港股结构与策略——被错杀的盈利增长

For the first time, readers will find at the bottom of this post a shorter, more compact summary of the main article. This will be the format going forward. The summaries will be held to ~1500 words or less, generally with 1-2 visuals at most, and written to the New York Times standard reading level.

Jump down to the summary

Thank you to all subscribers for your invaluable feedback on ways to improve Cassandra Unchained.

Subscribe now

So begins the second post in the Hong Kong Stocks: Structure & Strategy series. I highly recommend reading the first post, VIEs: Vulnerability, Virtue & Value before continuing.

Every one of these companies is superficially understood by many yet truly understood by few, it seems.

Besides VIEs, there is another bug (or feature) common to each of these stocks that must be understood. For this understanding, charts are necessary.

As mentioned in Part 1, Tencent was one of the very best stocks coming out of the mid-2000s as the stock price compounded at a ~35% annual rate through today. But the experience for some time now has been less enthusiastic. Tencent stock is roughly flat since November 2020.

[

](https://substackcdn.com/image/fetch/$s_!JZ9c!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F92c5a859-4bfc-4800-890c-e4b3395dd319_1804x1283.png)

This not uncommon. Baidu IPO’d August 5, 2005 on the NASDAQ, about a year after Tencent listed in Hong Kong. Many saw it, at the time, as the Google of China – Google was relatively young then too. Baidu had a stunning run from IPO to 2018, compounding at about 20% annually. There was an ephemeral meme-fueled resurgent spike in the early 2021 frenzy, but Baidu stock today trades where it did in 2010.

[

](https://substackcdn.com/image/fetch/$s_!cARa!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F11b281a1-713b-46aa-96ba-72873144848f_1805x1293.png)

Alibaba didn’t go public until 2014. Without the benefit of that mid-2000s start, Alibaba has returned only 4% annually. At the 2024 lows, Alibaba stock traded below its IPO price a decade earlier and today trades where it did in 2017.

[

](https://substackcdn.com/image/fetch/$s_!lM5R!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff1685074-c119-45df-b80b-dced694eba08_1798x1262.png)

JD.com went public in 2014 too. Since the IPO, the stock has only compounded at a 4% annual return and is not very far from its all-time lows.

[

](https://substackcdn.com/image/fetch/$s_!7fnJ!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fac2d0a25-6ed2-4ff8-b365-2685255e8a7e_1798x1276.png)

Trip.com, a Chinese online travel agency, went public in 2003 on the NASDAQ as Ctrip.com, a little earlier even than Tencent. Trip.com stock has compounded at 19.0% annually since that 2003 IPO. COVID, as we know, introduced significant volatility into the travel industry.

[

](https://substackcdn.com/image/fetch/$s_!cGVN!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9466bf60-71b3-4d86-bb41-2adf48880a67_1075x978.png)

Kuaishou is the #2 video-based social media company but did not go public until 2021. A rough start, Kuaishou is down 78%, a negative 26% annualized return.

[

](https://substackcdn.com/image/fetch/$s_!Gyrp!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4c5cea2a-1279-425f-b5b9-e0311c818205_1803x1285.png)

Meituan, China’s dominant food-and-other delivery company, went public in 2018. Meituan stock has fallen ~80% from its 2021 peak and has only compounded at a 5.5% annual rate since the IPO.

[

](https://substackcdn.com/image/fetch/$s_!aJvh!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe633b512-8c78-4462-b61a-ec620beb82a1_1802x1269.png)

Haidilao trades ~7% lower than its 2018 IPO price, down 81% from its 2021 peak. COVID lockdowns were hard on China’s restaurant industry, and consumer spending has been in check since.

[

](https://substackcdn.com/image/fetch/$s_!dVst!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcf4139b2-bdbb-409e-8e35-460e4d23a020_1807x1283.png)

So, what is happening here?

That green line is revenue. Take a look again at each chart. The reason for the terrible performance in no case involved a failing business. Quite the opposite. In almost every case revenue kept growing at a very healthy rate. For most, that revenue growth even powered through COVID, and the horrific Chinese lockdown.

What happened in 2021, and how come these stocks have been under so much pressure ever since?

Unfortunately, the explanation is multi-factorial. So, a list.

In October 2020, Alibaba Founder Jack Ma criticized the Chinese government for stifling innovation. Within days, Alibaba’s IPO of Ant Group, its financial arm, was scuttled. Thus began a two-year regulatory crackdown that focused on big tech companies and spanned gaming (Tencent), education technology (VIEs were not allowed for a set of education companies that wanted to go public), food delivery, e-commerce, and ride-hailing (Didi). Beijing exercised its anger month upon month and year upon year, though there has been change of late.

November 2021 was the month bubbles around the world popped. NFTs, crypto, watches, meme stocks, SPACs, you name it. Hong Kong stocks had lightly participated in the folly and were vulnerable as risk-off sentiment swept the world. While stock markets in the U.S. and other countries bottomed in 2022, China’s lockdowns were just getting started.

The Chinese property market bust alone would have created China’s version of the GFC, but in some ways it was even worse. The property sector was 30% of China’s GDP, 40% of local government revenue, and 2/3 of household assets at its peak. All far greater than the U.S. prior to the GFC. Property was the savings/investment account for the entire urban population of China. In August 2020, Beijing capped leverage ratios at the property developers to rein in speculation. Instead, the regulatory action revealed a Ponzi scheme of debt in the property sector. Evergrande’s liabilities of $335 billion were ~3x its annual revenue. Evergrande and 50+ other developers collapsed over the next few years. Household wealth destruction has been estimated at $5-$10 trillion. It is hard to fathom that much money lost in the U.S., let alone China. In fact, in the U.S., at the same time, the government was handing out $5-$10 trillion in forgivable loans, stimulus check and other gifts to its citizenry.

China’s COVID Zero policy led to lockdowns unlike any other country. Nightmarish scenes were filmed on phones and uploaded online. The crackdown/lockdown came in 2022, later than some other countries, but coinciding with just about everything else going wrong, as described above. Youth suicide rates doubled. In April 2022, retail sales fell 11% and industrial production crashed. Permanent damage included significant pay cuts for white collar and civil service employees alike. This triggered the already-high Chinese savings rate to spike higher still – household savings surged 81% in 2022 from 2021, directly stealing from consumption. Then, when China let up on the lockdowns, China endured an acute COVID pandemic anyway. Hundreds of millions of Chinese people caught COVID, and the hospital overload was yet another memorable nightmare that continues to suppress sentiment in the country and lead consumers to save rather than spend.

China continues to bully others in the South China Sea, and the risk of an invasion of Taiwan has hung heavy over foreign investors fearful of sanctions in the event of a China attack. Polymarket still predicts 30% chance of invasion by 2030, as discussed in Part 1.

So, China’s stock markets are suffering to a greater degree and for a longer time than the GFC in the United States. In fact, most of the stock charts above look like U.S. stocks after 1929 - at their lows, down 80% or more.

Still, the same charts show sales grew at almost every major company. A sign, maybe, that Chinese economic numbers are not as inflated, or as bad, as many say despite the appearance of a perfectly wretched storm.

The primary historical cause of major bear markets is multiple compression – collapsing PE, EV/EBITDA or EV/Sales ratios. Hong Kong’s multiple compression during the current bear market ranks with the worst in developed markets history.

Below, a comparison of multiple compression among major index bear markets from 1900 to present. Bear with me. The comparison chart below uses trailing PE ratio.

[

](https://substackcdn.com/image/fetch/$s_!GRQW!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F785237e5-53a7-4c13-a379-1e7cacbb14c5_1080x961.png)

Comparing crashes is difficult, especially across long periods. In this chart, the NASDAQ tech bust is the worst in history, with a very high PE ratio at the start. Generally, the chart reveals the worst multiple compressions start from the higher PE ratios. The Hang Seng Tech Index was up there too - irrational exuberance not so very long ago.

Below I will present another chart, this one using the Shiller Cyclically Adjusted PE (CAPE) ratio for the American indices. The Shiller methodology is better for comparisons across many decades, and good data on America’s great stock crashes span more than a century.

Shiller CAPE data exists for the S&P 500 and its historical analogues. Therefore, in the chart below, an S&P 500 Index analogue replaces the Dow Jones Industrial Index for the Great Crash of 1929. The chart also replaces the NASDAQ Composite Index with the S&P 500 Index for the Dot-com Bust of 2000. The S&P 500 Index is of course less volatile, and the Shiller methodology even more so.

Note, I say Dot-com here but as I have stated in The Cardinal Sign of a Bubble: Supply Side Gluttony, the 2000 NASDAQ crash was much more than “Dot-com.” Speculative lunacy encompassed epic overinvestment in data transmission, and it was this that gave the bust legs.

[

](https://substackcdn.com/image/fetch/$s_!JoKo!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4194bc4a-3a67-4d52-ba54-1c02d040936f_1080x1175.png)

This is interesting. Using the Shiller CAPE methodology instead of trailing PE ratios for the same study of American major bear markets, the Great Crash of 1929, represented here by an S&P 500 analogue, now takes the same place on the list that it occupies in our collective institutional memory - first.

The Hang Seng Tech Index crash remains #2 under this methodology.

Subscribe now

I present the final chart in this study below, and here I combines the two methodologies. The NASDAQ Dot-com Bust is back, as is the Dow Jones Great Crash of 1929, both calculated using trailing PE ratios. The CAPE methodology remains, and in fact, Shiller CAPE studies are added for the 1990s Nikkei crash as well as for the 1970s UK FT All-Share crash.

[

](https://substackcdn.com/image/fetch/$s_!Aguj!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdc4c59d6-a061-4a1c-af03-4e0566a8fc4c_1080x1653.png)

Students of bear market history will hopefully find the above juxtaposition of busts interesting. The NASDAQ’s fall from 2000-2002 ranks first, replacing the CAPE version of the Dow Jones Index Great Crash of 1929, again represented by the S&P 500 Index analogue.

The Hang Seng Tech Index slump, now in its 5th year, ranks second or third across all manifestations of this bear market multiple compression study.

Speaking of juxtaposition, the summary table below brings home the point of all this. Note the lonely green bar on the right.

[

](https://substackcdn.com/image/fetch/$s_!TuhY!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8fc541e5-5265-40e5-ba09-ba9948d4f28f_1747x911.png)

The Hang Seng Tech Index is the only one of the 11 major bear markets studied that has shown earnings growth during the stock market crash.

The Hang Seng Index – as opposed to the Hang Seng Tech Index – also ranks among the greatest bear markets under each methodology. This index is heavily weighted with financials, property firms and formerly state-owned enterprises. The PE compression there is from a lower level - 16 to 7, on trailing PE multiple – and not too surprisingly (for me) compares well with the 1929 Crash for the Dow Jones Industrials Index.

No other major index, major bear market in recorded history saw its constituent earnings grow through the crash, making the Hang Seng Tech Index collapse the only one that is due entirely to multiple compression, i.e. valuation and sentiment.

Of course, the Hang Seng Indices are also the only major bear markets currently in force. All of this mandates a hard, forensic look at Hong Kong stocks.

These will be deeper dives than I had originally conceived. The full picture of each company includes whether it is a VIE or not, the meaning of which deserves a refresh.

VIE Structure Review

I have previously assessed Hong Kong stocks’ relative value versus American technology and consumer stocks, and I explained the risks and realities of the VIE ownership structure often cited as the reason to avoid investing in Hong Kong.

To review briefly, I am comfortable investing in Hong Kong for a number of reasons, starting with the $200+ billion that Chinese mainlanders invest in VIE-structured Hong Kong stocks.

As well, the VIE structure is the bona fide structure of the entire company, with no separate “real” or “ true” shares. All shares of these major companies are shares in the Cayman offshore holding company that is the only gateway to ownership, even for Chinese mainlanders.

These are the largest tech and media companies in China, and they cannot be disrupted by Beijing without absolute chaos in Chinese society. The Sword of Damocles remains at Beijing’s hip, sheathed.

Finally, the U.S. is far more likely to act against VIEs than China, but the U.S. cannot touch the value of the VIE. Other than the Cayman Offshore Holding Company, the rest of the structure, including the VIE, is domiciled inside China. My detailed logic can be found in VIEs: Vulnerability, Virtue & Value.

With that, here we go.

Subscribe now

BYD (1211 HK)

BYD, a dominant electric vehicle (EV) and battery manufacturer based in China but expanding rapidly internationally, does not use the VIE structure. As BYD is not in a restricted industry, shareholders directly hold common shares. BYD trades its common shares, all with the same voting rights and the same dividend entitlement, in both Hong Kong and on the Mainland. The Hong Kong shares are called the H shares. The Mainland shares are called A Shares.

Founder Wang Chuanfu owns ~17%, and the founding team overall owns ~30%. Every one of these shares is the same exact share that you or I could buy. Berkshire Hathaway had a great return in BYD after a decade+ hold, but sold its last shares some time ago.

BYD is potentially the lowest-cost producer of vehicles in the world thanks to its vertically integrated model at scale. BYD controls its supply chains and saves margin at every step. It makes its chips, batteries, bodies, motors, etc.

BYD’s semiconductor arm is not limited to supplying auto companies. Baidu, Xiaomi, Huawei are customers as are industrial companies, renewables companies, and home appliance companies.

The Battery Innovation

BYD’s Blade battery is a leading-edge Lithium Iron Phosphate (LFP) chemistry battery. LFP batteries are chemically stable and do not release oxygen easily under thermal stress so do not easily burn in runaway fashion like Nickel-Manganese-Cobalt (NMC) batteries. LFP batteries are ~30% cheaper as well as safer, but pack ~30-40% less energy per kilogram than NMC batteries.

Across millions of vehicles, this is a significant cost advantage. Still, if NMC batteries have 50-75% more energy per kilogram, NMC might seem worth it. Tesla and others historically used NMC.

BYD’s Blade system eliminates the weight and space of the module to compensate. This innovative solution involves sliding long flat cells into a pack that acts as a structural part of the chassis. This increases its volume/energy efficiency by >50% according to BYD.

CATL is a direct battery competitor. It has an NMC battery system that, when paired to a chassis built for it, achieves 72% volumetric utilization and 255Wh/kg energy. This is all according to CATL, but seems favorable against BYD’s LFP new Blade system with 66% and 200Wh/kg.

CATL controls just under 45% of global EV battery share and is a pure third-party supplier that makes both chemistries, LFP and NMC.

BYD’s Day 1 platform design around the Blade system is comparable to Tesla’s Day 1 platform design around CATL’s battery system, but Tesla is a customer and negotiates away margin into the supply contract.

BYD not only captures that margin but sells the Blade batteries (branded FinDreams) to Toyota, Tesla (for its Megapack energy storage), Xiaomi, BorgWarner, LG and many more. BYD supplies all of these as standardized products without meaningful customization beyond dimensional variants.

The cell-to-body/chassis integration – that BYD retains for its own vehicles -minimizes the LFP-NMC energy differential and is a cost and performance advantage for BYD, which also retains the safety advantage by using LFP over NMC.

In fact, Tesla’s standard range Model 3 and Y are LFP, not NMC, in part because BYD legitimized LFP by innovating the cell-to-pack architecture that moved the industry in that direction. This is a global transition away from NMC.

CATL is still the dominant supplier. I will examine CATL (3750 HK) in more detail later, though I will note it listed in Hong Kong during May of 2025, and now trades ~40 times earnings, >5 times sales. Hold that thought.

BYD is a low-cost leader, a safety leader. Safety is especially important. Drive a nail into an NMC cell, it catches fire, runs away into neighboring cells, which catch fire, etc. Drive a nail into a Blade LFP cell, it gets a bit warm and damage is contained.

Financial Power

BYD showcases significant leverage with its third-party suppliers by carrying a significant float of payables (DPO is 100 days), which inflates operating cash flow. The dramatic drop in receivables days (~33 days sales outstanding or DSO) along with the expansion in payables days, as BYD moved from government to consumer/commercial, amounted to a ~$5-6 billion windfall and funded its global factory buildout to a significant degree. This may have been one of Berkshire’s key insights.

Note, BYD’s finance arm receivables, if included, would roughly double DSO, but these reflect loan asset receivables, not operations. The numbers above are focused on the manufacturing operations and the position of that business. The numbers in the chart below represent manufacturing, without finance included, as well.

[

](https://substackcdn.com/image/fetch/$s_!JXwd!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8af82254-161b-4448-be2a-12aa51180e3a_2324x1162.png)

The chart above shows BYD’s evolution over the last two decades.

2009-2013, batteries and handsets were the main products, and BYD expanded days payable outstanding (DPO) as BYD’s growing size gave it leverage over its suppliers.

2014, Beijing dramatically expanded EV subsidies, and pushed the government fleet to go EV. From 2014-2019, government purchasers sent BYD subsidy payments and payments for EV fleets - very slowly. Government receivables exploded faster than supplier leverage. Accounts receivable rose to 180 days sales outstanding (DSO), and BYD’s cash conversion cycle (CCC) rose to a rather toxic +144 days at the peak.

From 2020 on, BYD has focused on consumer/commercial sales, driving DSO to all-time lows and sending the cash conversion cycle negative the last three years. This is the working capital float that benefits BYD - the same ~$5-6 billion working capital release that continues to fund BYD’s international expansion, as mentioned above.

Now, BYD may face the reverse. In June 2025, an industry group pledged to reduce payables outstanding throughout the industry. If BYD complies, and it looks like it is unwinding its notes system with its suppliers, that could be a ~$10 billion one-time cash flow hit. Of course, it would hurt most Chinese car companies, but BYD has more to lose on a relative basis because the notes system it used is rather unique to BYD. Clarity on this could drop BYD’s stock, creating a potential buying opportunity. This is one of the negative catalysts I hope moves BYD stock to my preferred target buy price.

BYD expenses against operating income is an incredible 94% of R&D, and R&D is growing faster than revenue, so the R&D drag on earnings continues to grow. BYD’s margins appear like typical car companies at a little less than 5%. However, by only capitalizing 6% of R&D rather than the typical 30% or so at peers, BYD’s earnings sit about 25% lower than they would, apples to apples. For 2025 through the first three quarters, BYD’s total R&D nearly doubles BYD’s net income, $6.0 billion to $3.2 billion.

Overseas revenues are 29% of the total and are the growth driver. Factories are being built in Thailand, Brazil, Hungary, and Turkey. Europe put a 17% tariff on BYD cars, but 35% on other Chinese auto companies. Once local production is in gear, the tariff will disappear - a major future competitive advantage in Europe given BYD’s low cost advantage.

BYD has wisely not been targeting the United States. The China aversion here is bipartisan and helps Tesla keep its illusion going with U.S. investors.

BYD is facing increasing competition at home with over 125 Chinese EV brands. It has size and a significant durable cost advantage. BYD is profitable at its selling prices while most every major Chinese EV competitor is not. BYD’s R&D spend and lead is widening.

For now, the war of attrition continues – BYD’s 3Q2025 showed double digit quarter-on-quarter earnings declines in China. For the year, the decline is about 7.5%. BYD will retain its top position over the long run, and should remain profitable under any scenario. If BYD does not make money, competitors will go out of business, and BYD will improve its competitive position.

BYD has world-class disclosure, best in class with Tencent among Chinese companies I have examined, and no other Chinese company we are discussing in this series, other than CATL, is expanding overseas as fast as BYD is.

Geopolitical risk includes a pending 1260H designation. Strangely, BYD was added to the Pentagon’s Chinese military companies list in February 2026 – a list that was withdrawn just hours later. The Deputy Defense Secretary recommended BYD’s inclusion in October 2025, and that recommendation stands today. In time, the list may be published again, and BYD’s stock may take a hit. That would be a potential buying opportunity, as BYD has no Department of Defense/War exposure. Other stocks such as Tencent and CATL have bounced back from initial dips on 1260H inclusion.

As well, the EU is investigating subsidies for a Hungarian plant, and may yet cause trouble for BYD.

At ~HK$95/share, BYD’s ~$110 billion market capitalization is about 6.5x 2026 EV/EBITDA, 0.8x 2026 Revenue. BYD still shows net cash on the balance sheet despite the massive overseas spend.

Headline PE ratio is 17x 2026 earnings. Normalized R&D expense to the industry standard, the earnings ratio would be more like 13-14x. Top line growth is 9-10%, and may grow to 15% as investments come online.

Forward return on invested capital (ROIC) projects to 13-18% as those overseas factories kick in, but long-term auto businesses the last century have been terrible investments. They hang around well, but do not provide much return. I believe BYD is an exceptionally well-run and technologically advanced company today, and I believe long-term ROIC will settle above 10%.

BYD produced ~4.3 million cars in 2025, with over 1 million overseas, and operating leverage has really been ramping up since 2022. There will be years of improvement on that measure overseas.

BYD has ~29% more revenue than Tesla and ~1/12th the market cap. Tesla’s shareholders ask a lot of future robots. BYD sold 1.76 million battery electric vehicles (BEV) to Tesla’s 1.79 million.

I rate BYD a 7/10 at the current price. A rating of 7/10 is mild buy, small position, or wait for the 8/10 price, which is HK$75.

I would be an aggressive buyer at HK$75 or better, if the thesis still holds at that time. This implies a 10x normalized earnings multiple paired to a potential worst-case 10% long term ROIC. In addition to the investment case above and the excellent disclosure, that 10x/10% combination is potent in my experience, and leads to terrific upside, generally in the 15-20% annualized range. This would be because both the multiple and the ROIC overachieve in the future, if I have the investment thesis correct.

For a buy near the recent HK$95 price, if it were me I would shade the position light, say 4%, and leave room to double down aggressively to 8% if it gets to HK$75 and the thesis is still intact. BYD reports earnings later this month and releases its annual report. I will read it with interest.

Subscribe now

Haidilao International (6862 HK)

Founded by Zhang Yong in 1994 in Sichuan province, Haidilao is China’s only national hot pot chain. Hot pot is a staple in China and involves diners at each table cooking food in a communal pot of spicy hot broth. This spicy Chongqing/old Sichuan hot pot variation is the dominant hot pot today. Chongqing/old Sichuan peppercorn and a special fermented chili bean paste are the foundations for the characteristic flavor and spiciness.

Haidilao holds 2.2% of that market, almost three times the share of the nearest competitor at 0.8%. The secular growth runway for Haidilao’s offering is rather long. Haidilao is known for its service above all else. More on that later.

Haidilao International, the listed company is a Cayman Islands successor corporation founded in 2015. This Cayman Islands corporation is the full company, which owns all its subsidiaries. It is not to be confused with a VIE structure, which uses a Cayman holding company that is an empty shell, as was discussed in Part 1.

By buying the common stock 6862 HK in Hong Kong, one acquires pure common share ownership in China’s largest, and only national, hot pot chain, incorporated in the Cayman Islands.

Regarding structure, investors should know that the original predecessor company, Sichuan Haidilao Catering, grants Haidilao International royalty-free use of the brand, 海底撈, and is ~75% controlled by Haidilao International’s 70% owner and founder Zhang Yong.

海底撈 seems to mean “scooping from the bottom of the sea,” a phrase which fits hot pot rather well.

I can think of no reason why the founder, who controls ~70% of 6862 HK, would take the brand away from the operation. This structure existed long before the IPO, and is something of a vestigial tail. Harmless. I find nothing in this arrangement to worry about.

The Mistake?

Haidilao International is a successful, good Chinese company that did a very bad thing. This chart take us through it.

[

](https://substackcdn.com/image/fetch/$s_!9Fek!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F62d5ffc3-2a35-43dd-a624-3f9a9a623b6c_2324x1216.png)

2015-2017, Haidilao’s pre-IPO days were the picture of balanced, healthy growth. Free cash flow was invested back in the business.

2018, Haidilao IPO’d in Hong Kong, and management chucked the proceeds into a dramatic expansion, fatefully opening 540 restaurants during the first COVID year, 2020-2021.

Property, Plant & Equipment (PP&E) peaked at CNY 19.6 billion in December 2020.

By 2022, China was locking down hard, and Haidilao had written off 68% of its PP&E. Haidilao closed three hundred locations to stabilize the business. It was as much a dramatic mistake as it was catastrophic bad luck. It had ramped up to expand, and mistakenly kept the plan when COVID hit.

Also in 2022, Haidilao spun off its promising international operations as part of its restructuring. The spin-off, Super Hi, trades in Hong Kong with symbol 9658 HK.

Through 2023 and into 2024, the stock remained under pressure even as free cash flow ramped up fast and capital expenditures shrunk. For 2024, Haidilao generated ~$870 million in free cash flow despite a tough consumer spending environment.

During 2024, management paid out about 94% of its net income as a dividend to show confidence. At recent prices, that is about a 6% dividend yield, with no financial debt. Essentially every restaurant is leased, and Haidilao has termination options on most of them. This is how it was able to close hundreds of restaurants quickly.

2025 full year numbers are not out. We know that there was revenue pressure due to a weak Chinese consumer. On the other hand, food delivery wars of late may have given Haidilao a boost – more on that below.

Subscribe now

Top Hot Pot

Haidilao’s operating business benefits from a rather Asian manner of vertical integration. A good part of Haidilao’s needs (construction, seasonings, logistics) are met by an ecosystem of sister companies linked to Haidilao’s founder. This amounts to ~40% of its costs through connected parties. Related-party transactions (RPTs) are of course typically a negative in securities analysis.

[

](https://substackcdn.com/image/fetch/$s_!4Fnx!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd5a60e0e-199e-4e06-845e-05a4341f0329_640x245.png)

What is interesting here is the RPTs on balance seem to help Haidilao’s operating business. For example, Yihai International (1579 HK), the supplier of seasonings also controlled by Haidilao’s founder, has seen margins on its sales compress over time. That is not to say Yihai is suppressing its margins to help Haidilao, but it certainly does not appear to be gouging Haidilao. More on this in a bit.

China’s domestic market is full of counterfeit items and fake labeling, even in the food chain. One of JD.com’s competitive advantages, for example, is the wide array of trusted, authentic branded merchandise offered to customers. Haidilao similarly says this integration is the key to food safety, cheap construction costs, controlled logistics costs, etc.

Since the founder controls both sides, he could absolutely toggle margins one way or another. For now the data seems to advantage Haidilao. The founder controls ~70% of Haidilao shares, so even self-interest falls on Haidilao’s side. He owns, for instance, just ~35% of Yihai. It is also possible, one would hope, he is just being fair to both sides, which is a big advantage in China if the quality of the product is guaranteed.

Overall, these RPTs seem to give Haidilao a cost and quality advantage in some aspects, which can be important anywhere, but especially in China. It may even be necessary to scale.

Nevertheless, Haidilao’s big mistake – its doomed COVID buildout – benefited the founder’s brother’s construction firm Beijing Shuyun Dongfang. This is a private company with no public financial disclosure, and Shuyun Dongfang indeed built those 540 Haidilao restaurants very fast, receiving ~$620 million over 2020-2021, despite COVID and just before China’s lockdown.

Looking at other public competitors, the costs per restaurant seem roughly in line, which is heartening. The rushed nature of the expansion is suspect even so. An egregious grift, expensive mistake, or some parts of both.

Still, accounting for disclosed payments to related parties is good practice. In general, disclosure has been very good. The payments to the brother’s construction company, however, were not found with the others in the Annual Reports but rather in the Director’s Report. All such payments are collected in the table below.

[

](https://substackcdn.com/image/fetch/$s_!9Lqg!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F194faa01-fca0-42c1-8513-9fa005453954_1844x811.png)

The cash flows to related parties as a percentage of revenue has come down significantly from the 2020-2021 restaurant buildout. The founder’s brother’s construction/logistics company received significantly less in following years. If this was a wayward grift, the restructuring and closing of restaurants self-corrected. Such would not excuse the grift, however, and this is something that hurts the valuation a bit and requires monitoring.

I will watch the 2025 annual report at the end of this month for more color, and whether there is a resumption of payment disclosure to Yihai and Wehai. 2024’s related party transactions as a percentage of revenue is likely still around 12%, or higher. I note that real services and goods are provided – the question is the pricing relative to an arms-length relationship.

Related party transaction are not automatically negative, but should be evaluated in an even-handed manner. In this case, I do not see evidence for abuse of shareholders in the disclosed payments and engagements with related parties. Companies like this – and there are a ton of them throughout Asian markets – must be broken down like I did above. Every payment and conflict identified. To monitor well, one must fully account for each in the first place.

The company makes a significant effort to train its employees for consistent, excellent service and table experiences across 1300 restaurants. Many Chinese restaurants do not prioritize service, so design of the environment and service can set a restaurant apart, and Haidilao seems to succeed at that. Haidilao is known in China for its standardized product and for its extra service – such as 24 hour service, servers tying hair back before you eat or wrapping your phone in plastic. Staff replace with fresh hot towels quickly. If waiting, there are free drinks and snacks, and even free manicures, shoe shines.

Haidilao in some respects is like mixing a luxury hotel with a hot pot restaurant. It seems to have worked, and become part of the customer goodwill/brand. Customer-first excellence in work culture in the restaurant business is often admired but rarely replicated.

The food delivery war on which Meituan, JD & Alibaba spent ~100 billion yuan (~$14 billion, about 30% of Meituan’s total annual revenue and ~3x JD’s annual profit!) for the privilege of destroying each other’s margins is actually beneficial to restaurants in general. Hot pot is 14% of restaurants in China. Daily delivery orders exploded from 100 million to 250 million. I will say it again, 250 million daily delivery orders. Wow.

When JD offered merchants zero commissions to recruit premium chains, Haidilao and Starbucks were named as targets. That means these companies acquired customers on JD’s dime. This is a reason for Haidilao rallying late last year – other restaurant chains rallied as well for this reason.

A Fudan University study found that delivery wars increase food delivery but at the expense of dine-in volume.

When the cannibalization of dine-in ends, Haidilao will likely have more customers writ large. Some, probably a good portion, will become dine-in customers. Haidilao still paid nothing for it.

Better, Haidilao, as a leading national restaurant, was particularly highlighted in advertisements placed by the delivery warriors.

In this regard, the national chains got the most expensive exposure all bought and paid for by the delivery companies and will therefore lengthen their lead on regional and smaller competitors. The coming annual report should shed a lot of light on how this dynamic is developing.

Subscribe now

Pot of Gold

Haidilao gets paid up front but has significant payables to suppliers. During COVID, Haidilao expanded payables to a remarkable 57 days outstanding (DPO).

Could this be evidence that the founder uses the sister companies to help Haidilao? Yihai reports significant trade receivables from related parties. So the founder may not only toggle costs, but also cash flow, and in this regard, there is no evidence Haidilao is disadvantaged. Quite the opposite.

[

](https://substackcdn.com/image/fetch/$s_!muqd!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F96037661-b1fb-4072-bfb9-accb42df5cda_2337x1428.png)

Haidilao collects cash and credit card payments from diners immediately, and fresh ingredients turn over quickly, so the company holds only 21 days of inventory while taking 37 days to pay suppliers. Note, 37 days is not abnormal or an abuse of its suppliers – it is within the standard 30-45 days. Yet the business model generates a permanently negative cash conversion cycle.

Yihai and other connected parties may not even be helping Haidilao that much. It may simply be in the business model that suppliers – and in fact customers – create significant float and finances the entirety of Haidilao’s operations. Whichever way one slices it, Haidilao gets a cash flow boost, and the RPT concerns are essentially moot.

With its ~11 billion renminbi cash, Haidilao’s cash and cash flows are in great shape.

[

](https://substackcdn.com/image/fetch/$s_!9hky!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff3f583d4-1dfc-4a97-947a-7ea40ede4104_778x458.png)

The chart above shows Haidilao has a float of CNY 5 billion – an interest free permanent pool of capital that stands at almost 12% of revenue. The float will continue to grow with Haidilao, as it is part of the business model.

Note, prepaid gift cards are 20% of the float, and this has tripled since 2017.

Another wrinkle is that earnings are suppressed by a temporary depreciation overhang from the fateful restaurant buildout and reversal. Margins are lifting mechanically as that mistake rolls off the balance sheet through the income statement year by year.

[

](https://substackcdn.com/image/fetch/$s_!QHdD!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8c07708b-5a96-46cf-bb9a-6458dd015a74_794x458.png)

The chart shows the earnings tailwind is already happening, and will continue for a few more years.

Haidilao has started franchising its brand. It did not announce this, but in 2024, and now this past year, a recurring line item for proceeds from franchisees appeared in its filings. This is an economic opportunity but also a risk to culture and quality. The controlled supply chain should help maintain consistency on those factors.

Both the upside and the main risk is Haidilao’s status as a pure play on the Chinese consumer. For all the reasons already discussed, this is a dead end Japanification or a long-term hope/expectation depending on one’s view. Revenue declined to low single digits last year. They will pay a dividend.

The market cap is about $11.7 billion, and it trades at about 18x earnings, while cash flow runs higher – operating cash flow thanks to the negative cash conversion cycle is a bit less than $1 billion, and free cash flow ~$870 million. Net income margins at 9-10% are back to pre-expansion levels.

I like the company as a Chinese consumer recovery play. It is also a natural non-VIE pure China play that may benefit in the event of a U.S. attack on VIE structures.

The stock has been strong lately with other non-VIE stocks, perhaps validating this strategy. The company and stock are under no security-specific sanctions risk.

I rate Haidilao 8/10 at the current price ~HK$17, which is a buy for a normal position, 5-6%, depending on liquidity mostly. I would add more at HK$15 and bring it to an 8-10% position at HK$13 or lower, if the thesis still holds at that time.

Haidilao reports at the end of the month, March 31st, and releases its 2025 Annual Report, inside of which the weak China consumer faces off against Delivery Wars.

Subscribe now

PDD Holdings (PDD, formerly Pinduoduo)

Long before Temu, ex-Googler Colin Huang launched PDD, a VIE structured e-commerce company, currently the highest margin e-commerce company in China, with a remarkable financing strategy so extreme that short sellers and long investors alike thought it might be fraudulent.

In 2015, Colin used Tencent’s ubiquitous messaging app to group buyers to get low prices on perishable agricultural products for resale. PDD lost a lot of money doing that, so in 2017, Colin abandoned the strategy and became a pure web platform company arranging drop shipping from farmers to buyers. This drove gross margins from -47% to +56% in 3 quarters. An almost unbelievable pivot, clearly evident in the financials.

PDD would receive the order, arrange the drop ship, and then not pay its farmer suppliers for 500-600 days. That is no typo. Colin had a brand new lightly capitalized money-losing company, but even so, his/PDD’s leverage over its mostly rural and industrial suppliers was stunningly bold.

Next, PDD added industrial suppliers to the Pinduoduo platform – and still, remarkably, maintained payables outstanding at 500-600 days. PDD earned substantial investment income on some RMB 420 billion (nearly $60 billion) in payments held back from merchants and instead invested in short-term instruments.

July 2018, PDD went public, listing an ADR on the NASDAQ.

November 2018, short seller Blue Orca examined two of PDD’s China-based VIEs and put out a report that alleged total revenue per these regulatory filings only amounted to RMB 1.246 billion while PDD’s SEC prospectus reported revenues at RMB 1.951 billion, a potential 57% overstatement. Blue Orca also alleged the operating losses shown in PDD’s SAIC filing were 65% greater than losses filed with the SEC.

The payables float likely explains this. PDD apparently ran its cash inflows through a payment processor entity/VIE into restricted accounts for distribution to merchants 500+ days later. The cash holdings of PDD during Q2 2018 – the quarter Blue Orca examined – were an incredible 7x that quarter’s revenue, and were likely earning enough investment revenue and income to fill the gaps identified by Blue Orca.

The stock jumped 12% the day the short report came out. Perhaps investors saw the short report, investigated themselves, correctly identified the three VIEs or two VIEs and the payments processor, and were impressed. SAIC reports are no longer available for review by investors, so I cannot analyze those at this time.

Speaking of analysis, Blue Orca was not wrong to sniff around. PDD’s disclosure is subpar at best. PDD reports revenue with no breakdown – no country or segment details. The last period PDD reported segment revenue numbers was Q2 of 2018 – two, Merchandise Sales and Online Marketplace Sales. A “Transaction Services” line showed up only in 2019 and revealed 822.9 million renminbi revenue – this was likely the payment processor VIE. Ever since, no further information on revenue mix.

Hark back to Jack Ma’s single-handed transfer of Alipay’s assets from Alibaba’s VIE. PDD’s ephemeral disclosure tells me that “Transaction Services” is likely a particularly important part of PDD. It may process the income on the payables float. In the dark, in a poorly disclosed VIE perhaps, in Beijing.

When facing quirks in disclosure, it is important to do a comparative forensic analysis against similar companies. I do this by first reviewing PDD for red flags, then checking comps for those same issues. The table below summarizes this analysis for PDD, with JD.com and Alibaba as the comps.

[

](https://substackcdn.com/image/fetch/$s_!hxUf!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fccf49fb4-e65a-4ca4-951d-4730f281ea0e_1400x2095.png)

Per the table above, I cannot plant a Red Flag in the ground and proclaim fraud. Blue Orca likely did not examine one of three VIEs, or overlooked the payments processing arm as a source of revenue and cash flow. Nothing else in my comparative forensic analysis rises to an allegation of fraud. Rather, poor disclosure.

PDD’s problem with disclosure is vexing. Its unorthodox bootstrapped success almost seems to have been too exponential for adequate disclosure, if that is even possible. I hope this is not a reflection on the adequacy of internal controls.

The chart below shows interest income on payables outstanding was, as late as Q2 2021, the sum total of PDD’s profit. The pre-Temu payables float is epic.

[

](https://substackcdn.com/image/fetch/$s_!h8YU!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7f975d1a-c8f5-4b14-8942-c6c6f766560b_813x451.png)

Colin resigned from PDD completely in 2021, and September of 2022, PDD launched Temu, an international consumer-focused white-label drop-shipping business. With Temu’s launce, days payable immediately came down substantially before stabilizing from Q3 2023 through to today at what remains a significantly high level. By 2024, Temu was the most downloaded shopping app in the United States and operated in 70 countries.

Revenue grew in the 100% range into early 2024.

Then competition fought back. JD.com took its marketplace merchant count up 200%, and Alibaba launched Taobao Deals in tertiary markets that had been PDD’s owned territory.

[

](https://substackcdn.com/image/fetch/$s_!Sp8x!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2d0df863-09d0-4b14-bc1d-cffbccb862d0_3221x1281.png)

The chart above tracks the twin towers of PDD’s revenue growth, first Pinduoduo domestically then Temu internationally. The troubles of course were not limited to competition in China.

The U.S. in 2025 eliminated a de minimus loophole Temu had used to ship without tariff into the U.S.. Temu pivoted to shipping bulk to U.S. based warehouses, but Temu’s cheap limitless SKU advantage over Walmart and Amazon evaporated. These two had extensive physical logistics expertise that Temu simply does not have.

PDD barely grew EPS in 2025 on ~23% revenue growth, as margins took a hit. Gross margins peaked at 75% in 2022 but have fallen to the mid 50% range.

PDD however is not going away, not yet. The Wall Street outlook for 2026 is unclear – the expectation is revenue growth continues to fall in the face of competition (to ~10-15%), and earnings will potentially track that growth, but only if PDD can keep margins from compressing more. PDD’s gross margins are about 20 points lower than 2022’s 75% level.

I am not optimistic PDD’s earnings growth will hit double digits in 2026. I would acknowledge some value resides in the probability of Founder Colin Huang doing something impossibly great again. But he has not been involved as CEO or Chairman since 2021.

So, PDD has a $155 billion market cap, and $48B of cash – though we know it owes merchants a bit over $30 billion. So free cash is ~$15 billion, and the float is ~$33 billion. Float has grown as PDD has grown, and this appears likely to continue. How and where this float is invested is not disclosed.

PDD is making ~1 of every 3 dollars of net income on the $33 billion of float plus free cash plus an investment portfolio, all totaling to near $60 billion. This is not internally consistent. Short-term interest rates of 1-1.5% in China. PDD filings cite the use of “fair value assets” which implies assets producing capital gains rather than simple interest. This investment income is real, I believe, but is likely to be more lumpy and opaque than a simple bank deposit would be. Perhaps not as liquid as well. PDD of course does not disclose how it makes that $4.9 billion in investment income - out of $14.8 billion total income for the trailing twelve months as of 3Q2025. Dang disclosure.

This is why the company trades apparently cheap. Disclosure. New management did not change the poor level of disclosure. I have a feeling it is all on the up and up, but I cannot get confident about that.

A Hong Kong listing looks likely soon given a recent auditor shift. PDD switched its auditor from the Beijing Ernst & Young office to Ernst & Young HK. This shift is effective for the 2025 fiscal year. Ernst & Young HK brings PDD’s financials and documents out of the PRC such that U.S. regulators can more easily inspect them.

Significantly, a Hong Kong listing, would allow Southbound investors, through StockConnect, to invest in PDD finally. These flows might potentially add another pillar of demand for the stock. I look forward to evaluating the degree to which mainland Chinese support PDD by purchasing its stock.

PDD is on the Pentagon’s 1260H list, which puts it on the radar for sanctions but means little in itself. This seems mostly a negotiating tactic for the U.S. when engaging China diplomatically.

I plan to look at PDD again at the time of the Hong Kong listing. Because of the poor disclosure, especially with respect to revenue components, Temu’s competitive and regulatory headwinds hit harder. I hope the HK listing and the new auditor in HK bring additional disclosure.

I rate PDD 6/10.

I do not have a buy target price on PDD at this time. If I thought a position in PDD was worthwhile at this price, it could not be a full position due to the disclosure issue. I would keep it around 2%, roughly 1/3 of my typical full position, until disclosure is satisfying to me.

By tackling BYD, then Haidilao, and now PDD, I proceeded from best disclosure to disclosure of significant related party dealings and now to very poor disclosure. The middle one is my current favorite of the three for many reasons. One of those reasons is that hiding white collar shenanigans is worse than allowing investors to consider said shenanigans within the context of the total analysis.

PDD is expected to report earnings March 19th, and I will re-evaluate at that time. I will stay on top of this one, and look forward to the annual report shortly after earnings are released.

Subscribe now

PDD, JD.com & Alibaba: Valuation Metrics

The PDD Holdings analysis above cannot offer proper perspective without a quick comparative valuation look at the three companies that dominate e-commerce in China.

[

](https://substackcdn.com/image/fetch/$s_!u4Ra!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff719484a-11d3-4a0a-a250-9d6406f34a76_761x949.png)

There at the bottom, the table above references the Foundations: The Tragic Algebra of Stock-Based Compensation post, which is worth a read when evaluating these companies.

From the top, consider the scale. All are now major companies with JD.com and Alibaba both having more than $100 billion worth of revenue, and PDD Holdings no slouch at $60 billion.

The businesses could not be more different, and profitability, the margins of each, drives that home.

All are excellent businesses from a Return on Invested Capital perspective if one keeps in mind the Chinese consumer is saving, saving, saving.

[

](https://substackcdn.com/image/fetch/$s_!jYnT!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Feae39e74-a5f9-4881-9491-c9b53b4f87b6_833x599.png)

When consumer sentiment improves, spending should follow. ROIC and profitability should explode for all three. Current adjusted ROIC ~10% at a Great Depression-like trough is not so bad at all.

PDD’s cash is, as explained above, comprised of about $33 billion float and $15 billion free cash. But despite the stock’s severe fall, there have been no share buybacks. This gets to PDD’s opacity – as the cash will continue accumulating, the use of it is a real question.

PDD may be the best business of the three, but that judgment will have to wait. I already know it has the worst disclosure.

Scorecard

As for the 500-word bullet target for the companies covered herein, I missed, and the reader has my apologies. In consequence, I also overpromised about the full content of this note. The information presented in this post is perhaps a decent sequel to the first, but clearly no finale.

Part 3 of Hong Kong Stocks: Structure & Strategy series will cover Alibaba, JD, and Meituan. Those three companies make a good trio for a well-bridged analysis of the “Delivery Wars.” Part 3 will also cover other companies mentioned in Part 1, including Tencent, and maybe a few others.

To conclude the series, I will construct and explain my ideal “hedged” all-long Hong Kong portfolio.

I hope you learned something!

Until Next Time!

Cassandra Unchained is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

---

Hong Kong Stocks: Structure & Strategy

Common Stocks & Uncommon Value, Abridged

When investing in general, especially when investing overseas, the nitty gritty accounting and ownership details may seem extraneous, tedious, boring, but these things are often foundational to the value of a stock.

If one is monitoring a portfolio by looking at daily quotes, one is doing it wrong.

Monitoring a portfolio for long-term supranormal returns requires having performed a thorough forensic breakdown on each portfolio company, as above. Only then will an intelligent investor know what monitoring means.

However, I completely understand the value in analyses that get right to the point.

I heard you, and below you will find the above article written to New York Times reading level standard, and heavily abridged, with one visual (20+ in the main post). This format can be adjusted as we go and I look forward to hearing feedback.

The Big Picture

The largest Hong Kong companies are widely known and poorly understood. These stocks have had 1929-style crashes while the businesses kept growing.

This has never happened in market history going back to 1900.

[

](https://substackcdn.com/image/fetch/$s_!GhkA!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcb657032-9b76-4cee-ae24-0f2813d7ae2c_1747x911.png)

Five factors explain most of it.

Jack Ma criticized Beijing’s tech regulation in October 2020. Within days Beijing cancelled Alibaba’s $35 billion IPO of its financial arm and followed with two years of punishing its tech giants. Seems now at an end.

A global bubble popped in November 2021. Stocks fell along with watch prices, meme stocks, SPACs, you name it.

China’s property market collapse wiped out $5 to $10 trillion in household wealth. Savings rates skyrocketed, hurting consumption.

When China’s brutal COVID lockdowns ended, a heartless pandemic then pushed the Chinese consumer over the edge. No stimulus checks in China.

China’s potential invasion of Taiwan hangs over foreign investors as a realistic risk lately.

China stocks fell on souring sentiment – investors paid less and less for each dollar of the company’s earnings per share (EPS) of stock. When sentiment reverses, stocks will move up as investors increasingly are willing to pay more dollars for each dollar of EPS.

BYD (1211 HK) Rating: 7/10. Buy Full Position at HK$75.

BYD is China’s dominant electric vehicle maker. BYD has a clean ownership structure as investors own the same common shares that founder Wang Chuanfu owns.

BYD is dominant in China and is expanding internationally with great success.

BYD’s edge comes from its vertical integration and manufacturing expertise.

There are two special advantages that BYD possesses.

**The In-House Battery

BYD makes its own Blade battery, which employs lithium iron phosphate chemistry (LFP). LFP batteries are about 30% cheaper to produce than higher-energy nickel manganese cobalt (NMC) batteries used by Tesla and Korean manufacturers - and far safer**.

For example, if you drive a nail into an NMC cell, it will catch fire. Drive a nail into a Blade cell and it gets warm.

BYD closes some of the energy gap by building its battery pack module directly into the car’s frame. BYD, which sells batteries to Toyota and others, licenses this design technology to no other. Tesla has a similar approach but pays a high price to an outside supplier. BYD does not charge itself and so is much lower cost.

LFP became the world’s most common electric car battery chemistry in 2025. BYD led the way.

**The Cash Float

BYD sold mostly to the Chinese government, which could take > 6 months to pay its bills. When BYD pivoted to consumers, the company collected ~40 billion yuan (~$5-6 billion) in cash** as the government paid its bills. The windfall funded overseas factory growth.

BYD spends about 25% more on research and development (R&D) than Tesla. This is also double BYD’s own profit, which is understated by BYD’s conservative accounting of R&D costs.

If one applied industry-standard accounting, EPS would track 25% higher and reduce BYD’s price-to-earnings (PE) ratio from ~18x to 13-14x. I do expect BYD’s return on invested capital to continue to fall, and I am not sure where it stops. I believe it could be higher than 10%, but not sure how much. Historically, auto companies have been terrible long term holds.

BYD has ~29% more revenue than Tesla but one-twelfth the stock market value.

I do not own BYD currently. The company reports March 23rd and will issue its 2025 Annual Report.

Haidilao (6862 HK) Rating: 8/10. Buy full position at HK$17 (current).

Founded by Zhang Yong in 1994, Haidilao is China’s only national hot pot chain. Hot pot involves diners at each table cooking food in a communal pot of hot broth. Haidilao holds only 2.2% of that market, almost three times the nearest competitor’s share at 0.8%. Room to grow.

Like BYD, Haidilao stockholders directly own the same stock shares that the founders do.

Worth noting, Haidilao licenses its brand name for free from a predecessor Haidilao company also controlled by founder Zhang Yong. The brand name is secure and should not be an investor worry.

More concerning, Haidilao paid the founder’s brother’s construction firm roughly $620 million to build, rapidly, all 540 restaurants that Haidilao opened 2020-2021. This ill-fated expansion, just before China heinously locked down the country in 2022, resulted in Haidilao closing 300 restaurants and writing off 2/3 of the money spent. An expensive grift, or mistake.

Payments to the construction firm collapsed to $37 million by 2024. Total spending with companies related to the founder fell from 23 percent of revenue in 2020 to 12 percent (still about $700 million) in 2023.

The buildout bust paradoxically helps earnings growth. All that spending on restaurants is a balance sheet asset, which falls in value (depreciates) every year. This reduced earnings. The amount of depreciation is now falling, giving earnings a mechanical push higher. Investors should always be aware why margins are rising.

I will monitor, recognizing that real goods and services are provided by these related parties and that in China it can be better to deal with trusted suppliers and contractors. The build cost per restaurant does appear in line with industry peers, and Haidilao’s related party seasonings company appears to treat Haidilao fairly.

Free cash flow in 2024 was roughly $870 million, 1.5-2x greater than earnings, a good sign. Market capitalization is $11.7 billion. Management paid out 94 percent of earnings as a dividend.

Haidilao has no debt apart from rent payments on its restaurant buildings.

The food delivery war between Meituan, Alibaba and JD.com meant ~100 billion yuan (~$14 billion, roughly 1/3 Meituan’s total revenue and 3x JD’s earnings) worth of delivery subsidies. At the delivery companies’ expense, Haidilao acquired many new customers. Haidilao likely converts many delivery customers to dine-in over time.

Reports earnings March 31st. The 2025 Annual Report will follow thereafter.

PDD Holdings (PDD) Rating: 6/10. No target buy price.

Ex-Googler Colin Huang founded the most profitable e-commerce business in China and did so with unbelievable cunning. For years, as money came in from customers, PDD would hold that money for 500 to 600 days before paying its suppliers. In fact, as recently as mid-2021, interest income on that held money (also known as “float”) was ~100% of PDD’s profit.

PDD’s suppliers funded the company’s first 6 years of existence – at no charge.

In 2022, PDD launched Temu, its international shopping app. By 2024, Temu was the most downloaded shopping app in the United States, operating in 70 countries. In 2025, the U.S. destroyed Temu’s growth by closing a loophole Temu used to ship packages to its U.S. customers tariff-free.

In China, competition increased too. JD.com added 200,000 suppliers, and Alibaba launched a discount service aimed directly at PDD’s customers. PDD barely grew EPS in 2025 on ~23% revenue growth. The outlook is roughly 10-15% growth of both revenues and earnings if it can stop declines in its profit margins. I do not believe that is likely.

PDD shows $48 billion in cash, ~$33 billion of which is float (that money received from customers but not paid out to suppliers), and PDD earns about 33% of its profit from interest on that. If Chinese interest rates fall, interest income shrinks. If merchants demand faster payment, the float shrinks. Either way, ~33% of PDD’s profit is fragile.

The bigger problem though is disclosure. PDD’s revenue has no breakdown by geography, product, or business line.

Blue Orca, a short-seller, published a fraud report in 2018. I believe Blue Orca missed a key part of PDD’s structure that would account for the discrepancies they identified. Lack of disclosure makes certainty impossible.

Similarly, I found two anomalies in PDD’s financial statements - large swings in line items that both reversed without any reason given. Disclosure is bad enough to leave one questioning the quality of stated earnings.

Overall, I cannot call it fraud. I can call it bad disclosure, the worst.

Colin Huang, the visionary that started this company, has not been involved with the company since 2021, before Temu.

PDD switched auditors to Ernst & Young’s Hong Kong office, a typical move before a Hong Kong listing. Perhaps an IPO will bring more disclosure.

PDD is on the Pentagon’s 1260H list, which means little.

PDD reports earnings on March 19th and will issue its 2025 Annual Report.

The Opportunity

I am only buying one of these three stocks at the moment. As I mentioned at the start, the Hang Seng Tech Index is the only major index in recorded history to have seen its companies’ revenue and earnings grow during a major collective crash in share prices.

When sentiment changes, it will pair with rising earnings and rising revenue, creating significant upside potential in the shares of Hong Kong-listed companies. I am committed to finding the right companies at the right price.

Analysis will continue in Part 3, which will cover Alibaba, JD.com, and Meituan (the delivery wars trio) as well as other Hong Kong stocks such as Tencent. The finale will be the construction of my ideal conflict-hedged Hong Kong long portfolio.

Subscribe now