Part 8 · Sector foresight · Chapter 97
R&D as a forward indicator
R&D spend is an input on today's P&L; the revenue it creates is years away — so the forward signal is the output the spend produced and the productivity of it, never the spend itself, and what counts as 'a lot' inverts by sector.
13 min
Prerequisites not yet complete
This module builds on Chapter 92: Why statements lag and people lead, Chapter 93: The five classes of leading indicator. You can read on, but the sequence is load-bearing.
The question
Research and development is the rare cost that is supposed to buy the future. A company charges it against this year's profit, or capitalises it onto the balance sheet, but the revenue it is meant to create — the new drug, the new platform, the new model, the reformulated brand — arrives years later, if it arrives at all. That gap between spend and payoff is what makes R&D a leading indicator: read correctly, it tells you about revenue that has not yet been earned.
Read incorrectly, it tells you almost nothing. The number everyone quotes — R&D as a percentage of sales — is an input. It says how much money went in, not what came out. The whole skill of this module is to stop reading the input and start reading the output: the patents, the filings and approvals, the launch cadence, and above all the productivity — how much revenue the spend actually turned into. Two companies can spend the identical amount and have opposite futures.
Why the input misleads
R&D intensity feels like a quality signal. A company spending 12% of sales on research looks more serious about its future than one spending 3%, and higher-intensity firms are routinely called "more innovative". But intensity is money committed, not value created — it is the ticket price, not the prize. A company can spend heavily and produce nothing; a disciplined one can spend less and launch more. The spend line cannot tell the two apart, because it is measured at the wrong end of the pipe.
There is a second reason the input misleads, and it is a forensic one. R&D can be expensed — charged to this year's profit — or capitalised — parked on the balance sheet as an intangible asset and released to profit slowly. The choice moves reported profit without changing a rupee of cash, so a company that capitalises aggressively can show higher profit and a growing asset base that is really just deferred cost. The input line, taken at face value, hides this entirely. The output does not.
The mechanics
Reading R&D as a forward indicator has four moving parts.
- Intensity and its trend. R&D ÷ sales, tracked over years. The level is only interpretable against the sector (below); the trend is interpretable everywhere — intensity quietly falling while management talks up innovation is a tell that the pipeline is being starved to protect current profit.
- The output, not the input. What did the spend produce? Patents granted, product filings and approvals (a pharma exporter's , an agrochem's registrations), new SKUs and variants, and the cadence of launches. Output is often disclosed outside the financials — in a regulator's database, a product page, a filing count — which is where a careful reader looks.
- The lag. Output leads revenue by a sector-specific delay: a pharma molecule can be eight to sixteen quarters from filing to meaningful sales; an auto platform three to five years from spend to showroom; an FMCG reformulation two or three quarters. The lag is what makes R&D forward-looking, and it is why this year's spend and this year's revenue tell you nothing about each other.
- Expensed versus capitalised. Under Indian accounting, research is expensed as incurred and development may be capitalised only if it meets strict criteria. A firm that what peers expense reports higher profit today and carries the cost as "intangible assets under development" — the R&D twin of . If the development never launches, the balance is eventually written off as an : the delayed bill for research that did not convert.
The read is always the same shape: ignore the reassuring spend line, find the output the spend produced, allow for the sector's lag, and divide output by cumulative input to get productivity — then check that the accounting policy is not flattering the input in the first place.
The maths — R&D productivity
Intensity answers "how much went in". Productivity answers "how much came out", and it is the number that actually leads revenue. Two composite specialty-pharma firms illustrate it. Both spend 6% of sales on R&D — identical intensity — over a five-year window. [illustrative]
Meridian Labs illustrative spent a cumulative ₹900 cr on R&D across the five years, and products it launched within that window now generate ₹1,350 cr of annual revenue — a new-product vitality (share of revenue from recently launched products) of about 28%, and roughly ₹1.5 of new-product revenue for every ₹1 of cumulative R&D. [illustrative]
Corvus Pharma illustrative spent the same cumulative ₹900 cr, but its recent launches generate only ₹210 cr — a vitality of about 5%, and ₹0.23 of new-product revenue per ₹1 of R&D. [illustrative]
Identical intensity, a six-fold gap in productivity. Meridian's spend is building the next decade; Corvus's is a cost centre that happens to be labelled research. The productivity ratio — revenue from products launched in the last N years ÷ cumulative R&D over that period — is what the percentage-of-sales figure conceals, and it is the leading number to track. Rising productivity means the pipeline is converting; falling productivity, even at flat intensity, means the spend is decaying into the P&L with nothing to show for it.
Across sectors
Here is the inversion, and it is the heart of the module: R&D means a completely different thing in each sector, and what counts as "good R&D intensity" inverts. In pharma the pipeline is the future business, so high intensity is the price of survival and low intensity is the warning. In FMCG, R&D is small, incremental renovation, so high intensity is the anomaly. In IT it is platform and IP investment often buried inside ordinary operating cost. In autos and industrials it is a lumpy, multi-year platform cycle. The same 4% of sales that starves a pharma pipeline would bloat an FMCG P&L.
R&D IS the future business. The molecule and ANDA pipeline the spend builds becomes the revenue once today's products erode on price, so intensity of 8-14% is the norm and the inversion is the reading: low intensity is the warning, not the reassurance. Under Indian accounting most of it is expensed, so a research-heavy pharma's ROCE reads flatteringly high — the real asset, the pipeline, never appears on the balance sheet. Watch filings, approvals and launch cadence, not the rupee spend.
R&D is incremental renovation — reformulations, new variants, pack and format changes — and it is small, often around 1% of sales. High intensity here is the anomaly, not the virtue: it usually means either a genuine push into new categories or a misclassification. The output to watch is the cadence and success of new SKUs and the share of revenue they contribute, because the innovation is cheap and continuous, not a discrete pipeline.
R&D is platform and IP investment — building product, tooling and reusable assets — and it is frequently expensed as ordinary operating cost rather than shown on a distinct R&D line, so a services firm's 'R&D' can be nearly invisible in the accounts. For a product SaaS firm it can be capitalised as a software intangible instead. There is rarely a discrete filing pipeline; the output is product capability and platform revenue, which must be read from disclosure and commentary, not a spend ratio.
R&D is the new-platform cycle — a new vehicle architecture, an EV programme, a new engine — spent lumpily over three to five years and largely capitalised as development, then amortised across the model's life. Intensity swings with where the company sits in the cycle, so a single year tells you little. Watch the platform and model pipeline and the launch calendar; the spend leads showroom revenue by years, and the capitalised balance must eventually be justified by models that actually sell.
The inverting cell is pharma: it is the one sector where R&D is not a discretionary overhead to be minimised but the core productive asset, so the ordinary instinct — "lower cost is better" — reverses, and falling R&D intensity is a red flag rather than a sign of discipline.
Reading it live
A composite generics exporter, Ashcroft Pharma illustrative, reports a comfortable year: revenue up 9%, margins steady, and R&D held at 5.5% of sales, down from 8% three years ago. [illustrative] The headline reassures; the detail does not. Management has been protecting current margin by easing off research — intensity has fallen a quarter — and it shows downstream: ANDA filings have slowed from a dozen a year to four, and the share of revenue from products launched in the last five years has slipped from 26% to 14%. [illustrative] The pipeline that is supposed to replace today's eroding products is thinning, and the 9% growth is being harvested from an installed base, not renewed.
Set against the leading-indicator map (087), Ashcroft's forward signal is flashing amber while its financials look green. Meanwhile a peer that kept intensity at 9% and lifted filings is building the revenue of 2029 at the cost of today's margin — the classic shape of R&D as a forward indicator, where the company with the worse-looking current P&L has the better-looking future. An investor who reads only the intensity level, or only this year's growth, sees the two the wrong way round; one who reads the trend, the filings and the new-product share sees the pipeline turning before the revenue does.
The instrument — productive R&D versus money burned
The read reduces to a set of signatures. Productive R&D and burned R&D can carry the identical spend line; they separate on output, conversion and honesty of accounting.
| What you check | Productive R&D | Money burned |
|---|---|---|
| Intensity trend | Stable or rising, matched to the sector's norm | Quietly falling while innovation is talked up — pipeline starved to protect margin |
| Output | Filings, approvals, patents and launches growing | Spend rising, tangible output flat or vague |
| New-product revenue share | Meaningful and holding or rising | Small and shrinking — recent launches contribute little |
| Productivity (output ÷ cumulative R&D) | Steady or improving over the window | Deteriorating — more money in, less revenue out |
| Accounting | Expensed, or capitalised and steadily converting to launches | Capitalised, balance swelling, few launches — impairment waiting |
| Lag behaviour | Revenue emerges roughly on the sector's delay | The promised revenue keeps slipping another year |
No single row is a verdict. A pipeline can be lumpy; one slow year of filings is not decay. The confident read is a cluster: intensity, output, new-product share, productivity and accounting all pointing the same way.
What R&D cannot tell you
R&D output leads revenue, but it does not guarantee it. A filing is not an approval, an approval is not a sale, and a launch is not a profit — a generic can be approved into a market so crowded that leaves it barely worth making. Counting filings without asking about the economics of what is filed is the pipeline version of confusing motion with progress.
It cannot give you a precise lag. The ranges are indicative, approval cycles stretch, and platform programmes slip; the spend is forward-looking, but the horizon is fuzzy. Because the feedback loop runs in years, you often cannot tell productive from burned in real time — only the passage of several years, with output either appearing or not, settles it.
And R&D productivity says nothing about the size of the prize. A company can convert its spend efficiently into small products in a small market; high productivity on a trivial is still a small business. The instrument tells you whether the spend is converting, not whether what it converts into is worth having.
Where people get fooled
The first error is reading the input as the output — treating high R&D-as-a-percentage-of-sales as proof of innovation. , and the crowd fixes on the comfortable, quoted percentage while the filings, launches and new-product share — the numbers that actually lead revenue — go unread.
The second is ignoring the accounting policy. A company capitalising R&D that peers expense looks more profitable and asset-rich, when it has merely deferred a cost. The swelling "intangible under development" balance with no matching launches is money burned dressed as an asset, and it ends in an impairment that arrives as a nasty surprise to anyone who trusted the profit (this is expense-capitalisation from the forensics part, 056, seen through the R&D lens).
The third is applying one benchmark across sectors. "4% of sales on R&D" is starvation for a pharma exporter and excess for an FMCG staple; a single mental yardstick produces exactly the wrong verdict in one of them. The level is only meaningful against the sector's norm; only the trend travels across sectors.
The fourth is cheering the filing count. A pipeline of approvals into commoditised, eroding markets can be busy and worthless. Filings are the beginning of the question — what market, what competition, what margin — not the answer, and a high count read as automatic future revenue is a forecast built on activity rather than economics.
Decide
Test your reading, not your memory — short decisions under incomplete information. The answer only shows after you commit.
All figures are illustrative — constructed to demonstrate a judgement, not reported as fact.
Carry forward
- R&D is a forward indicator because spend precedes the revenue it creates by years — but only the OUTPUT leads: filings, approvals, patents, new SKUs and launch cadence. R&D as a percentage of sales is an input; on its own it measures the size of the bet, not whether it is winning.
- Productivity is the number that actually leads revenue: revenue from products launched in the last N years divided by cumulative R&D over that window. Two firms with identical intensity can differ manyfold in productivity — the ratio is what the intensity figure conceals.
- Capitalising R&D flatters current profit by parking the cost as an intangible asset; a swelling 'intangible under development' balance with no matching launches is deferred cost that ends in an impairment — the R&D face of expense-capitalisation.
- What counts as 'good R&D intensity' inverts by sector: in pharma the pipeline IS the business, so high intensity is the norm and low intensity the warning; in FMCG R&D is minor renovation, so high intensity is the anomaly; in IT it is platform investment often buried in opex; in autos and industrials it is a lumpy multi-year platform cycle. Read the output native to each, never one cross-sector benchmark.
Enables: 104 The sector playbook
Do not read the R&D spend line — read what it produced: the output, the trend, and the productivity of it, against the sector's own norm, because the same intensity that starves a pharma pipeline bloats an FMCG P&L.