Chief Executive Officer / Japan
Decisions that cost us something we valued
- Net sales, year to March 2026
- ¥271.4 bn
- Electronic Materials 53 per cent, Functional Resins 36 per cent, six months of Basic Chemicals 11 per cent
- Return on equity, year to March 2026
- 8.6%
- Up from 4.4 per cent in the year to March 2022
- Price-to-book ratio, end of June 2026
- 1.07x
- From 0.61x at the end of March 2023
- Cross-held listed stakes
- 41 → 8
- April 2022 to March 2026; ¥49.3 billion realised
Decisions & consequences
01Disclosing altered resin certificates before the investigation ended
In October 2022, four months after I became president, an employee at our Ichihara resin plant used the internal reporting line to say that certificates of analysis for two grades of laminate epoxy resin had been altered. Within ten days our internal audit team confirmed the core of it: for 2,140 lots shipped to 37 customers since 2014, gel time and epoxy equivalent values on certificates had been adjusted to fall inside customer limits, or reported without the agreed number of measurements. We had no evidence of failure in the field, but we could not yet show there was none.
The question was when to tell people. The head of Functional Resins and our general counsel argued for notifying customers individually and disclosing publicly only after an external investigation, perhaps three months later. Their reasons were serious: an early announcement might be inaccurate, might frighten customers whose products were unaffected and would invite comparison with the sector's wider history of data scandals. I decided that customers needed the facts we had at once, and that the market should hear them at the same time. Six days after confirmation we made a timely disclosure to the exchange and briefed every affected customer the same morning.
We gave up a quiet resolution. The share price fell 14 per cent in two days. We appointed a special investigation committee of three outside members, chaired by a lawyer with no previous connection to the company, and published its report in full in March 2023. It found that the customer limits had been agreed in 2012 at levels the process could not consistently achieve, and that the plant had never asked for a waiver. The total cost was ¥4.9 billion in retesting, compensation and systems. Two customers moved about ¥3.6 billion of annual business to other suppliers.
Customers' requalification found no lot that failed in use. We moved quality assurance out of plant managers' reporting lines, removed manual entry from certificates in all nine plants and renegotiated 23 specifications. I returned 30 per cent of my salary for three months.
I would change one thing. Internal audit had noted manual certificate entry at Ichihara in 2019 and again in 2021. I was a director by 2021 and read the second report. I should have asked why it was still open.
Selling thirty-three cross-held stakes, including in our own shareholders
In April 2022 Hatsuseki held shares in 41 listed companies, carried at ¥54.8 billion, about 20 per cent of our net assets. Most were customers, two were our main banks and several were suppliers. Many of them also held our shares; together, banks and business partners owned about 17 per cent of Hatsuseki. Our return on equity was 4.4 per cent, and capital that earned dividends of roughly 2 per cent was part of the reason.
I proposed to the board in early 2023 that we sell every holding that could not show a return above our cost of capital, counting both its dividend and any business we could fairly attribute to the relationship. That meant 33 of the 41. Our chairman, my predecessor Masaru Ebihara, disagreed. He had built several of the relationships himself and argued that the holdings protected us from a hostile approach and signalled loyalty to customers in a cyclical industry. The head of sales for Functional Resins agreed with him, and one laminate customer told us plainly that it would review its purchases if we sold.
The board approved a four-year programme rather than the two years I had proposed, which was the compromise that brought the chairman with us. Before each sale I either met or wrote to the chief executive of the company concerned, and where it held our shares we said we would not treat its sale as unfriendly.
We realised ¥49.3 billion. The laminate customer cut its purchases from us by about ¥1.1 billion a year, a cost I had judged possible and it proved real. By March 2026 we held eight positions worth ¥9.6 billion, and the share of Hatsuseki held by banks and partners had fallen to about 7 per cent. The proceeds funded ¥36.2 billion of buybacks and part of the Arizona plant. Our board now reviews each remaining holding every year with a written return calculation, and the holdings report has a column for the date we expect to sell.
I would do it differently in one respect. The slower timetable left us holding about ¥15 billion of shares into 2025, when an activist fund pointed to them as proof that we were not serious. Had I argued harder for two years, that criticism would have had less to point at.
Answering an activist's proposals without accepting its remedy
In October 2024 an overseas activist fund filed a report showing that it held more than 5 per cent of Hatsuseki, and by the spring it held 6.8 per cent. It put two proposals to our June 2025 general meeting: a special dividend of ¥45 billion, and an instruction to the board to pursue a sale of Functional Resins. Its letters argued that we still held excess cash and cross-held shares, that Functional Resins earned below its cost of capital across the cycle, and that our board was too close to management.
I thought the fund was right about the problem and wrong about the remedy. A ¥45 billion special dividend would have left us unable to fund the second phase of the Arizona plant without borrowing, and a forced sale of Functional Resins at the bottom of the laminate cycle would have sold it cheaply. Two of our directors favoured adopting an advance-warning takeover defence plan. I argued against it, because it would have told shareholders we cared more about the fund's presence than its arguments, and the board agreed.
Instead we accepted part of the case in public. We brought forward the completion of our cross-shareholding sales from 2027 to March 2026, published return on invested capital for each segment for the first time, and proposed a board with a majority of independent outside directors, including a former investment manager. For Functional Resins we set a test: return on invested capital above its cost of capital, averaged over the three years to March 2028, or the board would review its ownership.
Both proposals were defeated: the special dividend received 33 per cent support and the sale proposal 21 per cent. My own reappointment received 81 per cent, down from 97 per cent the previous year, which I took as a fair message. The fund has since reduced its stake to 5.9 per cent. Functional Resins earned an 8.1 per cent return on invested capital in the year to March 2026, and the test is public.
What I would do differently is speak to shareholders earlier. Our first detailed engagement came after the proposals were filed. Most of what we eventually announced could have been said a year sooner, and saying it then would have been more persuasive than saying it under pressure.
Selling control of the business that founded the company
The Toyama chlor-alkali business founded Hatsuseki in 1936. By the year to March 2024 it had sales of about ¥62 billion, an operating margin between 1 and 3 per cent and electrolysers that needed about ¥30 billion of renewal within five years. Electricity was its largest cost, and the business carried most of our exposure to power prices and to the cost of reducing emissions. It also produced the high-purity hydrochloric acid feedstock on which part of our electronic chemicals business depended, and it employed 1,080 people in a city where Hatsuseki was one of the largest employers.
I concluded in 2023 that the business deserved an owner whose whole strategy was soda and chlorine, and that we could not justify ¥30 billion of renewal at its returns. The Toyama branch of our union opposed a sale. So did several retired directors, and the city of Toyama asked to be consulted. Within management the head of Electronic Materials worried about losing control of his feedstock. The trade-off was clear: we would give up our history and some security of supply for capital we could use better and a smaller exposure to power prices.
We carved the operation out as Jinzu Electrochemical Co., Ltd. and ran a sale in which employment terms and feedstock supply carried as much weight as price. Hinomori Soda Co., Ltd., a regional soda maker, bought 66 per cent for ¥18.6 billion, completing on 1 October 2025. We kept 34 per cent, a seat on its board and a ten-year contract for high-purity feedstock at a formula price. All 1,080 employees transferred with pay and conditions protected for three years. I met the union branch eleven times and the mayor three times before signing.
A higher bid of about ¥21 billion came from a financial investor that would not commit to employment terms beyond one year; we declined it. Our consolidated electricity consumption fell by about 70 per cent, and the renewal capital now sits with an owner that can spread it across three soda plants.
I would do one thing differently. I told the union before the board had approved the process, which was right, but I told retired employees only after the announcement. Many of them learnt about it from the local newspaper. They deserved to hear it from me.