Value and growth: two bets on two different sources of return

What each style is actually betting on

Value and growth are bets on two different sources of return. Value buys a claim it judges to be priced below what the business is worth, and is paid if that gap closes. Growth buys expanding earnings, and is paid only if the expansion beats what the price already assumed.

Start with what neither of them is. Neither is a judgement that the business is good. A company can be admirable in every respect and still be a poor thing to have bought, because the price you paid contained everything admirable about it and then some. In both styles you are making a bet against a price, not a bet on a company.

The names get in the way here, and it is worth saying how. “Value” does not mean the business is valuable; it means the buyer thinks the price sits below an appraisal of it. “Growth” does not mean the company will grow; almost everyone can see that it will, and that is exactly the problem. Both words describe a relationship between the price and something else, and neither describes the something else on its own.

What follows is the trade-off each one accepts. The value buyer takes on the risk that the discount was deserved and the risk of waiting. The growth buyer takes on the risk that the expansion was already paid for. Those are different risks, and avoiding one does nothing at all about the other.

Growth investing is not a bet that the company will grow

This is the sentence the rest of the article rests on, so take it slowly. If a company is widely expected to grow quickly, that expectation is already in the price. The accounts are public, the guidance is public, and every participant can read both. Nobody is paid for noticing the obvious.

So the growth buyer is not paid when growth arrives. They are paid when growth arrives in a larger quantity than the price had assumed — and the price has assumed a quantity whether or not anybody wrote it down. Growth already inside the price is not return; it is the entry fee.

Here is that in arithmetic, with figures chosen to make the mechanism visible rather than to represent anything typical — there is no typical case, and claiming one would be a claim. Two illustrative companies each earn ₹10 a share today. Company G trades at ₹400, so 40 times earnings. Company V trades at ₹100, so 10 times. Hold them for five years.

Suppose both end the five years on the same multiple, whatever that multiple turns out to be. Then for the two to return the same, G's earnings must grow by four times as much, in total, as V's — because G cost four times as much per rupee of earnings at the start. That requirement is four times as much growth, and it does not depend on what the exit multiple is, only on both arriving at the same one.

That last condition is doing real work and is worth naming rather than hiding. If G still commands a higher multiple at the end than V does, the growth G needs falls, and the comparison moves in its favour by whatever that difference is worth. Nothing here says a fast-growing business must finish on the same multiple as a slow one. The point is narrower: whichever exit multiples you use, you have assumed them, and an assumption about the exit belongs beside the growth rate in the open, not underneath it.

Put numbers on it. If V's earnings grow at 3% a year, five years takes ₹10 to ₹11.59. Four times that growth means G must reach ₹46.37 — which is 35.9% a year, compounded, for five straight years.

Now give G a genuinely fast five years: 25% a year, compounded, in every one of the five. Its earnings reach ₹30.52. That is a company which tripled its profit, and it has fallen well short of ₹46.37. Price both at an exit multiple of 20 and G returns 8.8% a year against V's 18.3%.

Nothing went wrong in that example. The growth was real, large, and delivered. It simply came in below what ₹400 had asked for. Run it the other way and the result flips just as cleanly: give G 40% a year and its earnings reach ₹53.78, and on the same exit multiple it returns 21.9% a year against V's 18.3%.

That is the honest shape of the thing. The example does not show that one style beats the other; it shows that the entire outcome turned on whether one number — delivered growth — landed above or below one other number that was already embedded in a price. Both examples used the same two companies.

Where a return comes from, as an identity

There is an exact way to see what the previous section did, and it is one line of algebra rather than a theory. The price of a share is its earnings multiplied by the multiple the market puts on those earnings. So the price at the end, divided by the price at the start, is exactly the change in earnings multiplied by the change in the multiple. Dividends arrive on top, separately.

Two terms, then. One is what the business did. The other is what the market decided to pay for what the business did. That second term is expectations, and nothing else — it moves when the market's assumption about the future changes, which can happen on a day the business does nothing at all.

Go back to the illustration and hold the multiples still. If G exits at 40 times and V exits at 10 times — each ending where it began — then G returns exactly 25% a year and V exactly 3%, which is precisely each one's earnings growth and nothing else. Every difference between that result and the previous section's came from one place — the multiple, not the earnings.

Which gives each style a clean description. A growth position is mostly long the earnings term and exposed to the multiple compressing — the fall in a multiple when the market lowers what it assumes, which has a name worth knowing: a de-rating. A value position is mostly the other way round: the earnings may do very little, and the return has to come from the multiple rising as the market revises what it assumed.

The identity also explains a result that surprises people. A holding can grow its earnings every single year for five years and still lose money, if the multiple fell by more than the earnings rose. The two terms multiply; they do not take turns.

What the value bet needs, and the clock nobody prices

The value bet has two parts, and collapsing them into one is where most of the trouble starts. The first is an appraisal: an estimate of what the business is worth, built from the same public accounts everyone else has. The second is the closing: the market coming round to something nearer that estimate.

The appraisal is hard and it is honest work — what the earnings are made of, how much of the profit turns into free cash flow, what the assets would fetch, how much capital the business consumes to stay where it is. A ratio is not an appraisal. A low price-to-earnings ratio — the PE — is a rank, not an appraisal, and what actually sits in its denominator is the subject of PE and PB.

The closing is the part with no estimate attached, and it decides most of the return. Take an illustrative case: a share bought at 60% of an appraised worth, with the appraisal eventually proved right in full. If the gap closes within a year, that is a 66.7% gain. If it takes five years, the same correct appraisal pays 10.8% a year. If it takes ten, it pays about 5.2% a year.

Same analysis, same discount, same vindication — three quite different outcomes, separated only by a date the analysis had no way of producing. Being right and being paid are different events, and the gap between them is measured in years.

There is a second clock, and it charges through the same arithmetic. Suppose the appraised worth itself erodes at 5% a year while you wait, and the gap closes fully at year five. That pays about 5.2% a year — near enough to what the ten-year wait on a static appraisal paid, though the two are not the same function and the near match at these particular figures is an accident of the numbers chosen.

What is general is the shape. An annualised return is one gain spread across the years, so a shrinking appraisal and a lengthening wait reach the same result from opposite directions, and neither of them appears anywhere in the analysis that produced the discount. Which is why the question “how long can this business sit here without getting worse?” belongs inside the appraisal rather than after it.

Now name the specific failure, because it has a name and it is the one that catches people. The value trap is a share that is cheap on every ratio and stays that way, and the mistake underneath it is assuming the discount is an oversight — a cheapness nobody has got round to noticing. It cannot be that. The ratio is computed from two public numbers that every other participant can divide just as easily, so a low reading is the market's stated view that the earnings will not persist, or are of poor quality, or carry a risk that demands a larger discount.

So a discount is a stated view, not an oversight, and the value case is the argument that this particular view is wrong. That argument has to be made. It is not supplied by the ratio that drew the attention in the first place, which is the market's side of the disagreement rather than yours.

What a screen means by the words, which is less than you think

Most people meet these two words as labels — on a fund, on an index, on a column in a screener — and the labels mean something much narrower than the bets described above.

A rank-and-slice scheme sorts a universe on one or more ratios and takes the cheapest portion into the value bucket and the rest into the growth bucket. What membership tells you is where a company sat in the cross-section on the day of the sort. It is a statement about the other companies as much as about this one. How that sorting is done, and the ways a screen quietly answers a different question from the one asked, are in the screener guide.

Two consequences follow that are worth holding on to. The first is that every company lands in one bucket or the other, including the many about which nobody has formed any view at all. The second is stranger: because the price is in the numerator of the sorting ratio, a share that falls far enough becomes a value constituent without anything happening inside the business. The label follows the price, on a schedule set by the rebalance date.

It also runs the other way. A business whose earnings grow quickly and whose shares are priced modestly relative to those earnings can qualify on both descriptions at once, and the fact that the taxonomy has no room for it is a fact about the taxonomy.

So a screen ranks; it does not appraise. Running a filter and buying the output is not the value bet described in the previous section — it is the first ten minutes of it, with the appraisal left out. The screen's job is to reduce a universe to something a person can read, and that job is worth doing. It is just not the same job.

The two bets, side by side

Setting them out together makes the symmetry visible — and the symmetry is the point, because each column's blind spot is the other column's subject.

The row that repays a second reading is the last one. Value's screens are built to find a low price and are indifferent to what the assets earn — the measure that addresses that gap is return on equity and on capital employed. Growth's screens are built to find expansion and are indifferent to what it cost. Each style's standard toolkit is precisely blind to the other's subject, which is why a screen run from either side leaves the harder half of the work undone.

Both can be wrong, in ways the other is not

The two errors are not the same error wearing different clothes, and this is the reason a preference between the styles protects you from less than it seems to.

Value's error is about the business. The appraisal said the earnings would persist and they did not, or the assets were carried at a figure the market would never pay. The analysis was of the wrong thing, and the price was giving a signal that got read as noise.

Growth's error is about the price. The analysis of the business can be exactly right — the earnings expanded, the products worked, the market share arrived — and the position can still lose money, because what it had to beat was a quantity already inside the price, and no amount of watching the business would ever have revealed it. Being right about the company is not the same as being right about the trade, and this failure mode is unusually painful because nothing in the business disconfirms it.

Which means the two are wrong in opposite directions. An investor who has decided value is safer has swapped a pricing risk for a business risk. One who has decided growth is where returns come from has swapped a business risk for a pricing risk. The swap is real. The reduction in risk is not.

And now the mistake that sits underneath both, stated precisely enough to recognise yourself in it. The common failure is holding a position whose owner cannot say what has to happen for it to pay — not “the company does well”, but which quantity has to exceed which other quantity, and by when. If the answer is “the PE is low”, that is a ratio reading, not a bet. If it is “this business is growing fast”, that is a description the whole market shares. A bet you cannot state has no way of being wrong, which sounds comfortable and is the opposite.

The question this article will not answer

So which one is better? The article refuses that question, and the refusal is not caution — it is that the question has no answer in the form it is asked.

Separate two kinds of claim here, because they get run together constantly. That each style has a rationale is a claim about mechanism: value's rationale is that prices and appraisals can diverge and sometimes reconverge, growth's is that expansion compounds and the market can under-assume it. Both mechanisms are coherent, and neither needs any evidence to be stated, because neither asserts anything about what happened.

That one has done better than the other is an entirely different claim. It is empirical, it belongs to a dataset, and it is only meaningful once you have named the period, the universe, the sorting ratio, the rebalance rule, the weighting, and the treatment of costs and taxes. Change any one of those and the answer can change with it. An article that writes “historically, one has outperformed” without all six has not made a finding; it has made a sentence.

The general rule underneath is worth carrying beyond this topic. A rationale is never evidence for a result, and a result is never proof of the rationale. Anyone offering you the first as though it were the second has skipped the only step that was difficult.

What can be said flatly is the mechanical part, and it is enough to be useful. Your return is the change in earnings multiplied by the change in the multiple. Each style is a statement about which of those two terms you expect to do the work. Neither statement is a forecast, and this article makes none — not about any company, any index, any style, or any period. FNOTrader is not a SEBI-registered investment adviser or research analyst, and nothing here is a view on any share or on any level of any ratio.

Making the bet explicit before you take it

Everything above points at one practical move: write down the quantity your position needs, before you own it. For a growth position that is the growth rate the price is already asking for, which the break-even arithmetic in this article computes from three inputs you can read off a page. For a value position it is the appraisal and the horizon, stated as two separate numbers rather than one feeling.

FNOTrader's Stocks app screens across roughly 2,390 stocks and 17 NSE sector and size indices, and filters on the reported fundamentals rather than only on the ratios computed from them, so a screen can be built on the input you actually meant instead of the output somebody else derived from it. What that produces is a shortlist and the figures behind it, side by side, which is where reading the profit and loss account starts.

What no screen does is the appraisal, and no screen decides which of the two terms in the identity you are betting on. It can put the same calculation across every company so the numbers are at least comparable. The bet remains yours to state.

Common questions

Continue reading

More in Stocks & Screening · App: Stocks · Definitions: glossary · Free tools: calculators · All: every article