Company Research vs Investment Research: Where They Part

Nothing in the method tells them apart. Both open the same accounts and run the same arithmetic. The parting happens at the very end of the work. One kind of work sets down what a business is and what it did, and it is done once the evidence is there. The other attaches a judgement to a price, and it is not done until that judgement exists. The work below takes the first route and halts.

Four things sit under that answer, and everything below is those four worked out.

The line runs through the output and not through the method. Both kinds of work open the same accounts, read the same notes behind them, run the same arithmetic and arrive at the same figures. Nothing in the method separates them. Nothing in the effort separates them either. The separation is in what the finished output is permitted to say. The working is where the two agree, so hunting for the difference there turns up nothing.

A piece of company research is finished when it becomes reproducible, and a piece of investment research is finished when it stops being so. Handed the same sources, a stranger who knows the trade writes the lines out again. If those lines match the ones already written, the first kind of work is complete, and that sameness is the proof rather than a worry. A conclusion every reader of the same public evidence would reach identically is not doing the job it exists to do, so the second kind needs something a stranger would not reproduce. Two opposite finishing tests, and they are why the two kinds of work go wrong in different ways.

A label is part of a figure rather than an ornament attached to it. A split that reconciles to the rupee against a published total is still somebody's working, if the source carrying it says so, and lifting the figures while leaving those six words behind converts an estimate into a disclosure without altering a single digit. The duty about labels has nothing to do with conclusions at all. Company research is therefore not a junior version of investment research. Most of the damage a note does downstream is done by a label somebody dropped rather than by a judgement somebody made.

And the second kind of work needs one input the first never touches, a price. A price appears in no set of accounts. A price is set outside the business, by other people, continuously, and a conclusion is attached to it. So a piece of company research can be complete without ever meeting one, and a piece of investment research cannot begin without one. The missing price is the cleanest single statement of the separation, and it is why the work below can do the first kind in full and stop.

Two notes on one desk, built from the same figures. Where do they stop being the same?

Meeting the two names first is what causes the trouble, so hold both back for a moment. A reader who is given the two names before the two objects will assume the difference is one of seniority, or depth, or rigour, or how long somebody spent, and every one of those readings survives a definition intact. So start with one line that moved, and two people writing it up.

Anjani Stationers Private Limited, an invented stationery business, published one year in which the part of its costs that stands still whatever it sells, its standing baseThe slice of a business's costs that stays the same size whether it sells a lot or a little in the year. Shed rent, salaried pay and insurance sit in it., rose Rs 24,40,000/-, from Rs 49,60,000/- to Rs 74,00,000/-. Nobody had to be told that. The rise falls out of two published columns by subtraction, and anybody with those columns gets the same answer.

Two people sit down with those columns on the same afternoon.

The first note writes that the rise is Rs 24,40,000/-. The note writes that the built record in these notes names three things the money went on, in that record's own words: more people, more space and a binding works the business bought into. The note writes that another account in these notes supplies an estimated split of the rise, stamped there an estimate rather than a disclosure: Rs 6,00,000/- of it set against employee benefits, Rs 11,40,000/- against the standing share of other operating costs, put down mostly to a second warehouse taken in the year, and Rs 7,00,000/- against depreciation and amortisation arising on assets bought. The note writes that the three amounts add to Rs 24,40,000/- exactly. And it writes, in the same breath as the amounts, the sentence printed in the source's own caption: the split is an estimate, not a disclosure. Then the first note stops.

The second note carries every one of those lines, in the same order, with the same figures and the same label. Then it keeps going. The second note was asked for something the first note was never asked for. The two notes share every figure and differ only in where they end.

The same split is familiar without the name research attached to it. Two people watch the same cricket match. One writes down what happened, over by over, and somebody who missed it can then rely on the account. The other writes down who is going to win. The second person is not watching harder, and is not further through the same task. The two were asked different questions and will be judged by different standards. The first person's notebook, handed to a stranger who had also watched the match, should agree with the notebook that stranger would have kept.

So what is each kind of work actually for?

The objects the two names attach to are already in view, so the names can arrive now.

Company research exists so that somebody else can rely on a description of a business without having to rebuild it themselves. Its product is evidence with its sources attached, and its virtue is that it can be checked. Someone who doubts a line in it can open the same document and settle the doubt. Checkability is a low bar to state and a hard one to clear. Clearing it means every figure carries where it came from and what kind of figure it is.

Investment research exists so that somebody can act, at a price, on a particular day. Its product is a conclusion with its basis attached, and its virtue is that the basis is visible enough to be argued with. Investment research normally consumes a description of the business as its raw material. The two therefore sit in one document more often than not, and the join between them is easy to miss.

The two kinds of work are not two stages of one process, and they are not a ranking. The second usually consumes the first, and the first is complete without the second ever happening. Say the thing readers quietly assume and then take it apart: the difference is not depth, not seniority, not rigour and not how long anybody spent. A description can take a fortnight and a conclusion can take an afternoon, and the reverse happens just as often. Neither is the serious one.

The everyday version is a building survey. A surveyor walks a house, opens the loft, tests the damp, and writes down what is there, and anybody reading the report then knows what they are dealing with. Somebody else decides whether to buy the house at the asking price. The surveyor is not doing an unfinished version of the buyer's job, and a surveyor who ended the report with a view on the asking price would be doing something they were not asked to do, however sensible the view.

So what is each one allowed to conclude?

Permission is the centre of the subject, and the two kinds of work are best taken one at a time, specifically rather than generally.

Company research may conclude four kinds of thing. First, that a figure is published, and where it is published, named specifically enough that a stranger could open the same place. Second, that a figure is an estimateA figure somebody worked out rather than one the business itself stated. Whose working it is matters as much as what the number says., and whose estimate it is, with that label carried alongside it. Third, that a purpose, a cause or a split is not disclosed anywhere, said only after a search rather than instead of one. And fourth, that a reader outside the business can go no further on the evidence available. Stopping there is a finding rather than an apology.

Investment research may conclude something about a business at a price, and it then carries the obligation to show what that conclusion rests on and what would change it. The extra permission is exactly that much, and the obligation attached to it is not a formality. A conclusion whose basis is invisible cannot be argued with, and a conclusion nobody can argue with is not research.

The line runs through the output and not through the method. Both kinds of work open the same accounts and may run the same arithmetic. The separation sits in what the finished output is permitted to say. The consequence that makes this usable rather than merely tidy is that a sentence sits inside or outside the line regardless of who wrote it. A description written by an equity research analyst is still a description. A judgement written into the third paragraph of an internal company profile is still a judgement, and the fact that the document was never called research does not move it back across the line.

When is a piece of company research finished?

The test runs on a draft in a single afternoon, with nothing but a colleague and half an hour.

The sources go to a stranger who knows the trade. If that stranger would write the same lines, the work is finished. Company research is finished when it becomes reproducible and investment research is finished when it stops being so. The first half of that is comfortable. The second half is the one worth sitting with, and it repays working rather than being left as a paradox. A conclusion that every reader of the same public evidence would reach identically is adding nothing. The evidence was already doing that work on its own. So the second kind of work is expected to contain a step somebody else might not take, and it is expected to show that step rather than bury it. The step is the product. Hiding it is the failure.

Which gives the practical version. Running down a draft line by line, the mark falls at the first line a stranger handed the same sources would not reproduce. Everything above that mark is one kind of work. Everything from it down is the other. Most drafts that go wrong go wrong because nobody knew where that mark fell, so a judgement arrived three paragraphs early, wearing the same typeface as the figures around it, and every reader after that treated it as one of them.

A published rise, three named limbs and an estimated split. How far can that be taken?

Take the rise as far as it honestly goes, in order, and stop where the evidence stops. The stop matters more than the distance travelled before it.

First, the published total. The part of Anjani Stationers' costs that stands still whatever it sells rose Rs 24,40,000/- in one year, from Rs 49,60,000/- to Rs 74,00,000/-. Two published columns, one subtraction, and any reader gets the same answer.

Second, the three things named against it. The built record in these notes describes the rise as falling on more people, more space and a binding works the business bought into. Three named limbs, and no amount attached to any of them there.

Third, the estimated split, and it never appears without its label in the same breath. Another piece in these notes divides the rise, and stamps its division an estimate rather than a disclosure. Rs 6,00,000/- of it is put against employee benefits. Rs 11,40,000/- is put against the part of other operating costs that does not move with sales, and that source attributes the amount mostly to a second warehouse taken in the year. And Rs 7,00,000/- is put against depreciation and amortisation arising on assets bought. The three add to Rs 24,40,000/- exactly. The same source also prints, in its own words, that the components are on the face of the statementWhere a financial statement actually prints its named line items, as opposed to the supporting notes filed alongside it. even though the split is not, and its own caption carries the stamp: the split is an estimate, not a disclosureA thing the business put into its own published document. It carries weight because the business said it, which is a separate question from whether it turns out to be accurate..

Fourth, the stop. Another piece in these notes puts the reason in one line: from outside, an innovation spend and an ordinary expansion spend are the same line in published accounts. Company research can report the total, can report the three limbs, can report an estimated size for each with its label, and still cannot say what the money was for. The distance between a reported estimate and a classification is the whole of the separation.

And now the harder half, the part readers resist: more evidence does not close the distance. The distance feels as though it should close. Enough published lines feel as though they must eventually pin a purpose down. Published lines never do. The missing item is a statement of purpose, and no quantity of figures is one. A hundred more figures give a hundred more things that can be reported and leave the right hand side exactly where it was.

What kind of conclusion is refused outright, and where is that refusal written down?

A refusal quoted from somewhere else is evidence. A refusal merely announced is a claim. Two were written down elsewhere in these notes, and both can be opened.

The first sits in the account of what an equity research analyst actually does. Written into it is a refusal: no conclusion about any company, no target for a share price, no fair valueA stated view of what something is worth at a point in time, arrived at by working rather than read off a market screen., no ratingA short label attached to a piece of investment research so a reader can see the direction of its conclusion at a glance. and no recommendation, and no view on the business it describes or on anything else.

The second sits in the comparison of two ways of making money. The comparison computes no value, no multiple and no return, and it will not say which of the two models is the better business. With its last selector set to ask for a ranking, it refuses in words, in the drawing and in the readout, rather than quietly producing one.

A sentence can be skipped and a control has to be answered, so a refusal built into a working control is stronger than a refusal written into a sentence. A demonstration of a conclusion is a conclusion, whether in miniature, hypothetically, with invented figures, inside a control or labelled as an illustration, and that label does not travel into whatever a reader writes down afterwards.

So where does a price come into it?

The cleanest single statement of the separation is one sentence long.

A price appears in no set of accounts. The published year of Anjani Stationers carries revenue of Rs 2,70,00,000/-, materials of Rs 1,48,50,000/-, gross profit of Rs 1,21,50,000/-, employee cost of Rs 42,00,000/-, other operating expenses of Rs 26,00,000/-, depreciation and amortisation of Rs 12,00,000/-, operating profit of Rs 41,50,000/-, finance cost of Rs 3,50,000/-, earnings before tax of Rs 38,00,000/-, total tax expense of Rs 8,00,000/- and Rs 30,00,000/- reaching the residual claimWhat is left over for the shareholders after every earlier claim on the year, including materials, pay, interest and tax, has been met.. Eleven lines and no price among them. A price is set outside the business, by other people, and it moves while those eleven lines stand perfectly still.

The second kind of work needs an input the first kind never touches. Three useful things follow from it, and each is worth stating. A piece of company research can be complete on the day the accounts are published. A conclusion attached to a price cannot be, and the reason is that the price has already moved. The same piece of company research is still true a month later. A conclusion may not be. Neither statement is a compliment to the first or a criticism of the second. And these notes can carry the first in full without going anywhere near the second.

The honest limit on this material, in one line: not one of the businesses in these notes carries a price of any kind anywhere, so the second kind of work could not be demonstrated here even if it were permitted to be. The day-to-day of that kind of work, and what it will not produce, is covered separately under Equity Research: What the Job Actually Involves.

Which sentences look like description and are actually a conclusion?

Three of them, and one question separates all three from the correct lines sitting around them: which published line would a stranger point at?

The first. The business put Rs 24,40,000/- into building capability. Every digit in that is right, and the figure is published, and the sentence still fails. The word capability is a purpose. Nothing published anywhere states a purpose, so a stranger has nothing to point at, and the sentence has quietly classified a total that nobody classified.

The second. The spending should earn that back before long. Nothing anywhere names a period, and before long is a period. The sentence is about a stretch of time that nobody observed and nobody published, dressed as a remark about a figure that was.

The third. The fall in profit shows the decision was wrong. The fall in the year's operating profit is published. The decision is published. The word shows is the reader's own step, taken inside a single year that cannot settle it. A commitment lands on one clock and whatever it was aimed at lands on another, and the year happens to hold only the first of the two.

All three read as description, all three carry a conclusion, and one question separates them from the lines around them. A school report saying a child sat every paper is a different kind of object from one saying the child could do better. The second is not a longer version of the first, and no amount of care with the attendance register turns one into the other.

What four lines should travel with any figure in a company note?

Here is how somebody doing this for a living actually keeps the line where it belongs, and it is not vigilance. The method is a fixed order that four things go in, every time, and dropping one then becomes visible rather than silent.

One, what is it, in plain words, specific enough that a stranger could find it without being told where to look. Not the rise, but the rise in the part of the costs that stands still whatever the business sells.

Two, where is it published, named so precisely that somebody can open it. A source that cannot be opened is a claim about a source.

Three, is it a disclosure or an estimate, and if an estimate, whose, and where the label sits. For the split, this line reads: an estimate, published under operating leverage in these notes and stamped there an estimate rather than a disclosure. Twenty words that do more work than the three amounts they follow.

Four, what would it take to go further, written as a named disclosure that is missing rather than as a shrug. For the split, this line reads: a statement from Anjani Stationers of what the money was for. A figure with the third line blank has been promoted from an estimate to a disclosure by silence. And line four turns a stopping point into a finding. The line hands whoever reads next the exact thing to request, instead of leaving them to guess whether anybody looked.