Essay · 28 July 2026
Most families don't collapse when capital is lost. They collapse when the system that produces it breaks.
Of the more than fifty ethnic Chinese family dynasties that once dominated Thailand's economy before 1997, only four remained standing after that year's financial crisis, according to a figure cited by the Straits Times in 2001. The external shock was the same. What differed was how each family had been built underneath.
They lacked a kind of capacity that only shows up under stress, and is extremely hard to build once conditions turn.

Most families don’t collapse when capital is lost. They collapse when the system that produces it breaks
A Pattern That Repeats
Expansion tends to mask fragility. During long growth cycles, leverage, connections, and speed can look like real capability. You only see the difference when conditions tighten and those supports disappear.
As Morgan Housel has pointed out, risk often stays hidden until the moment it matters. It doesn't distribute evenly; it concentrates in the rare moments that matter most. Between two families with the same starting conditions and comparable competence, one surviving and one not says nothing about superior judgment. It reflects which layers of capital each had built before the storm, and which layers each had left exposed.
Risk stays hidden until the moment it matters.
The Family Capital Stack
Family wealth in Southeast Asia builds across five layers. When families blur these together, they tend to pour too much into financial structures while the layers underneath quietly weaken.
Financial capital is the most visible: cash, equity, credit lines. It's also the most exposed to a single shock, a currency float, a regime change, a market cycle.
Institutional capital sits closer to power: regulatory access, political proximity, preferential treatment. In Southeast Asia it can drive enormous growth, and unwind just as quickly, since it's fragile in direct proportion to how much it depends on who currently holds power.
Operational capital is the infrastructure a family controls directly: factories distribution networks, logistics. It survives shocks that financial and institutional capital don't, but only when it's embedded in something the surrounding economy cannot easily replace.

Relational capital determines who continues to transact when contracts stop working, who extends credit without collateral, who keeps ordering when a bank won't clear a payment, and who absorbs a shock alongside the family rather than walking away from it.
Cultural capital is what makes relational capital transferable across generations rather than dying with the person who built it. It's the reason a supplier still extends trust to a grandson who has done nothing yet to earn it individually, purely on the strength of a family name.
Only the first layer appears on a balance sheet. The others build more quietly, and matter most when the first one breaks.
Only one layer of capital is visible. The rest matters most when it breaks.
Liem Sioe Liong left Fujian for Java in 1938 and started out selling cloves to small traders, many of them shopkeepers other merchants had refused to extend credit to. That decision was an early deposit into relational capital, made decades before it would matter. By the 1990s, Salim Group had become Indonesia's largest conglomerate, and a significant share of its growth traced back to Liem's personal relationship with President Suharto, institutional capital, layered on top of the financial capital it enabled.
When the 1997-98 crisis brought Suharto down, institutional capital collapsed first, as the political relationship that had underwritten decades of preferential access disappeared with the regime. Financial capital followed, as leverage built on that access became unserviceable and Liem was forced to surrender his most valuable assets, including Bank Central Asia, to settle the group's debts.
Operational and relational capital did not collapse, because neither had ever depended on Suharto. The flour mills kept running. The distributors who had extended decades of mutual credit with Salim kept ordering. That continuity in cash flow gave Anthony Salim the capacity to rebuild. He didn't restore the institutional layer. He didn't need to. Indofood still leads the world in instant noodles today, built on the two layers that had never been exposed to Jakarta politics in the first place.
About 2,500 kilometers away, the Horungruang family in Thailand ran the same sequence with a different result. Also Chinese migrants, from Guangdong, they built a steel empire worth 200 billion baht largely on leveraged financial capital, dollar-denominated debt made cheap by a pegged currency.

When the baht floated in 1997, financial capital collapsed the same way it had for Salim. But the operational capital underneath it, steel mills and industrial plant, was fungible, and so was the demand behind it: steel is bought on price and specification, not on who is selling it. There was no relational layer to fall back on, because steel manufacturing had never required one. Tata Steel Thailand and G Steel absorbed the family's assets without needing anything the Horungruangs had built, because there was nothing embedded to absorb. More than a hundred of their companies were sold off or dissolved. One son, Pravit Horungruang, spent years in bankruptcy before being discharged in 2014, and returned to the steel industry in 2015 not as an owner but as an employee managing operations for someone else's company.
The difference came down to structure, not intent or effort. Salim retained operational and relational systems that kept functioning after financial capital collapsed, because those systems had never been separable from the specific relationships that built them. Horungruang's operational capital survived the collapse too, but as fungible infrastructure that any buyer with sufficient financial capital could absorb, because it had never been bound to anything beyond itself. What survived in one case was not the assets. It was the system that had produced them.
What survived was not the assets, but the system that had produced them
Economic historians have a name for the relational layer at scale: the Bamboo Network, hundreds of thousands of ethnic Chinese businesses across Southeast Asia held together not by notarized contracts but by shared dialect, kinship, and word-of-mouth credit that operates alongside formal courts, and at times proves more enforceable than them.

Western commerce tends to build trust through contracts and institutions. The bamboo network builds it through reputation tested across generations, closer to guanxi than to any legal clause, where a reputation for keeping one's word functions almost like a form of currency, one that's hard to dilute. This is the specific mechanism Liem was building the day he extended credit to shopkeepers everyone else had refused. It is also, precisely, the layer a Singapore trust structure cannot manufacture on a family's behalf.
Trust, built over generations, functions like a form of capital.
The Singapore Paradox
Singapore had 400 single family offices receiving tax incentives in 2020. By the end of 2024, that number had passed 2,000, a fivefold jump in four years, according to the Monetary Authority of Singapore.
Singapore preserves financial capital efficiently. It does not recreate the conditions that produced it. In practice, the move often shifts capital from one form into another: financial capital, once embedded in operating businesses and relationships across the region, becomes abstracted into fund structures, professionally managed, geographically detached, financialized. Operational, relational, and cultural capital do not travel with it. They stay behind, tied to the people and places that built them, whether or not anyone is left to tend them.

What emerges is a mismatch across capital layers: a family's financial capital is fully modernized and institutionally protected while its operational, relational, and cultural capital, the layers that actually determine whether the estate survives a shock, are left in the jurisdiction the family just moved away from. For a long time, nothing looks missing on paper. Then the gap becomes visible, often all at once.
Financial capital can be preserved. The conditions that produced it cannot be recreated.
The Next-Gen Dilemma
A bulletproof trust deed is not the constraint. Whether the underlying layers are still being built is.
Most next-gen principals do not avoid the home market's operating environment by accident. They avoid it because families signal, through compensation, through status, through which roles get discussed at the dinner table, that finance is the senior track and operations is the junior one. A next-gen principal who can read a private bank's allocation report but cannot explain why a supplier in Jakarta or a distributor in Manila still extends credit on the family name has not been prepared to operate the system that sustains the estate. They have been prepared to hold its financial layer, because that was the layer the family taught them to respect.
They are prepared to hold capital, but not to operate the system that sustains it.
Families rarely lose everything when capital takes a hit. They lose it when the system that produces it is no longer under their control.
Financial capital is portable. The rest is not.
Singapore preserves capital efficiently. It does not recreate the conditions that produced it.
Financial capital is portable. The rest is not
Figures on ethnic Chinese families in Thailand after the 1997 crisis are drawn from a Straits Times report cited in January 2001. TheSalim Group account draws on ISEAS Publishing and The Jakarta Post. The Horungruang account draws on Bangkok Post (2019). Singapore family office figures come from the Monetary Authority of Singapore, as of end-2024.
— Dao Nguyen Anh