Why annual-report text can reveal tunneling risk

"Tunneling" is the academic term for the transfer of assets and profits out of a company for the benefit of its controlling shareholders, at the expense of minority investors — for example through related-party transactions priced away from fair value, asset sales to affiliated entities, or preferential loan guarantees. The term and its foundational analysis come from Johnson, La Porta, Lopez-de-Silanes & Shleifer, who documented how legal systems with weaker minority-shareholder protections make this kind of expropriation both easier to execute and harder to detect from the outside.

Nanayojana's related-party-transaction signals (disclosed-without-value, unusually large relative to company size, disclosure gaps, and quiet cessation of prior disclosures) are a direct, annual-report-based operationalisation of this concept — the transaction types the literature identifies as tunneling channels are exactly what those signals look for.

Johnson, S., La Porta, R., Lopez-de-Silanes, F., & Shleifer, A. (2000). Tunneling. American Economic Review, 90(2), 22–27. doi.org/10.1257/aer.90.2.22

Claessens, Djankov, Fan & Lang studied East Asian companies where a controlling family or group holds voting control well in excess of their actual cash-flow (ownership) stake — a structure common across the CSE via pyramid and cross-holding arrangements. They found firm value rises with the controlling shareholder's cash-flow ownership (the incentive effect), but falls as the gap between control rights and cash-flow rights widens (the entrenchment effect) — because insiders who control more votes than capital they've put in have both the means and the motive to extract value at minority shareholders' expense.

This is the theoretical basis for treating governance structure — board composition, director overlap across related companies, and audit qualifications — as a distinct risk category alongside sentiment and RPT signals, rather than folding everything into a single generic "risk" number.

Claessens, S., Djankov, S., Fan, J. P. H., & Lang, L. H. P. (2002). Disentangling the Incentive and Entrenchment Effects of Large Shareholdings. The Journal of Finance, 57(6), 2741–2771. doi.org/10.1111/1540-6261.00511

Easley & O'Hara showed that the split between public and private information directly affects a company's cost of capital: investors demand a higher return to hold stocks where more of the relevant information is privately held rather than publicly disclosed. In a market like the CSE, where most listed companies have public free floats well under 25%, this asymmetry is structural rather than incidental — insiders control both the underlying information and the pace and framing of its disclosure.

This is why Nanayojana treats the mandatory annual report — not social media or retail sentiment, which the low free float makes an unreliable signal — as the primary evidence source: it's the one channel where disclosure is legally required regardless of whether insiders would prefer to stay quiet.

Easley, D., & O'Hara, M. (2004). Information and the Cost of Capital. The Journal of Finance, 59(4), 1553–1583. doi.org/10.1111/j.1540-6261.2004.00672.x

Two findings underpin the sentiment side of the score. Tetlock showed that pessimism measured in financial media predicts downward pressure on stock prices and earnings, later reverting — establishing that the tone of text about a company, independent of its stated facts, carries real predictive information. Loughran & McDonald then showed that generic sentiment word lists (built for psychology or general text) badly misclassify financial and legal language — words like "tax", "cost", or "liability" register as negative in a generic dictionary despite being routine, neutral terms in a 10-K.

Together, these motivate using a finance-domain sentiment model (FinBERT) rather than a generic one, and motivate the Sentiment Divergence signal specifically: if the optimistic, publicly-quoted Chairman's Statement and the technical, rarely-quoted notes to the financial statements diverge sharply in tone, that gap between the public narrative and the fine print is itself informative, in line with Tetlock's finding that tone carries signal beyond the stated facts.

Tetlock, P. C. (2007). Giving Content to Investor Sentiment: The Role of Media in the Stock Market. The Journal of Finance, 62(3), 1139–1168. doi.org/10.1111/j.1540-6261.2007.01232.x

Loughran, T., & McDonald, B. (2011). When Is a Liability Not a Liability? Textual Analysis, Dictionaries, and 10-Ks. The Journal of Finance, 66(1), 35–65. doi.org/10.1111/j.1540-6261.2010.01625.x

The theories above were developed largely from US, European, and broader East Asian evidence. Nazliben, Renneboog & Uduwalage's history of Sri Lankan corporate governance — tracing it from the colonial period through the post-civil-war era — documents the specific ownership and board structures (family-controlled conglomerates, cross-holdings, ethnically and socially concentrated boardrooms) that shape how tunneling risk actually manifests on the CSE, as distinct from a generic emerging-market template.

This is part of why Nanayojana's risk classification is sector- and business-group-aware (see the tunneling-risk classification methodology) rather than applying one uniform threshold across every listed company.

Nazliben, K. K., Renneboog, L., & Uduwalage, E. (2024). Corporate governance from colonial Ceylon to post-civil war Sri Lanka. Journal of Management & Governance, 28(1), 265–335. doi.org/10.1007/s10997-023-09678-5