Interest on Reserve Balances and Rate Control

Before 2008, the Fed hit its fed funds target the hard way: it bought or sold small amounts of securities daily to nudge the supply of scarce reserves until the market rate landed where it wanted. That approach only works when reserves are scarce enough that small supply changes move the rate. Once quantitative easing flooded the banking system with reserves, that lever stopped working — reserves became too abundant for tiny supply tweaks to matter. The Fed switched to a different tool: paying interest on reserve balances (IORB) directly.

Why banks won't lend below it — except when they do

The logic is simple for any bank: why lend reserves overnight to another bank at, say, 5.25% when you could earn 5.30% risk-free by just holding the reserves at the Fed? So IORB should be close to a floor for bank-to-bank lending. But the fed funds market includes non-bank participants — chiefly the Federal Home Loan Banks (FHLBs) — who are not eligible to earn IORB themselves. FHLBs have every incentive to lend fed funds for any positive return, even below IORB, since the alternative is earning nothing. Banks, in turn, are happy to borrow from FHLBs below IORB and pocket the spread by redepositing at the Fed. This is exactly why the effective fed funds rate typically settles a few basis points below IORB rather than exactly at it.

Worked example

IORB is set at 5.40%. An FHLB lends $500 million in fed funds overnight to a bank at 5.33%, since 5.33% beats earning nothing. The bank redeposits that $500 million as reserves at the Fed, earning 5.40%. The bank's overnight profit on the trade is $500{,}000{,}000 \times (0.0540-0.0533)/360 \approx 972$, i.e. about $972 — a small, close-to-riskless arbitrage that keeps recurring daily and is exactly why the effective fed funds rate averages a few basis points under IORB rather than sitting on top of it.