In a pharmaceutical acquisition, the licences are the company
Buyers of generics manufacturers in Southeast Asia still build their integration cases on plants, headcount and purchasing. Most of the value, and most of the delay, sits in the marketing authorisations, and only regulators decide when those can move.
When Lumaren's board approved the purchase of PT Sindhara Farmasi in 2022, the investment paper ran to sixty pages. Eleven of them described synergies: purchasing savings on active ingredients, a shared enterprise system, fewer roles in finance and administration, and the transfer of about forty Sindhara products to our plant in Johor, which had capacity to spare. One page dealt with registrations. It listed how many products Sindhara held approvals for and noted, correctly, that those approvals would remain valid after the change of ownership.
I signed that paper, and I now think its proportions were upside down. The single page on registrations described the asset we were actually buying. The eleven pages on synergies described benefits that depended, to a large degree, on regulators in several countries agreeing to changes on their own timetables, and in one important case on a change that Indonesian practice made unrealistic. We had valued the buildings and the payroll of a company whose real substance was a set of permissions to sell.
Why the plant saving arrives last
A generic medicine may be sold in a country only under a marketing authorisation granted by that country's regulator. The authorisation does not describe the product in the abstract. It names the manufacturer and the site, the formulation, the specifications, the suppliers of key materials and the stability evidence behind the shelf life. Change any of those and the holder must usually obtain the regulator's approval for a variation before selling the changed product. The region's regulators use broadly similar categories for such changes, but each sets its own evidence requirements and works to its own timetable.
A buyer's favourite saving, moving production from the acquired plant to its own, is precisely the kind of change that triggers the most demanding variations. Each product needs a change-of-site approval in every country where it is registered, usually supported by stability data from batches made at the new site. A product registered in five markets means five approvals, each arriving when it arrives. Until the last is granted, the old site must keep making the product for at least some markets, so the saving from closing a line or a plant cannot yet be taken. The saving the model books in year two often arrives in year four, if it arrives at all.
In Indonesia the difficulty was more basic. The country's registration practice and its public procurement both favour medicines made in Indonesia, and its national health insurance scheme buys a large share of the generics sold there. Moving Sindhara's products to Malaysia would have turned locally made medicines into imports, weakening their standing in public purchasing and, for some products, leaving no practical route to registration. That saving existed only in the spreadsheet.
Counting variations, not headcount
Once the deal had completed, we rebuilt the integration plan around the registrations. Every one of Sindhara's roughly 310 products, and every Lumaren product we might want to make in Indonesia, received a line in a register showing where it was approved, what change we proposed, which variations or new registrations that change would need, the evidence required and our best estimate of timing. The register replaced headcount as the integration's main measure of progress.
It reversed the direction of the plan. Instead of moving Indonesian products out, we moved 22 Lumaren products into the Cikarang plant, each registered as a locally made medicine. Instead of putting Sindhara onto our enterprise system in the first year, we aligned the two companies' quality systems first, because a regulator reviewing a change at a site expects to find one coherent quality system rather than two in the middle of a merger. Sindhara kept its president director and its management, because the people who knew how its files had been built were the only ones who could change them without introducing errors.
The cost was visible. Savings in the first two years fell about S$9 million short of the investment paper. One of Calderwick Capital's directors on our board argued, reasonably, that the paper's numbers had been the basis of the price, and he asked for his disagreement to be minuted. By 2025 eighteen of the 22 transferred products had been approved and listed for public purchase. Annual savings had reached S$15 million against the S$24 million the paper had promised for its second year, and Sindhara's revenue had grown from S$188 million to S$236 million.
What I ask before signing
I now put four questions to any acquisition in this industry, and I would put them as a director of any company buying a regulated manufacturer.
First, what exactly are we buying the right to sell, where, and in whose name? The answer should be a product-by-market register of authorisations, with renewal dates and pending variations, checked against regulators' own records wherever those are public.
Second, which synergies need a regulator to approve something, and what is a realistic timetable for each, based on that regulator's recent behaviour rather than its published service standards? Savings that depend on variations belong in a later year of the model and in a different risk category from savings that management controls.
Third, who in the acquired company built the registrations, and will they stay? Licences are maintained by people who know their history. Losing those people in the first year of an integration is the quickest way to turn a portfolio of approvals into a portfolio of queries.
Fourth, can the price carry the timetable risk? Had we proposed part of Sindhara's price as deferred consideration, payable as the transferred products were approved, the sellers would have shared a risk that we carried alone. I did not propose it, and next time I would.
There is a fair objection to all of this. Registrations are not the whole of a company. Plants matter, brands matter, and relationships with hospitals and pharmacies matter; some acquisitions are justified by a plant or a sales force rather than a portfolio. I accept that. My argument is narrower. In a generics business spread across several regulators, the authorisations set the timetable for everything else, and a valuation that does not start from them is guessing at when the rest will be worth anything.