The Central Bank: The Mandate and the Toolkit

A central bank is the institution that issues a currency and manages its price and quantity. Its mandate is the written statement of what it is trying to achieve, and putting that in writing is what makes it accountable rather than merely powerful. The toolkit is small: a policy rate, operations that add or drain money, requirements on the banks it supervises, and words.

Underneath that answer sits one idea that carries everything else. A central bank has exactly one ability that nobody else has, and every tool it holds is a use of that ability. The institution can create money in its own unit, and it can withdraw money in its own unit, and it does not have to earn, borrow or tax that money from anyone first. A government has to raise what it spends. A bank has to fund what it lends. A household has to be paid before it can buy. The issuing institution is the single point in the system where money can come into existence, and that is what makes four otherwise ordinary activities into instruments of policy.

Monetary policy itself, its aims and the meaning of a stance are set out separately, along with inflation and the movement of the cycle. The institution comes next: what it is, what document holds it to account, which four levers it has, what each lever actually touches, what independence does and does not mean, and where the honest limits are. An authority can still act when one of its levers is stuck, and blaming that authority for the wrong thing is very easy to do.

One rate carries the arithmetic that follows, the Sankhya policy rate of 6.00 per cent, and Sankhya is a republic built for teaching that exists nowhere. The arrival speeds attached to the four tools are stated assumptions, open to argument, not measurements. The bodies listed at the foot publish their own rates, decisions, dates and targets, and those published figures are the ones to work from.

What is a central bank, and what can only it do?

Start with an ordinary banknote. A note marked Rs 500/- is worth Rs 500/- because an issuing institution stands behind it and because the state has declared it legal tenderMoney that a creditor is obliged to accept in settlement of a debt within a jurisdiction. Legal tender is a status conferred on a particular kind of money, not a statement about how valuable that money is.. The tea stall that took Rs 2,000/- in cash this morning did not check who issued the notes, and did not need to. The unexamined confidence behind that transaction is the product the institution actually makes, and everything else it does is in service of keeping it.

Now the part that matters mechanically. A central bank can add to the quantity of money in its own unit, or take money out of it, without having to get that money from anywhere first. Nothing else in the system can do that. A household must be paid before it spends. A shop must sell before it restocks. A government must tax or borrow before it spends. A bank must fund itself before it lends. Every other participant faces a constraint on where the money came from, and the issuer does not. The asymmetry between the issuer and everybody else is the source of the whole toolkit, and every tool below is a particular way of using it.

Two other duties usually sit with the same institution, and they follow from the same ability. Because an institution that can produce the unit can always supply it at short notice, the central bank is the lender of last resortThe institution that will lend to a solvent bank which cannot borrow anywhere else at that moment, so that a shortage of cash does not turn into a collapse. How that lending is priced and secured belongs to banking.. Accounts held at the issuer are the one place where a payment between two banks becomes final, so the institution also settles the payments that banks make to each other. Both duties are covered separately.

Hold on to the shape rather than the list. The ability sits at the root, and the tools hang off it. Take the ability away and there is no toolkit left, only an institution with opinions.

What is a mandate, and why does putting it in writing matter?

A mandate is an objective, written down. Not a mission line, not a stated preference, and not what the people currently in charge happen to believe. A mandate is text, usually in law or in an arrangement published under a law, that names the objective the institution is answerable for.

Think about a school that says it is committed to good outcomes. What does the parents committee measure at the end of the year? Nothing was named, so nothing can be measured. Now think about a school that publishes, in advance, that every child in the leaving class will sit the state examination and that the pass rate will be reported publicly in April. The second school can now fail. Its head can be asked a question with a right answer. An unwritten mandate cannot be failed, and being able to fail in public is the exact thing that turns power into accountability.

A written objective creates three things at once. A stated aim can be compared with an outcome, so the writing creates a test. The writing creates a reporting duty, and somebody has to say whether the test was met. And the writing constrains the institution itself. Moves that do not serve the stated aim now have to be explained. An institution that can pick its own objective every quarter faces none of those three, and its independence would be a very different and much harder thing to defend.

The direction of the benefit is worth noting. Writing the mandate down is not mainly a service to the people watching. Writing the mandate down is a service to the institution. A decision that is costly today and useful in two years is very hard to take. The institution can take it only by pointing at a piece of paper that says this is what it was set up to do. Without the paper, every unpopular move is a matter of opinion against opinion.

What tools does a central bank actually have?

Here is where most explanations go wrong. Most explanations present a list of four things, and a list invites the reader to treat them as four flavours of the same action. Four flavours of one action is exactly what the tools are not. The four tools reach four different targets, and separating them by what each one moves is the difference between memorising a list and being able to predict which lever is even relevant. Go through them one at a time.

The policy rate moves the price of money. It is the rate at which the institution itself will deal with banks, and because it will always deal at that rate, no other short-term rate in the system can drift far from it. Price is the entire lever. The policy rate changes what money costs to hold for a short period, and everything short-dated reprices against it. The rate does not create or destroy anything. It puts a price on money. Where the policy rateThe rate at which the monetary authority itself deals with banks for very short periods, which is the anchor every other short-dated rate in the system is quoted against. How it is decided and what it targets is covered separately. is set, and against what target, is covered under monetary policy.

Operations in the market move the quantity of money. The institution buys or sells securities and government paper, and in doing so it puts cash into the money marketThe market where banks and large institutions lend to each other for very short periods, often overnight. The money market is where the shortest and safest borrowing in the system happens, and its rate is the one closest to the policy rate. or takes cash out of it. Buying puts money in. Selling drains it out. Such purchases and sales are usually called open market operationsThe purchase or sale of securities by a monetary authority in the open market, done to add cash to the system or remove it. The mechanics, the instruments used and the way each one is unwound are covered separately., and how they are actually conducted, what instruments are used and how each one is unwound is covered separately. The target is quantity, not price.

Requirements on banks move lending capacity. The institution can require that a bank hold a certain proportion of what it takes in, in a specified form, rather than lending all of it out. Raise that reserve requirementA rule obliging a bank to hold a stated share of what it takes in, in a specified form, rather than putting all of it to work. How a bank actually manages that holding, and what it costs the bank, belongs to banking. and less can be lent for the same deposits; lower it and more can. A reserve requirement reaches neither price nor quantity in the market directly. It reaches how much lending is possible at all. How a bank then manages that holding, what it costs the bank, and how the bank prices the loans it does make are questions about banking, and are covered there.

Words move expectations. The institution says what it expects to do, or what conditions would make it act, and everybody who lends or borrows over longer periods reprices immediately, before a single rupee has moved. Such a statement is usually called forward guidanceA statement about what the authority expects to do later, or what conditions would make it act, issued so that longer-dated rates move now. The credibility of such a statement, and what happens when it is broken, is covered separately., and what makes such a statement work at all is a proper subject of its own, covered separately. Note only that it is the cheapest tool, the fastest, and the one that can be undone simply by saying something else.

Now the payoff, and it is worth sitting with. Because the four tools reach four different things, an authority whose rate is already very low, or already very high, or politically impossible to move, is not disarmed. The authority still has quantity, still has capacity and still has words. None of that is a clever trick. The four levers never pointed at the same target in the first place.

What does the Sankhya toolkit look like set on one timeline?

The Republic of Sankhya holds the same four tools through its monetary authority, and its policy rate is 6.00 per cent, stated to two decimal places. The Sankhya rate sits in the middle of a narrow band: 5.75 per cent at the bottom, 6.25 per cent at the top, a band 50 basis pointsOne basis point is one hundredth of a percentage point, so 100 basis points is 1.00 percentage point and 50 basis points is 0.50 of a point. Rate moves are quoted this way because the differences that matter are small. wide and 25 basis points either side of the policy rate. Why a band of that shape does anything useful is a subject of its own and is covered separately. The rate is one number inside an arrangement, not a number floating alone.

The four tools of Sankhya can now be walked across a single timeline. The arrival times below are assumptions, stated openly so that they can be disagreed with rather than absorbed. Nobody has measured them.

Read the last column as carefully as the third. Speed and reversibility travel together here, and that is not a coincidence. Nothing has actually happened yet except a change of mind about the future, so the tool that lands instantly is the one that can be withdrawn instantly. The tool that takes four months to land is the one that has restructured what banks are holding, and unwinding that means a second disruption rather than a simple reversal. The four tools are not four speeds of the same action; they are four different actions that happen to be held by one institution.

One more thing the timeline hides, and it matters more than anything on it. Every arrival time in that table is an arrival at a market, not an arrival at a person. The Sankhya rate lands in the money market on day 3. Whether it lands on the instalment a household actually pays, and how much of it survives the journey, is a separate and much longer question that is covered where transmissionThe journey a policy change makes from the authority to the rates people and businesses actually face. How far it travels, how long it takes and how much of it never arrives is a subject in its own right. is taught. Do not read a fast arrival as a fast effect.

Independence from what, exactly?

The word independence gets used as though it were a general quality, like courage. It is not. Independence is a specific freedom from two specific pressures, and naming them makes the whole argument tractable.

The first pressure is the demand to fund government spending directly. An institution that can create money, sitting beside a government that would like to spend more than it raises, is an obvious temptation. If the institution simply produces what the government wants to spend, there is no meaningful limit on that spending and no meaningful limit on how much money exists. Independence in this sense means the institution is not obliged to be the funder, and that the government has to raise what it spends the ordinary way.

The second pressure is the electoral timetable. Some monetary decisions cost something now and pay something later, and the two do not land inside the same term. A body that has to face voters in eighteen months has a real reason to prefer whichever decision is comfortable for eighteen months. An institution that does not face voters can take the uncomfortable version. The argument, stated plainly, is not a claim that unelected people know better but a claim about the horizon each body answers to.

Now the distinction that readers collapse, and the reason this block exists. In most arrangements the institution gets instrument independence and not goal independence: somebody else sets the objective, publicly and in writing, and the institution is free only in how it pursues that objective. The elected side decides what the aim is. The institution decides which lever, how large and when. Without that split the argument turns into a defence of an unelected body choosing its own purpose, and almost nobody actually argues for that.

Once that split is held in view, several things stop being confusing at once. The split explains why a written mandate and independence are not in tension but are two halves of one design: the writing fixes the aim so that the freedom can be about method. The split explains why criticism of a decision is ordinary and criticism of the objective is a different conversation aimed at a different body. And because process is the part the institution was actually given, an institution defends its process far more fiercely than its results.

What can a central bank not do?

Four things sit outside the reach of every tool described above, and no size of move brings any of them inside.

A central bank cannot make an economy productive. Productivity comes from what people know how to do, what equipment they have, how quickly goods move, how easily a business can be started and how reliably a contract is enforced. None of those responds to the price of money. Cheaper borrowing may make it easier to buy a machine, but somebody still has to build the machine, know how to run it, and be able to get the output to a buyer.

A central bank cannot decide who gets credit. It can make lending cheaper or dearer in general. A central bank cannot direct a particular rupee to a particular workshop, and in most arrangements it is deliberately not allowed to try. Choosing who receives credit is an allocation decision that belongs with bodies that answer to voters. A tool that changes a price for everybody is exactly not a tool for choosing between people.

A central bank cannot fix a supply shock. If a harvest fails, if a shipping route closes, if an input that has to be brought in becomes dearer, then fewer goods exist and each one costs more to produce. A rate changes what it costs to borrow. A rate change does not make more goods exist. The tool aims at how much people want to spend, and the problem is on the other side of the market entirely.

A central bank cannot make a rate reach a borrower who fails a lending test. Suppose the rate falls and a bank has plenty to lend. The lender applies its own test and the authority does not, so a workshop with three months of thin orders and no security to offer will still be turned down. How far a rate change actually travels toward a borrower, and how much of it never arrives, is covered separately where transmission is taught.

Now the sentence worth carrying away. Most disappointment with monetary policy is disappointment that it is not fiscal policy or industrial policy. Those are different levers held by a different body, and they are compared properly elsewhere. Roads, schools, power supply, land rules, skills and who receives a subsidy are all decisions somebody makes, and none of them is made by moving a rate. Being clear about that is not making excuses for an institution; it is a precondition for holding it to account for the things that are genuinely its own.

Why does an analyst read a mandate rather than forecast a decision?

Here is the practitioner move, and it is the most immediately useful one in the subject. Most people who follow a monetary authority spend their effort predicting the next decision. Predicting the next decision is a coin toss dressed as analysis, and a wrong call leaves nothing to build on.

An analyst covering a lender, a credit desk deciding how to price a facility, or an investor thinking about how sensitive a business is to rates does something different. Each of them reads the written objective first, and uses it as a filter over everything the statistical system publishes. The question the mandate answers is not what will the authority do. The question is which readings this institution is obliged to respond to, and which readings it may look at and move past.

A mandate states what the institution must answer for, and a statement of that kind is far more durable information than any forecast of the next move. If the written objective names a price reading, then a surprising price reading forces some response and a surprising reading about one industry does not. None of that is a prediction. The obligation is a structural fact about the arrangement, and it holds across many decisions rather than one.

Two practical uses follow. A household deciding between a fixed and a floating arrangement on a long borrowing gets something more solid from knowing which readings will drive future decisions than from anybody guessing the next move; which arrangement a particular household should choose is a separate question. And a credit team stress testing a portfolio can build its scenarios around the readings the mandate names rather than around a single guessed rate path. Scenarios built that way are a far more honest way to be wrong.

Where is a real mandate written down?

Everything above was mechanism, and mechanism is the same everywhere. The specific document is not. In India, the institution is the Reserve Bank of India, it acts through a committee constituted for the purpose, and both the objective it works to and the arrangements around how decisions are taken are published under the statute that establishes it. A summary of an objective is not the objective, so the objective is worth reading in that document itself.