Upside: What the Number Means and What It Assumes

Upside is normally the distance from today's price to a level somebody has named, divided by the price. No level is produced anywhere on this platform, so no percentage can be either. The honest version of the same idea survives: the gap between the growth a price already carries and the growth a record supports. On the figures below, 24.2 against 13.9 per cent, a distance of 10.3 points a year.

Underneath that sits one observation that decides everything else here. A quoted price is a fact anybody can look up on a stated date. A level is a sentence somebody wrote. Put the two together and the result is a single tidy percentage in which the fact and the sentence are no longer distinguishable, and the reader inherits somebody's assumption while believing a measurement has been handed over. Taking that percentage apart means throwing away the half nobody can check and rebuilding the useful half in a form two people can actually disagree about.

What does the number actually measure?

Take it literally for a moment. A note says upside of 30 per cent. Somewhere behind that sentence sit two numbers. The first is the price the share last traded at, public, timestamped and identical for everybody looking at the same screen. The second is a level the writer named for some date in the future. Subtract the first from the second, divide by the first, and out comes 30 per cent.

So the percentage is two numbers wearing the appearance of one, and only the first of the two was ever observed. That is not a criticism of the writer. Naming a level is exactly what a great deal of published research is for. The split is a statement about where the arithmetic came from, and the arithmetic came half from a screen and half from a person.

The shape is familiar from somewhere less abstract. A shopkeeper says a jacket is forty per cent off. One of those two numbers, what is paid today, is real and can be counted out of a wallet. The other, the price it was supposedly reduced from, is a number the shop wrote on a card. The discount percentage is the honest arithmetic of one real number and one written one, and every argument anybody has ever had about a sale is an argument about the second number rather than about the subtraction.

So why can the percentage not be produced at all?

Because producing the percentage means producing the level, and no level for any share is produced anywhere on this platform. The reasoning behind that refusal is set out in full elsewhere in this material. The consequence matters more here than the reasoning. If the second number is unavailable, the percentage built on it is unavailable too, and quietly picking a level would be the same act with the working hidden.

The interesting question is not whether the percentage can be done without, but whether anything is actually lost when it goes. Most readers assume something is. Worked through, what the percentage was carrying survives intact, provided the same idea is stated in a slightly longer form.

What is the honest version of the same question?

Here is the restatement. Instead of asking how far the price is from a level, two questions with checkable answers take its place. First: what growth is the current price already carrying, given the assumptions somebody is prepared to state out loud? Second: what growth does the published record actually support? The two then sit side by side, and the distance between them is stated in percentage pointsThe plain arithmetic difference between two percentages. Going from four per cent to six per cent is a rise of two percentage points, and also a rise of fifty per cent. The unit has to be said out loud. a year.

The first of those is settled separately. Starting from an illustrative price of Rs 486/-, a reader who wants 12 per cent a year and who assumes the shares will be valued at twenty five times earnings in five years is carrying an expectation of about 24.2 per cent compoundA rate applied each year to the result of the year before, so the base it works on already contains every earlier year's growth. Two rates that look close diverge a long way over five years. earnings growth a year. Both of those inputs belong to the reader and are labelled that way wherever they appear. If either one changes, the 24.2 per cent changes with it.

The restatement loses nothing a reader could have used, and it gains the one thing the percentage never had, something to disagree about. Two people looking at upside of 30 per cent can only trade opinions about whether it feels high. Two people looking at 24.2 per cent required against 13.9 per cent supported can argue about the required return, about the exit ratingThe number of times earnings a share is assumed to be valued at on some future date. Where such an assumption comes from, and what it holds inside it, is taken up separately in this material., about which parts of last year repeat, and about what evidence would settle each of those. Every one of those arguments is winnable. None of them is available inside a single percentage.

How is what the record supports worked out?

The second of those two questions is the harder half, and it is where most of the mistakes live. The price side is mechanical: two stated assumptions go in and one rate comes out. The record side requires a decision about which parts of what a company already did are the kind of thing that happens again.

The tempting answer comes first. Profit after tax at Sarvani Coatings Limited, an invented maker of decorative paints and industrial coatings whose shares have never traded, went from Rs 197 crore in year two to Rs 278 crore in year three, a rise of Rs 81 crore, or 41.1 per cent in one year. Taken as the supported rate, 41.1 per cent does not merely close the gap, it sails past the 24.2 per cent the price is carrying, and the whole exercise ends in about four seconds with the wrong answer.

Splitting the Rs 81 crore shows why. Revenue rose 13.9 per cent over the same year, from Rs 2,120 crore to Rs 2,415 crore. Had net marginProfit after tax expressed as a percentage of revenue, so it carries every cost the profit ladder deducts rather than only the direct ones. How the ladder is built is taught in the accounting material. stayed exactly where it was in year two, that revenue alone would have produced Rs 224.4 crore of profit. Net margin moved from 9.29 per cent to 11.51 per cent in the same twelve months, and the year produced Rs 278 crore instead.

So about 66.2 per cent of last year's profit growth came from a margin that moved, not from a business that got bigger. And a margin is a ratio that stepped up once, not a rate that repeats. The largest single piece of that Rs 53.6 crore is the gross marginRevenue less the direct cost of what was sold, expressed as a percentage of revenue. The contents of that direct cost, and the way it is measured, belong to the accounting material. gain of 2.0 points in year three, from 44.0 to 46.0 per cent, which is worth Rs 48.2 crore of gross profit and roughly Rs 36.1 crore once the year's effective tax rateThe tax charge shown in the accounts divided by profit before tax. The effective rate is an outcome of the year rather than the statutory rate, and the two are rarely the same number. of 25.1 per cent is taken off, or 44.6 per cent of the whole Rs 81 crore rise. Employee and other costs grew more slowly than revenue did. The rest of the margin move sits below the gross line, and cost growing slower than sales is operating leverageWhat happens to profit when costs that do not rise in step with sales are spread over a bigger revenue line. Operating leverage flatters a growing year and works just as hard in reverse. rather than anything new about the business.

The level shift comes out of the record before the record is turned into a rate. The same thing stated as a ten second test: take the ratio that moved, repeat its move five more times, and see where it lands. Gross margin rose 2.0 points in year three. Repeated, that gives 46.0, then 48.0, then 50.0, 52.0, 54.0 and 56.0 per cent five years out. No further evidence is needed to reject that path. The shape alone is enough to reject it.

One warning about periods is easy to trip over. Over the two years from year one to year three, gross margin rose 3.0 points, from 43.0 to 46.0 per cent. The 3.0 points is a two year figure and it does not belong beside a one year profit growth rate. The 41.1 per cent above is one year, so the margin move that sits beside it is the one year move of 2.0 points. Mixing the two periods is the quietest way to get a wrong answer that still looks carefully worked.

The gap, worked once, on one set of figures

The price side first, carried in rather than built again. At Rs 486/-, wanting 12 per cent a year, the price would have to reach Rs 856.50/- in five years. At an assumed twenty five times, that needs earnings of Rs 34.26/- a share. Year three earnings per share is printed as Rs 11.58/-, and Rs 278 crore over 24.00 crore shares is Rs 11.5833/- before rounding. Growing the unrounded figure to Rs 34.26/- over five years takes 24.22 per cent a year; growing the printed Rs 11.58/- takes 24.23 per cent. Both are 24.2 per cent to the decimal printed here. The answer does not depend on which version of the per share figure was picked up.

Now the record side, built the careful way. Revenue grew 13.9 per cent last year. Hold net margin where the record left it and earnings grow at the same 13.9 per cent. The 13.9 per cent is the defensible rate, defensible precisely because it assumes nothing about the margin question left open above.

The price carries 24.2 per cent a year, the record supports 13.9 per cent a year, and the distance between them is 10.3 percentage points a year. The distance measures how far apart two rates are, and nothing more. A distance is not a verdict, not a level, and not a return anybody could earn.

What would actually have to happen to close 10.3 points?

A distance stated in percentage points is still slightly abstract. Convert it into the things the earlier material actually measured and it stops being abstract immediately. There are two clean routes, and they demand completely different things.

Take the revenue route first. Hold margin flat and let revenue do all the work. Revenue then has to grow 24.2 per cent a year, five years running, from Rs 2,415 crore to about Rs 7,143 crore. The field those sales come from is growing 11.0 per cent a year, so over the same five years it goes from Rs 48,300 crore to about Rs 81,388 crore. Divide one by the other and the share of the fieldOne maker's revenue as a percentage of the revenue of every maker in the same market added together. How such a market is defined and measured is settled in the sector material. has to go from 5.00 per cent to about 8.78 per cent.

The move needed is roughly 5.8 times last year's measured share gain, repeated every year for five years, in a market with an established leader in it. Notice what the conversion did. The conversion took a number nobody can argue with, 10.3 percentage points, and turned it into a claim about share that anybody who knows the market can immediately push back on. The 0.76 points a year is an average across a compounding path rather than a straight line, so the later years demand more than the earlier ones.

Now the margin route. Let revenue grow at the 13.9 per cent the record shows and make the entire gap out of margin instead. Revenue reaches about Rs 4,633 crore in five years, profit after tax has to reach about Rs 822 crore, and net margin therefore has to reach 17.75 per cent, about 17.8 to the decimal the record prints, where year three left it at 11.51 per cent. Same ending earnings of Rs 34.26/- a share, completely different demand on the business.

Is a wide gap good news?

The correction most readers need is worth being blunt about. A wide gap says the price is carrying a great deal. The width describes the price, not whether the business will get there, and it is certainly not a reason to be pleased.

A wide gap is equally consistent with a demanding price and with a rating assumption that was set too low by whoever chose it, and nothing inside the gap can separate those two. Lower the assumed rating five years out and the required growth goes up and the gap widens, with the business untouched. Raise the required return and the same thing happens. The gap is a joint product of the price, the record, and two numbers the reader supplied, and it does not come with a label saying which of the four moved.

Here is the household version. A friend would have to walk fifteen kilometres to arrive by six o'clock. The fifteen kilometres is real arithmetic. The six o'clock was fixed by whoever set it, and the distance says nothing about whether the friend will arrive. Move the deadline to eight and the same walk is comfortable. Anyone reading the fifteen as though it were a fact about the friend has forgotten that half of it came out of their own head.

Why does this side get written about more than the other?

Across enough published research, a gap in this direction is written up more often, at greater length, and with more conviction than a gap in the other direction, treated separately. The usual explanation, that everybody is being dishonest, is both unkind and not very useful. The mechanism is worth understanding instead.

A gap in this direction gives a writer something to say and a reader something to do, and a gap in the other direction gives neither. A note arguing that a record could stretch towards what a price already carries writes itself: there are drivers to list, evidence to gather, a case to build. The mirror image ends in a sentence with nowhere to go for most readers, who cannot act on it and did not want to hear it. So one side attracts effort and the other does not, and the imbalance appears without anybody deciding to create it.

Seen as structural, it can be corrected for cheaply. At the end of a note written on this side, the question is what the same author would have had to write about the other side, and whether that section exists at all. A missing section says something about how the note was written rather than about the company.

What does a usable statement have to contain?

Pull all of it together and a statement of this kind that is actually worth writing down has five parts. Four of them are arithmetic and appear in almost every serious attempt. The fifth is the one that separates research from a snapshot, and it is missing far more often than the other four combined.

The two inputs, named, with whoever chose them attached. The growth the price requires under those inputs. The growth the record supports, with any level shift stripped out. The distance between the two, in percentage points a year. And then the fifth: the evidence that would move either side, written down before anybody goes looking for it.

A statement missing that fifth part is a snapshot rather than research. Nothing in it names what to watch, and nothing can ever update it. Three months later a set of quarterly figures arrives and the four part version has no idea what to do with them. The five part version named two lines in advance, back when there was no result to protect, and reads those two first.

How does anyone actually use a gap like this?

Three uses, and none of them ends in a verdict. An analyst covering several makers in one market runs the same arithmetic on each of them and sorts by the gap, not to find the best one, but to find where the burden of proof is heaviest. A wide gap means the price is asking a lot. The reading time goes there, and the evidence had better be strongest there. Meghna Iyer, the analyst whose working this material follows, treats the gap as a queue for her own attention rather than as a score.

The second use is the one that pays for itself. The gap is written down with the date and then left alone. When the next set of figures arrives, the supported side is recomputed and the direction the distance moved is recorded. The habit converts a static number into a series. A series shows within a few quarters whether a record is closing on what a price carries or drifting away from it, and no single reading of any percentage will ever do that.

The third use is defensive, and it is the one a household investor gets the most from: before any percentage is accepted, it is converted into the growth it requires, and where the note does not carry enough to do the conversion, the percentage carries no information. That is not cynicism. The same instinct already applies to a builder's quotation with no line items in it. The quotation might be perfectly fair. A quotation without line items still cannot be checked, so it cannot be compared with the one next to it, and ranking two of them by the total is ranking the confidence of the people who wrote them.

Sarvani Coatings Limited, Nandivarman Paints Limited, Kesaria Surface Solutions Limited, Thottam Chemicals Limited and the analyst Meghna Iyer are invented.
Educational material. Not advice on any investment, tax, budget or market position.