The Hidden Intermediary: How Bankers’ Banks Price Local Liquidity
This chapter examines bankers’ banks as the institutional link between community and regional banks and the narrower overnight market that produces the bankers’ federal funds rate (BFFR). The key question is not simply what a bankers’ bank is, but how its position in the funding chain turns pooled customer reserves into a price signal. Understanding that chain lets readers explain why BFFR can diverge from the effective federal funds rate (EFFR) without either rate being incorrectly calculated.
Why a separate institution exists
A community bank can be important to local households and businesses while lacking the scale to provide every payment, settlement, funding, and liquidity service internally. Historically, many smaller banks depended on large correspondent banks for those functions. Correspondents could provide access to payments systems, check clearing, reserve management, and federal funds trading, but the relationship carried a structural tension: after interstate branching restrictions were relaxed, large correspondents could also compete for the retail deposits of the community banks they served.
Bankers’ banks emerged as a way to reduce that dependence without simply reproducing the same competitive relationship. They are specialized financial institutions that support community and regional banks with funding, settlement, liquidity-management, and related services. Their role is to let member banks pool resources and obtain a level of market access that would be difficult or expensive to achieve one by one. They are not retail banks serving the general public in the same way as their customers; their lending is primarily to depository institutions.
The arrangement solves a scale problem. A single community bank may have too few transactions, too little spare liquidity, or too narrow a set of counterparties to participate efficiently in every short-term market. A bankers’ bank can aggregate the needs and resources of many member institutions. That aggregation is useful operationally, but it also creates an observation point.
The correspondent relationship explains why that observation point became institutionally valuable. When a large correspondent both supplied essential services and competed for deposits, a smaller bank could face a trade-off between access and independence. It might need the correspondent’s infrastructure while watching the same correspondent become a stronger rival for customers and funding. Bankers’ banks were designed to provide a similar service function without making retail deposit competition the organizing relationship. Their purpose was not to isolate community banks from the market, but to give them a collective way to reach it.
That history also explains why the phrase “bankers’ bank” should not be reduced to a lender of last resort. The institutions provide ordinary market access, settlement, and liquidity services; they are part of the funding architecture used before a crisis as well as during a period of strain. Their value comes from making routine transactions more feasible for smaller institutions and from aggregating information about the conditions those institutions face. A price observed at that interface can therefore become informative before anyone can conclude that a bank is in acute distress. The institution is close enough to its members’ reserve and funding conditions that its lending decisions can reveal changes that a market-wide measure dominated by other lenders may not show clearly.
The Federal Reserve note reports 12 active bankers’ banks at the end of 2024, with approximately $10 billion in combined assets. Those numbers establish the scale of the institutions themselves, not the full reach of the networks they serve. A channel can be small on a national balance sheet and still be consequential to the banks that rely on it. The point of studying bankers’ banks is therefore not that they are large enough to set every market price. It is that they are positioned at a useful interface between pooled community-bank resources and overnight borrowers.
That interface should not be confused with the Federal Home Loan Bank system. A bankers’ bank and an FHLB may both be relevant to the same community bank, but they perform different functions. The FHLB system provides advances to members under its own collateral and credit arrangements. A bankers’ bank provides services and can intermediate customer reserves in the federal funds market. The distinction matters because the price and availability of liquidity depend on the route through which funds arrive, not only on the fact that a bank has a relationship with a liquidity provider.
It also matters for interpretation. If a community bank’s funding need is met through a bankers’ bank, the resulting transaction carries information about that bank’s network and the intermediary’s pool of available reserves. If the bank instead obtains an FHLB advance, the relevant constraints may center on collateral, exposure capacity, and the advance process. A calm rate in one channel cannot automatically certify comfortable access in the other. The later comparison of these routes will make that limitation concrete; first, the bankers’ bank mechanism needs to be made explicit.
From pooled reserves to BFFR
Bankers’ banks commonly lend their customers’ excess reserves rather than relying only on their own balance-sheet liquidity. “Pooled” does not mean that every member bank faces exactly the same condition or that the intermediary erases all differences among them. It means that the bankers’ bank brings together customer resources and manages lending to approved depository-institution borrowers. The pool creates a larger and more visible funding interface than any one member bank would create alone.
The flow can be understood in four steps. First, member banks have reserve balances they do not currently need for their own immediate purposes. Second, the bankers’ bank makes those customer resources available for appropriate overnight lending. Third, a domestic bank borrower seeks funds through the channel and pays a rate that reflects the balance between lending capacity and demand. Fourth, the observed transactions are used to calculate a segment rate. Each step matters because the final price is not an abstract opinion about small banks; it is the outcome of a particular set of institutions, reserves, borrowers, and overnight loans.
Consider a stylized evening in the network. One member bank has more reserves than it needs, another expects a temporary shortfall, and the bankers’ bank has rules for lending the pooled resources to an approved borrower. The intermediary does not make the two banks identical. It makes their positions visible to a common funding process. If several borrowers seek funds while the pool is limited, the price paid by the marginal borrower can rise even though the national market still contains abundant reserves elsewhere. If demand is light relative to the pool, the same process can produce a calmer price. The example is not a claim about a particular day; it shows why the pool connects reserve positions to a rate.
This is why the note describes bankers’ banks as marginal lenders. Their lending can sit at the margin of the funding decision for community-bank customers: when pooled reserves are readily available relative to demand, the channel may price calmly; when the pool is thinner or borrowers’ demand is stronger, the marginal rate can rise. The claim is not that one variable mechanically determines every BFFR movement. The claim is that the intermediary’s reserve position and customer network give its price a closer connection to smaller-bank liquidity than a rate shaped mainly by other lender populations.
The BFFR makes that connection measurable. The Federal Reserve note identifies overnight federal funds loans made over the Fedwire Funds Service in which a bankers’ bank was the lender, using transaction information and the FR 2420 report to isolate the relevant federal funds activity. The study covers October 2015 through August 2025. BFFR is then calculated as the daily volume-weighted median of transactions between bankers’ banks and domestic bank borrowers.
Several words in that definition do important work. “Overnight” limits the measure to a particular maturity. “Bankers’ bank was the lender” limits it to a particular intermediation channel. “Domestic bank borrowers” limits the borrower population. “Volume-weighted median” means that the measure summarizes the center of the observed transactions rather than reporting the highest or lowest price. These boundaries are what make BFFR interpretable; removing them would not make the indicator broader in a useful way, only less specific.
The construction also creates a clean comparison with EFFR without making the measures identical. Both are transaction-based rates, both summarize overnight activity, and both use a volume-weighted median. The difference is the filter applied before the summary is calculated. EFFR reflects a broad set of federal funds transactions; BFFR first asks whether a bankers’ bank was the lender and whether the borrower was a domestic bank. A reader can therefore compare the rates as two views of related activity, while still asking which transactions were admitted to each view.
That filter is particularly important when a rate is used as evidence rather than as a number to quote. A high BFFR is not merely “a higher federal funds rate.” It is a higher center of the transactions that passed through the bankers’ bank channel. The interpretation must follow the filter: the rate says something about the cost of funds in that channel, and the size of the difference from EFFR says that the channels were not facing identical conditions or were not weighted equally in the aggregate market.
The diagram shows why BFFR is a signal rather than a census. The member bank contributes to the network, the bankers’ bank intermediates, a borrower receives an overnight loan, and the transaction enters the data used to calculate the rate. An FHLB relationship can exist alongside that chain, but it does not become part of BFFR merely because the same community bank could use it.
BFFR can therefore diverge from EFFR for several related reasons. The lender populations differ: EFFR is heavily influenced by non-bank lenders, while BFFR is restricted to loans made by bankers’ banks. The borrower populations differ as well: BFFR focuses on domestic bank borrowers in that channel. The reserve positions differ because bankers’ banks lend from pooled customer resources, while the aggregate market includes other balance sheets and funding relationships. Finally, the market shares differ: a change in the bankers’ bank channel may be meaningful for its customers without having enough volume to move the market-wide median.
These differences are not four separate explanations that must be proved one by one before BFFR can be useful. They are four interfaces at which the two measures can see different conditions: who supplies the funds, who receives them, what resources are available to the lender, and how much the transactions count in the aggregate distribution. The more carefully those interfaces are kept distinct, the less temptation there is to treat a divergence as a contradiction. A divergence is often the observation to be explained.
This is the mechanism behind the phrase “local liquidity.” It does not mean that BFFR is a local retail rate or that every borrower has a unique private price. It means that the rate is closer to a defined network of institutions whose funding needs and reserve positions are connected. If those institutions begin paying more at the margin, BFFR can rise while EFFR remains relatively steady because the two rates are looking through different interfaces.
The narrowness is also the main limitation. BFFR is not a direct survey of every community bank. It does not establish that every smaller institution is under pressure when it rises, or that a stable BFFR proves every member bank is comfortable. The Federal Reserve note presents it as a cleaner and more direct measure of smaller-bank funding conditions than EFFR, especially for institutions less active in repo markets. “Cleaner” here means better aligned with the specific segment being examined, not universally superior.
Nor should the construction be used to make a stronger causal claim than the evidence allows. A high BFFR may coincide with declining reserves at bankers’ banks and with greater demand for funds, but the rate alone does not identify every cause of the movement. It is best read alongside reserve balances, broader overnight rates, repo conditions, collateral availability, and information about institutional access. The next chapter uses two episodes to test what this segmented signal revealed when the aggregate market and the bankers’ bank channel moved differently.