The Federal Reserve

The institution, and the committee inside it 

The name covers three things that are routinely spoken of as one. The Board of Governors sits in Washington and is an agency of the federal government: its members are nominated by the President, confirmed by the Senate, and serve staggered terms of fourteen years, a length chosen so that no single administration replaces the board. The regional Federal Reserve Banks, twelve of them, each serve a district, hold their own boards and their own research staff, and are not government departments. They exist in that number because the country that built the system distrusted a single central financial authority and split it geographically on purpose. The third thing is the Federal Open Market Committee, which is not a building or a staff but a committee that meets, and it is where the decision on the policy rate is taken. When a headline reports that the Federal Reserve raised or lowered rates, the body that did so is the third of these.

That structure is not trivia, because it determines who speaks in public and what standing their words carry. A regional bank president giving a speech in their own district is a genuine part of the institution talking, and is not the committee. A governor is a member of the committee whatever the calendar says. One institution can therefore sound like several at once without anybody having contradicted anybody, which is a property of how it was built rather than a sign of disarray.

Who votes 

The committee seats nineteen participants when every seat is filled, and casts twelve votes. All seven governors vote, at every meeting. The president of the Federal Reserve Bank of New York votes at every meeting as well, because that bank operates the desk that carries out in the market whatever the committee decides, so the office that executes policy sits permanently at the table where policy is set. The remaining eleven regional presidents share four votes on an annual rotation. A regional president therefore votes in some years and not in others, while attending, speaking and submitting projections in all of them.

Two consequences follow, and both are checkable against any published document. First, the projections the committee publishes are not the votes it casts: projections are submitted by every participant, so a set of nineteen individual views sits beside a decision taken by twelve people. Second, disagreement is a matter of public record rather than of inference. The statement names each participant who voted against the decision and states, in a line or two, what that participant preferred instead.

What the mandate says 

The objectives are set by statute rather than chosen by the committee, and that is the important structural fact about them. Congress instructs the system to pursue maximum employment, stable prices and moderate long term interest rates. Convention calls this the dual mandate, treating the third objective as following from the second, and the phrase is worth knowing because it is the phrase everybody uses, not because the statute has two limbs. The committee also publishes its own numerical definition of price stability, as a longer run inflation objective, in a statement of goals it revises rarely. The figure itself sits in that document rather than in this one: a curriculum that restates another institution's number inherits the job of correcting it, and the primary source is where the current wording always is.

A mandate written in two terms rather than one is what makes the language of a statement readable at all. A committee accountable for prices alone has one objective and nothing to weigh it against. A committee accountable for prices and for employment has two, and when the two point in different directions, which of them a statement leads with, and how each is characterised, is the substance of what the reader of hawkish and dovish language is reading. Other central banks are constituted differently, some with a single objective, and the next lesson takes three of them in turn.

The meeting cadence 

Eight scheduled meetings a year, spaced roughly six or seven weeks apart, on dates published more than a year ahead. A meeting runs over two days, and the decision is announced on the afternoon of the second at a minute that does not move. Additional meetings can be convened at any time, and occasionally are, which is the standing reminder that a published calendar is a list of scheduled information and of nothing else. In the days before each meeting the participants observe a blackout and make no public comment on policy, so the flow of speeches that runs between meetings stops, and whatever the committee has to say arrives in its documents rather than around them.

The announcement minute also has a position in the trading day. It falls in the New York afternoon, after the European session has largely finished, so the decision is priced in a window in which many of the participants who quote the other side of the world's most heavily traded pairs have gone home. That is a statement about who is at a desk, not about what any decision does to any price.

The documents, in the order they arrive 

A decision is not one publication. It is a sequence of them, spread over weeks, each produced under different constraints. Knowing what each one is, and in what order it lands, is most of what following this institution consists of.

1. The policy statement, released at the announcement minute. A few hundred words, and in large part the previous statement with amendments. The convention among participants is to read it against the previous version phrase by phrase: a clause deleted, a characterisation of the labour market softened or hardened, a single adverb moved. It also records the vote and any dissent.
2. The Summary of Economic Projections, published alongside the statement at four of the eight meetings. It collects each participant's projections for growth, unemployment and inflation, and the policy rate each considers appropriate at the end of each of the coming years. The last of these is plotted as the dot plot: one anonymous dot per participant per year.
3. The chair's press conference, conventionally beginning about half an hour after the statement. A prepared opening, then unscripted questions from journalists. It is the least controlled document of the day, and the one most often associated with the largest movement in a price.
4. The minutes, published three weeks after the meeting. A fuller account of the discussion, attributed to unnamed participants and graded by a vocabulary of quantifiers, from a few through several and many to most and all, which the institution uses consistently enough for the words themselves to carry weight.
5. The full transcripts, released after a lag of several years. They have no use in a market by the time they appear, and every participant speaks in the knowledge that the whole of it will eventually be published under their name.

The dot plot is worth stating carefully, because it is the document most often described as something it is not. Each dot is one participant's own view of an appropriate rate, conditional on that participant's own projections for the economy, submitted on one day, anonymous, and binding on nobody. It is not a vote, not a plan, not a commitment and not a schedule, and the accompanying material restates that every time it is published.

The order matters more than any one document. The statement is drafted collectively and agreed line by line, which makes it the most deliberate and the least revealing. The press conference is one person answering a question they were not shown in advance. The minutes arrive three weeks later, long after prices have absorbed the first two, and sometimes place an emphasis that the market read differently on the day. A reader who treats the three as a single message will find them appearing to contradict one another. They are three kinds of document produced by the same committee under three different constraints, and the apparent contradiction is usually the constraint rather than the committee.

Why this reaches instruments with no American exposure 

The dollar sits on one side of the large majority of foreign exchange turnover, is the currency a great deal of international trade is invoiced in, and is what much cross border borrowing is denominated in. The return available on money held in dollars is therefore an input to the price of many things that have no other American connection, and it arrives by three routes worth keeping separate.

• Quotation. Metals, oil and many other commodities are quoted in dollars, so the quoted number changes when the dollar changes even though nothing whatever has changed in the physical market for the commodity itself.
• Discounting. The yield on United States government debt is the reference against which a wide range of other assets are valued, so the step that converts expected future earnings into a value today is frequently performed at a dollar anchored rate even when the earnings themselves are entirely non American.
• Administered exchange rates. Where a currency is held at a fixed rate against the dollar, the policy setting attached to the dollar is imported rather than chosen. That is a structural feature of this region, and it has two lessons of its own later in this module.

The first of those routes is small enough to write out in full, and doing so separates the two things a single quoted number contains.

The dollar price of the ounce is identical in both cases. Nothing was mined, shipped, stored or consumed differently between them. What changed was the currency the price is quoted in, and a holder whose account is denominated in that other currency records the difference. This is the mechanism by which a decision about the dollar reaches an instrument that is not American in any respect, and it is the same mechanism whichever direction the currency moves in.

This institution comes first among the institutions in this module for the reason just given, and for no other. The dollar's position in quotation, in invoicing and in discounting is a fact about plumbing rather than a ranking of importance, and a reader in this region is affected at least as directly by the two central banks that appear two lessons from here as by this one. Nothing in this module treats American policy as the default lens through which a market is read.

Where practitioners disagree 

The first argument concerns the dot plot. One tradition treats the median dot as the committee's collective intention, and reads a shift in it as the nearest thing available to a stated path. Another points out that the dots are anonymous, that each is conditional on its author's own economic projections, that they are submitted by nineteen participants of whom twelve vote, and that they commit nobody, so a committee can and does end a period somewhere other than where its own dots at the start of it sat. Both traditions are describing the same chart. The disagreement is about whether a collection of individual conditional projections becomes a forecast by being plotted together, and it has not resolved.

The second concerns which document deserves the weight. One view holds that the press conference is the most informative event of the day precisely because it is unscripted, and is therefore where a committee's thinking is least edited. Another holds that an extemporaneous answer to a journalist's framing is the least considered sentence the institution produces all day, that its emphasis is routinely adjusted by minutes published three weeks later, and that treating an off the cuff clause as policy is reading a transcript as though it were a statute. Both positions are held by people who watch every one of these events.

The third runs underneath the section above it. One position treats this institution as the world's central bank in practice, on the grounds that the conditions attached to dollar funding reach everywhere regardless of what any other committee decides. Another holds that the relationship is variable rather than structural, that other institutions set policy for their own economies and have repeatedly diverged from this one, and that stretches in which everything appears to trade on a single committee's decision are followed by stretches in which it plainly does not. The strength of that relationship is itself a variable, which is why the module gives the other institutions their own lessons rather than footnotes to this one.