Investing Classics: Zhang Lei & HillhouseLong-Termism and Its Limits

The economic bedrock of long-termism is compounding: time is the only variable that opens a chasm between a great business and a mediocre one. But there is a premise that is easily forgotten — only when a company's intrinsic value is itself growing is "holding long" the same as compounding; otherwise, holding long merely stretches a small mistake into a large one. Zhang Lei's long-termism is therefore always bound to a hidden condition: first pick a business that can keep creating value, and only then does holding it long make sense. It is a conclusion reached after selection, not a starting point to apply in advance.

Hillhouse's very origin was an extreme bet in long-termism. In 2005, Zhang Lei took roughly $20 million from the Yale endowment (run by David Swensen) to found Hillhouse, and concentrated most of it in the recently listed Tencent. At the time Tencent still earned mainly from QQ and SMS value-added services, with a market cap under $2 billion; over the following decade he essentially held throughout, watching it grow into a several-hundred-billion-dollar giant. That "understand it, then don't move" hold is the most persuasive footnote to Hillhouse's long-termism.

Long-termism's biggest trap is being used to gild being "trapped": you lose money, then say "I'm a long-term investor," dressing an unwillingness to admit error as faith. For it to hold, the premise is that the value is still growing; once the fundamental logic is falsified, "holding long" degenerates into a value trap. A true long-termist, paradoxically, needs an explicit exit condition — "under what circumstances do I concede I was wrong."

• Is my reason for "staying bullish long-term" that the business is improving, or just that I already bought and don't want to admit error?
• Can I write down "which signals would make me concede the long-term thesis is falsified"?
• For this decade-scale judgment, have its key assumptions quietly loosened?
• Is my "long term" an active choice, or a passive narrative adopted after getting trapped?

Value discovery (Graham-style) earns the one-off spread of "price reverting to value," capped by the size of the market's mispricing; value creation earns the money of "the value itself getting bigger," with, in theory, no ceiling. The divide is whether the investor genuinely gets into operations: using capital, talent and digitization to remake the enterprise, rather than waiting for someone else to bid it up. The bar is therefore high — value creation requires control, operating ability and long-term capital, and the vast majority of secondary-market investors have none of the three.

In 2017, Hillhouse, together with CDH, took the Hong Kong-listed footwear king Belle International private for about HK$53.1 billion (~US$6.8 billion) — the largest privatization in the history of the Hong Kong exchange at the time. Belle's core women's-shoe business was being battered by e-commerce and had stalled; the market treated it as a sunset business. Rather than wait for "value to revert," Hillhouse rebuilt the supply chain, installed a digital mid-office, and poured resources into the sports-retail arm; in 2019 it spun off that sports business, Topsports, in an IPO whose market value at one point exceeded the entire Belle take-private price. This is the template for "buy it, then get your hands dirty."

Value creation sounds more sophisticated, but it is also harder and more prone to failure. It needs control, and a team that can actually remake operations; ordinary investors who copy it will simply lose money. More to the point: even with Hillhouse at the helm, Belle's core women's-shoe business has still not truly revived — what was rescued was more the sportswear track. Value creation can improve a single business, but it cannot change an industry's larger tide.

• Does this investment earn the money of "market correction" (discovery) or of "the business growing" (creation)?
• If I'm counting on value creation, do I (or management) truly have the ability and control to remake this business?
• Am I mistaking a small "discovery-type" bargain for a large "creation-type" opportunity?
• Even if the transformation succeeds, will the industry's structural headwind cancel out the effort?

Traditional institutions divide labor by stage and stay walled off from one another: VC for the early stage, PE for growth, the secondary market for public shares. Hillhouse's "public–private linkage" breaks that line — it can be an early-stage shareholder, add to the same names in the secondary market, and use its network to make matches for portfolio companies. The upside is a compounding of information and resources: seeing more across stages, betting earlier. The cost is that it demands enormous scale and network, and brings conflicts of interest and information-wall problems — what you know in the private market cannot be traded on in the public one.

JD.com is the textbook of this playbook. In 2010, Liu Qiangdong came to Zhang Lei for a new funding round, hoping to raise $75 million; Zhang replied, "Either I give you $300 million, or nothing at all" — because he judged that a heavy model like building your own logistics would fail if underfunded. Hillhouse ultimately invested about $265 million. Four years later, in 2014, Hillhouse brokered Tencent's stake in JD, folding Tencent's e-commerce assets in and paving the way for the IPO. From the private bet, to brokering the integration, to the listing, one player ran through JD's entire growth.

The whole-lifecycle model sounds perfect, but it is a game only a handful of mega-institutions can afford, and is nearly impossible for individuals to replicate. Beware its dark side: straddling public and private creates conflicts of interest (who sets the price? is the information wall truly sealed?), and a highly concentrated bet means that if you're wrong, exposures across every stage take the hit at once. Scale and linkage are both a moat and an amplifier of risk.

• Is the "linkage" playbook I envy something that belongs to institution-grade resources I simply can't reach?
• When one player straddles public and private, am I, as a minority shareholder, structurally at an information disadvantage?
• With a concentrated, cross-stage heavy position, if I'm wrong, will my exposures be pierced all at once?
• Am I learning its "see early, see whole" mindset, or blindly imitating moves I can't reach?

Behind "betting on China" is a respectable structural judgment: consumption upgrade, an engineer dividend, and industrial digitization would bring decades of growth — a judgment largely validated from 2005 to 2020. But taken to the extreme, "going all in" becomes pressing a huge exposure onto one and the same macro and institutional variable — and policy and regulatory risk is precisely what you cannot diversify away by "buying a few more stocks." The stronger the conviction, the slower one reacts when it is challenged: between structural conviction and structural blindness there is often only a hair's breadth.

2021 was a brutal lesson. In July, the "double reduction" policy landed and K-12 after-school tutoring was zeroed out overnight; education stocks like TAL and New Oriental fell about 90% within months. In the same window, platform antitrust tightened — Alibaba was fined RMB 18.2 billion, and China's internet leaders JD, Tencent and Meituan came under collective pressure. Hillhouse held heavy positions in both education and internet platforms, taking a hit to net value and reputation at once. Long-termism wasn't wrong, and the growth logic of betting on China wasn't all wrong either — but when the risk comes from the institutional level and is highly concentrated, "long term" cannot save you from permanent exposure losses.

This is not "don't invest in China," but a reminder: any conviction to "go heavy on some nation's fortune or track" must leave room for "what if the variable I care about most suddenly reverses." Being structurally bullish isn't the error; making it a single wager with no exit is: is your return over-dependent on one macro variable you can neither predict nor hedge?

• Does my portfolio press too much exposure onto one and the same macro/policy variable?
• For my most core "structural conviction," how much room have I left for "it gets falsified"?
• Can I tell "business risk" from "institutional/policy risk" — the latter being something I simply cannot hedge?
• Where my conviction is strongest, am I in fact stress-testing it least?