Bay Bridge Bio · Pharma’s dual mandate, part 1

R&D doesn’t pay pharma shareholders. What does?

Pharma serves society by making medicines, but each company’s duty is to its shareholders. How much of pharma’s shareholder returns come from the medicines it invents?

Over the last 15 years, pharma has generated healthy shareholder returns. At the same time, the industry’s return on capital fell by more than a quarter, to about its cost of capital. R&D productivity stagnated well below the cost of capital.

If pharma can not invest profitably in its own R&D, how does it deliver shareholder returns? Through three main mechanisms: 1) higher valuation multiples, 2) growth through acquisitions, and 3) dividends + buybacks (all of which were funded through profits from existing products plus an increase in net debt).1

This may be a healthy ecosystem, where innovative startups provide new products, and big pharma acquires and commercializes them. Or it may be the sign of an industry with declining fundamentals, relying on accommodative financial conditions to sustain investor returns.

1Shareholder returns and cost of capital exceed R&D productivity

Top 20 pharma companies: shareholder return and R&D productivity, % a year

Shareholder return is the trailing three-year annualized total return of the companies Deloitte measures. R&D return is Deloitte’s projected return on the same companies’ late-stage pipeline. Companies included based on criteria in Deloitte’s R&D productivity report: 15 large companies to 2019, then the 20 largest R&D spenders (note 2).

2Where do investor returns come from?

If investor returns don’t come from R&D, where do they come from? We decompose shareholder returns into seven parts: revenue growth, margins, valuation multiple (EV / operating income), net debt, cash distributed to shareholders (dividends + buybacks), pre-revenue companies, and a residual.5 Revenue growth is then split into organic (invented by the company that sells them), or acquired (bought and in-licensed).6

Decomposition of pharma industry shareholder returns, % a year

Companies: the shareholder return covers drug and biotech companies listed in the US (NYSE, Nasdaq, or ADRs over the counter), plus Roche, Merck KGaA and UCB, excluding Japan: “Top 25” in the revenue bars means the 25 largest by the previous year’s revenue, so membership changes year to year.

GLP-1, PD-1 & COVID-19 vaccines everything else (red if it fell)

The four revenue bars count only growth in revenue. Revenue from drugs a company already sells — in-house or acquired — also produces the cash that pays for dividends and buybacks, so part of its contribution is in that bar, not in the revenue bars.

3Pharma increased debt to fund spending

The top 25 pharma companies spent more than they earned from 2008 through 2025 on acquisitions, dividends and buybacks.7 They increased debt to cover the difference, while their cash balances barely moved.

The 25 largest companies: debt and cash, $ billions

Companies: the 25 largest by 2025 revenue, the same 25 in every year (also the next chart).

Cash and net debt of the top 25 pharma companies ranked by 2025 revenue. Debt is total borrowings; cash includes short-term investments.

The 25 largest companies: spending on R&D vs. acquisitions and shareholder payouts

Companies: the 25 largest by 2025 revenue, the same 25 in every year.

4Can it last?

Despite fundamentals that can be viewed as deteriorating, the industry has supported shareholder returns through multiple expansion, increasing leverage, and sending profits from successful drugs to investors in the form of dividends and buybacks.

In one sense, this is rational. If the industry cannot profitably invest cash into their R&D engines, shareholders are better off having that money distributed as dividends or buybacks. From an investor perspective, the R&D engine doesn't have to get fixed, as long as pharma can buy enough growth to support shareholder distributions. Assuming those acquisitions are accretive, and the supply of acquirable companies remains plentiful.

But is that aligned with what is best for society? Is that an industry strong enough to survive competition with China? If we extrapolate this same pattern out 10-20 years, what does it look like? The answer to that relies in part on whether those two assumptions — is pharma paying prudent prices for acquisitions, and will the supply of acquirable companies remain plentiful — are accurate.

The next posts in this series will explore both of these: whether pharma overpays when it buys biotech companies, and whether biotech IPOs deliver the value investors pay for.

Methods and sources

  1. Universe. Drug and biotech companies with a US listing — NYSE or Nasdaq shares, or ADRs traded over the counter — that Tiingo covers (US, European and other SEC filers and ADR issuers), with Roche, Merck KGaA and UCB taken by hand from their annual reports. Companies listed only on their home exchange are not included, so large Chinese, Korean, Indian and Australian drug makers (Jiangsu Hengrui, Celltrion, Sun Pharma, CSL) and several European ones are missing, and the 16 Chinese companies in the panel are US-listed ones. Japanese companies are left out because their filings reach us only from 2011–13. Spin-offs are counted once (AbbVie, Zoetis and Organon sit inside Abbott, Pfizer and Merck before they separate), Mylan’s history is counted once under Viatris, and Shire is left out (it left the universe when Takeda, which is excluded, bought it). Conglomerates such as J&J and Abbott are counted whole. Fiscal years are the filer’s own; years ending Jan–Jun are not shifted. ↩
  2. R&D productivity. Deloitte, Measuring the return from pharmaceutical innovation (annual since 2010): the projected internal rate of return on a fixed group’s late-stage pipeline, latest published value per year. The chart uses Deloitte’s headline group — 15 companies to 2019 (Amgen, AstraZeneca, Bristol Myers Squibb, Eli Lilly, GSK, Johnson & Johnson, Merck, Novartis, Pfizer, Roche, Sanofi, Takeda, AbbVie, Biogen, Gilead), then the top 20 by 2019–20 R&D spend (adding Astellas, Bayer, Boehringer Ingelheim, Novo Nordisk and Regeneron) — and the shareholder return of the same listed companies (cap-weighted, dividends reinvested; Boehringer is private and left out), switching group in 2020 the same way. The 2021 value includes COVID-19 products (Deloitte’s ex-COVID figure for that year is 3.2%). Deloitte’s original 12-company group (2010–2020) tells the same story: 10.1% in 2010, 0.7% in 2019. Cost of capital: Damodaran’s January weighted average cost of capital for US drug companies. ↩
  3. Return on invested capital, R&D capitalized. After-tax operating income over average invested capital (equity + debt − cash and short-term investments), with a clipped aggregate tax rate, and R&D treated as an investment: each company’s R&D is amortized straight-line over 10 years, the unamortized balance is added to invested capital, and R&D minus that year’s amortization is added back to operating income. Expensing R&D understates both the income and the capital of an R&D-heavy company, so this is the fairer measure of the return on everything the industry has invested; it averages 10.9% in 2009–15 and 7.8% in 2016–25, roughly equal to the 8.0% average cost of capital of 2016–25 (5-year amortization: 11.5% and 8.2%). The capitalized R&D balance is $1.1T in 2025 against $1.4T of conventional invested capital. As reported (R&D expensed), ROIC falls further, 15.0% to 9.2% (−39% rather than −29%). ↩
  4. ↩
  5. The waterfall. Each year the change in combined market value of companies present in both years is split (in logs) into revenue, gross margin, operating income per dollar of gross profit (where rising R&D and selling costs show up) the multiple of enterprise value to operating income (enterprise value = market value plus net debt, where net debt is debt minus cash and short-term investments), and net debt. Cash returned = dividends plus net share buybacks minus net new shares, over start-of-year market value. “Pre-revenue companies” is the difference between the industry’s return with and without companies that have no revenue yet (statements showing no revenue that year and the year before; mostly clinical-stage biotechs, at most 3% of the industry’s weight in the return, in 2021); their value cannot be split into revenue, margins and a multiple, so it is one bar. “Other” is the remainder against the cap-weighted total return: calendar vs fiscal years, quarterly rebalancing, and other companies outside the same-company set (IPOs, and acquired companies in their final year). The parts are scaled in proportion to their contributions so they add to the annual return. The multiple leg depends on the start year: 2009 was a trough multiple after the financial crisis, and almost all of the re-rating happened by 2015. ↩
  6. Organic vs acquired revenue (an estimate). The revenue part is split into four bars. Top 25 = the 25 largest companies by the previous year’s revenue. Each one’s revenue is divided using its own annual reports: non-drug segments (generics, consumer, devices, animal health) go to “other revenue”; every drug with at least $1B of sales in any year (325 drug series, about 3,500 sourced product-years) is classed invented in-house (including merger-of-equals lineage and majority-owned units such as Genentech), bought (the seller acquired the company that discovered it, at any stage: Keytruda came with Schering-Plough, Opdivo with Medarex) or licensed in (Dupixent at Sanofi, Comirnaty at Pfizer). “Acquired” in this post = bought + licensed in. Those shares are applied to the panel’s revenue. Brands under $1B, royalties and alliance revenue: revenue that arrived with an acquired company (203 sourced deals and asset transfers) counts as acquired; the rest is split in proportion to the company’s classed drugs, and the hover shows the range if all of it were organic or all acquired. Innovator biotechs outside the top 25 are companies whose revenue is mainly drugs they discovered or developed (1,058 companies classed); specialty pharma built on bought products, generics, royalty buyers and service companies are in “other revenue”. ↩
  7. Both charts follow one fixed set: the 25 largest companies in the universe (note 1) by 2025 revenue, tracked back to 2008 — the same companies every year, so debt and cash are comparable levels from year to year. Carve-outs enter when they separate (AbbVie in 2011; before that its business is inside Abbott), which is why fewer than 25 report in 2008–10. A company these 25 bought has no row of its own; its debt and cash join the buyer’s on the deal. Debt = total borrowings (current + non-current); cash = cash and equivalents + short-term investments; net debt = debt − cash. Free cash flow = operating cash flow − capital spending. R&D is an expense, so it is already deducted from operating cash flow; “cash generated before R&D” adds it back so R&D can be compared with the other uses. Upfront payments for licensed-in drugs and acquired in-process R&D that is expensed sit inside R&D or operating cash flow rather than in acquisitions, so part of the “bought” pipeline is in the R&D bar. Cash acquisitions are cash paid for businesses net of cash acquired; deals paid in shares, and debt taken over with an acquired company, are not in that line, which is why net debt can rise by more than spending exceeds cash generated. Buybacks are Tiingo’s net equity line per company-year (repurchases where net negative), so option-exercise issuance is netted against buybacks. Net debt / EBITDA is over the companies that report EBITDA (Roche, Merck KGaA and UCB do not in our data). The loss-making figures in the text are the whole universe, each company-year classed by the sign of its net income. ↩