Why the next deposit decision requires more than a competitor-rate comparison.
When rates move quickly, banks can learn the wrong lesson from aggregate deposit flows. Two customers with similar balances may have very different alternatives, relationship depth, and sensitivity to an offered rate.
A useful pricing model separates behavior by product, tenure, balance band, and customer context. It asks not only whether a customer moved funds but when, to where, and after which interactions. A simple average beta hides the customers most likely to respond.
The relevant counterfactual is also difficult: some balances would have stayed without an incentive, while others would have left even with one. Research and modeling should make those uncertainties visible rather than presume every retained dollar was saved by a rate change.
The result is a more selective strategy: where to compete on price, where service and relationship matter more, and where the cost of retention is unlikely to be justified.