Extreme DLOM & Control-Premium Cases, and Hong Kong–Mainland Method Divergence — Hong Kong Valuation Quality Series (Part 3)

On 17 August, I unveiled the full picture of this ten-year study for the first time. In the audience sat directors of listed companies, financial advisers, accounting firms, lawyers and other professionals, along with friends from various regulators, and the response was far warmer than I had expected. Yet on the platform and in the room alike, what made people pause was never the average, nor the trend, but the two ends of the distribution. How extreme are the most extreme cases, and what reasons were written into the reports?

That was exactly the trailer at the end of my previous column. This piece opens up the two ends of the distribution, one case at a time. The old rule first: every figure and every quotation below comes word for word from a public circular, and can be traced and checked. This set of data was compiled with the help of artificial intelligence, with my team and I performing sample checks, and the limitation was set out honestly in the previous piece, so I will not repeat it here. One more thing must be said first. This column states only the facts recorded in public documents, and makes no judgment on whether any valuer's approach is right or wrong. Every valuation differs in its subject, its purpose, its timing and its method, and the figures cannot be compared directly. The only question I want to ask is this: when a single number can push a valuation up by half, or cut it in half, do we not need a fuller and more rigorous explanation? Please read every case below with that question in mind.

Why must there be gatekeeping at all? Because these percentages reside in the independent valuation report, the very document that tells minority shareholders whether a transaction is "fair and reasonable". This is true of an ordinary transaction, and truer still of a connected one. When the counterparty is the controlling shareholder or its associate, the one thing minority shareholders can rely on before they vote is the fairness of that report. And a percentage twenty points higher or twenty points lower is enough to rewrite those four words, "fair and reasonable". Gatekeeping is what guards precisely this gate.

The ceiling of the discount: 56.8%

Over the ten years, the highest adopted discount for lack of marketability was 56.8%.

The case was a 2017 acquisition by EJE (Hong Kong) Holdings Limited (stock code 08101, formerly Jia Meng Holdings Limited), and the subject of the valuation was the entire equity of a bespoke furniture company in Guangzhou. The valuation report cited studies from Mainland M&A databases (Wind Info and CVSource), comparing the price-earnings gap between Mainland unlisted acquisition targets and their listed comparables, and then wrote: "the average discount from this study is 56.80%". Whatever the study's average was, the adopted discount was the same, taken at its original figure, with no further adjustment in the report for the subject's own circumstances.

A single discount, cutting more than half the value away.

Look next at two cases of 50%. In a 2021 acquisition by Tong Kee (Holding) Limited (08305), the valuation report first cited an American Stout study (mean 20.6%, median 15.8%), and then, on the grounds that the target company's "historical turnover and profitability… are relatively volatile, it is not unreasonable to have a high discount", adopted 50%. The median of the study it cited was 15.8%, and the figure adopted was more than three times that. A 2019 case at Mayer Holdings Limited (01116) was more direct still. The report first noted that there was no widely adopted empirical study for Mainland private companies, and after referring to the American Mergerstat, closed with a single line: "50% is considered reasonable for this valuation".

The floor of the discount: 2%

At the other end, the lowest adopted discount was 2%.

In 2016 Yuexiu Property (00123) sold a 49% stake in a Guangzhou project company, and the report wrote: "2% DLOM was considered to be sufficient in covering the expenses incurred to sell… transaction cost, commission, administration expenses, legal fee etc.". In this valuer's hands, "marketability" was the cost of selling the asset, the commission, the administrative expenses, the legal fee.

There was also 2.55%. In a 2024 acquisition by Chuan Holding Limited (01420) of a Singapore company holding industrial property, the report filtered the Pluris discount database down to the real estate investment trust (REIT) category and adopted its median of 2.55%.

For these two cases, once you understand that the valuation subjects were property companies, holding "heavy assets", a low discount does not look unreasonable.

One term, three meanings

Placing the two ends side by side, the finding I consider most worth the profession's reflection appears. This 28-fold gap is not, in fact, a single disagreement. It is three different concepts sharing one name.

The Mainland price-earnings gap studies read "marketability" as the whole valuation gap between unlisted and listed companies, and so gave 56.8%. The American restricted stock tradition read it as the discount of restricted shares relative to freely traded ones, and so gave 15.8% to 50%. The transaction-cost logic read it as the cost of selling the asset, and so gave 2%. Three readings, three orders of magnitude. A reader who sees only the six characters "discount for lack of marketability" will never know which concept was actually being priced.

And what decides the outcome is often a choice so quiet as to be almost silent: which research universe is opened. The same Pluris database, filtered to REITs, gives a median of 2.55%. Open a Mainland price-earnings gap study, and the mean is 56.8%. Both reports can honestly write "with reference to studies", yet the two differ by a factor of 22, and that gap was settled before any analysis began.

So the question for a gatekeeper has never been whether a study was cited, but rather why this study, and why this particular slice of it.

The two ends of the premium: 50.05% and 5%

Discounts have a ceiling and a floor, and control premiums have them too, with the distance between the two ends no less striking.

One listed company, both ceilings

The reader may already have noticed a detail that makes every listener pause for half a second when I mention it. The highest discount of the decade (56.8%, in 2017) and the highest control premium of the decade (50.05%, in 2019) come from two circulars of the same listed company, 08101, first named Jia Meng Holdings Limited and later renamed EJE (Hong Kong) Holdings Limited. The same valuation firm, the same industry, the same habit. Both extremes are the raw means of Mainland database studies, adopted at their original figures, with no adjustment for the subject's own circumstances visible in the report. One listed company, holding both ends of the whole market's ten-year distribution.

Hong Kong and the Mainland: a difference of method, not of scope

Having read the two ends of the distribution, there is a broader dividing line, the difference in approach between Hong Kong and Mainland valuers.

Let me clear up a misconception first. This is not a case of "Mainland firms only do property valuation". The data shows clearly that Mainland valuation firms take on a great deal of business and equity valuation. Counting business and equity valuation subjects alone, China United Assets Appraisal Group (中聯) has 51, 天健興業 (translated as Tianjian Xingye) has 44, and China Enterprise Appraisals (中企華) has 34. The scope is the same. The difference lies in method.

Mainland firms' conclusions are mostly reached by the income approach (discounted cash flow) or the asset-based approach, and their subjects are mostly controlling or wholly owned equity. The value this path arrives at is already at the control level. There is no minority interest to which a premium can be added, and by Mainland convention no separate discount for lack of marketability is applied either. Where discounts and premiums truly reside is the market approach, which starts from the minority interest price of listed comparables, first adds a control premium, then subtracts a discount for lack of marketability. And the data shows that Hong Kong and international valuers use the market approach in about 51% of their cases, while Mainland firms use it in only 11%. The same market, the same kind of transaction, and two method traditions, so that one side's reports are full of discounts and premiums, and the other's have them almost nowhere.

This brings out a gatekeeping question I consider worth the whole profession's discussion. Should whether an asset carries a discount or a premium turn so completely on which method the valuer chose?

There is a more hidden possibility still. A line with no discount does not mean there is no discounting effect. It can be tucked inside the discount rate, in the form of a company-specific risk premium. In fact, some Mainland appraisers explain it in exactly this way, that the liquidity consideration is already reflected in the discount rate. That account can be entirely valid, but whether it is valid should not rest on a single sentence. It should be settled by the data. The good news is that this research pipeline extracted parameters such as the discount rate and the size premium as well. A future extension of the study can test precisely this: for valuations that strike no discount, and Mainland appraisers' discounted cash flow valuations in particular, is the discount rate systematically raised? If so, that is the same discount that has merely changed its hiding place, and the reader should have the right to see it. If not, then it is a gap that genuinely lacks an explanation.

Thinking one layer deeper, there is a question I consider worth the whole profession, and the regulator, reflecting on together. Should Hong Kong's financial market allow foreign appraisers to carry out valuation work for the transactions of Hong Kong listed companies by their own standards and conventions? I must state first that this is not protectionism. Hong Kong is an open market, and that openness is precisely our strength. But business valuation has no statutory licensing regime in Hong Kong, so within one and the same market two parallel method traditions in fact operate at once. The same transaction, given to one tradition, is a page full of discounts and premiums, and given to another, need not mention them at all, and the difference lies not in the facts of the subject but in the valuer's school. When a market's "fair and reasonable" can be built on several mutually unconnected sets of standards, and the rules say nothing about it, is this the gatekeeping structure we want? I set no answer here. But at the least one thing can be asked: that every report state clearly which set of standards it follows and why they apply to this case, so that the person reading the report has something to compare, and a choice to make.

Within the ecosystem, no peer review

By this point, the reader may ask: with so many extremes and divergences, has no one seen this over the ten years?

That question is precisely where this study began. Behind every transaction stands a whole gatekeeping ecosystem, and an imposing one at that. The 297 transactions carrying a discount or premium were all approved by boards. There are close to a hundred valuation firm brands in the market issuing reports. Some 482 circulars named a reporting accountant, 314 carried an independent financial adviser's opinion, and a further 118 engaged another financial adviser, while every circular was reviewed by the Exchange before publication. Each post discharged its own duty. But no single post's duty was to place these 477 discounts and premiums together and read them across. Case by case, every figure has a source. Side by side, you see that 56.8% and 2% share one name, that a single company holds both ends, and that the choice of method decides whether a figure exists at all.

Gathering the evidence to this point, the conclusion is really just one line. There are many gatekeepers, and the one thing missing is "peer review". Academia has peer review, and medicine has peer review, yet this profession, whose signed figures routinely run into the hundreds of millions, has never had a mechanism for peers' figures to be set against one another. The market does not lack gatekeepers. What it lacks is a common ruler placed at every gatekeeper's hand. My team and I undertook this study precisely to supply that ruler, and we are glad, as an independent third party, to go on maintaining it.

Next time

The next piece, which also closes this series, turns to what this ruler, once in place, means for the regulator, for the profession, and for every gatekeeper who signs. How is consistency to be measured, how is transparency to be required, and where should the division of labour between artificial intelligence and people fall?

(Part three of the series “Gatekeeping Valuation Quality in Hong Kong.” To be continued.)