Disclose More, Float Less: Free-Float Measurement and MSCI's 2026 Review of Indonesia

Free float decides how much of a company an index holds and, through minimum thresholds, whether it is held at all. It is treated as a measurement, but it is an estimate that depends on the rule applied (which holders count as strategic), on the information available (what owners must disclose) and on estimation error; a minimum threshold turns these differences into decisions about which companies an index holds. Applying the rules of MSCI, FTSE Russell, S&P Dow Jones Indices and the Indonesia Stock Exchange to the shareholder register of 961 of the 962 companies listed in Indonesia, we find that the share of market value in companies with free float below 15% ranges from 27% to 41% across definitions, and companies holding nearly a fifth of the market's value pass under some rules and fail under others. The rules disagree most where more of a company is held in stakes between 1% and 5%, the band Indonesia began disclosing in March 2026. In a reconstruction that holds ownership fixed, seeing those holdings roughly doubles the share below 15% under the providers' rules, from about 17% to between 34% and 41%; this is an upper bound on the effect of the March 2026 disclosure reform. A simple error model implies that a two-point estimation error would leave some 40 companies on the wrong side of 15%. Free float is informative about trading but leaves most of its variation unexplained. Six of the eight companies MSCI removed from its standard index in 2026 remain in funds tracking other providers. These findings bear on MSCI's October 2026 proposal to move Indonesia to Standalone status. We show how much a fund could trade in MSCI's nine-stock index and in broader alternatives, and set out graded, company-level options. The evidence describes associations and does not establish causes. AlphaBlock Working Paper 2026-04. Disclosure: the author is the founder of AlphaBlock Technologies Inc., which develops systematic index methodologies and licenses them, including as model portfolios, to investment funds and institutions, some of which hold Indonesian equities. Use of AI tools: during the preparation of this work the author used Claude (Anthropic) to help write and check the analysis code, draft and edit text, and typeset the paper. The author designed the study, chose the data and methods, reviewed the code, results, citations and text, and takes full responsibility for the content.