Case studies › KBE · Rates
Bank stocks in November are a net-interest-margin story with a credit tail
A payrolls beat of 165,000 against 59,000 expected, with wages at 3.1%, does two things to a bank book in the same week. It supports the December hike, which helps asset yields, and it lifts deposit competition, which hurts funding.
The first is quoted. The second is the one that decides the margin.
Deposit beta is the whole trade
Deposit beta is the share of a rate move passed through to depositors. At a beta of zero, a hike is pure margin. At a beta above one, the funding cost rises faster than the asset yield and the margin compresses even though rates went up.
| Scenario | Asset yield | Deposit cost | Net margin |
|---|---|---|---|
| Beta 0.0, +25bp | +25bp | 0bp | +25bp |
| Beta 1.0, +25bp | +25bp | +25bp | flat |
| Beta 1.4, +25bp | +25bp | +35bp | -10bp |
A bank reporting a rising margin in a rising-rate quarter is reporting a low beta. A bank reporting a flat or falling one is reporting the opposite, and it is telling you about its deposit franchise rather than about the rate.
The tell is the deposit line before the margin line. Net outflows with a rising cost of deposits means the franchise is paying to keep balances, and that is a beta of more than one in practice.
Why November specifically
A December hike at better than 80% probability means the market has already marked the asset side. The deposit side reprice is contractual and lagged: promotional rates, savings tiers and CD rollovers move over one to three months, after the market has moved.
So the fourth quarter contains the asset benefit and the first quarter contains the funding cost. A bank that looks good in Q4 and worse in Q1 has not changed its business, it has changed its repricing schedule.
The credit tail
Provisions have been benign because employment has been. Those are the same variable. Every payrolls print is a provision input with a two-to-three quarter lag, and 4.1% unemployment is why loss rates are where they are.
A strong employment print is therefore margin-positive and credit-negative in sequence: it supports the hike that compresses the margin now, and it holds the unemployment rate that keeps provisions low later. Reading either leg alone gets the direction wrong.
The segment detail matters for the tail. Professional and business services added 70,000. That is the white-collar cohort that services commercial real estate and leveraged lending exposure. Manufacturing added 19,000, which supports industrial credit.
What a screen should sort on
Not the yield. The order is: cost of deposits, then loan-to-deposit ratio, then non-performing accruals, then the securities book duration. Duration is the fourth item because a long book is a mark-to-market item, and marks are the mechanism through which a margin problem becomes a capital problem.
A bank with a high loan-to-deposit ratio and a long securities book is the worst combination into a hike: it needs wholesale funding at the same time as its held-to-maturity book is under water.
Where the case breaks
A soft CPI on Tuesday reduces the hike probability and removes the asset-yield leg while leaving the deposit cost in place, because promotional rates are already contracted. That is the bad version: margin compression without the offset.
The credit tail breaks if unemployment turns before provisions do, which is the normal sequence. It would show as a rise in non-performing accruals two quarters out, at which point the margin question is the smaller one.
What the arithmetic says
A 25 basis point hike with a beta above one is a margin loss. October's print supports the hike and supports the beta. The screen is a beta read and a deposit-franchise read, and the credit tail sits two quarters behind the employment number that keeps it quiet.
A hike at a beta above one loses ten basis points on a 25 point move. November pays the asset yield. January pays the deposit cost. Same business, different calendar.
Sources
Payroll, unemployment and sector figures are Bureau of Labor Statistics data with consensus from the Dow Jones estimate. Rate probability is third-party market pricing. The margin table is arithmetic on the stated betas and can be re-run. The repricing lag and screening order are analytical.
Published for information only. Nothing here is investment advice or a recommendation. This note is not prepared in accordance with the legal requirements designed to promote the independence of investment research. No position is held in any security named.