What a central bank actually does
M4 — Macro & rates
Central banking sounds enormous and is, mechanically, quite narrow. A central bank sets one price. Everything that follows — mortgages, corporate borrowing, currencies, equity valuations — is transmission from that single number.
Getting clear on what the number is removes most of the mystery.
The one price
The policy rate is the interest rate on overnight lending of reserves between banks — the Fed funds rate in the US, Bank Rate in the UK, the deposit facility rate in the euro area. Banks hold reserve accounts at the central bank; at the end of each day some have more than they need and some less; they lend to each other overnight; that market has a price.
The central bank controls that price, and only that price, directly.
Note what this is not. It is not the rate on your mortgage, a corporate bond, or a government bond. It is the shortest, safest rate in the system — the anchor from which all others are quoted at a spread.
How the rate is actually enforced
Textbooks describe open-market operations: the central bank buys or sells securities to change the quantity of reserves, and a scarce supply meeting demand determines the price. That was broadly true before 2008 and is not how it works now.
Since the crisis, reserves are abundant — QE created trillions of them — so scarcity cannot set the price. Modern implementation is a floor system:
the central bank simply PAYS INTEREST ON RESERVES at its target rate
⟹ no bank lends to another bank below that rate, because it can
earn the target risk-free by leaving the money at the central bank
⟹ the rate the central bank pays becomes a FLOOR that drags the
whole overnight market to it, regardless of quantity
Supporting facilities catch the edges — a standing repo facility above and a reverse-repo facility below form a corridor for entities that cannot earn interest on reserves. But the essential mechanism is that the central bank sets the rate by announcing what it will pay, not by rationing supply. It is administered, not traded into place.
Why one overnight rate moves everything
An overnight rate is, by itself, irrelevant to a thirty-year mortgage. It matters because of an arbitrage-flavoured relationship you already have the tools for:
A long rate is approximately the average of expected future short rates, plus compensation for the risk of being wrong about that path.
Lending for two years should pay roughly what rolling one-year loans is expected to pay, or one of the two is mispriced. So the entire curve of interest rates is, to a first approximation, the market's forecast of the central bank's future decisions — which is what M4.4 is about.
That has a consequence which surprises people watching their first rate decision:
the DECISION is usually already priced — markets have inferred it
from data and speeches for weeks
what moves prices is the CHANGE IN THE EXPECTED PATH:
· the statement's wording
· the "dot plot" of officials' rate projections
· the press conference
· anything implying the path is higher/lower for longer
A 25bp hike that was fully expected can leave markets unmoved, while an unchanged rate with a hawkish sentence can move them violently. This is why "forward guidance" is a policy tool in its own right: talking about future rates changes long rates today, at no cost, without touching the policy rate at all.
When the policy rate runs out
The policy rate cannot fall far below zero — depositors can hold cash instead, so there is an effective lower bound. In 2008 and again in 2020, central banks reached it and needed something else.
QE — quantitative easing
buy long-dated government (and other) bonds with newly created
reserves. Removes duration from the market, which compresses the
TERM PREMIUM (M4.4) and pushes long yields down even with the
policy rate stuck at zero. Also a signalling device — it commits
the bank to a stance more credibly than words.
QT — quantitative tightening
the reverse, usually passive: let bonds mature without
reinvesting, shrinking the balance sheet gradually.
QE is often described as "printing money". The mechanically accurate description is an asset swap: the central bank exchanges reserves for bonds, changing the composition of what the private sector holds rather than its net wealth. Whether that is inflationary is genuinely contested — 2009–2019 saw enormous QE and persistently below-target inflation, which is a real problem for the simple story, and 2021–22's inflation followed fiscal transfers and supply shocks at least as much as balance-sheet growth.
The mandate, and what it isn't
Federal Reserve DUAL mandate — stable prices AND maximum
employment. Explains why US decisions hinge
on payrolls data as much as on CPI.
Bank of England inflation target (2%), with a secondary
growth/employment objective.
ECB price stability, primary and near-singular.
None of them has a mandate for the stock market. Equities matter only through the wealth effect and financial conditions — as a transmission channel, not a goal. "The Fed won't let stocks fall" is a claim about reaction functions, not about mandates, and it is a claim that 2022 tested and falsified.
That is also the honest framing of central bank independence: the whole design assumes politicians want lower rates before elections and someone else should be able to say no.
Which leaves the question the next lesson answers — if the mandate is inflation, why does every equity trader watch the decision?
Source: for mechanics, any central-bank primer on implementation — the Fed's own "Policy Implementation" explainers are short and unusually clear about the floor system, which most textbooks still describe wrongly. Matt Levine's Money Stuff is the roadmap's pick for absorbing this by osmosis; start reading it now rather than after the module.