The decision is not the news. The surprise is.

The rate everything else is priced from

A central bank sets one price: what banks pay to borrow overnight. Every other interest rate in the economy is that rate plus compensation — for lending longer, for credit risk, for holding a currency. Move the anchor, or move what people expect the anchor to be, and everything priced off it has to be re-priced.

That is the whole mechanism, and it is worth being precise about the two halves of it, because almost all of the market's attention goes to the wrong one.

The level of the overnight rate today sets the cost of money today. It is a single number and it affects a single point on the curve — the very front of it. Nearly nothing an investor holds is priced off that point alone.

The expected path of the overnight rate over the next several years is what prices everything else. A two-year bond yield is, to a close approximation, the average overnight rate the market expects over those two years, plus a premium for tying money up. A ten-year yield is the same idea stretched over ten. So the market does not wait for a central bank to act. It holds a continuously updated view of what the bank will do, and that view is already inside every price on the screen.

Which produces the result that confuses people most: a decision everyone expected changes nothing, because it was already in the price before it happened. This article traces how the anchor reaches an asset price, why the guidance usually carries more information than the number, and where the chain is weaker than the standard telling admits. Transmission into your own loan and deposit rates is a different and equally important chain, owned by how a rate change reaches you.

One rate, a whole curve

Start with why a single overnight rate can price a ten-year bond at all.

A lender choosing between rolling overnight loans for two years and buying one two-year bond is comparing the same two years of money. If the two-year bond paid much less than the overnight rate is expected to average, nobody would buy it; much more, and nobody would roll. The two are tied together by arbitrage, and the tie is what makes the curve a statement about expectations rather than a list of unrelated prices.

So the conventional decomposition of any government bond yield is two terms: the average overnight rate expected over the bond's life, plus a term premium — the extra compensation demanded for locking money up and bearing the risk that the expectation is wrong. The first term is where a central bank operates. The second is set by everyone else.

Now the arithmetic that makes the article's point. Take an illustrative policy rate of 6% — an invented round number, not a market level — and ask what happens to a two-year yield under two different events. Ignore compounding and day-count conventions; this is deliberately simple arithmetic you can redo.

In that illustration, the bank that did nothing moved the two-year anchor roughly fifty times as far as the bank that cut. Not because the cut was unimportant, but because a single month is one twenty-fourth of a two-year average, while a change of mind about the path is all twenty-four of them at once. Stretch the same logic to a ten-year yield and the ratio gets more lopsided still.

That is a mechanical claim — it follows from what an average is, and no market behaviour is required for it to be true. It is also the reason the shape of the curve is read as a statement about expected policy, which is the subject of yield curve inversion.

The decision, and the surprise: two different objects

If prices already contain the expected path, then the information content of a policy announcement is not the decision. It is the difference between the decision-plus-guidance and what was priced going in.

Keep those two apart and a great deal of otherwise baffling market behaviour stops being baffling. Here are four cases, all of them ordinary.

Read the second row again, because it is the one that catches people. A rate cut can be hawkish news. There is no contradiction in that sentence once you hold the two objects apart: the bank eased, and it eased less than the path that was already inside prices. Both are true at the same time, and only the second one is news.

This is where the commonest mistake in reading a policy day lives, and it is worth naming so it is recognisable. Call it the priced-in trap: treating the announcement as the event, when the event was the weeks of repricing that led up to it. Someone reading only the headline sees a cut and an unmoved market and concludes the market is irrational. The market had simply finished doing the thing before the headline arrived.

The honest limitation of that framing is that the priced path is not directly observable. It is inferred from instruments — interest-rate swaps, policy futures, the near end of the government curve — each of which contains a term premium and a liquidity premium of its own that has to be assumed away. So "the market expected a cut" is an inference from prices, not a measurement of anybody's belief, and two careful people can read the same instruments and disagree about what was priced. That is a judgement, and it should be held as one.

Why the words usually matter more than the number

The curve arithmetic above explains, without any appeal to psychology, why the statement typically carries more weight than the decision.

A decision moves one month of a multi-year average. Guidance moves the average. When a central bank changes how it describes the conditions under which it would act — the inflation it is willing to tolerate, the growth it is watching, the sequencing it intends — it is not adjusting the front point. It is adjusting the market's model of every future point at once, and the curve is a weighted sum of all of them.

Which is why the following are all information events in their own right, quite separately from the rate:

• The statement's wording. A dropped or added conditional phrase changes the mapping from data to future action.
• The vote split and any dissent. A unanimous decision and a narrowly carried one describe different distributions of the next decision.
• The published projections, where a bank publishes them — a set of numbers describing the path itself rather than the level.
• The minutes, released weeks later. A second information event from a meeting that already happened.
• The stance and the liquidity framework. A bank can leave the policy rate alone and still change the effective cost of overnight money by draining or adding liquidity — which is the point the next section turns on.

None of that tells you where any price goes. It tells you which parts of a policy day are capable of carrying new information, and which part — usually the headline number — mostly is not.

The trade-off in this way of reading is real and worth stating. Guidance is qualitative, so extracting a path from it is interpretation, not arithmetic. You are trading a precise number that carries little information for an imprecise reading that carries most of it. That is usually the better trade; it is not a free one.

From the anchor to the price of an asset

Once the expected path has moved, the transmission into asset prices runs along three routes. Each is a mechanism, and none of them says what any price will do.

Discounting. A share is a claim on money arriving over many years, and converting future money into today's money is division by a rate built on the same anchor. Discounting is the arithmetic; the further out the money, the more a change in the divisor matters, which is the same sensitivity that duration measures for bonds. The full version of this channel, including the part where a rupee cash flow is not discounted at a dollar rate, is the yields article.

The hurdle for capital. When safe money pays more, the return demanded from anything risky rises with it. That is the conventional reading, and note what carries it: the reallocation runs through mandates, committees and rebalancing dates rather than through a screen — a slow channel sitting behind a fast one.

The rate differential and the currency. A currency pair is priced partly off the gap between two countries' rates. Narrow the gap and the compensation for holding the higher-yielding currency narrows with it. This is the machinery behind the yen carry trade, where the borrowing currency's policy rate is the input, and behind the rupee's effect on your portfolio. It is also why a single bank's meeting can move assets in countries it has no mandate over.

Notice that all three run through the expected path, not the delivered decision. Nothing in the chain requires a bank to have acted.

Two central banks price an Indian asset, and only one of them is the RBI

An Indian share is discounted at a rupee rate and funded, at the margin, by capital that has a dollar alternative. So two policy anchors reach it, and they are set by institutions with different mandates and no obligation to agree.

The RBI sets the domestic anchor, for domestic conditions — Indian inflation and Indian growth. It is not a follower of US policy, though the currency creates a link that is real and that policymakers watch. One mechanical detail matters for reading India specifically: the RBI's operating target is a market-determined overnight rate, and the policy rate works by steering it inside a corridor. The bank can therefore change the effective cost of overnight money by managing liquidity without changing the policy rate at all — and the market rate can drift away from the policy rate when liquidity is tight or flush, with no decision having been taken.

The Fed sets the anchor for global capital. Its rate is the base from which the world's funding currency is priced, so it enters an Indian asset through the dollar, through the hurdle rate global allocators apply, and through the dollar index that summarises both.

Two anchors, and here is the consequence. The rate differential between them is an object in its own right, and it can move when neither bank does anything. If the market's expected path for one shifts while the other's stays put, the gap has changed — and the gap is what prices the currency leg. Reading only the domestic meeting calendar means missing half of what moves the rupee cost of capital. On a day when both are moving the same way at once, the result is the pattern described in what a risk-off day looks like.

Where the chain is weaker than it looks

Four places this argument is thinner than the standard telling admits.

The surprise is inferred, never observed. Everything in the decision-versus-surprise framing depends on knowing what was priced, and what was priced is backed out of instruments that carry premia of their own. The framework is sound; the input to it is an estimate.

Attribution on the day is not measurement. A policy statement lands in a market that is simultaneously absorbing data, flows, earnings and everything overseas. Saying an index moved because of the statement is an attribution, and a same-session move is consistent with several explanations at once.

Transmission is slow, partial and uneven. A policy rate reaching bank lending rates, deposit rates and actual borrowing behaviour takes quarters, and it does not arrive equally across borrowers. The financial-market repricing described above happens in seconds; the economic effect it is supposed to represent takes far longer and may be smaller than the price move implies. The transmission half of that is its own article.

The rate is not the only lever. Balance-sheet operations, liquidity management, reserve requirements and regulatory settings all change financial conditions, and none of them shows up as a change in the headline policy rate. An observer watching only the rate is watching one instrument out of several.

Reading this on the Macro page

FNOTrader's Macro page carries a central-bank card, and how it is built is worth stating plainly — partly because one of its choices is an honest compromise, and partly because the card relates to the rest of the page in a way that is easy to get wrong.

Four banks, India first. The card shows the Reserve Bank of India, then the US Federal Reserve, then the European Central Bank, then the Bank of Japan. India leads because it is the home bank. The People's Bank of China is absent, because the data source carries no clean series for it — an omission by design, not a gap.

The India card is a proxy, and the limitation is worth carrying in your head. A market-determined overnight rate can sit away from the policy rate when liquidity is tight or flush, so a change on that card is not proof that a decision was taken, and a decision is not guaranteed to show up cleanly. It is also monthly, which means it cannot show you a meeting-day move at all. We use it because it is the honest series available rather than the ideal one, and saying so is better than implying a precision the feed does not have.

The "Hiking / Cutting / On hold" label is arithmetic, not the bank's own stance. The page takes the latest observation, walks back to the most recent one at least ninety days older, and labels the difference: above 0.02 percentage points reads Hiking, below −0.02 reads Cutting, anything between reads On hold. That is a backward-looking difference over a fixed window, computed from a 220-day pull. A bank that has just signalled a change and not yet acted still reads On hold, and on a monthly series the label can lag a decision by weeks. If a card is missing entirely, its data fetch failed — the page drops it rather than showing an empty tile.

Now the part most readers will not expect: none of these rates enters the page's composite score. The score is built from the market tiles, which the page groups as currencies; rates, bonds and credit; commodities; volatility; and global equities — plus foreign institutional cash flow. No policy rate is among them, and that is deliberate. It follows directly from this article's argument. A policy rate is the realised past of something markets have already priced. Adding it to a score built from market-priced inputs would double-count the expectation and add the stalest input on the page. Policy reaches the score the way it reaches everything else — through the US 2-year yield, the dollar and the yen cross, which are the instruments that carry the expected path. The card is context, and it is placed where context belongs.

The score itself is score = 100 × Σ(w_ic_i) ÷ Σ(w_i), over twenty weighted tile inputs summing to 1.33, of which the five largest are the dollar index at 0.20, the US 10-year yield at 0.15, Brent at 0.12, and the yen cross and the US VIX at 0.10 each — plus foreign flow folded in at 0.12 and saturating at ±₹5,000 crore, which is not in the tile table at all. Twenty-one inputs, then, and a divisor of 1.45 on a day when every feed reports and the flow number is in.

That divisor is worth one more sentence, because it is the part people misread. When a feed fails, its weight leaves the numerator and the denominator — so 1.45 is a maximum, not a constant, and the score is always an average of whatever actually reported rather than a fixed-scale reading. Two days with the same headline score can rest on different sets of inputs.

Regime boundaries sit at ±20, with a stress override at a score of −35 or lower, or a one-day US VIX rise of 20% or more, or a one-day USD/JPY fall of 1.2% or more. The stress test is evaluated first, so either of those two one-day moves labels the day Stress even when the composite is positive — a deliberate choice that lets a single fast move overrule the average. Every one of those numbers is FNOTrader's modelling judgement — a considered view about what matters to Indian equities, not a measured constant and not a fitted coefficient. A different reasonable weighting produces a different score from identical inputs.

Two more page conventions, both counter-intuitive on first read. Colour means impact on Indian equities, not the direction of the number — a falling dollar shows green because a softer dollar loosens global conditions, and a rising USD/JPY also shows green, because a weak yen means the carry trade is intact. And the page carries no swap or policy-futures series, so the priced path this article turns on cannot be read off it directly; the closest thing available is the US 2-year yield tile, which is a market instrument and not a survey. How the tiles are meant to be read together rather than one at a time is the macro pillar.

Common questions

Continue reading

More in Macro & Markets · App: Options Analytics · Definitions: glossary · Free tools: calculators · All: every article