Bank Depositors vs. Deposit balances by products
Internal data in bank deposit warehouses contain customer IDs. This has allowed the understanding that the product reporting of NMDs distorts what's really happening. And it is this.
Banks of successfully been marketing multiple deposit accounts to their customers because retail bankers know attrition rates are negatively correlated with the number of accounts. Simultaneously the internet and nature of the products has allowed depositor to move funds among their multiple accounts at no out of pocket cost. The result: product level balance analyses are biased. Customer accounts are far more stable than deposit accounts. Banks are far more liquid than measured and currently in the banks we've evaluated, the balances of existing customers are growing faster than customers are leaving the banks. One other finding: customers with multiple accounts more than occasionally will close one account and move the funds to their other accounts. The impact on liquidity is precisely zero.
Implications: the measurement of the duration of product level NMDs is biased. EVE based on "account attrition" significantly understates the longevity of the customer relationships. And what matters for bank's liquidity is not product based behavior, but customer based behavior.
Most banks of any size have customer IDs in their data warehouses, obtained (as we were told anecdotally) that this was a side benefit associated with regulator based initiatives to understand aggregation of credit risk by borrowers.
Fed economists do not have access to these data, nor do academics. Both rely on call report data which can't and doesn't capture the dynamics from having access to the internal data.
And what makes deposit balances at bank A different from bank B? The heterogeneity of the deposit base. Do they have mostly single account customers or do they have customers with multiple accounts. On the commercial size what one sees is that a pizza parlor in Seattle manages in multiple balances similarly to a pizza parlor in NYC.
The implications to FTP and LCRs should be obvious.
This is really interesting, thanks very much for posting. Official sector definitely doesn't have this data, the best they have is FR2052a, which will break down by categories but not by depositors
Fed economists are looking for the key under the street light.
The data I've reviewed at different banks -- which were obtained in files used to conduct "attrition analyses" -- was (a) easy to see what was going on and (b) confirmed by the marketing depts in commercial banks. In two different banks over half the depositors had multiple accounts . And it was easy to show that the balances aggregated by customer were growing. If Fed economists are interested they can query marketing departments because these people are looking at customer behavior. They know retention rates are higher when a customer has more accounts. That's one of the many reasons they are constantly marketing more accounts to existing customers . And to track the success of a marketing campaign, they need "customer IDs."
My conclusion: commercial banks with sizable NMD accounts are far more liquid than currently measured. Some know it, but are forced by regulators to understate it.
What did the data also show: the upsurge in NMD accounts following the great recession was from maturing time deposits. And (I can't prove but I suspect) there are significant threshold effects playing out in depositors' behavior: what's the point of shopping for higher rates with rates at this low level? The implication is that should rates ever rise to those levels that motivate moving funds out of the banks, the banks (and regulators) may be find a change in trend.
I expect ultimately the banks will start looking at their own deposit data. And maybe some academics will be provided access. Deposit data is longitudinal when customer IDs are present, presenting significant opportunities for publication in refereed journals.
To the extent rates remain low, it could finally change the pricing of NMD accounts from an account basis to a customer basis. Banks are doing it by "exception" now, but only using rules of thumb to adjust the deposit rate for a large customer.