Downside: The Case That Gets Less Attention and Matters More

A downside case is the same arithmetic as the main case, run with every assumption pushed the way that hurts. A downside case arrives by two routes: the business does less than was assumed, or it does exactly what was assumed and other people pay a lower multiple for the same earnings. Reverse Sarvani Coatings Limited's 3.0 point gross margin gain and the growth the price carries rises from 24.2 to 29.7 per cent a year.

Everything below rests on four things already settled. The backward arithmetic starts at a quoted price and asks what that price assumes rather than what it should be. The two assumptions that drive that arithmetic, a 12 per cent yearly return and an exit rating of 25 times, both of which are the analyst's own and are labelled as chosen every time they appear. The gap between the growth the price carries and the growth the business has shown, worked separately. And the published ladder for Sarvani Coatings Limited, an invented paints and coatings maker, whose year three revenue was Rs 2,415 crore and whose year three profit after tax was Rs 278 crore.

So what is a downside case?

A downside case is not a mood, a warning or a second opinion. The case is the same sheet of assumptions already built, with each line moved in the direction that hurts, and the same arithmetic run over it again. A downside case is a procedure, not a temperament, and a procedure takes an hour rather than a week.

Think of a caterer pricing a wedding season. The main case says onions stay where they are, the hall takes the bookings it took last year, and the staff cost holds. The downside case is the same spreadsheet with onions up, two bookings gone and one extra hand on the payroll. Nobody has changed their view of the business. The caterer has changed four cells and pressed the same total.

The finance version is identical in shape. For Sarvani Coatings the assumption sheet holds a gross marginRevenue less the cost of the materials that went into the product, stated as a share of revenue. Gross margin is the first line of the profit ladder and sits above every operating cost., a volume growth rate, the costs below the gross line, and the rating other people are assumed to apply at the end. Move any of them and the same three steps run: down the ladder to profit before taxWhat is left after every operating cost, depreciation and interest, but before the tax charge. The line the tax rate is applied to., across the tax charge, and out into earnings per shareA whole year of profit after tax, spread across every share in issue. Earnings per share exists so results from businesses of very different sizes can be set side by side on one scale..

Why does almost nobody write one?

Because nothing in the process asks for it. The absence of a prompt is the whole answer, and it is worth stating as mechanism rather than as a complaint about analysts. The fix that follows from mechanism is a procedure. The fix that follows from complaint is a resolution to try harder, and no resolution survives a busy week.

Four things are true at once. Writing one is unpleasant. The case argues against work just finished. The case has no natural audience. The person who asked for the note wanted the view, not what would break it. Writing one weakens a case still being assembled, at exactly the moment an author is least willing to weaken it. And nobody is scored on it. No reader ever came back to say the downside case was excellent. All four are true of a careful writer as much as a careless one. The answer is therefore a rule about when the case gets written, not an appeal to discipline.

What are the two ways a downside actually arrives?

There are exactly two, and telling them apart is the most useful habit here. The two need different evidence, they arrive at different speeds, and only one of them involves anybody at the company doing anything wrong.

Route one is the business. Volumes come in lighter, the margin does not hold, a cost line runs away. Earnings end up below what was assumed, and the price follows the earnings down. Route one is the route everybody pictures when they hear the word downside, and it produces news somebody can point at.

Route two is the rating. The business delivers exactly what was assumed. Revenue, margin and earnings all arrive on the line drawn for them. And the people who set the price decide, for reasons that have nothing to do with this company, to pay a lower multiple for the same rupee of earnings than they were paying before. Route two needs no bad news about the company at all. Route two is therefore the one that catches a careful analyst by surprise.

What happens to the earnings if the margin gain unwinds?

Now take the one assumption everything earlier in this material deliberately left open. Between year one and year three, Sarvani Coatings' gross margin rose 3.0 points, from 43.0 per cent to 46.0 per cent, and the published statements do not separate a sector wide pricing environment from the company's pricing or from a mix shiftA change in what a company sold rather than how much. Selling more of a higher priced line lifts the average even if no individual price moved. towards industrial work. Three explanations fit and nothing on file rules any of them out. The missing explanation is not a criticism of the company. The missing explanation is a gap in the record, and a gap in the record is exactly what a downside case is for.

So reverse the whole of it, and keep everything else exactly as published. Holding the rest still is a simplification chosen on purpose. Three points of gross margin on year three revenue of Rs 2,415 crore is Rs 72.45 crore. Rs 72.45 crore comes straight off gross profit. Nothing below the gross line is being touched, so the same amount comes straight off profit before tax as well: Rs 371 crore becomes Rs 298.55 crore.

Tax next, and here the arithmetic has to be handled carefully. The record prints the year three effective tax rateThe tax charge stated as a share of profit before tax. Allowances, past losses and items taxed differently pull the effective rate away from the headline rate. as 25.1 per cent, but 25.1 is a rounded number. The published absolutes are a tax charge of Rs 93 crore on profit before tax of Rs 371 crore. Those two figures give 25.07 per cent, and 25.07 per cent is the figure used here. The test is simple: run the reversal dial back to zero and the arithmetic has to hand back the published Rs 278 crore exactly. On 25.07 per cent it does. On the rounded 25.1 per cent it returns Rs 277.88 crore, and that small error compounds into the wrong printed digit further down. Tax on Rs 298.55 crore is therefore Rs 74.84 crore, and profit after tax is Rs 223.71 crore against the published Rs 278 crore.

On 24.00 crore shares, that is Rs 9.32/- against Rs 11.58/-. Reversing a gain the record never explained takes earnings per share down by Rs 2.26/-, or 19.5 per cent, without anybody at the company having done anything wrong.

Then the second half, and the second half is the part that surprises people. The price has not moved in this scenario. The price is still Rs 486/-. The two stated assumptions have not moved either: still a 12 per cent yearly required returnThe yearly return an investor settles on before putting money at risk. A required return is a chosen preference, not a measurement anybody publishes. and still an exit rating of 25 times, which together mean the shares would need to reach Rs 856.50/- in five years and therefore need earnings per share of Rs 34.26/- to get there.

The destination has not changed. The starting point has. Getting from Rs 11.58/- to Rs 34.26/- in five years needs 24.2 per cent a year. Getting there from Rs 9.32/- needs 29.7 per cent a year, a rise of 5.5 points. The same journey now starts further back, so a lower earnings base raises the growth the price carries rather than lowering it. That is the shape worth carrying away: earnings fall, and the demand hidden inside the unchanged price goes up.

Reversing the whole gain is the extreme end of a range, not a prediction. A partial reversal scales down smoothly, and the middle of the table below is where an honest case usually sits. One point back gives Rs 10.83/- and 25.9 per cent. Two points back gives Rs 10.08/- and 27.7 per cent. Nothing in the record says which of these, if any, is the right one to write down, and that is precisely why the case is written as a range with a reason attached to each step.

If the business is fine, what would the rating have to do?

Route two is worked backwards, and it has to be. The moment route two produces a price it has produced a valuation, and a valuation is not what the backward arithmetic exists to make. So the question is put in the only form that stays on the right side of that line: if the business does something entirely ordinary, what would the rating have to be at the end for the stated return to be met?

Ordinary means the sector rate. The sector is growing 11.0 per cent, so let earnings do the same for five years. One detail decides the answer here: the record prints earnings per share as Rs 11.58/-, but Rs 278 crore over 24.00 crore shares is Rs 11.5833/-, and the compounding below runs on the unrounded figure. Rs 11.5833/- compoundedGrowth applied to a base that already includes the earlier growth, so the fifth year builds on the fourth rather than on the start. at 11.0 per cent reaches Rs 19.52/-, where the printed Rs 11.58/- would have given Rs 19.51/- and quietly moved the answer. The stated 12 per cent required return on Rs 486/- needs Rs 856.50/- in five years. One divided by the other puts the rating at the end at about 43.9 times, against about 42.0 times today.

Read that finding carefully. Over reading it is easy. The finding does not say the shares are expensive. The finding does not say anything about what they are worth. Under the two stated assumptions, a perfectly ordinary five years is consistent with today's price only if the rating at the end is higher than it is now. The higher rating is a written assumption open to testing, not a judgement anybody has made.

There is a neat way to see why. If the rating is held exactly where it is, the price moves in step with the earnings, so getting 12 per cent a year needs earnings to grow at 12 per cent a year. The sector grows 11.0 per cent. One point a year of shortfall compounds over five years into 4.6 per cent, and 41.96 times lifted by 4.6 per cent is 43.9 times. The whole of the gap is that single point a year.

One route through an argument is never enough, so a cross check runs a second way. Suppose instead the company repeats last year exactly: revenue growth of 13.9 per cent with the margin held where it is. Earnings per share reaches Rs 22.21/- in five years. Satisfying the same 12 per cent then needs a rating of about 38.6 times, below today's 42.0 times rather than above it. So the finding is not that the rating must always rise. The finding is that the answer sits on a knife edge between two entirely reasonable growth paths, and which side it lands on depends on an assumption nobody has settled.

Does a high rating make this sharper, and if so how?

Yes, but not for the reason usually given, and the difference matters enough to work through. The common claim is that at a high rating more of the required return depends on the rating holding. Checked against the arithmetic, that one does not survive. The proportional rise route two demands is 1.12 to the fifth over 1.11 to the fifth. The answer is 4.6 per cent whatever the starting rating. At 15 times it is 4.6 per cent. At 20 times it is 4.6 per cent. At 30 times it is 4.6 per cent. The level cancels out.

The level genuinely governs one thing: how much of the earnings the price needs already exists. The stated assumptions require earnings per share to reach Rs 34.26/- by year five. Year three produced Rs 11.58/-. So 33.8 per cent of what the price needs has already been earned, and 66.2 per cent of it has not happened yet. At a rating of 20 times the same arithmetic leaves 70.9 per cent already in hand. At 15 times, 94.6 per cent. The higher the rating, the larger the share of the required earnings that is still in the future, and the future is the part a downside case can actually reach.

There is a second, plainer consequence of the level. The exit assumption of 25 times sits 17.0 times below today's rating, a fall of 40.4 per cent, and that fall is already built into the 24.2 per cent growth figure. A reader starting from a rating of 20 times with the same exit assumption would be assuming the rating rises rather than falls. Same arithmetic, opposite sign, purely because of where the starting rating sits. None of that is a comment on how good the business is. The arithmetic is a description of the price.

What has to be in it before it is usable?

Three things per assumption, and a case missing the third is not a case. First, the assumption moved and where it moved to. Second, why that far and not further. A move with no size argument behind it can be dismissed by anybody who prefers a smaller one. Third, and this is the one that gets left out, the thing a reader could actually watch that would tell them this route is opening.

The observable is what turns a case into something that can be updated. Reverse the margin by 3.0 points and the observable is not the share price. The observable is realisationRevenue per unit sold. Realisation rises when prices rise or when the sales mix moves towards costlier lines, and it is the figure that has to outrun input cost for a margin to widen. in the quarterly numbers no longer running ahead of input cost, or the peers reporting the same margin turning at the same time, or a discounting season starting in the trade. A downside case with no observable attached is a mood. Nobody can ever tell whether it is coming true.

Who writes one of these, and when does it earn its keep?

Four kinds of reader use a downside case, and each uses it slightly differently. An analyst uses it as a dated file entry: the case exists before the event, so when the event arrives the question is whether the observable fired, not whether the story can be retold. A fund manager sizing a position uses it as the input to how much of the book goes into one name. The size of a holding is a judgement about the downside case and not about the main one. A lender looking at the same company uses only route one. A lender is repaid out of cash rather than out of the rating, so a de-rating that halves the shares may leave the loan entirely untouched.

And a household uses the same habit without any of the vocabulary. A household running on one salary that writes down, on a quiet Sunday, what the next six months look like if that salary stops has built a downside case with an observable attached. The observable is the first missed appraisal cycle or the first round of quiet exits at the office. Written in advance it changes what the household does. Written after the salary has stopped, it is just an account of what happened.

When is it written?

On the same day as the main case, and dated. The date is the whole rule, and the date is worth more than any amount of care taken over the contents. A downside case written after the price has moved has been shaped by the move. The writer already knows the answer the case has to accommodate.

Sarvani Coatings Limited is invented.
Educational material. Not advice on any investment, tax, budget or market position.