Jakarta Is the World's Worst-Performing Market: What MSCI Found in Indonesia's Share Registers

On a single day in late January 2026, Prajogo Pangestu — Indonesia’s richest man, whose fortune runs to roughly US$29.9 billion and who sits at 84 on the Forbes list — lost about $9 billion.

No factory burned down, no regulator raided an office, and no customer went anywhere. What happened was that MSCI, a private index company headquartered in New York, published a warning about how it intended to count shares.

By the middle of 2026 the Jakarta Composite Index had fallen 34% in six months, and roughly 37% from its peak — the worst performance of any major equity index in the world, on Bloomberg’s reckoning. Foreign investors pulled some $3.4 billion out of Indonesian equities. The rupiah touched a record low beyond 17,500 to the dollar.

The proximate cause was technical. The underlying cause is the structure of Indonesian capitalism, and specifically a habit that thirty years of domestic regulation never disturbed.

What free float means, and why 7.5% was always a problem

A company’s free float is the proportion of its shares actually available for outsiders to buy and sell. Everything held by founders, families, holding companies and affiliates is not float, however impressive it looks in the market-capitalisation line.

Indonesia’s minimum was 7.5%. A family could list a company, sell a fourteenth of it, keep the rest, and enjoy a market valuation applied to the whole.

That arrangement has two consequences, and they compound. A thin float is cheap to move: relatively small buying can lift a share price a long way, and with it the paper value of the 92.5% nobody is trading. And a thin float is hard to leave, because there is no depth on the other side when a large holder wants out.

For a global index provider this is a specific, practical problem. MSCI exists to give passive investors replicable exposure to a market. If a stock carries a large index weight derived from a market capitalisation inflated by a sliver of tradeable stock, every passive fund tracking that index is compelled to buy something expensive and illiquid. The index stops describing the market and starts distorting it.

The rule change

MSCI’s adjustment sounds like bookkeeping and functions like a repricing. Rather than accept a company’s own account of its float, MSCI began taking the lower of the company’s reported figure and the data held by KSEI, Indonesia’s central securities depository, which classifies holdings booked as “corporate”, “others” and “scrip” as not genuinely floating.

The gap between those two numbers was the discovery. Shares that companies counted as public turned out, in depository records, to sit in vehicles that behaved like insiders.

The method was tested through 2025 and applied in the May 2026 index review. MSCI froze foreign inclusion factors, blocked new Indonesian additions, halted upward migration between size segments, and deleted 19 Indonesian companies from its Global Standard and Small Cap indexes without adding a single one. It also introduced a high-shareholding-concentration framework and signalled it would use Indonesia’s new disclosure data to refine float estimates further.

Passive money does not deliberate. Removal from an index is an instruction to sell, executed by machines, regardless of whether the underlying business is sound. One analysis put the mechanical outflow at around $2 billion; as much as $13 billion was reported at risk while the market waited on subsequent decisions.

Why no domestic pressure ever built

Here is the part that explains the decades of inaction, and it is a matter of incentives rather than negligence.

If a family holds 92.5% of a company, the share price is close to irrelevant to how that company is run. Nobody can accumulate a stake large enough to demand a board seat. No activist can force a strategic review. A depressed valuation costs the family nothing they intend to realise, because they are not selling. Listing served as a financing event and a prestige marker rather than as an accountability mechanism, and the market’s opinion could be received as commentary rather than instruction.

Domestic institutions had little appetite for the fight either. Indonesian pension and insurance money is a modest presence beside the groups themselves, and the exchange competes for listings rather than policing them. So the discipline that public markets are supposed to impose — the pressure to disclose, to widen ownership, to explain capital allocation — was simply never applied. The governance research house Glass Lewis, reviewing the episode, treated Indonesia’s stress as a case study in what thin floats and opaque ownership do to governance quality.

An index committee turned out to be the constituency that could compel a change no Indonesian regulator had managed. MSCI has no mandate over Indonesian corporate governance and no interest in it as a moral question. It simply declined to keep counting shares that were not really available, and in doing so priced control as a liability for the first time.

Jakarta’s answer is to adopt MSCI's demands as law

The official response has been swift and unusually candid about its motivation.

Indonesia’s financial regulator OJK, with the exchange, is raising the minimum free float from 7.5% to 15% — a doubling — with an exit policy for companies that fail to comply in time. The threshold for disclosing a major shareholding has been cut from 5% to 1%. A high-shareholding-concentration framework has been built jointly with the depository, explicitly citing low float, opaque ownership structures and indications of coordinated trading. The listing rules themselves have been rewritten under a new Regulation I-A, which law firms have been busy explaining to clients, and a pilot has been aimed at 49 companies representing some 90% of market capitalisation.

The exchange has also asked MSCI to apply its new rules fairly, which is the reasonable request of an institution that has discovered where authority actually sits.

Doubling the float requirement is a serious reform, and it will do something uncomfortable: to comply, controlling families must sell shares into a market that has just fallen by a third, or lose their listing. Prajogo Pangestu has already been reported selling stakes as ownership rules tightened. The reform is correct and its timing is punishing, which is what happens when a change is made under external pressure rather than in advance of it.

The irony in the new money

There is a final turn worth noting, because it complicates the easy reading that this is an old-economy problem.

The fortunes that grew fastest in Indonesia over the past three years came from the energy transition. Prajogo’s Barito Renewables operates 886MW of geothermal capacity, close to 38% of the national market, and accounts for more than a third of his wealth; it earned $53 million in the first quarter of 2026. The Lim family’s Harita processes Halmahera nickel into the sulfates that go into electric-vehicle batteries. Both are creatures of hilirisasi, the state’s downstreaming policy, which rewards whoever can finance a smelter or a power plant — and that is these groups.

So Indonesia’s green industrial policy generated a new cohort of enormous listed valuations, built on the same thin floats as the old ones, and it was those valuations that made the index problem acute. The transition did not modernise the ownership model. It scaled it.

What a foreign investor should take from this

Three things, none of them the headline.

First, a rich-list number in Indonesia is a statement about a share price applied to a large holding, not a measure of industrial strength. A fortune that can move $9 billion in a day on an index announcement was never $9 billion of anything solid.

Second, the reform is real and the direction is right. A 15% float, 1% disclosure and a concentration framework are the foundations of an investable market, and Indonesia is building them. It is doing so under duress, which means implementation will be contested and the timetable will slip, but the ratchet only turns one way.

Third, and most useful: for the next few years the question to ask about any Indonesian listed company is not what it earns but who owns the shares nobody trades. That figure now determines whether the stock is in an index, whether it can be exited, and whether its price means anything at all. For thirty years that question had no consequences. It has them now.