes only 20 cents of every dollar of rate hike to its depositors — keeping the other 80 cents as widened margin. A bank with deposit beta of 80 percent passes most of the hike through and sees little margin expansion when rates rise. Sticky, low-beta deposits (checking accounts, small-business operating accounts) are the most valuable funding a bank can have.

Correlation is not causation in the curve-NIM relationship. Curve inversion typically reflects expectations of future rate cuts, which usually accompany economic weakening, which usually brings higher loan losses. Some of the NIM compression observed during inversions actually reflects credit deterioration showing up alongside the rate dynamic. When analyzing a bank during inversion, separate the pure rate effect (visible in NIM) from the credit effect (visible in net charge-offs and provisions).

Banks are not a uniform bet on the yield curve. A bank with sticky low-beta deposits, well-laddered asset maturities, and disciplined credit underwriting can hold NIM in a flat curve while peers compress. When the curve normalizes, the bank with retained margins is also the one whose loan book did not deteriorate — and that combination produces the cleanest cycle-over-cycle compounding among bank stocks.