The Zombie Pattern: How Distressed Companies Drain Before They Fall

1. Two numbers, two ends of the same process

On 10 June 2026, the Bank of Korea released its 2025 Corporate Business Analysis. Among 34,456 externally audited non-financial companies, 39.9% had an interest coverage ratio below 100% — up from 38.5% a year earlier, and the highest figure the series has recorded. The share of firms running an operating loss reached 28.2%, also a record. Separately, the Federation of Korean Industries reported on 30 June that 27.6% of listed Korean companies now qualify as marginal firms — three consecutive years of interest coverage below 1 — against 11.8% in 2017.

On 9 July, Tokyo Shoko Research reported that Japan recorded 5,346 corporate failures in the first half of 2026, up 7.1% year on year and above 5,000 for the first time since 2014. Ninety percent of those firms employed fewer than ten people. Total liabilities of failed firms reached ¥734bn, up 6.3% — the first increase in four years — and large failures (over ¥1bn in liabilities) hit 114, the most in six years for a first half.

These look like two different stories. They're the same story, photographed at different moments.

2. What RII actually measures

The Reinvestment Intensity Index doesn't ask whether a company is paying its debts. It asks whether the company is still buying its own future.

The mechanics: RII combines the reinvestment rate (capex relative to operating cash generation), the coefficient of variation in capex over time, and the investment gap — the divergence between what a firm's revenue trajectory implies it should be investing and what it actually invests. A firm in genuine distress rarely announces it. What it does is close the investment gap in the wrong direction: revenue is defended, headcount is defended, interest is paid — and capex is quietly deferred, quarter after quarter, until the deferral becomes the business model.

This is why insolvency statistics are a lagging measurement. By the time a firm appears in Tokyo Shoko Research's count, the RII signal has been degrading for years. Japan's 2026 numbers are not information about 2026. They are the settlement of decisions made when money was free.

3. Korea parallel — the entrance, not the exit

Japan's zombie share peaked around 14% in 2011, per Teikoku Databank, following the 2009 SME Financing Facilitation Act (the "Kamei Act"), which approved over 90% of loan-modification requests during its 3-year-4-month run. The policy prevented an immediate collapse and produced a two-decade delay.

Korea's 39.9% is nearly three times that peak. For further scale: US zombie shares ran 4–6% after the dot-com bust and 8–10% after 2008.

What makes the Korean figure harder to see is aggregation. Exceptional results at a handful of exporters — Samsung Electronics, SK Hynix — lift the aggregate corporate numbers while four in ten audited firms cannot service their interest from operations. The index level tells you nothing about the distribution beneath it. This is precisely the condition under which relational risk compounds: capital stops being deployed into the business and starts circulating to keep the structure alive, and the individual investor — holding the small-cap, reading the headline aggregate — is the last to learn which side of the distribution they own.

4. Academic frame

Caballero, Hoshi and Kashyap (2008, American Economic Review, "Zombie Lending and Depressed Restructuring in Japan") established the core finding: the damage from zombie firms is not their own failure but the congestion they create — depressed investment and job creation among healthy competitors in the same sector. Banerjee and Hofmann (2018, and the 2020 BIS Working Paper No. 882, "Corporate zombies: Anatomy and life cycle") extended this internationally, documenting that zombie firms have become more likely to survive over time — and that the mechanism running through the balance sheet is a reinvestment collapse well ahead of any default event.

5. What this means if you hold the stock

Screening for default risk finds companies at the exit. Screening for reinvestment collapse finds them at the entrance — while the disclosure still reads clean and the revenue line still holds.

So the practical question isn't "which of my holdings might fail?" It's the one nobody asks: which of my holdings has stopped buying its own future, and how long has that been true?

This is a general market observation based on published data and is not investment advice.

#RaymondsRisk #RelationalRisk #CorporateGovernance #ZombieFirms #InterestCoverage #KoreaEquities