When rates fall, your bank fees don't: the repricing gap treasurers
Operating-account yields and earnings credit rates reprice on the bank's schedule, not yours. The drag rarely shows up in a forecast because almost nobody reconciles the account analysis statement to the rate cycle.
Every easing cycle produces the same quiet arithmetic problem inside corporate treasury. The yield on operating cash falls the moment the policy rate moves. The price of the banking services that cash is paying for does not. The gap between those two curves is real money, it lands in interest income and bank fees rather than in any line a business unit owns, and in most companies nobody is explicitly assigned to close it.
The asymmetry sits in the account analysis statement
Most large corporates pay for cash management two ways. Hard-dollar fees are invoiced and visible. Soft-dollar compensation runs through the earnings credit rate - the notional yield a bank applies to collected balances in an operating account, which is then used to offset service charges for wires, lockbox, ACH origination, positive pay, and account maintenance. The mechanics are documented monthly in the account analysis statement, a file most treasury teams receive, few load into anything, and fewer still reconcile against the terms in the pricing agreement.
The asymmetry is structural. Earnings credit rates are administered rates. Banks generally move them down promptly when the policy rate falls, because the funding value of the deposit has fallen. Service unit prices, by contrast, are set in a negotiated schedule that typically resets on an annual or multi-year cycle, and frequently carries an escalator. So the value of the balances a company leaves on deposit to pay for services erodes faster than the cost of the services themselves. The result is a growing hard-dollar residual - the portion of the fee bill the credit no longer covers - that arrives as an invoice nobody forecast.
There is a second-order version of the same problem on the interest-bearing side. Rates on hybrid and interest-bearing operating accounts are administered too. When the cycle turns, the deposit beta that mattered on the way up - how much of each hike the bank passed through - matters again on the way down, and it is rarely symmetric.
The diagnostic almost nobody runs
The work here is not sophisticated, which is part of why it goes undone. Pull twelve to twenty-four months of account analysis statements across every bank in the structure, normalize the service codes so the same activity is comparable across providers, and plot three series against the policy rate: the effective earnings credit rate actually applied, total service charges, and the hard-dollar residual. If the first line has tracked policy down while the second has been flat or rising, the company has an unpriced repricing gap.
Then ask what the balances are actually earning on a risk-adjusted basis. A dollar sitting in an operating account to generate earnings credit is only rational if the credit it produces exceeds what the same dollar would earn in a government money market fund or a Treasury bill, net of the incremental hard-dollar fees the company would then have to pay in cash. That break-even moves every time the earnings credit rate moves. In a falling-rate environment it moves against the deposit - and against the target balance that was set two cycles ago and never revisited.
Timing is the last piece. Money market fund yields lag the policy rate because portfolio weighted-average maturity carries older paper for weeks after a cut; direct bill purchases reprice immediately at auction. Neither behaves like an administered deposit rate. A treasurer who models all three as a single blended assumption will be wrong in both directions, and will be wrong at exactly the points in the quarter when interest income is being explained to the audit committee.
Renegotiation has a window, and it is now
Bank pricing conversations are easier to open before the fee residual becomes visible than after. The credible ask is not a blanket discount. It is a floor on the earnings credit rate, or an explicit formula tying it to a published benchmark with a stated spread, so that the pass-through on the way down is contractual rather than discretionary. Failing that, a multi-year freeze on unit prices for the highest-volume services buys back most of the same economics.
Volume is the leverage. Treasurers who can show a bank the full picture - deposits, payment volumes, card, FX, trade, and the credit facility the relationship is priced around - negotiate differently from those presenting one account's worth of activity. That argues for doing the analysis at the group level before the first meeting, not bank by bank as invoices arrive.
It also argues for knowing which services are genuinely priced above market. Benchmarking is imperfect, but standardized service codes make cross-bank comparison possible within a company's own footprint, and that internal spread is often wide enough to make the case without external data.
The concentration question nobody re-asked
The 2023 regional-bank episode pushed a lot of treasury organizations into multi-bank structures, additional custody relationships, and reciprocal deposit arrangements that spread balances across insured limits. Those structures carry ongoing cost: more accounts, more connectivity, more reconciliation, more people. When the perceived risk recedes, the cost stays, and the quiet path of least resistance is re-consolidation back to one or two relationship banks - usually without a documented decision.
That is worth surfacing deliberately rather than by drift. A CFO should be able to answer, from a single page, how much operating cash sits with each counterparty, what share of it is uninsured, how quickly it could be moved, and what the concentration looks like relative to the policy the board approved after the last scare. If the answer requires a week of work, that is itself the finding.
The related question applies to invested cash. Cash parked in funds is only as conservative as the fund's underlying holdings, and prospectus flexibility varies more than the label on the vehicle suggests. Treasury policy should state not only which vehicle types are permitted but what the treasurer is expected to look through to - and how often.
Key takeaways
- Earnings credit rates fall with the policy rate; negotiated service prices do not. The residual arrives as an unforecast hard-dollar fee bill.
- Reconcile 12 to 24 months of account analysis statements against the rate path before opening any bank pricing conversation.
- Recalculate the break-even between earnings credit and market yields each time rates move - target operating balances set in the last cycle are probably wrong.
- Ask for a contractual floor or benchmark-linked formula on the earnings credit rate, not a one-off discount.
- Re-examine whether post-2023 multi-bank structures have quietly re-consolidated, and document the decision either way.
