Deposit Betas Across Three US Rate Cycles, 2015–2026: Partial Adjustment, a Closed-Form Beta Path and the Value of the Deposit Franchise

A deposit beta is the share of a change in the policy rate that a bank passes on to its depositors. I measure betas for US national average deposit rates published by the FDIC against the effective federal funds rate over three cycles: the 2015 to 2019 hikes (fed funds up 2.28 points), the 2022 to 2023 hikes (up 5.25 points) and the 2024 to 2026 cuts (down 1.70 points). Savings-account betas were 0.017, 0.070 and 0.047; 12-month CD betas were 0.174, 0.309 and 0.176. I fit a partial-adjustment error-correction model and prove that it implies a cumulative beta path of beta plus (gamma minus beta) times (1 minus lambda) to the power t, which a simulation reproduces to four decimals. On the FDIC series from April 2021, the long-run beta is 0.079 for savings and 0.362 for 12-month CDs, with gap half-lives of 2.3 and 4.8 months; on the earlier series, savings rates adjusted with a half-life of 42.9 months. I derive the value of a deposit franchise when deposits run off at rate delta and show that its sensitivity to the policy rate is ((1 minus beta) delta plus c) over (r plus delta) squared, positive for all parameter values. For a savings beta of 0.070, a 10% run-off rate and a 2% operating cost, a rise in rates from 1% to 5% raises the franchise value by 27 cents per dollar of deposits. Low betas are the reason banks gain from higher rates, and the reason their deposits are less stable than the low beta suggests.