Where a regional generics chief executive has to decide

Lumaren holds about 2,140 marketing authorisations across eight countries, each granted by a national authority: the Health Sciences Authority in Singapore, Malaysia's National Pharmaceutical Regulatory Agency, Indonesia's BPOM, the Drug Administration of Vietnam and their counterparts in Thailand, the Philippines, Cambodia and Laos. The ASEAN common technical dossier gives us one format for most of a file, but each authority sets its own data expectations, review times, labelling language and rules for changes after approval. I treat registration as a portfolio rather than a queue. Every product has a plan showing where it is filed and approved, when renewals fall due and which variations are pending, and we decide filings on expected margin per market rather than on the number of approvals we can announce. Where a regulator offers an abridged route for products already approved by a reference authority, we sequence filings to use it. In 2020 we let about 230 registrations lapse because keeping them cost more than they earned.

About 37 per cent of our revenue comes from public buyers, and no two countries buy the same way. Vietnam runs central procurement rounds for some high-volume generics alongside hospital tenders that group products partly by the manufacturing standard and approvals behind them. Indonesia's national health insurance scheme buys a large share of its generics through an electronic catalogue. Malaysia's Ministry of Health and hospital purchasers elsewhere tender on their own schedules. Since Vietnam's 2021 round, Lumaren has bid against a full-cost floor: plant time, laboratory load, safety stock to the contract's service level, the cost of keeping each registration alive and a modest return. A bid committee of commercial, supply, quality and finance leaders signs each floor; only I can approve an exception, in writing, and the board sees every exception each quarter. We track on-time, in-full delivery for every public contract, because a missed delivery costs more in the next tender than any price concession wins.

Our three plants make about 4.6 billion doses a year. The Malaysian and Indonesian sites are inspected by regulators that participate in PIC/S, the international scheme for co-operation between inspectorates, and the Vietnamese site is held to the standard that qualifies its products for the higher groups in hospital tenders there. I decide plant questions on three numbers together: the cost of the quality system a site needs, the products it is approved to make in each market, and the time a regulator will take to approve any move. That is why closing Jurong cost S$21 million rather than the S$48 million a rebuild would have taken, but needed about 420 variation filings across eight regulators. Capital spending runs at S$35 million to S$60 million a year, and every proposal must show the laboratory, warehouse and registration work its added output will require. Quality decisions belong to our qualified people; my job is to fund the conditions in which they can make them.

In regional generics, an acquired company is a set of marketing authorisations held in its own name and tied to its own sites, and every change a buyer wants needs a regulator's approval in each market. When we bought PT Sindhara Farmasi, I rebuilt the integration plan around a register of its 310 products: where each is approved, what change we proposed and what evidence and time that change required. Indonesia's preference for locally made medicines, reflected in registration practice and in local-content scoring for public procurement, meant the investment paper's largest promised saving, moving production to Johor, was not realistic. We reversed it and brought 22 of our products into Cikarang. I also learnt to sequence integration: quality systems first, because a regulator reviewing a change expects one coherent system; enterprise software later; and the people who built the files kept, because only they could change them safely. A warranty claim for a pre-acquisition tax assessment, settled from escrow, taught me to value the protections in the purchase agreement as highly as the price.

Lumaren's board brings together a founding family with 70 per cent, an investor with 30 per cent and consent rights over certain decisions, and three independent directors. I have helped build the structures that let those interests disagree without paralysing the company: a family charter, a family council separate from the board, one employment standard for family and outside managers, and a Nominating and Remuneration Committee chaired by an independent director that runs succession, including mine. Every director receives the same board pack at the same time, and I share with the whole board any answer I give a family member outside it. I am now preparing the board's choice on how Calderwick's stake should be realised, a listing or a sale to a long-term holder, which touches every one of those arrangements. The work has given me a practical view of what the Singapore Code of Corporate Governance asks of board composition and succession, and of where a family company's informal habits must become written rules.