The 3,000-dollar account

The median household checking balance at Fifth Third is about 3,000 dollars. Chief financial officer Bryan Preston brings that figure to the debate over artificial intelligence agents moving customers’ money to accounts offering higher interest. I start with what the money does. A balance supporting everyday payments serves a different purpose from excess cash that a customer can readily move elsewhere. Preston similarly describes much of the commercial deposit base as operational money attached to transaction services. Management’s defence rests on those elements of the customer relationship beyond price. The debate therefore includes the payments supported by a balance, alongside the capabilities of the agent proposing to move it.[1]

I take seriously the argument that an agent making rate comparisons and account selection easier can intensify competition. Spence identifies online banks offering value mainly through price as more exposed. The strongest version of that argument needs no assumption that every customer moves an entire balance at the same speed. Operational money and surplus cash seeking yield can respond differently. My inference is that understanding a bank’s funding cost requires attention to the use of balances as well as their size. A strong alternative remains: customers can become more sensitive to interest differences without changing their payment services. Management’s emphasis on relationships leaves that possibility open.[1]

More deposits, higher funding cost

The quarter offers a more concrete starting point for the cost of retaining deposits. Fifth Third’s average core deposits increased 3 per cent sequentially to 236 billion dollars. Over the same period, deposit costs rose 10 basis points to 1.64 per cent. The cost of interest-bearing deposits increased 11 basis points to 2.24 per cent. Read together, those figures make an automatic equation between quantity growth and cheap funding difficult to sustain. The bank reported more deposits while also paying a higher rate for them. Management associated this year’s stronger deposit competition with banks seeking funding for loan growth.[1]

My narrow reading is that a deposit carries both an amount and a cost on the bank’s balance sheet. Money in a customer’s account is the bank’s liability to that customer, with interest an expense attached to the relationship. Retaining the balance therefore leaves its funding terms open to change. Net interest margin falling 2 basis points to 3.34 per cent is also part of this quarter. I attribute no unique causal role to artificial intelligence agents in that movement. Deposit mix, pricing decisions and other interest income and expense components are alternative explanations. The report contains no measurement of an actual agent-driven deposit outflow; existing funding competition sits alongside discussion of a new tool.[1]

The terms of a service relationship

The bank’s present boundary matters: it does not allow artificial intelligence agents to initiate transactions from customer accounts. Preston wants authority and accountability clarified. At the same time, chief executive Tim Spence is considering subscription review and budgeting functions inside the bank’s own application. Providing account-management conveniences internally fits management’s emphasis on strengthening the relationship. That connection is my interpretation, rather than an announced result for customer behaviour. An alternative is that customers use those conveniences while placing excess cash elsewhere. More functions in an application leave the question of where the money sits unresolved.[1]

I consequently read this debate through the terms of a funding relationship. Spence says low-cost deposits require structural investment, and identifies branch expansion and new banking services among the sources of gains. The expense of service provision and interest paid on deposits are separate costs of maintaining that relationship. A median checking balance of 3,000 dollars offers a clue about customer behaviour, but it cannot stand for the entire 236 billion dollar core deposit base. The distinction I find useful is the mix of cost and services supporting a growing balance. This quarter deposits increased and their cost rose too; that is where I begin assessing management’s relationship argument.[1]