When the Federal Funds Rate Is Calm but Smaller Banks Are Not
The subject of this series is the relationship among the effective federal funds rate (EFFR), the bankers’ federal funds rate (BFFR), bankers’ banks, and funding conditions at smaller U.S. banks. Its central question is straightforward but easy to miss: how can the market’s best-known overnight rate remain calm while a narrower channel used by community and regional banks records tighter funding? The practical value of answering it is learning to ask what a financial indicator actually represents before treating it as a reading of the whole system.
A calm rate is still a view from somewhere
The EFFR is a daily summary of overnight federal funds transactions. It is not a survey asking every bank whether funding feels easy or difficult. It is a volume-weighted median: transactions with enough volume to sit near the center of the market have the greatest influence on the number that readers see.
A median is already a selective statistic, and the volume weighting makes the selection economically meaningful. Imagine two groups borrowing overnight. One group trades in large size at fairly steady rates; another trades in smaller size but must pay more as its reserve cushion thins. The second group can be experiencing a meaningful change in funding conditions without moving the center of a market dominated by the first. The example is stylized, not a reconstruction of any single day, but it captures the measurement problem: the rate can be calculated correctly and still answer a narrower question than readers assume.
The relevant question is therefore not whether EFFR is a good rate in the abstract. It is: good for whom, and under what market conditions? A benchmark can be excellent for tracking the price of the transactions that determine it while being less responsive to a thin network of borrowers whose needs are different. That is why the identity of lenders and borrowers belongs in the explanation, rather than being treated as background plumbing.
That distinction became more important after the Global Financial Crisis. Before 2008, the Federal Reserve operated with scarce reserves and used daily open-market operations to steer an active interbank market. Banks routinely borrowed and lent reserves overnight to manage their end-of-day positions. After the Fed’s large-scale asset purchases expanded reserve balances, the system moved toward an ample-reserves framework. The Fed could anchor short-term rates through administered rates, especially interest on reserve balances (IORB) and the overnight reverse repurchase agreement (ON RRP) rate, without adjusting reserve quantities every day.
This framework has kept the EFFR stable even when reserves have been historically abundant. That is an important success if the question is whether the Federal Reserve can control a broad overnight benchmark. It is a different question from whether every type of bank can obtain funds on equally comfortable terms.
The composition of trading is the missing link. A January 2026 Federal Reserve note reports that non-bank lenders, especially Federal Home Loan Banks (FHLBs), account for more than 90 percent of lending in the federal funds market. The residual bank-to-bank market is much smaller; daily interbank volumes reported in the note range from roughly $2 billion to $6 billion. A rate dominated by the first group can remain steady even when conditions change in the second.
That does not make EFFR inaccurate. It makes the rate selective. A thermometer placed in one room can be precise without describing the temperature in every room of a building. The financial version of that problem is more consequential because the “rooms” have different balance sheets, collateral, access routes, and reasons for borrowing.
The difference between an aggregate and a segment measure is also a difference in time horizon and decision pressure. A large lender with broad access may continue to transact near a stable market rate while a smaller institution is deciding whether it has enough immediately usable reserves, whether its collateral can support another advance, or whether it must pay more for funds later in the day. A market-wide number may register the first condition more clearly than the second. That is not a defect in aggregation; it is a reason to pair aggregation with a measure designed around the second condition.
The dashed relationship in the diagram is not a causal arrow. It represents the possibility that the two observations can coexist because they summarize different populations. The rest of the series follows the institutions and reserve flows that make that coexistence possible.
The institutions behind the apparent contradiction
The smaller-bank signal examined here is connected to bankers’ banks. These are specialized institutions created to provide community and regional banks with funding, settlement, liquidity-management, and related services. Their historical purpose was to help smaller banks pool resources and obtain scale without relying entirely on large correspondent banks that might also compete for their retail deposits. The Federal Reserve note reports 12 active bankers’ banks at the end of 2024, with about $10 billion in combined assets.
Those figures are small compared with the national banking system. Their importance comes from their position, not their aggregate size. Bankers’ banks sit close to a network of community-bank customers that may need access to short-term funding but do not trade with the same frequency or scale as the largest institutions. They lend primarily to depository institutions, not to retail customers, and can make the smaller-bank funding channel visible through the prices they charge.
A bankers’ bank is not an FHLB. The two can serve the same member institution, but they are different channels. An FHLB is part of a system that provides advances to members, typically against collateral. A bankers’ bank intermediates for its customers and can lend pooled customer reserves in the federal funds market. A community bank may therefore have more than one potential source of liquidity while still facing different conditions in each channel.
The distinction matters because “access to liquidity” has several layers. A bank may have a formal relationship with a provider, yet face a limit on how much it can borrow, what collateral it can pledge, or when a request will receive the most favorable treatment. A bankers’ bank can also be relevant later in the business day, when an institution has a clearer view of its settlement position. These are not yet conclusions about which source is preferable; they are reasons to avoid treating membership, capacity, and immediate usability as the same thing.
The BFFR is designed to observe the bankers’ bank channel. It is the daily volume-weighted median rate on identified transactions in which bankers’ banks lend to domestic bank borrowers. The rate is narrower than EFFR by construction. That narrowness is the reason it can carry information that an aggregate measure does not: it is closer to the reserve demand and funding choices of the institutions using that channel.
This also sets a boundary. BFFR is not a direct survey of every community bank, and a movement in it does not prove that one particular reserve decision caused every bank’s costs to change. It is a segment-specific market signal. To read it well, we need to know both what is inside the segment and what remains outside it.
Why the two episodes matter
The Federal Reserve note provides two useful contrasts. On September 17, 2019, a day associated with a sharp spike in repo rates, the EFFR briefly traded above the Federal Reserve’s target range. The BFFR moved little. The note interprets that pattern as evidence that bankers’ banks and their clients did not appear to face comparable liquidity constraints and that reserves remained ample among those institutions.
That episode is not evidence that no smaller bank experienced pressure. It is evidence that the particular bankers’ bank channel did not show the same movement as the market conditions reflected in EFFR. The distinction matters: one indicator may be telling us about a disturbance in a part of the short-term funding system, while the other is telling us whether a different network of borrowers and lenders is tightening.
March 2023 presents the reverse lesson. The EFFR traded about 7 basis points below IORB, while the BFFR reached historic highs after rising for several months. Reserves held at bankers’ banks had declined. The note treats this combination as evidence of tightening funding conditions among smaller banks and reports that BFFR fell only after the Bank Term Funding Program was introduced.
Here “funding stress” should be read carefully. In the evidence at hand, it means higher marginal funding costs and tighter conditions in an observed smaller-bank channel. It does not mean that every small bank was insolvent, that BFFR alone diagnoses a banking crisis, or that a calm EFFR is irrelevant. The value of the comparison is precisely that it prevents a single rate from carrying more meaning than its market population can support.
The comparison also shows why direction matters. In 2019, the aggregate disturbance did not pass through visibly to the bankers’ bank channel. In 2023, the smaller-bank channel tightened while the aggregate measure remained relatively contained against IORB. Neither pattern says that one rate should replace the other. Together they show that different rates can be most informative at different points in a market’s structure, and that interpretation depends on the question being asked: is the concern broad short-term-rate control, repo-market pressure, or the reserve position of smaller-bank borrowers?
The next step is to understand the mechanics behind the signal: whose reserves are being pooled, when those reserves are lent, and why a bank might choose one liquidity route over another. Only then can the difference between a calm aggregate rate and a high segment rate become more than a statistical curiosity.