Why We Automatically Fail Every Chinese VIE, No Matter How Cheap

← All Articles

A company can clear every quantitative test our screen runs (a price-to-earnings ratio deep in Graham territory, a price-to-book below our ceiling, a current ratio above 1.5, a decade of stable earnings) and still receive an automatic FAIL. The trigger is structural, not financial. If the business is organized as a Variable Interest Entity, or VIE, we reject it before valuation ever enters the conversation. The reason is simple enough to state in a single sentence: with a VIE you are not buying the business, you are buying a contract that entitles you to the business, and those are not the same thing. What follows is what a VIE actually is, why it is a Graham problem rather than a China-politics opinion, and the five-year record of investors who learned the difference the expensive way.

You own a contract, not the company

The structure is worth understanding, because the whole argument rests on it. Chinese law restricts foreign ownership across many sectors, including much of technology and education, so the operating company cannot legally be owned by outside investors. The workaround is a piece of legal engineering and whoever devised it was too clever for their own good. A founder incorporates an offshore holding company, usually in the Cayman Islands, which controls a wholly foreign-owned enterprise inside China, which in turn signs a stack of contracts with the actual Chinese operating company and its nominee shareholders. Those contracts are designed to route the operating company's profits and control back up to the offshore shell.

Here is the part a value investor needs to understand. The shares that trade in New York are shares in the Cayman shell, and that shell owns no equity in the Chinese business. Your claim on the earnings is contractual, not proprietary. Two facts compound the point. The Chinese government has never formally approved the VIE structure, and the enforceability of these control contracts has never been clearly tested in Chinese courts (Investor.gov, Norton Rose Fulbright). The arrangement works right up until Beijing decides it doesn't.

Margin of safety assumes you own the asset

This is where the structure meets Graham directly, and why we treat it as methodology rather than a market call. Margin of safety is the distance between the price you pay and the value of what you own. If ownership itself is contingent on a contract that a sovereign has never endorsed and a court has never upheld, then the thing you are measuring against price is unstable at its foundation.

Our other screening rules discount for risk. A thinner current ratio, a shorter record of stable earnings, a richer multiple: each of these lowers a score. A VIE is a different kind of animal. The danger is not that the business goes bankrupt. It's that the legal claim to the business can be nullified outright, sending the equity toward zero regardless of how the company is actually performing. There is no honest discount rate for that. You cannot price a coin flip you are not permitted to observe, so we do not try. You would have better odds at the roulette table betting on red.

The record is not hypothetical

The last five years supplied the evidence for us to stand behind this decision with confidence. Consider DiDi. The ride-hailing company raised 4.4 billion dollars in a June 2021 New York listing. Days later, on July 2, China's cyberspace regulator opened a data-security investigation and ordered app stores to remove DiDi, choking off new users. The shares fell roughly 90 percent, and something on the order of 60 to 70 billion dollars of market value evaporated. By May 2022, shareholders had voted to delist from the New York Stock Exchange (Fortune, Bloomberg). The operating business didn't collapse. The state acted, and the equity followed it down.

The education sector told a broader version of the same story. In July 2021, China's "double reduction" policy banned for-profit tutoring in core school subjects effectively overnight. An industry estimated near 100 billion dollars was legislated out of existence, and the US-listed names built on it (New Oriental, TAL Education, Gaotu) lost the great majority of their value in a matter of days (CNBC, Class Central). These were profitable, growing companies whose entire revenue base disappeared, not because customers left but because a policy changed. The SEC's response was telling. It began requiring VIE issuers to disclose precisely this risk (WilmerHale).

Luckin Coffee shows the structure's second weakness, the distance it puts between a shareholder and the operating entity. From 2019 into early 2020, the company fabricated more than 300 million dollars in sales, overstating revenue by roughly 28 to 45 percent across two quarters. Nasdaq delisted its American Depositary Shares in July 2020, and the company later paid a 180 million dollar penalty to settle SEC accounting-fraud charges (SEC). Fraud is not unique to VIEs, but the layers between the New York share and the Chinese company make it harder to detect and harder to recover from.

Two more episodes round out the pattern. In November 2020, regulators suspended Ant Group's record 34.5 billion dollar IPO two days before it was due to price, after summoning the company's leadership (Forbes). And from 2021 into 2022, the entire cohort of Chinese ADSs faced forced delisting under the Holding Foreign Companies Accountable Act, until US regulators secured audit-inspection access in December 2022 and reset the clock (Orrick). The threat receded, but it did not disappear. The three-year compliance requirement simply restarted. The lesson across all of it is that the risk is structural and recurring, not the story of a single bad year.

We fail it on the screen so you never have to price it

The obvious objection is that several of these names looked genuinely cheap on the way down, and some looked cheap for years before. It's a trap. A quantitative screen rewards a VIE for the very cash flows a single policy stroke can erase, so the cheaper the stock looks, the more confidently the numbers mislead. The false signal grows with the apparent bargain.

So, we make the decision structural rather than discretionary. If the 20-F discloses a VIE, the company fails before we run the qualitative filing review and before valuation matters at all. That is a bright line by design. Bright lines have one great virtue: they cannot be talked out of in the moment a name looks too good to pass up. The rule applies the same way to every screen, every cycle, which is the entire point of having it.

The case against our own rule

Intellectual honesty requires stating what the rule costs us, because nothing is free and choices have consequences. Not every VIE blows up. Plenty of Chinese ADSs have delivered strong long-run returns, and a blanket FAIL means we will never own the winners among them. The case studies above are the failures by selection, and a fair reader should hold that survivorship in mind. The rationale also has a soft edge. If Beijing ever formally recognizes the VIE structure, or a Chinese court enforces these contracts cleanly, part of our argument weakens. At that point we may reconsider our rule.

We accept those false negatives on purpose. Missing good companies is the price of eliminating a specific tail risk we cannot price, and we would rather pay it. This is a values choice about capital preservation, consistent with Graham, and reasonable investors can weigh it differently. It comes down to the Buffett line we keep on our methodology page: "Rule number 1: Never lose money. Rule number 2: Never forget rule number 1." A rule that occasionally costs us a winner but never lets a state-voidable claim into the portfolio is that principle applied literally.

What this means for the next report you read

The rule, restated now with the evidence behind it: a VIE structure earns an automatic FAIL, on structural grounds, independent of any view on China's markets or its growth. When a strikingly cheap Chinese ADS turns up in the financial press, our screen has already set it aside, and now you know the reasoning rather than only the verdict.

Three things are worth watching, because any of them would change the analysis. Whether Beijing ever formally recognizes the VIE structure. Whether PCAOB audit access holds through its three-year test. And whether any Chinese court actually enforces contractual control when it is challenged. If those facts move, we will say so and revisit. Until then a bargain you cannot be certain you own is not a bargain.