Chief Financial Officer / Singapore / Southeast Asia
Costs accepted early so that larger ones never arrived
- Total assets, December 2025
- RM46.2bn
- 64 per cent of financing to households, 36 per cent to businesses
- CET1 ratio, December 2025
- 13.6%
- Up from 12.4 per cent in December 2023 under a stress-tested payout
- Current and savings accounts
- 31%
- Share of customer deposits, up from 24 per cent in 2021
- Finance colleagues led
- 240
- Reporting, balance sheet management, capital, tax and zakat
Decisions & consequences
01Hedging the salary-deduction book before the policy rate rose
In December 2021, RM7.4 billion of our financing was salary-deduction personal financing to public-sector employees, written at fixed profit rates for up to ten years and funded mostly by term deposits that repriced within twelve months. It was our most reliable credit asset and also our largest rate exposure. Under a 200 basis point rise, the economic value of our equity would have fallen by 17.8 per cent of Tier 1 capital, and the overnight policy rate had been at a record low of 1.75 per cent since July 2020.
I recommended two measures to the asset-liability committee in January 2022: Islamic profit rate swaps covering RM3.0 billion of the book for five years, and, for new financing, a 40 basis point increase in pricing with the maximum tenor cut from ten years to eight. Our head of consumer banking objected strongly. Salary-deduction financing was the franchise that brought public-sector customers to us, and competitors would keep offering ten years. Our treasurer supported hedging but wanted RM5.0 billion. I chose the smaller figure because a full hedge would have locked in a large cost if rates had stayed low, and our new pricing would gradually shorten the book.
We executed the swaps in three tranches between February and April 2022 at an average fixed rate of 3.2 per cent, after the Shariah Committee confirmed the documentation. The first year's carry cost was RM19 million. New personal financing disbursements fell by about RM1.1 billion over 2022 and 2023, and our share of new public-sector financing slipped.
Between May 2022 and May 2023 the policy rate rose by 125 basis points. Our net financing margin fell from 2.34 per cent in 2021 to 2.18 per cent in 2023; unhedged, our model showed 2.02 per cent. Value sensitivity fell to 9.6 per cent of Tier 1. Fixed-rate personal financing is now 21 per cent of the book, down from 26 per cent, and every new fixed-rate product reaches the committee with its value sensitivity shown beside its expected credit loss.
I would do two things differently. I would have started hedging in tranches during 2021, when swaps were cheaper. And I would have shown the consumer team the branch-level economics before the pricing change, because they learnt about it from a rate sheet.
Recommending a smaller dividend than our listed parent had promised
In 2021 our listed parent told its shareholders that it aimed to distribute about 60 per cent of earnings, and nobody had asked the bank whether its capital plan could support that. Almost all of the parent's income came from our dividends. By late 2023, preparing the recommendation on the bank's 2023 dividend, I could see the problem in our own stress test. With a 60 per cent payout and financing growth of 8 per cent a year, our CET1 ratio would fall below the board's internal floor of 12 per cent in the severe scenario, reaching 11.4 per cent at its trough.
I recommended a 40 per cent payout from the bank to the parent, and suggested that the parent offer its own shareholders a dividend reinvestment plan, with the cash retained from reinvestment subscribed back into the bank as new ordinary shares. The parent's managing director disagreed firmly, and so did two directors appointed by its largest shareholder. Their argument had force: the market had priced the parent on the 60 per cent signal, and breaking it would damage trust in management. The chair of our Board Risk Management Committee supported my analysis.
The board approved a 45 per cent payout for 2023, with the reinvestment plan. The bank paid RM176 million to the parent instead of RM235 million. Shareholders holding 61 per cent of the parent's shares chose new shares over cash, and the parent subscribed RM107 million for new bank shares. Its share price fell 9 per cent in the week after the announcement, and I spent much of that quarter with its investors.
CET1 rose from 12.4 per cent at December 2023 to 13.1 per cent a year later and 13.6 per cent at December 2025. The severe-scenario trough now sits at 12.6 per cent, above the floor, and the payout for 2025 was 50 per cent. We have issued no additional tier 1 capital. The board now approves a capital plan before the dividend paper is written, and the dividend is shown as what the plan leaves over.
What I would change is my own silence in 2021. I was CFO when the 60 per cent aspiration was announced, and I let it go out without a capital plan behind it. Challenging it then would have cost an awkward meeting instead of a share-price fall.
Keeping RM96 million of credit overlays for a year longer
By December 2022 we carried RM214 million of management overlays above our modelled expected credit loss. They had been built during 2020 and 2021, while repayment moratoria and later assistance programmes suppressed the arrears that our models used to detect deterioration. In 2023 the arrears data was flowing again, and our parent's managing director asked that the whole amount be released that year. The release would have lifted profit before zakat and taxation by RM214 million and, not by coincidence, supported the dividend the parent wanted.
I agreed that most of the overlay had done its job, but not all of it. Personal financing and home financing cohorts that had made twelve months of on-time payments after their assistance ended were behaving like their pre-pandemic peers. Three SME cohorts were not. In hospitality, construction subcontracting and land transport, accounts thirty days or more past due were still running at 2.3 times their 2019 level. I recommended releasing RM118 million in 2023 and keeping RM96 million against those three cohorts, each with a written condition for release, and recalibrating our probability-of-default models with data from 2020 to 2023 during 2024 so that no overlay would be needed after that.
Two directors felt I was being cautious at the shareholder's expense. The chair of the Board Audit Committee asked for the release conditions to be minuted and tested each quarter, and the committee approved the split. Our external auditors reviewed the cohort analysis as a key audit matter.
In 2024 two of the three cohorts deteriorated as several construction subcontractors lost payments from main contractors. RM61 million of the retained overlay was absorbed as exposures moved into stage 3, rather than arriving as a fresh charge. Our gross impaired financing ratio rose from 1.31 per cent at the end of 2023 to 1.55 per cent at the end of 2024, and eased to 1.42 per cent in 2025. The remaining RM35 million was folded into the recalibrated models in the fourth quarter of 2024.
I would do one thing differently. The release conditions should have been written in 2020, when the overlays were created and when I was deputy CFO. Writing them in 2023, under pressure to release, made a sound judgement look like a negotiating position.
Letting RM2.6 billion of institutional deposits leave
In mid-2022 our twenty largest depositors provided 34 per cent of customer deposits. Most of that money was institutional term funding from statutory funds, government-linked companies and large corporates, placed as commodity murabahah deposits for one to six months. It was easy to raise and expensive to keep: every rise in the policy rate passed through within weeks, and a handful of treasurers could move billions at a single maturity. Our liquidity ratios were comfortable, but our funding depended on a few people's view of us.
I recommended to the asset-liability committee that we let RM2.6 billion of the most rate-sensitive institutional deposits run off over eighteen months by declining to match competitors' rates at maturity, and replace them with operating balances: payroll accounts from public-sector employers whose staff already held our salary-deduction financing, and cash-management accounts for SMEs. That required RM74 million of investment in a transaction banking platform and a payroll product. Our treasurer argued that institutional money was the cheapest source to raise and the fastest to replace if needed. Our head of corporate banking warned that two large clients linked their deposits to their financing.
He was right. Two corporate clients moved about RM640 million of financing to other banks during 2023. Our financing growth slowed to 4 per cent that year, and the liquidity coverage ratio fell from 162 per cent to 139 per cent at its low point in the third quarter of 2023, still well above the requirement but closer than the board liked.
By December 2025 current and savings accounts had risen from 24 to 31 per cent of customer deposits, and the top twenty depositors' share had fallen to 21 per cent. The net stable funding ratio rose from 104 to 112 per cent. Our cost of funds, 2.38 per cent in 2025, now responds more slowly to the policy rate in both directions, which helped when the rate was cut in July 2025. The asset-liability committee now receives a monthly concentration report alongside the liquidity ratios.
I would sequence it differently. Onboarding payroll employers took fourteen months longer than I planned, so for a year we were running deposits down before the replacement was ready. I would build the cash-management capability first and start the run-off only when the first employers were live.