CB2 · explained

How monetary policy reaches the economy: the transmission mechanism

A central bank changes one short-term interest rate. Somehow that is supposed to alter what millions of households and firms spend, more than a year later. This page follows the actual channels, puts numbers on three of them, and explains why the effect is uncertain in size and late in arriving.

CB2 · Business Economics · 12 min read

After this page you can

Trace a policy rate change through each transmission channel to spending, and quantify its effect on a borrower, a firm's investment and a bond price

The gap the mechanism has to cross

A central bank does something narrow. In India the Reserve Bank of India sets a policy rate — the repo rate — at which banks can borrow from it against collateral, and it operates in the money market to keep overnight rates near that level. That is the whole instrument: a price for very short-term money, in one market, between a small number of institutions.

What it is aiming at is enormous and distant: the prices households pay, a year or two from now. The transmission mechanism is the chain of real behaviour that connects the two, and the reason it is worth studying rather than assuming is that every link in it can weaken — and when one does, the same rate change has a different effect.

Take a rise in the policy rate throughout, and reverse every sign for a cut.

Nobody borrows at the repo rate. What the policy rate moves first is the rate at which banks lend to each other overnight, and from there the rest of the money market, then bank deposit and lending rates, then the yields on bonds of longer and longer maturity.

This first link is where most of the leakage happens, and three things decide how much gets through.

How banks fund themselves. A bank funded largely by market borrowing feels a policy change almost at once. A bank funded by deposits whose rates it adjusts slowly feels it later and less.

Competition and the health of the banks. Banks competing hard for lending pass changes through quickly. A bank short of capital may widen its margin instead, absorbing a cut rather than passing it on — which is why monetary policy is weaker after a banking crisis than before one.

Floating against fixed. Only borrowers on floating rates feel a change on their existing debt. Where most borrowing is at fixed rates, a policy move reaches only new borrowing, and the effect is far smaller and slower.

Longer rates are a separate matter again. They depend on where markets expect short rates to go over years, so a rate rise that is expected to be reversed quickly may barely move ten-year yields. What the central bank says about the future can matter more than what it does today, which is why communication is treated as an instrument in its own right.

Once rates have moved, five distinct mechanisms carry the change into demand. They work alongside each other and they do not all point with equal force.

The cost of credit. Dearer borrowing raises the hurdle an investment project must clear, so projects between the old and new hurdle are dropped. The same logic postpones household purchases that are usually financed — a house, a vehicle. The size of this effect depends on how thickly projects are clustered near the hurdle rate, which is why an identical rate rise bites hard at some times and barely at others.

Cash flow. This one is often larger than the first and is routinely forgotten. A household with a floating-rate home loan has a higher monthly payment, which leaves less to spend on everything else — whether or not it was planning to borrow anything new. The firm with existing floating-rate debt is in the same position. The effect falls hardest on the most indebted, who typically have the least spare saving to cushion it.

Asset prices and wealth. A higher interest rate means future cash flows are discounted more heavily, so bonds, shares and property are worth less. Holders feel poorer and spend a little less; firms find equity dearer to raise. The bond arithmetic is the cleanest case and is worked below.

The exchange rate. Higher domestic rates make domestic assets more attractive to hold, which tends to strengthen the currency. A stronger currency makes imports cheaper — pulling inflation down directly — and exports dearer, pulling demand down. In an open economy this is one of the fastest-acting channels, because currency markets move in minutes.

Expectations. If people believe the central bank will act to keep inflation near its target, they set wages and prices on that basis, and the belief does part of the work. Credibility is a genuine input to the mechanism rather than a public-relations matter: a central bank that is believed needs smaller rate changes to get the same result.

Why it takes so long, and what that forces

Each of those channels needs real decisions to be revisited — a loan to be repriced, a project to be reappraised, a contract to be renegotiated, a house purchase to be deferred. Those things happen over quarters, so the effect on output builds over several quarters and the effect on inflation later still.

That lag has one unavoidable consequence: a central bank cannot react to inflation it can see. By the time today's inflation is in the data, the moment to have acted on it has gone. It must act on a forecast of where inflation will be when today's decision is still working — which is why monetary policy is argued about in terms of projections, and why it can turn out to have been wrong without anyone having been foolish.

When the short rate cannot fall any further

A policy rate can be cut only so far. Near zero the main instrument is exhausted while demand may still be too weak, and the channels above have to be entered from somewhere other than the short rate.

Quantitative easing does that. The central bank creates reserves and buys longer-dated bonds, mostly government bonds, in quantity. The intent is to push their prices up and their yields down — compressing the extra return investors demand for holding long rather than short — and to push the sellers into other assets, lifting their prices too. Long rates and asset prices are exactly the links the five channels run from, so the mechanism downstream is the familiar one; only the entry point has changed. It also signals that short rates will be kept low for a long time, which works through expectations.

Forward guidance is the same idea with words instead of purchases: a commitment about the future path of rates, intended to pull down the long rates that depend on expectations of it. Both tools rely on credibility, and both are weaker in a banking system unwilling or unable to lend.

Questions students ask

Why not control the money supply directly?

Because the relationship between the quantity of money and spending turned out to be too unstable to steer by, as financial innovation kept changing what counted as money and how fast it circulated. The quantity of money in a modern system is also mostly created by banks as they lend, so it responds to the demand for credit rather than being set by the central bank. Fixing the price of short-term money and letting the quantity adjust proved far more reliable, which is why nearly every central bank now does it that way.

Does a rate rise always reduce inflation?

It reduces demand-driven inflation, with a lag. Against a supply shock — a jump in energy or food prices — it has no power over the cause at all, and its only real work is on expectations: preventing a one-off price rise from turning into a continuing wage-and-price spiral. That is why central banks distinguish headline from underlying inflation, and why they sometimes let a supply-driven spike pass while watching expectations closely.

Why does the same rate change have different effects at different times?

Because the strength of every link varies. How much household debt there is and how much of it floats; whether banks have the capital to lend; how many investment projects sit just above the hurdle; whether the currency is free to move; and above all whether the central bank is believed. Those conditions change over years, so the response to an identical move changes with them. That is the honest answer to why monetary policy is judged rather than calculated.

A worked example

An invented case, built for this page. The policy rate rises by 100 basis points — one percentage point — and the banks pass it through in full. Every rate and amount below is illustrative, chosen for clean arithmetic; none of it is a current market figure. Follow the rise through three channels.

  1. Channel 1: cash flow — a household with a floating-rate home loan

    A family in Pune has ₹50,00,000 outstanding with 20 years to run. Their floating rate moves from 8.5% to 9.5% a year. The level monthly instalment on a loan of P over n months at monthly rate i is

    EMI = P × i(1 + i)n ÷ [(1 + i)n − 1]

    At 8.5%: i = 0.085/12 = 0.00708333, n = 240.

    (1.00708333)240 = 5.4412426
    EMI = 50,00,000 × 0.00708333 × 5.4412426 ÷ 4.4412426 = ₹43,391

    At 9.5%: i = 0.095/12 = 0.00791667.

    (1.00791667)240 = 6.6360614
    EMI = 50,00,000 × 0.00791667 × 6.6360614 ÷ 5.6360614 = ₹46,607
    rise in the monthly payment = 46,607 − 43,391 = ₹3,216, or ₹38,592 a year

    Two things to notice. First, the rate rose by about 11.8% of itself (one point on 8.5) but the instalment rose by only 3,216/43,391 = 7.4%, because part of every instalment is principal and that part is unaffected by the rate.

    Second, and more revealing, look at where the extra ₹3,216 goes. First-month interest at 8.5% is 50,00,000 × 0.00708333 = ₹35,417, leaving 43,391 − 35,417 = ₹7,974 of principal repaid. At 9.5% interest is 50,00,000 × 0.00791667 = ₹39,583, leaving only 46,607 − 39,583 = ₹7,024 of principal.

    extra interest = 39,583 − 35,417 = ₹4,166
    less principal repaid = 7,974 − 7,024 = ₹950
    4,166 − 950 = ₹3,216 — agreed.

    So the family pays ₹3,216 more each month, has ₹3,216 less to spend on everything else, and is paying the loan down more slowly than before. That is the cash-flow channel, and it reaches a household that was not planning to borrow anything at all.

  2. Channel 2: the cost of credit — a firm's investment list

    A manufacturer has five projects costed at ₹20 crore each, with these expected returns:

    ProjectExpected returnHurdle 11%Hurdle 12%
    A — new line16.0%approveapprove
    B — automation14.0%approveapprove
    C — second shift12.5%approveapprove
    D — warehouse11.5%approvedrop
    E — depot10.5%dropdrop

    The hurdle is the firm's borrowing cost plus a margin, so the 100 basis points moves it from 11% to 12%.

    capital spending at 11%: four projects × ₹20 cr = ₹80 crore
    capital spending at 12%: three projects × ₹20 cr = ₹60 crore
    investment cancelled = ₹20 crore

    One percentage point cancelled a quarter of this firm's capital programme — and it did so because exactly one project happened to sit in the one-point band. Had the returns been 16, 14, 12.5, 12.4 and 12.2 per cent, the same rate rise would have cancelled three projects and ₹60 crore. The size of this channel depends on how thickly projects are packed just above the hurdle, which nobody can observe in advance. That is a large part of why the effect of a rate change is uncertain in size.

  3. Channel 3: asset prices — what a bond does

    A ten-year bond with a face value of ₹100 pays a 7% annual coupon and is priced at par when the yield is 7%. Let the yield rise to 8%.

    1.0810 = 2.1589250, so v10 = 0.4631935
    a10 at 8% = (1 − 0.4631935) ÷ 0.08 = 6.7100814
    price = 7 × 6.7100814 + 100 × 0.4631935 = 46.9706 + 46.3194 = ₹93.29

    A one-point rise in the yield has taken 6.71% off the price. Check it roughly a second way: this bond's modified duration at par is a little over seven years, and duration predicts a fall of about 7% for 100 basis points — slightly more than the 6.71% actually computed, which is the expected direction of the approximation's error. Anyone holding this bond is poorer by 6.71% of its value, and every long-dated asset has moved the same way for the same reason.

  4. What the three add up to, and what they do not

    One rate change, three separate pressures on demand: a household with ₹3,216 a month less to spend, a firm with ₹20 crore of cancelled capital spending, and bondholders 6.71% poorer. Add the exchange rate strengthening and import prices falling, and the fifth channel of expectations working on wage and price setting, and that is the mechanism.

    Three honest qualifications belong with the arithmetic. None of these effects is instant: the family adjusts its spending over months, the firm's cancelled projects would have been built over years, and the inflation consequence arrives later than either. None is certain in size: change the composition of debt between floating and fixed, or the clustering of projects near the hurdle, and the same numbers come out quite differently. And the figures are a single household, a single firm and a single bond — the aggregate effect depends on how many of each there are, which is the step from this arithmetic to a forecast, and the step where the real uncertainty lives.

Where this fits in CB2

This page explains one idea from CB2 Business Economics. The full CB2 syllabus, with what each part is examined on, sits on the course page, and past-paper style mock exams with full solutions are where you find out whether the idea has stuck under time pressure.

Other CB2 topics explained

Stuck on a term rather than the idea? The actuarial glossary defines the notation used above. If something here is wrong or unclear, tell us — these pages get corrected.