The Theme: How It Differs From a Sector and Why It Attracts Capital

A theme is a claim that one change will reach the results of many companies, together with a rule saying which companies it reaches. A sector is a set of companies doing similar work right now. A sector is therefore a description of the present, and a theme an argument about the future, and nearly everything else about how the two behave follows from that one difference.

What does this build on?

Three things, all settled earlier. Sector research and its limits, meaning what it produces and what it cannot, were settled first. Demand can have a secularDemand that keeps rising for reasons unlikely to turn around next year, such as more households coming into existence. Settled earlier in this sequence. part and a cyclicalDemand that goes up and down with how busy things are, so a run of strong years tends eventually to be followed by a poor one. Settled earlier in this sequence. part, and telling them apart is the hard bit. And from the very first sequence, a view that cannot be checked is not research, whatever else it may be.

One object is added to that stack here: the theme itself. Not thematic research as an activity, and not the comparison of sector work against thematic work; both are covered separately. Just the thing being researched, taken apart carefully. A great deal of confused analysis comes from people using the word theme for something that has none of the properties a theme is supposed to have.

What is a theme, exactly?

Start on a street rather than in a market. Suppose every shop on one road sells vegetables. The vegetable statement is a description of what is there today, and writing it down would mean walking the road and looking. Now suppose a storm goes through and every roof on that road is damaged. The storm is a different kind of statement altogether. The claim is that one event will show up in the lives of many people, and the people it picks out have nothing in common except that the claim reaches them: the vegetable sellers, the tailor, the man who runs the photocopy shop, and the roofing contractor two roads away who is going to be very busy.

The first statement is a sector. The second is a theme. A sector is a classification, and a theme is an argument, and almost everything else about how the two behave follows from that single difference.

Unpacked, the second statement has two parts welded together, and both must be present. The first part is the change: roofs are damaged. The second part is the reach: whose position is different because of it. Drop either one and what remains is not a theme. A change with no reach is just news. A list of companies with no change attached is just a list.

So the working definition, and it is worth writing down in these exact words: a theme is a claim that a particular change will matter, together with the set of companies it would matter to. Both halves. Every failure set out below comes from somebody keeping the first half and quietly abandoning the second.

How is a theme different from a sector, taken slowly?

The difference is not that one is bigger or vaguer than the other. The difference is in how membership is decided, and once that is seen, four other differences fall out of it without needing to be remembered separately.

Membership of a sector is decided by what a company does now. Sarvani Coatings Limited makes paints and coatings, so it sits in the coatings sector, and it will keep sitting there until it stops making paint. Nobody has to agree with anybody for that to be true. Sector membership is checkable by looking at what comes out of the factory, it changes slowly, and two analysts who dislike each other intensely will still produce the same sector list.

Membership of a theme is decided by whether the claim applies. The claim is a claim about the future, so theme membership is a completely different kind of test: whether the claim applies to a given company is a judgement somebody has to make and defend. Two analysts will produce different lists. The same analyst will produce a different list next quarter if the argument moves. The disagreement is not a defect. An argument is exactly the kind of thing two careful people can differ on.

But it does create a hole that a sector never has, and the hole is where most bad thematic work lives. If membership is decided by the person making the claim, then the person making the claim can decide membership after the fact. A theme with no membership rule stated in advance is not testable. Anything at all can be added to it later.

Why does capital move to a theme so quickly?

Capital moving to a theme gets discussed cynically and it should not be. There is a real structural reason, and understanding it makes for better work rather than merely more suspicion.

A theme has four properties that a company view does not. A theme is a story with a mechanism inside it, so it can be understood rather than merely believed. A theme can be said in one sentence, so it survives being repeated. A theme explains several companies at once, so one hour of thinking covers a lot of ground. And, the important one, a theme does not require reading any of those companies. Capital can therefore move faster than analysis can be done.

The fourth property is doing all the work. Reading one company properly takes days: the statements, the notes, three years of what management said against what then happened, the working capital, the peers. Reading a theme takes ten minutes, and at the end of those ten minutes there is a reason to be interested in eight companies. There is no arrangement under which the careful thing and the fast thing arrive at the same time, so the fast thing arrives first. The order is not a moral failing of the people involved. The order is arithmetic about how long reading takes.

The same thing happens without a market in sight. A rumour goes around that a metro line will run down a particular road. Within a fortnight, enquiries pick up in four localities along the route, and not one of the people making those enquiries has read a single title deed. The claim reached them. The claim was easy to state. The claim covered four places at once. And the checking, if anyone does it, will happen afterwards, at the pace checking happens.

What must a theme assert before anyone can test it?

If a theme is an argument, then like any argument it can be well formed or badly formed, and there is a short list of things it has to say out loud. Four items. Miss any of them and the theme is not wrong exactly. A theme that cannot be wrong is in a worse condition than a wrong one.

First, the change: what is actually different now, stated so that somebody could disagree with it. Second, the mechanism: how that change reaches a company's results. Not that it will be good for them, but the route it travels, through volume or through price or through cost. Third, where it lands: which line of the profit ladder it should show up in first. Naming the line makes it clear where to look and when to give up looking. Fourth, the membership rule: which companies the claim applies to, written down before any company is named.

A theme without a membership rule can absorb any company that does well and shed any that does not, and the missing fourth item is the commonest defective form a theme takes. The other three are usually present, because they are the interesting part to write. The fourth is boring to write and it is the only one doing any protective work.

How does a theme cut across sectors?

Here is what a theme is genuinely good for, and it is the one property that earns the whole apparatus its keep.

A change rarely stops at one step of the chain a product travels along on its way to a buyer. If households repaint more often, the paint maker sells more paint, but the maker of resins and additives sells more resin, the dealerA shop that buys stock from a maker and sells it on to households. Because it holds inventory, its orders can move well before household demand does. turns more stock, and whoever applies the paint has more work. Maker, resin supplier, dealer and applier sit in four different sectors. A sector classification is built on what each one makes, and they make entirely different things, so no sector classification will ever group them together.

A theme cuts along the chain rather than across it, and a sector reading cannot. A change travelling down a chain is invisible to any reading that stops at one step of it.

The same sorting happens at a wedding. Suppose weddings in a town start being smaller and arranged faster than they used to be. The change in weddings reaches the caterer, the tent supplier, the printer of invitation cards, the jeweller and the man who hires out chairs. Asking which sector they are in returns five answers. Asking which of them is affected by the change returns one list. The list is the useful object, and no sector map would have produced it.

What does it look like when the claim is built in the right order?

Now do it properly, on the invented record, and pay attention to the order. The order is the lesson.

Step one is the claim, and step two is the membership rule, and both of those happen before a single company is named. The claim: households are repainting their homes more often than they used to, so coatings demand is being lifted by something that does not depend on how much new construction is happening. The membership rule, written next: any company whose revenue rises if households repaint more often.

Only now open the list of companies and test each one against the rule.

Kesaria Surface Solutions Limited sits in the same sector as two of the members and the rule still leaves it out. The exclusion is the single clearest sign that a theme and a sector are genuinely different objects. If a theme happens to contain exactly the sector, one of two things has happened: either the change really does reach everybody in the sector equally, which is rare, or the rule was written to match a list already drawn up.

Now test the claim against the record

A theme that has been written properly makes a prediction, and this one does. If repainting frequency is lifting demand for reasons unconnected to construction activity, then sector volume growth should hold up reasonably even in a year when construction is weak. The prediction is checkable against the invented five year record, and here is what the record says.

Across the five years, sector volume printed 2.1 per cent, then 6.8, then a fall of 1.4, then 7.2 and finally 4.5. The five average 3.84 per cent, and the single year that went backwards sits 5.24 percentage pointsThe gap between two percentages, counted in units of one per cent. A move from a 4.87 per cent share to a 5.00 per cent one is a move of 0.13 of these. Growing by 0.13 per cent is a quite different thing. below that mean.

So what does the test settle? Arithmetically, a swing of that size applied to a mean of 3.84 gives exactly minus 1.40, and applied to a mean of nothing at all it would have given minus 5.24. The record is therefore consistent with a persistent component of demand sitting underneath the swing and holding the worst year up. But being consistent with a claim is not the same as establishing it, so the theme has survived a test that it could have failed, and that is all it has done. One year, one record, one field: this is enough to keep going and nowhere near enough to be sure.

When does a theme stop being a theme?

Themes end. Themes do not usually end by being disproved, and that is what makes the ending hard to notice.

There are two ways out. The first is that the claim gets settled, one way or the other: the change did not arrive, or it arrived and did not reach results the way the mechanism said it would. The first ending is clean and rare. The second, far more common, is that the change fully arrives, everybody sees it, and it is sitting plainly in the reported numbers of every company on the list.

The moment a theme is visible in reported results it has stopped being a claim about the future. Its usefulness ends at exactly the point where it becomes easiest to say with confidence. That is an uncomfortable shape, and the reason it happens is worth stating plainly. A theme is useful when it says something the numbers do not yet say. Once the numbers say it, the theme is a description, and descriptions are freely available. Meanwhile the person stating it has more evidence every quarter, and so sounds more certain every quarter. Confidence and usefulness move in opposite directions along the same line.

What a theme does not settle about any company inside it

Thematic thinking does the most damage here, and it does the damage quietly, by feeling like an answer when it is only a shortlist.

Rain is certain. Two men sell umbrellas on the same road. One is standing outside the station with forty umbrellas and the other is a fifteen minute walk away with six. Both are exposed to the rain. One of them is going to have a good afternoon. Knowing about the rain does not say which.

Theme membership fixes exposure and settles nothing whatever about capture. Sarvani Coatings Limited passes the repainting rule, Nandivarman Paints Limited passes it too, and that shared fact does not separate the two of them by a hair. Nor does reaching for the market shareHow big a slice of everything the field sold in a year belongs to one company. A share climbs only when that company grows faster than the field around it does. numbers help. Sarvani Coatings finished the year holding 5.00 per cent of the field where twelve months earlier it held 4.87 per cent. The movement is 0.13 percentage points, off revenue growth of 13.9 per cent in a field that managed 11.0 per cent. Small, real, and pointed at an entirely different question from this one. The share move says nothing about where that growth came from, and therefore nothing about which of the two makers is turning the repainting change into anything.

There is a second thing membership does not settle, and it is the one people skip. Membership does not say whether the exposure is already priced inAlready built into the share price, because everybody else worked it out first. When the whole market expects a change, paying today's price for it buys nothing on the day it arrives.. A theme that everybody agrees with is a theme that everybody has already acted on, and the company work is what settles whether there is anything left in it.

So theme membership is where company work begins and never a substitute for it: it narrows the list, and every name on the list still has to be read.

Who actually uses a theme, and what for?

An analyst uses a theme as a shortlist generator and nothing more. The output of an hour of thematic thinking is a list of names to read, not a view on any of them, and the discipline is to write the membership rule into the note so that a reader six months later can check whether it was ever applied. If the note says which companies were considered and rejected, the theme was real. If it names only members, it was a story.

A person running money uses a theme mostly as a risk instrument, and that surprises people. If four positions all pass the same membership rule, then those four positions are not four bets, they are one bet held four times, and the membership rule is the cleanest way to see that. The rule that settles what to include is the same rule that shows where the concentration sits.

A household investor meets themes in their marketing form, in the name of a product. A well written theme always sounds right, so the useful question there is not whether it sounds right. The question is whether anybody has stated what the product would exclude, and what would make the claim wrong. Where a product's naming and description sit under a regulator's rules, those rules are worth reading in the original.

All three uses run through the same fact: the membership rule is the part that does the work, and it is the part nobody prints. Notice also what none of the three did. Not one of them concluded from a theme that a company would do well. The step to a company view, if it ever comes, comes from reading the company itself. Company analysis is where that work is set out.

One last note on what the field evidence can and cannot do here

The margin figures are tempting to reach for, so they are worth a note. Over the two years in this record, gross marginWhat survives out of every hundred rupees of sales after paying for the stuff the goods were made of. How it is built from the statements sits in the accounting material. rose at all three makers: 3.0 percentage points at Sarvani Coatings, then 2.4 at Nandivarman Paints, then 3.6 at Kesaria Surface Solutions. A shared move like that is tempting to read as the theme showing up in results, and it is not that at all. The margin gains came from realisationThe average amount a company ends up collecting for each unit it ships, once discounts and the mix of things it happened to sell are allowed for. Worked out in the earnings material. outrunning input cost, and rising realisation is a pricing story rather than a repainting story. Put side by side, the numbers make the direction plain: the two members averaged 2.70 points between them, the excluded maker put on 3.60, and the gap of 0.90 points runs in favour of the one company the membership rule throws out. Evidence that reaches non members hardest is not evidence for the theme, and the ledger above files a move like that under evidence pointing elsewhere, never under support.