Follow the Cash: The Day a Shareholder Had to Price a Look at the Ledger

On 25 August 2026, Gabia — a KOSDAQ-listed hosting and cloud company — disclosed that Align Partners Capital Management and one other applicant had filed for an injunction at the Anyang branch of the Suwon District Court. The case number is 2026카합10135. The relief sought is unusually concrete: thirty days of access to the company's accounting books and records at its head office, starting three business days after the ruling is served; permission to bring lawyers and certified public accountants along; and ₩5 million for every day the company fails to comply. The application was filed on 10 August. Per the company's own disclosure, it was served on 25 August. Gabia says it will respond through legal procedure.

What makes this worth a Wednesday column is not the dispute. It is what the dispute reveals about where public data ends.

The request moved.

Six months earlier, in February, the same fund had asked for something far smaller. It submitted a recommendatory shareholder proposal — non-binding by design — asking the board to publish, once a year from fiscal 2026, the detailed basis on which directors and senior executives are paid. Align submitted the same proposal to five companies. Gabia explicitly declined to table it, arguing that recommendatory proposals have no basis in the company's articles and that the prevailing academic view is that matters outside statutory or charter resolutions are not proper subjects of a shareholder proposal at all. Align went to the same court in late February seeking an order to table it, and its counsel argued the opposite: that Article 363-2 of the Commercial Act, read with the narrowly enumerated grounds for refusal in Article 12 of its Enforcement Decree, does not permit refusal merely because a doctrinal question is contested.

Set the legal merits aside — they are genuinely unsettled. Look at the object of the two requests. In February: create a document. In August: let me see the document you already have. The first asks the company to publish. The second asks a court to grant access to something the company already possesses and does not publish.

Balances and movement.

That is the observability boundary, and it runs through the middle of every cash-governance measurement.

Financial statements publish balances: cash at a date, short-term instruments at a date, proceeds raised and proceeds stated as deployed. Every input into a cash-governance index — idle cash ratio, short-term instrument share, conversion rate of raised funds — is computed on that side of the line. Movement is on the other side. Who authorised what, to whom, on which day, against which resolution: that lives in the ledger, and the ledger, this filing shows, is not free. It costs a 3% stake, counsel, a filing, a service of process, and — between filing on 10 August and the market learning on 25 August — fifteen days that nobody could compress.

The correct inference is not that measurement fails without the ledger. It is that inaccessibility of the ledger is a permanent property of the object being measured, so the useful question is what consecutive balances force to be true about the interval between them. A snapshot is only a snapshot; a sequence of snapshots constrains the set of paths. An idle-cash ratio that will not fall while proceeds are booked as deployed, short-term instruments that absorb a raise and never release it, a conversion rate that never converges — none of these name a transaction. All of them narrow what can have happened. That is the entire claim available without a court docket, and it is worth making precisely because it is modest.

Korea parallel, and a Japanese counterexample.

RaymondsIndex is built on exactly this constraint: it reads Korean issuers off published aggregates, because published aggregates are what exists for every name, not just for the handful where a fund is willing to litigate. Coverage is the point. A ledger inspection is precise and singular; an aggregate-based signal is imprecise and universal, and for an individual investor who will never file anything, universal is the only category that is actually reachable.

Japan shows the same boundary from the other end. Hedgeweek, summarising Bloomberg data on 4 August 2026, reported that activist-held Japanese stocks underperformed TOPIX by an average of four percentage points over the first seven months of 2026, ending four consecutive years of outperformance, and that the success rate of PE-linked activist campaigns has fallen to roughly 36% from nearly three-quarters in 2022. Portfolio managers quoted in the report describe why: the first phase of Japan's governance reform targeted excessive cash balances, non-core assets and inefficient capital structures — that is, everything visible in published aggregates. Those targets have largely been worked through. Roughly one-third of TOPIX constituents now trade below book, down from about half in 2022. What remains are companies that simply do not answer; Elliott has reportedly received no public response at all from Kansai Electric Power. When the fights that can be won on published numbers are exhausted, the remaining fights move to the records — and those are slow and expensive.

Academic frame.

This is the classical disclosure problem, and the literature is unambiguous about the mechanism. Verrecchia's survey of the disclosure literature (Verrecchia, R. E., 2001, Essays on disclosure, Journal of Accounting and Economics) organises voluntary disclosure around the fact that managers choose what to publish under a cost, which is precisely why aggregate publication and ledger detail are not substitutes. Diamond and Verrecchia (1991, Disclosure, Liquidity, and the Cost of Capital, Journal of Finance) established the corollary that matters for a retail holder: reduced information asymmetry lowers the cost of capital through liquidity, which means the asymmetry is not merely an inconvenience for the uninformed but a price everyone pays.

Conclusion.

An injunction application is not a finding. Gabia may prevail; the doctrinal question it raises is real. But whichever way case 2026카합10135 goes, the boundary it maps does not move. Some parties can eventually buy a look at the ledger. Most cannot. If your process for judging where a company's cash actually goes depends on eventually seeing the records, it is a process available to almost no one. What would it take to make it available to everyone?

#RaymondsRisk #RelationalRisk #CorporateGovernance #BooksAndRecords #ShareholderActivism #KoreaEquities