Macro Transmission: How a Variable Reaches a Company's Numbers

Transmission is the path a macro variable takes to arrive inside a company's own numbers. A variable arrives along named routes: what the company sells, what it buys, what it owes, and what its customers can afford. A variable touching none of those four does not arrive at all. Operating profit is a small difference between two large numbers, so what does arrive lands magnified.

The panel above opens on one worked chain, written out here in full so it can be read without touching anything. A policy move of 200 basis points enters it. The first link, from the policy rate to the benchmark the lender publishes, passes 80.00 per cent and takes three months, so 160.00 basis points leave it. The second link, from that benchmark to the rate on Nirvi Engineering's own borrowing, passes 62.50 per cent and takes six months, so 100.00 basis points leave it. The third, from the rate on the borrowing to the interest actually charged in a reported period, passes 50.00 per cent and takes three months more, so 50.00 basis points arrive. Multiply the three shares and 25.00 per cent of the move survives the chain; 200 basis points at 25.00 per cent is 50.00 basis points, and the three leaks of 40.00, 60.00 and 50.00 add to the 150.00 basis points that never arrived. On Rs 200 crore of borrowing that takes the rate from 9.00 per cent to 9.50 per cent and the interest charge from Rs 18 crore to Rs 19 crore, a rise of Rs 1 crore. Against operating profit of Rs 100 crore that is 1.00 per cent, against revenue of Rs 1,000 crore it is 0.10 per cent, and the ratio between those two readings is 10.00 times. Ten times is one divided by the 10.00 per cent margin. Three months plus six plus three is twelve, so all of it lands in month twelve and none of it lands before.

Two separate things are happening there, and running them together is how a reader draws the wrong conclusion from a perfectly ordinary set of accounts. Leakage decides how much arrives. Lag decides when. Set the reading window in the panel to six months and the whole account changes character. The interest line shows nothing, not one rupee, and every part of the chain stands exactly as it was. The move has cleared the first link, the benchmark has moved 160.00 basis points, and the borrowing has simply not reached its reset dateThe date written into a floating rate loan on which it takes the current benchmark. Between two reset dates the rate on the loan does not move at all, whatever the benchmark does. yet. A nil reading taken before the lag has run is not evidence that the mechanism failed, it is evidence that the reading was taken early, and telling those two readings of the same blank line apart is a large part of what a transmission account is for. That is why the panel prints what is still in transit beside what is showing, and why the two always add to the same total.

The variables themselves are covered separately: the rate a central bank sets and the corridor it sits in, the price level and the basket it is measured on, the exchange rate, the output measure, a commodity move, and the vintage a released figure carries. Transmission begins one step after all of them, at the moment a variable has already moved and somebody has to say what it does to a company. Asking what a move does to one company is a different question, and it is the one an analyst is actually asked.

Five kinds of quantity run through the arithmetic below, and mixing them is the fastest way to get a wrong answer that looks right. Money is in rupees, grouped in crore. A move in a variable is a per cent. The gap between two per cent readings is a point. A move in an interest rate is often quoted in basis pointsOne hundredth of one percentage point. A rate moving from 9.00 per cent to 9.50 per cent has moved 50 basis points.. And a share is a per cent of a stated base. The base is the part people drop. A share with no base named is not a usable number, and the failure this guide exists to prevent is somebody quietly changing the base halfway through a sentence.

By what routes does a macro variable reach a company?

Start with an ordinary household. A household already runs this calculation without calling it anything. Suppose the price of cooking gas moves. The question is how it reaches that household. The move does not reach the household by being announced. It reaches the household at the next cylinder bought, along the buying route. For someone who drives for a living, it also reaches them along the selling route, through what customers will pay for a ride. If a loan was taken against the vehicle, a change in lending rates reaches them through the instalment, along the owing route. And if the people who hire them have less left over at the end of the month, that reaches them too, through what their customers can afford.

Now suppose the price of gas moves and that household cooks on an induction plate, walks to work, holds no loan and sells nothing to anybody. The move happened. It was reported. The move did not reach them, and no amount of reading about it will make it reach them.

A macro variable arrives at a company along exactly four routes: what the company sells, what the company buys, what the company owes, and what the company's customers can afford, and a variable that touches none of the four does not arrive at all, however large the move and however loud the headline. Most macro news is exactly that for most companies, so the last clause is the part worth holding on to. Not small. Absent.

The company used throughout this guide is Nirvi Engineering, invented, sitting inside the invented Republic of Sankhya. The company is minimal: revenue, costs, and one debt line, and nothing else. Every figure attached to it can therefore be published in full and recomputed independently. Nirvi Engineering has revenue of Rs 1,000 crore and costs of Rs 900 crore. The difference is operating profitWhat is left of revenue after the costs of running the business, taken before interest and tax. It is the line a macro move usually lands in. of Rs 100 crore. Of the revenue, Rs 300 crore is exported and Rs 700 crore is sold at home. Of the costs, Rs 360 crore is imported inputs and Rs 90 crore is oil. The debt is Rs 200 crore at 9.00 per cent, giving interest of Rs 18 crore.

Why is the effect on profit larger than the effect on revenue?

Profit is where macro reaches a company at all, and the reason is arithmetic rather than opinion, so it is worth doing rather than accepting. Notice the shape of Nirvi Engineering before any variable moves: revenue and a cost baseEvery rupee a company spends to produce and sell what it sells in a period, added together. It is the base a cost share is measured against. both near a thousand crore, and a gap between them of Rs 100 crore.

Now one of them moves. Suppose Nirvi Engineering's whole cost base rises by 1.00 per cent. Taken in steps, with every step written out: one per cent of Rs 900 crore is Rs 9 crore. Nothing about revenue changed, so revenue is still Rs 1,000 crore. Costs are now Rs 909 crore. Operating profit is Rs 1,000 crore less Rs 909 crore, or Rs 91 crore. The profit fell from Rs 100 crore to Rs 91 crore, and Rs 9 crore on a starting Rs 100 crore is 9.00 per cent.

Set against revenue instead, the same Rs 9 crore on Rs 1,000 crore is 0.90 per cent. Same rupees, same event, same company, and two readings that differ by a factor of ten. Dividing 9.00 by 0.90 gives exactly 10.00. Ten is not a coincidence: it is one divided by the margin, and the margin here is 10.00 per cent.

Operating profit at Nirvi Engineering is a tenth of revenue, so a 1.00 per cent move in the cost base is Rs 9 crore. Rs 9 crore is 9.00 per cent of profit and only 0.90 per cent of revenue, and the magnification is exactly ten times because ten is one divided by the margin. The magnification is arithmetic on a thin residual and nothing more: not a claim that profit is volatile, and not a claim that anything is likely to move. The arithmetic says only that when one large number is subtracted from a slightly larger one, a small proportional change in either shows up as a large proportional change in the difference.

Why can the same variable reach one company and leave another untouched?

Two food stalls stand outside the same office building. One fries its snacks and buys oil every week. The other sells only cold drinks from a chest. The price of frying oil moves, and one of them has a problem while the other has only a headline. Neither stall is better run than the other. The two stalls simply buy different things, so the same event has a route into one and no route into the other.

Put two invented Sankhya companies side by side on exactly that basis. Nirvi Engineering buys Rs 360 crore of imported inputs, or 40.00 per cent of its Rs 900 crore cost base. Tarasi Handlooms is identical to it in every published respect except one: its input shareThe portion of a company's total costs that one particular purchased item accounts for. Naming a share without naming the base it is a share of makes it unusable. of imported inputs is zero. Tarasi Handlooms buys nothing at all from outside Sankhya.

Now let the Marut unit, the invented Sankhya currency, move 5.00 per cent in the direction that makes imports dearer. For Nirvi Engineering, 5.00 per cent of Rs 360 crore is Rs 18 crore of extra cost, and against operating profit of Rs 100 crore that is 18.00 per cent. For Tarasi Handlooms, 5.00 per cent of zero is zero. Not a small effect. Not a delayed effect. Zero, exactly, and it stays zero however far the Marut unit travels.

A share of zero multiplied by any move whatsoever is zero, so a currency move is not a macro event for Tarasi Handlooms at all, and that is why arithmetic done on a company's own exposures beats arithmetic done on the economy when the question is about a company. The same arithmetic is why an analyst reads by sectorA grouping of companies that buy, sell and borrow in broadly similar ways, so that one macro move tends to reach them through the same routes. before reading by economy. Companies inside one sector tend to share their routes in. Shared routes make the grouping useful and also approximate: two companies in one sector can still differ on the shares that decide the size.

What decides how much of a move actually arrives?

Three things stand between a headline and a rupee figure in a company's accounts, and each of them can only shrink the move, never enlarge it. The first is whether the route exists at all, the yes or no set out above. If the answer is no, nothing that follows matters. The second is what share of the base the route touches. The third is how much of the move can be handed on to somebody else, either passed to buyers in a selling price or absorbed by a supplier in a purchase price.

One case worked through. The material on reading global oil and commodity price moves established a compounded rupee oil move of 26.00 per cent, and Nirvi Engineering buys oil, so the route exists. Oil is Rs 90 crore of the Rs 900 crore cost base, a share of 10.00 per cent, so the move meets Rs 90 crore rather than Rs 900 crore. Then 26.00 per cent of Rs 90 crore is Rs 23.40 crore of extra cost. Against operating profit of Rs 100 crore that is 23.40 per cent, almost a quarter of the profit from a single input at a tenth of costs. Sizing how much of that Rs 23.40 crore could be handed on needs a measure covered separately, so the third gate is left unsized.

All three gates shrink a move and none of them enlarges it, so what arrives inside a company is always less than the headline, and a transmission estimate that comes out larger than the move that caused it has skipped a gate rather than found an amplifier. The magnification set out in the block above is a separate thing and it is worth keeping the two apart. Shrinking happens on the way in, as the move meets a smaller and smaller base. Magnification happens at the very end, when whatever survived is finally set against a thin residual. A move can shrink at every gate and still land as a large share of profit. The oil case does exactly that.

Why do the margin and the profit move by different amounts?

Here is a trap that catches careful readers. Both numbers are correct and they disagree. Operating profit is a rupee amount. The operating margin is that amount divided by revenue. When a macro move touches costs alone, revenue holds still and the two move together in a fixed way. When the same move touches revenue as well, the margin's denominator moves too, and the two readings come apart.

Three cases on Nirvi Engineering make it plain, and the table publishes every intermediate so any row can be rebuilt from it. The first moves costs alone. The second moves revenue and costs by the same 2.00 per cent. The third moves them by different amounts, revenue by 2.00 per cent and costs by 3.00 per cent. Read the profit column and the margin column across the three rows and watch where they stop agreeing.

The second row is the one to keep: operating profit moved 2.00 per cent while the operating margin did not move at all, so a transmission account quoting only one of the two has reported half of what happened. Read that row slowly. A reader given only the margin would report that nothing happened to Nirvi Engineering, and Rs 2 crore of extra profit says otherwise. A reader given only the profit would report a gain and miss that the company is no more profitable per rupee of sales than it was. Neither reader is wrong about their own number. Both are wrong about the event.

What does one company look like carried through every route at once?

Everything above has been one route at a time. Set out together for Nirvi Engineering, sized in rupees, the relative weights become visible. Each row states its own base rather than assuming it is remembered. Oil and imported inputs are not stated as separate parts of the cost base, so the shares are never added to each other.

The owing route is stated in basis points rather than per cent because a rate is quoted as a level, and a move in a level is a point rather than a proportion. Its arithmetic is the panel at the top of this guide carried to its end, and one thing about that last column needs saying plainly: the extra interest is a charge below operating profit rather than a cost inside it. The extra interest is set against operating profit in the table because operating profit is the yardstick every route here is measured against, not because it reduces the operating line.

Now change one thing and nothing else. Bhadra Ceramics is a third invented Sankhya company with the same revenue of Rs 1,000 crore but a cost base of Rs 800 crore, so its operating profit is Rs 200 crore and its margin is 20.00 per cent. Put the identical 1.00 per cent cost move through both, and read each step of it off the figure below.

One macro event, identical in every respect, arrives as 9.00 per cent of profit at a company with a 10.00 per cent margin and 4.00 per cent of profit at a company with a 20.00 per cent margin, so how much of a move a company feels is decided by its own shape and not by the size of the move. The general form is easy to hold: the share of profit is the cost share of the move divided by the margin. Doubling the margin roughly halves the arrival. Two companies reading the same headline can therefore reach very different numbers without either having made an error.

What does an analyst establish before estimating anything?

The order somebody experienced works in is the reverse of the order that feels natural. The natural order is to read the headline, feel that it matters, and start estimating. The working order is to establish which of the four routes exists first, and only then to reach for a number.

The reason is that the three gates run in sequence and the first one is binary. An estimate built on a route the company does not have is not a rough estimate, it is a wrong one, and it is worse than no estimate because it carries a rupee figure and therefore looks like work. A lender sizing whether a borrower can still service Rs 18 crore of interest asks first whether the borrowing is at a rate that moves at all. An equity analyst sizing an inflation move asks first whether the company buys the thing that got dearer. Somebody running a household budget against a fuel move asks first whether they drive. In each case a wrong answer to the first question makes every later decimal place meaningless.

Establishing which routes exist costs an hour and reading a company's own disclosure of what it buys and where it sells. A careful estimate built on a route the company does not have can be defended for months because it looks like arithmetic. The order also protects against the opposite error, the quieter one. A route that exists but was never listed produces an estimate that is too small, and nobody ever queries an estimate for being too small. Listing the routes exhaustively before sizing any of them is the only step that catches both.

What can no transmission account deliver?

Three things, and the third is the one that gets forgotten. A transmission account cannot say what a variable will do. Everything in this guide began with the words suppose it moves, and a stated assumption carried through arithmetic is not a prediction about the world however carefully the arithmetic was done. A transmission account cannot say what a company will report either. The gates above were sized on published shares that themselves change, and a company facing a cost move does things about it that no chain of shares can anticipate.

And it cannot deliver a value. The step from a profit effect to a value is exactly the step a reader will try to take unaided, so the limit is worth stating flatly rather than leaving as an omission. A lower interest rate tends to raise a valuation, and that much is a direction. Computing one is a different matter. Turning a rate into a value needs a model with its own assumptions, its own inputs and its own failure modes, and that model belongs to valuation rather than to macro. The model is set out under How Interest Rates Feed Into Equity Valuation.

A transmission account gives a direction and a size, expressed in rupees and as a share of a named base, and it stops there, so the work is finished at the point where a value would begin. The boundary is not modesty but what keeps the arithmetic checkable. Every figure above can be recomputed from components printed in this guide, and the moment a value entered, the answer would depend on assumptions nobody printed and nobody could check.