The answer to a low price-to-book ratio is in the plant register

In the spring of 2023 the Tokyo Stock Exchange asked listed companies to do something that sounds obvious and proved uncomfortable: to analyse their cost of capital and the valuation the market placed on them, and to publish what they intended to do about it. The request was addressed to every company on the Prime and Standard markets, but everyone understood whom it was really for. Roughly half of the larger listed companies in Japan then traded below the book value of their equity. Hatsuseki was one of them. At the end of March 2023 our shares traded at 0.61 times book.

A price-to-book ratio below one is a blunt statement. It says that investors believe the company's assets, in the hands of its present management, are worth less than the accounts say they cost. It is not a statement about the share price being unfair, though many of us read it that way at first. It is a statement about the plants, the inventories, the cross-held shares and the cash, and about what each of them earns.

What the exchange actually asked

The exchange did not ask companies to raise their share prices. It asked them to understand their own cost of capital, to compare it with the returns they earned, and to explain to shareholders what they would change. That is a request for analysis before action, and it is easy to skip the analysis.

Our first internal draft response, in early 2023, skipped it. It proposed a larger dividend, a buyback and a more ambitious medium-term sales figure. It was a reasonable document, and it would have moved our ratio for a quarter or two. It did not say which of our businesses earned less than our capital cost, because at that time we did not measure our businesses that way. We measured sales and operating income by segment. We knew the operating margin of every plant to one decimal place and the capital employed in almost none of them.

So we started again. Our finance team estimated a cost of capital for the company of about 7 per cent, and higher for Functional Resins, whose earnings rise and fall with the circuit-board cycle. Then we calculated return on invested capital for each segment and, more roughly, for each major plant. The results were not surprising to anyone who had worked in the plants, but they had never been in one table. Electronic Materials earned well above its cost of capital. Functional Resins earned about its cost of capital in good years and well below it in bad ones. Basic Chemicals, the chlor-alkali business at Toyama that founded the company, earned between 1 and 3 per cent on sales and far less than its capital cost, with ¥30 billion of electrolyser renewal approaching. And ¥54.8 billion of cross-held shares, about a fifth of our net assets, earned dividends of roughly 2 per cent.

A buyback buys time, not returns

We did buy back shares: ¥36.2 billion over three years. I do not regret it, because we had more equity than our plans required and returning it was right. But I have come to distrust the way buybacks are discussed in Japanese boardrooms, as if they were the answer to the exchange's request rather than one consequence of answering it.

A buyback raises return on equity by reducing equity. If the underlying businesses earn the same as before, the improvement is arithmetic. It also uses cash that might otherwise fund the plants that earn more than their capital cost. A company that buys back shares while leaving a low-return business untouched has made its ratio look better and its portfolio no better at all. Investors notice. The ratio may rise for a while and then drift back, because nothing that produced it has changed.

The more useful question is what the buyback is paid with. Ours was funded mainly by the sale of cross-held shares, which converted capital earning 2 per cent into a smaller share count. That is a genuine improvement in capital efficiency, and it is different from borrowing to buy back shares or spending cash that the Arizona plant needed. When a board approves a buyback, it should be able to say which low-return asset paid for it.

Reading the plant register line by line

The work that actually moved our ratio was slower and less visible. It consisted of reading our register of plants and holdings one line at a time and asking, for each, whether it earned more than its capital cost or had a credible path to doing so.

For the cross-held shares the answer was mostly no, and we sold 33 of 41 positions over four years. For Basic Chemicals the answer was that it could earn its capital cost only in the hands of an owner whose whole strategy was soda and chlorine, so we sold 66 per cent of it to such an owner in October 2025. For Functional Resins the answer was uncertain, and we said so publicly: if its return on invested capital does not exceed its cost of capital on average over the three years to March 2028, the board will review its ownership. For Electronic Materials the answer was yes, and we reserved most of our growth investment for it, including a phased plant in Arizona whose first phase began supplying a customer in August 2026.

None of these decisions was popular with everyone. The sale of Basic Chemicals ended our ownership of the business my grandfather worked in. The cross-shareholding sales cost us about ¥1.1 billion a year of business with one customer who saw the sale as disloyal. The public test for Functional Resins worries its managers, who fear that customers will read it as a signal that we are leaving. I think that worry is reasonable, and I think the alternative, leaving a business's future unexamined because examining it is awkward, is worse for the people who work in it.

By the end of June 2026 our shares traded at 1.07 times book value and our return on equity for the year to March 2026 was 8.6 per cent. I do not claim the ratio will stay above one. A downturn in semiconductors would test it, and the market is entitled to change its mind about us. What I do claim is that the ratio now reflects a portfolio we have examined line by line, rather than a buyback we used to postpone the examination.

For directors of other companies trading below book value, my suggestion is modest. Ask management for return on invested capital by business and by major plant, against a cost of capital the board has approved. Ask which assets earn less than that cost and what the plan is for each of them. Ask what any buyback is paid with. If management can answer those three questions, the response to the exchange will write itself. If it cannot, no buyback will answer it.