Guinea Value's Jingshu Zhang on $EDU
Jingshu Zhang sees New Oriental Education ($EDU) as a battle-tested compounder whose 2021 collapse obscures a stronger competitive position. From its 2006 IPO through July 2021, the stock compounded at roughly 28% annually before China’s “double reduction” policy drove a 95% drawdown; it has since rebounded about 350%. Zhang calls the crackdown the company’s “ultimate stress test” and believes the shares remain materially undervalued.
Zhang’s edge is operational: he has spent roughly 12 years in the industry and competed directly with New Oriental through his overseas-admissions consulting firm. His firm once served about 3% of the US-graduate-school study-abroad market, but he sold it after Trump’s election because the outlook had become “endless pain again.” While smaller competitors are now being “decimated,” EDU guides overseas consulting and services to decline only 4–5%—an outcome Zhang considers extremely impressive and evidence of exceptional resilience.
The crackdown may have converted political trauma into a structural moat. New school-running licenses are “rarely issued, if not at all,” advertising is heavily restricted, and EDU and TAL retain the strongest brands; meanwhile, China’s exam-driven culture remains intensely competitive. Parents simply moved toward scarcer, more expensive tutoring, leading regulators to tolerate some return of supply—so EDU’s profitability has already surpassed its pre-crackdown level even though its stock has not.
The valuation case rests on assets and cash generation rather than merely a low earnings multiple. Against a roughly $7.8 billion market capitalization, Zhang counted about $5 billion of cash, or $3 billion after deducting $2 billion of deferred revenue, plus a 57% East Buy stake worth approximately $3 billion. That implies roughly $1.8 billion for the remaining operations, which he expects to generate more than $500 million—closer to $550 million—of next-year free cash flow.
Andrew Walker’s central objection is that EDU may be “writing risk”: collecting years of steady returns before another regulatory or geopolitical axe falls. Zhang counters that 30–40 times earnings before the crackdown was riskier than today’s asset-backed valuation, because his primary definition of risk is “permanent capital loss.” EDU’s fortress balance sheet and negative-working-capital model materially reduce restructuring or bankruptcy risk, even if they cannot eliminate volatility.
East Buy’s agricultural livestreaming business was an accidental survival mechanism that became both valuable and strategically useful. Founder Michael Yu created alternative work rather than dismissing teachers after the crackdown; the platform then took off as TikTok and livestream commerce grew. Besides producing a publicly traded stake worth about $3 billion, East Buy lets EDU discuss books, life, agriculture and tourism—effectively sustaining brand awareness when direct education advertising is constrained.
AI improves EDU’s junior-high economics but also creates the episode’s clearest competitive risk. Adaptive devices use recorded teachers, backend support and decades of student data while eliminating classrooms, leases and much of the teaching cost; Walker responds that the same digitization lowers entry barriers and empowers superstar teachers to capture the economics themselves. Zhang concedes junior-high retention is only 60–70%, versus roughly 80% in senior high, making this “something that we need to keep a close eye on.”
1. Zhang’s operating history turns a weak headline into a share-gain signal
New Oriental was founded by Michael Yu in 1993 and listed on the New York Stock Exchange in 2006. Zhang calculates that an IPO investor holding through July 20, 2021 compounded at about 28% annually—until the regulatory shock erased 95% of the stock’s value. Its subsequent roughly 350% recovery supports his description of EDU as a “storied battleground stock,” but not yet, in his view, a fully valued one.
The 2021 collapse was not EDU’s first credibility test. Muddy Waters published a short report in 2012, wiping out about 35% of the market capitalization that day; Zhang says the resulting investigation found nothing wrong. Yu bought shares personally, the company repurchased stock, and employees received additional options—an episode Zhang offers as evidence that Yu is “one of the most ethical entrepreneurs” he has encountered globally.
Zhang’s edge comes from roughly 12 years in the industry. While completing an undergraduate degree at Cornell and a PhD at MIT—“poor, hungry, and driven,” as he jokes—he co-founded a business guiding Chinese students through exams, essays, the Common Application and US admissions. It eventually served about 3% of the US-graduate-school study-abroad market and competed directly with New Oriental.
That business also supplies Zhang’s variant data point. After COVID forced it to borrow for survival, he and his partners sold at a “severe discount” following Trump’s election; without the sale, he believes it would now be losing millions of dollars monthly. Against that backdrop, EDU’s guidance for only a 4–5% decline in overseas consulting and services looks less like weakness than exceptional resilience.
2. The 2021 crackdown constrained supply without changing parental demand
Zhang remembers seeing China’s “double reduction” Document 42 while quarantined for 14 days in Shenzhen in July 2021. The policy sought to reduce both student homework and after-school tutoring; in US premarket trading, EDU and TAL fell roughly 70% and 90%, respectively. Like banks after the global financial crisis, the surviving education companies retained a stigma long after the immediate event.
Walker’s challenge is that this stigma may be rational. EDU has faced a government crackdown that temporarily killed its core K–12 market, disruption to overseas education from visa and political changes, and an earlier short-seller attack. His “risk writing” analogy is the turkey fed safely for a thousand days before the axe falls: years of compounding do not prove the absence of a hidden terminal risk.
Zhang’s counterintuitive response is that the crackdown made EDU more defensible. Operators now require school-running licenses that are “rarely issued, if not at all,” while conspicuous advertising invites immediate punishment. With bus-stop education ads largely gone, incumbents EDU and TAL possess brand awareness that new entrants cannot readily purchase, creating tobacco-like restrictions on supply and customer acquisition.
Demand, meanwhile, survived because China’s bureaucratic and educational systems have rewarded examination performance for more than a thousand years. Borrowing anthropologist Clifford Geertz’s phrase, Zhang calls culture “a web of significance that we ourselves have spun”—and one that is extremely difficult to escape. Parents remained competitive, tutoring became more expensive, and regulators responded with “one eye open, another eye closed” as EDU rebuilt.
3. The balance sheet reframes volatility as survivability
Zhang does not equate a violent share-price move with fundamental risk. EDU traded around 30–40 times earnings before “double reduction,” which he considers genuinely dangerous; immediately afterward, the market effectively valued only the overseas business while assigning nothing to K–12 operations or balance-sheet cash. Perceived risk peaked just as valuation risk diminished.
His preferred definition, borrowed from value investing, is “permanent capital loss”: an investment entering restructuring or bankruptcy and destroying the principal. That makes EDU’s balance sheet central. At a share price around $48, Zhang estimated a market capitalization near $7.8 billion and approximately $5 billion of cash; deducting roughly $2 billion of deferred revenue still leaves about $3 billion of adjusted net cash.
The business naturally creates cash because customers pay before services are delivered. Zhang’s own company could collect roughly $10,000 from a freshman, provide services over three or four years, and pay consultants afterward. That negative-working-capital “float” can be invested in the interim; he notes that earning 10% on it could turn a 30% free-cash-flow margin into something closer to 40%.
Add EDU’s 57% holding in publicly traded East Buy, which Zhang valued at about $3 billion based on the prior close, and cash plus that stake total roughly $6 billion. Against the $7.8 billion equity value, the implied price for the remaining operations is about $1.8 billion. Zhang expects those operations to produce more than $500 million, perhaps $550 million, of free cash flow next year.
4. East Buy turned a layoff problem into an accidental second franchise
East Buy began after the crackdown left New Oriental with teachers it could no longer deploy conventionally. Yu did not want them forced back to rural hometowns without work, so the company experimented with livestream e-commerce, principally selling agricultural products. Zhang stresses that the outcome “was really unintentional”: management was trying to preserve jobs and salaries, not deliberately construct a multibillion-dollar public business.
As TikTok-style livestreaming surged in 2021–22, the livestream became extraordinarily popular and profitable. The apparent conglomerate drift therefore has an origin Walker initially missed: it was emergency labor redeployment that unexpectedly found product-market fit. Selling farm products and promoting local tourism also strengthened relationships with local governments by generating income for farmers and fiscal activity for smaller cities.
The education business receives a subtler benefit. Direct tutoring advertisements are restricted, but East Buy can advertise agricultural products while its hosts discuss books, life and related topics under the New Oriental umbrella. Zhang therefore sees genuine brand synergy: a seemingly unrelated commerce platform keeps EDU culturally visible without overtly advertising the regulated product.
5. Excess cash is both a drag and the scar tissue of two near-death events
Walker accepts the fortress balance sheet but argues that cash equal to roughly 60% of market capitalization can suppress shareholder returns. Even if EDU distributes 50–60% of net income, the cash pile may keep expanding. His preferred structure would preserve a smaller emergency reserve, rebuild it gradually when needed, and return the rest rather than permanently capitalizing the company for another once-in-a-century storm.
Zhang answers with his own near-death experience. After his consulting company’s most profitable year in 2019, the partners paid out all dividends because seasonal customer prepayments seemed dependable. COVID then closed their Shanghai and Hangzhou offices for more than two months, senior employees departed, and the dividend could not be recalled; a low-rate China Construction Bank revolver “saved our ass,” while Zhang had to cut much of the marketing team.
Yu endured an even larger liquidity shock in 2021. Chinese employees dismissed after seven years could demand “N plus one”—eight months of salary in that example—and courts generally favored workers. Zhang estimates New Oriental paid roughly $1–2 billion to departing employees, making Yu’s conservatism understandable even if it is no longer economically optimal.
Zhang and 12 other institutional investors, collectively owning more than 10% of EDU, have politely urged management to retain about $3 billion but distribute more. Management’s three-year policy returns 50% of net income, while recent actions included buybacks and a $100 million special dividend. Zhang expects gradual improvement: “We try to resolve it peacefully,” and Yu’s mindset is becoming more capital-return-oriented as EPS continues growing at a mid-double-digit rate.
6. AI raises margins precisely where it lowers barriers
For junior-high students affected by K–9 restrictions, EDU sells devices containing vocabulary, recorded lessons and access to backend teachers. Decades of behavioral data let the software move a student from easy questions toward more sophisticated and cross-disciplinary ones along an individualized learning curve. With fewer learning centers, landlords, classrooms and live teachers, Zhang says the segment’s operating margin is higher than 30%.
Walker’s pushback—worth keeping—is that removing physical infrastructure also removes an incumbent barrier. He and Zhang could theoretically launch a competing AI tutor at a fraction of EDU’s price, while celebrity teachers increasingly resemble MrBeast or Joe Rogan—personal brands able to own distribution and demand a larger revenue share. His newspaper analogy is pointed: free digital distribution looked wonderful until unlimited competition destroyed legacy economics.
Zhang concedes that star teachers have always left EDU to open studios and capture more profit. His defense is system-level rather than contractual: EDU offers audience scale, deep question banks, student data, hardware, software, marketing and a recurring pipeline of graduates whom it can train. Yet Beijing remains fragmented—EDU and TAL together hold only about 15%—and junior-high retention of 60–70% trails senior high’s roughly 80%, so AI disruption remains a real monitoring item rather than a dismissed risk.
Full transcript
With me today, I'm happy to have on from Guinea Value, Jingshu Zhang. How's it going?
I'm doing great. Thank you for having me, Andrew.
I'm really excited about this. I was telling you before, I think your background on this company is super unique. It's not a company I would normally follow, but with your background, I think it's a really interesting idea.
Before we get there, a quick disclaimer: Nothing on this podcast is investing advice. That's always true, and maybe particularly true today because we're going to be talking about an international stock, which carries all sorts of extra risks.
The company we're talking about today is New Oriental. It trades domestically, and the ticker is EDU. Obviously, it's New Oriental, so people might be able to guess that this is Chinese edtech. I'll toss it over to you. What is New Oriental, and why are they so interesting?
New Oriental is an education-services company based in China. They were founded by Michael Yu back in 1993, and they went public on the New York Stock Exchange in 2006. If you bought them at the IPO and held them until July 20, 2021, you would have compounded your capital over those 15 years at about 28% per year.
It's a very successful compounder, but then, if you look at the chart, the stock collapsed 95% because of a policy that came out of China. That's the ultimate stress test of this company. From the low, it has risen about 350%, and I believe it is very undervalued.
It's quite a saga. Other than the episode in 2021, which we can get into in much more depth, there was another incident in 2012. This is actually a pretty famous Chinese ADR in the U.S. because Muddy Waters shorted the stock. They came out with a short report, and the stock lost 35% of its market capitalization that day. On a pre-split basis, I think it went from $20 to the high single digits.
However, the investigation found nothing wrong. Michael Yu is one of the most ethical entrepreneurs, not just in China, but of all the companies that I cover on a global basis. He's probably one of the most ethical. Nothing was found to be wrong, so the stock recovered, and he bought back shares. He personally bought a lot of shares at the bottom as well. He issued more options to his employees to incentivize them.
This is a storied battleground stock that has compounded very successfully. I believe it will continue to do so.
I've got a lot of questions. I had forgotten about the Muddy Waters short, but I do want to dive into that. What's particularly interesting here is that you've got some background and some interesting insights into the space.
I mean, if I were having a cancer researcher on the podcast to talk about cancer research, and we didn't talk about a cancer stock or mention that they knew what they were talking about, I'd be remiss. I'd love to talk about your background because I think it lends so much credence and interest to the argument. Do you want to talk about that?
Of course. I personally founded a business. New Oriental started off helping students in China come to the United States, so it was an overseas test-prep business—GRE, GMAT, TOEFL, IELTS—and an overseas-consulting business.
I did my undergraduate degree at Cornell University, and I did my Ph.D. at MIT. In my first year at MIT, as the acronym Ph.D. implies—poor, hungry, and driven—my current wife had just become my girlfriend, and I wanted to make some extra money. My partner and I started an overseas-consulting business that did exactly the same thing as New Oriental. I know this space very well. We grew the business very successfully from 2012 all the way to 2019.
Can you just say exactly what your business was doing? You said overseas consulting, but I don't think you said what you guys were doing.
Because of the cultural and language differences, Chinese students have a lot of difficulty crafting essays and preparing for exams. They don't know how to apply through the online Common App system, and so on. We helped them streamline their entire application process, providing them with guidance.
In the U.S., most students probably apply for colleges on their own. But in China, it's very difficult. It's a very high-barrier-to-entry type of task that needs help.
We grew that business to about 3% of the study-abroad-for-U.S.-graduate-school market, and we competed directly with New Oriental. I respect Michael Yu and all the consultants there tremendously. One of the chief operating officers we hired came from New Oriental, and he's top caliber, with very strong execution skills and excellent service.
At the end of last year, we sold the business because, after Trump got elected, I directly called my partner, who was the CEO of the company, and I said, “We should get rid of this thing because it's not going to work.” We were able to sell the business at a severe discount to what it would be worth because the COVID times were very difficult. We even took on a revolver to keep the business afloat. It helped me tremendously to manage working capital, the marketing team, and so on.
It was extremely tough, and then Trump got elected. I was like, “This is endless pain again. I just want to get out of it completely.” So we sold the business.
This year, we're seeing IDP, which is an Australian business that administers more than 50% of the global IELTS exams. The stock utterly cratered; from where it went public, it's down more than 70%. If we still had the business this year, we would lose millions of dollars a month.
What's impressive about New Oriental is that the market is not happy that they are guiding overseas consulting and services to negative 4% to 5% growth for this year, despite all the visa craziness, the possibility of Trump, and everything else. Yet other players—myself and a lot of the colleagues and peers I know in that industry—are getting decimated in absolutely brutal fashion.
So, declining 4% to 5% is extremely impressive. That's why I'm very bullish on the stock. They will come back stronger than anyone else, just like they used to.
I think that's such a unique insight. Chinese tech is something you don't understand, and then you hear that it's down 5%, and you're like, “Oh, that's bad.” But it's one of those classic things: The market is down 30%, and you're down 5%. You're taking huge amounts of market share, and if the market ever stabilizes, normalizes, or, heaven forbid, grows, imagine what that does.
Let me start with a question I like to ask every guest. The market's a competitive place, right? This stock is actually pretty well covered. I was surprised—JPMorgan has research on it.
I think I saw a Goldman analyst, and there were several other really big-bank analysts on the call. So, just to ask you: The market’s a competitive place, right? The stock is well covered. What are you seeing that the market is missing that makes this an alpha opportunity?
Yeah. I have talked with many people on the buy side and the sell side, not just in the U.S. but in China as well. Their understanding of this industry is probably not as granular as someone who has been in this industry for about 12 years, like myself. In addition, they don’t seem to have a lot of the sources of information, such as how difficult it is for the smaller players. The smaller players aren’t public, so that information isn’t as well covered.
In addition, this is a stock that went down 95% because of a policy tail risk back in 2021. What happened back then was that I had just gone back from the U.S. to China, so I was isolated in a hotel room for 14 days—2 weeks—in Shenzhen. On July 24th or 26th—I forgot which—there was a document called Document 42. That document basically said, “We want to have a double reduction to lessen the burden for these students.”
What “double reduction” means is, first, we’re going to reduce the amount of homework that students have to do. Secondly, we’re going to reduce after-school tutoring. I remember vividly that it was around 4:00 p.m. in Shenzhen, which was about 4:00 a.m. in the U.S. In the premarket session, EDU and TAL went down 70% and 90%, respectively. That collapse was just so epic.
I think once you have something like that, it’s sort of like the banking industry. For many years after the Great Recession and the global financial crisis, people still had this stigma associated with that industry. They didn’t want to touch it because they were afraid that it might happen again.
My variant perception is that precisely because it happened in 2021, precisely because the Chinese consumer economy is not doing well, and precisely because of a lot of the policies that came out of Document 42, it makes EDU more resilient. I’ll give you a couple of examples.
Nowadays, in order to operate a school or learning center in China, you have to have something called a school-running license. Those licenses are rarely issued, if at all, in China nowadays. That’s the first thing. Secondly, just like the tobacco industry, if you advertise in a noticeable fashion, you immediately get killed. People can’t advertise much at all compared with before 2021.
Before 2021, you could go to a bus stop and see ads all over the place: “Send your kids here. Send your kids there.” The government said, “No, the kids are being burdened here, so stop all that.” Now there are no advertisements, so brand awareness is extremely important. TAL and EDU have the highest brand awareness in China. There are no more licenses, so supply is significantly constrained.
Therefore, it’s sort of like tobacco. Another thing is that this industry is very, very sticky. I say that because it has been around for more than 1,000 years, since the Tang Dynasty. The Chinese bureaucratic system is constituted by taking exams and climbing one ladder at a time. To become top officials, you have to take exams. It’s always exam- and preparation-oriented.
During Renaissance times, I remember Voltaire and a lot of great thinkers from the West specifically said that this kind of meritocracy was exactly what we needed in the West. This system of taking exams and excelling, rather than holistic review like we have in the U.S., will always be here. It’s very resilient.
I remember an anthropologist, Clifford Geertz, said, “Culture is a web of significance that we ourselves have spun.” Once we have spun that web of significance, it’s very difficult to get out of. It’s extremely sticky, and we’re seeing that.
After the crackdown, what the government noticed was that parents were just as competitive as ever in sending their kids to all sorts of tutoring schools. Because you no longer have the scale that TAL and EDU provide, the price is actually higher for middle-class families in China. It’s more burdensome, not less, and it’s just as competitive.
The government is sort of keeping one eye open and one eye closed, saying, “You guys can come back and do this again.” The profitability of EDU has eclipsed what it was before the crackdown, but the stock is nowhere even close to where it was. That’s because people are scared.
True. I’ve got to be honest: I think this is the first time someone has used Renaissance history and Renaissance theory to pitch a stock, but I love that pitch. It’s really great, and you’ve obviously got deep insight into the sector.
Let me push back slightly here. We started this podcast by saying, “Let’s put Muddy Waters aside,” right? We started by talking about 2 tail risks that have hit the sector. The big one—the one where I remember waking up and seeing these stocks down 70%—was Bill Hwang from Archegos in this stock. I think he might have been, but I can’t remember for sure.
Do you know?
I don’t think he’s in this one. He’s in Tencent Music.
Okay, I know he’s in Tencent Music. I thought he was in this one too, but regardless, we’ve talked about 2 different tail risks hitting this company. The first, and the big one, was the 2021 government crackdown on advertising and everything else. The second, smaller one was the overseas consulting business that you talked about. I do hear you, but I worry.
I’ve got this concept of risk writing, right? What it is is that you buy a company, and it goes up 15% per year for 8 years. You think, “This is a great business. It’s a compounder,” whatever. Then, in the 9th year, there’s that old story: The turkey thinks the farmer is its friend for 1,000 days, and then it gets its head chopped off.
Risk writing is kind of like that. You get 15% per year, you think it’s a great compounder, and then, in the 9th year, the axe comes down. That axe could be Google entering your market, or the government changing the regulations, or all that sort of stuff.
In this case, in 2021, it was the government changing the regulations. In 2024 or 2025, it was Trump trades. I just have to wonder: This might generate alpha if you look at your screen and hold it for 2 or 3 years, but are you actually generating alpha? Or are you kind of doing that risk writing, where in 2027 the government comes and says, “Actually, we’re changing the rules again. We don’t like this”? Does that make sense?
Yes, absolutely. I guess we can view volatility as risk, but to me, before the Double Reduction policy came out, the stock was trading at 30 to 40 times earnings. To me, that’s risk, because the valuation was so high.
After it got butchered, the market only assigned a valuation to the overseas business. The K–12 business was completely wiped out in the valuation, and all the cash on the balance sheet wasn’t included either. At that time, although the perceived risk was greatest, the actual risk was minimal.
When we look at a business, my view, like that of many value investors, is that the greatest risk is so-called permanent capital loss. We might invest in this thing, get killed, and it could go into restructuring or bankruptcy court, in which case we lose all the capital.
I think analyzing the balance sheet is important for this business. As Buffett said at the last annual meeting, “I spend a lot more time analyzing the balance sheet than the income statement,” which is different from Wall Street.
The balance sheet of this business is really a fortress balance sheet. When I reached out to you, I think the business was trading at a $7.1 billion or $7.2 billion market capitalization. Today, I think it’s $7.8 billion at a $48-and-change share price. Of that $7.8 billion market cap, $5 billion is cash.
One thing about this business is that it has one of the most beautiful business models, which is a negative-working-capital business model. When I was doing Broadsy [?], which is my own company, we used to have a freshman student come to us and pay $10,000 for the service upfront. We provided that service over a 3- or 4-year time period, and after we provided the service, we paid the consultants, especially the foreign consultants.
The money comes upfront and the costs go out later. You have this negative working capital, with a huge amount of so-called float that you can use to invest. Back then, I was investing in the stock market using that capital. If you just do 10%, your free cash flow margin changes from 30% to 40% per year.
It’s a very cash-generative type of business with a negative-working-capital model, and they have $5 billion. Of course, if you want to be conservative, you exclude the deferred revenue—the things for which you have not yet provided the service. That’s $2 billion.
So it has net cash of $3 billion.
After the Double Reduction policy, a lot of the teachers had to go home or whatever. Michael, being a very ethical entrepreneur, wanted to think of something to accommodate these remaining teachers as much as possible. So they started a separate business called East Buy Select.
It’s a livestream e-commerce business that primarily sells agricultural products to various smaller cities and local areas. Through that livestreaming platform, they also started a tourism business.
It helps the local governments tremendously because now the farmers can sell their agricultural products. They can attract more tourists to their cities and bring in fiscal revenue. Through those 2 prongs, it further enhances its relationship—a friendly relationship—with the local government.
Just to add 1 more thing: they hold 57% of East Buy, which is a publicly traded company on the Hong Kong Stock Exchange. Based on yesterday’s close, that 57% stake is about $3 billion.
I don’t know. I didn’t even see that when I was looking at it. That’s crazy.
So if you look at East Buy, which is $3 billion, and you have net cash—excluding all the deferred revenue—of $3 billion, add them together: $6 billion. So you have this $6 billion, and the market cap is $7.8 billion. It was even lower when I reached out to you. So it’s like $1.8 billion of net market cap excluding those 2 divisions, and next year they’re going to generate free cash flow of more than $500 million, more like $550 million. So—
No, look, it all sounds incredible. Let me try and push back on a few things. The first thing is, I did not know about their agricultural livestream business. Let me ask you: this is a business that’s New Oriental Education, and you say, “Hey, this is great for them. They take these people who would have been out of work because of largely the Trump policies and everything, and give them a job.” But I hear, “Hey, a company expanding into agricultural livestreaming from education?” I don’t get it.
Now, I do know that Chinese companies, kind of like Japanese companies, all have their fingers in every pie and everything. But when you say it, it sounds, “Oh, cool. They’re doing this,” and then you think about it and you’re like, “Wait, what? Why?” So I’d love to ask that.
If you recall back in 2021, the policy killed K-12 temporarily, at least for a year or 2. Michael didn’t want to fire all those teachers and wanted to find something for these teachers to do. So it was really unintentional. He was not thinking of building, in terms of market cap, this big a business; he was just trying to solve the problem for the teachers so that they still got a job and a salary, so they didn’t have to go back to the rural areas where many of them came from.
However, there was 1 guy who had already left the company. TikTok was taking off in 2021 and 2022. Remember, Meta was killed—there was all this fear. TikTok was doing everything right over those 2 years, and TikTok was growing so fast. This livestream just went phenomenal, and there were so many people buying. It just became an incredibly profitable business for this company that grew very quickly.
So it was something that was done unintentionally, just to try to solve the K-12 issue with the teachers. However, it is not a focus for them. Remember, there’s another thing: you can’t do a lot of advertising, so Michael and all the executives leverage this platform to do brand advertising for themselves. You can advertise agricultural products on your platform; that’s fine, but it associates with New Oriental, and they can talk about books, talk about life, and so on, to make people more aware of their brand. So it actually has some synergy with their education business, which you can advertise.
So let’s go to the balance sheet. You are right: this is a fortress balance sheet. They’ve got all this cash, and they’ve got the investment. I hadn’t even picked up on it. I just saw short-term investments. I assumed that was short-term; I didn’t think that was 55% of a publicly traded company on the Hong Kong Stock Exchange. But they’ve got $4.5 billion, $5 billion, whatever, of cash and investments on a $7.5 billion, $8 billion market-cap company. That sounds great.
And to their credit, they have bought back shares in the past. They came out with their new 3-year plan, and they said, “Hey, 50% of net income going forward, we’re going to return it to shareholders in some way, shape, or form.” So it’s not like they don’t do capital returns.
But if I was being critical—and I have been critical of things in the past—I like financial engineering. I like debt. I like share buybacks. I have friends who do not. When I look at this and say, “Hey, you’ve got an $8 billion market-cap company with $5 billion of cash on the balance sheet,” and yes, there’s negative working capital and everything, but if I just looked at it, I’d say, “Hey, your biggest issue here is actually that, yes, the stock might look cheap on kind of an EV-to-EBITDA basis, but the cash drag is really big here.”
And that’s to say nothing of—I’m sure most investors remember 2012 to 2014, all the Chinese companies. You mentioned Muddy Waters with this short report, but just the overall Chinese market where you’d find these companies. It was like, “Hey, there are $2 billion market-cap companies with $1.4 billion of cash on the balance sheet. Why is that?” Well, because all they do is raise money, and their whole operations are fraud. I’m not accusing this of that by any extent, but I do think that overhang plays when you think, “Oh, $8 billion market cap, $5 billion of cash, big cash drag.” This rings a bell of a lot of these stocks that blew up 10 years ago.
So I threw out a lot of thoughts there. I’d love to get your thoughts on my thoughts.
Yeah, definitely. That’s a very valid critique. I’d like to offer a personal experience.
Please.
My partners and I, in 2019, had the most profitable year for our previous education consulting business. We made so much money. For us, at a small scale—
Anytime anyone says, “We made so much money,” I’m very happy for them. Whether it’s $20 or $20 million, I’m very happy.
I even had the pleasure and privilege to go to the central tower in Beijing to be interviewed by one of the very famous hosts for the 70th birthday of the Communist Party, as a representative of a small education company in the education space.
We made a lot of money, and we paid out all the dividends. We didn’t like the cash drag. At the end of every year, because we had been expanding and it was a negative-working-capital business model, we were like, “Why do we need so much cash?” Every March, the season comes and the students’ cash just turns up. So we paid out all the dividends in January and February of 2020, and then COVID hit.
I got left out, from my perspective, because when I was a PhD student at MIT, I covered the commodity structural market. So I understand the oil and gas market very well. I put all that capital into the oil and gas space in March, when the oil price went negative.
But the business wasn’t that lucky because the money had already all been divvied out, and our chief operating officer left the company after COVID hit. He was like, “This industry is done for.” Two of the top salespeople left the company because 1 of them went to become a government official, and the other went to music consulting because they thought studying abroad was the end of it.
In Shanghai and Hangzhou, 2 of the major cities in China, our offices were shut down for more than 2 months. Students and parents couldn’t come to visit. That was the first incident.
At that time, we realized the dividends had already been paid out. We couldn’t get them back. We had to take out credit on a revolver. To the credit of the government, they made the interest rate really low. We took out the revolver from China Construction Bank, and that saved our ass.
In 2021, the business bounced back, and we started to pay out the dividends again. But that was a near-death experience. I had to cut a lot of people on the marketing team because of that.
The second one is with New Oriental. In July 2021, when that Double Reduction policy came up, in China, if you’re an employee of a certain company and you’ve stayed with that company for, let’s say, 7 years—if I work for you for 7 years and you say, “Sure, I’m going to fire you,” if you want me to go, you have to pay the so-called N+1 salary. You have to pay my salary that’s worth 7 months because I’ve worked for you for 7 years, plus 1, which is 8 months of salary.
There were a lot of old employees at New Oriental who had worked their tails off for Michael, whom he had to fire, and he had to pay all these folks. You can’t say force majeure because these people—and I’ve had this unfortunate experience—some of them would go to the arbitration court. The court in China will always favor the consumers and the employees; the employer is very unfortunate in China. So you have to pay out.
Michael had to pay out something like, I think, $1 billion or $2 billion because all these employees were leaving with that N+1. As someone who has gone through that kind of near-death experience, he probably has this conservative mindset.
So you have a valid critique here. Actually, the 13 of us—we’re 13 institutional investors, including myself. I’m the small fish. There are some really big folks out there. We own more than 10% of this company, and we all admire Michael tremendously. We wrote a very polite letter to their executive team, and we said, “We understand you want to preserve $3 billion of cash on your balance sheet.”
We did the most stringent stress test of 2021. We think you can survive with that $3 billion. So perhaps you can consider paying out a bit more than 50% of your net income back to us shareholders?
Of course, I've already talked with one of their executives on this matter, and she said, “Once we hike it to 50%, we are not going to lower it ever. It will be there forever.” That's the first thing. The second thing is that they are usually conservative, so when they get to, like, 5%, it will usually come out to be 70% or 80%.
They want to be conservative. Last year, they paid a special dividend, bought back about $7 million of their shares, and paid a $100 million special dividend on top of a market cap of $78 billion. So that's quite healthy capital return.
I think we are making progress, and they are likely to hike the dividend payout ratio, but it will take a little bit of time. We hope it will be a win-win. In China, we don't engage in that kind of activist-type situation where you just go in and want to fight everyone. We try to resolve it peacefully with a win-win type of situation.
So, to echo your concern, yes, we try to get the payout ratio a little bit higher. But we also understand the balance sheet from Michael's perspective. We can understand, as someone who has gone through that kind of experience in 2020, that preserving some cash will be good.
First, thank you for that story. It's really unique to hear somebody who can say, “Hey, look, I went through this, and I think I see it. It makes total sense to me.”
My only worry is that, let's say 2021 was a once-in-a-hundred-year storm for the industry or something. As we said, this industry dates back to Renaissance times and beyond. I worry that any company that says, “Hey, we've built this company to weather a once-in-a-hundred-year storm” sounds really good. There is the Buffett idea that any number times zero is zero.
You often find that companies that do this are just so over-reserved that it's ridiculous. At some point, it does become a drag on the business. That cash drag is huge because if you're holding cash equal to 60% of your market cap, then the business really needs to perform for you even to get stock-market-like returns.
I worry that if they're so wedded to this, and they're returning 50% of net income for the next 3 years to shareholders, as you said, they tend to be conservative, so they'll probably beat that. Even if they pay back 60%, that cash balance has just grown bigger, and it's like, “Man, I don't know.” Could they run this with $2 billion and then, every year, throw another $250 million onto the balance sheet until they hit $3 billion, and then draw it back down or something?
It feels very conservative to weather, “Hey, what if we have to lay off half our sales force because of regulatory changes and pay them all huge change-of-control and termination fees?” It just feels pretty slow.
Yeah, go ahead.
Understood. Yeah, so it probably has something to do with the personality of the founder. Michael is really a legendary guy.
In China, there's the so-called national entrance examination for colleges, the Gaokao. You have to take this exam. It's sort of like the SAT, but a lot more important than the SAT.
He took this exam the first time, and he's not the brightest of his type, let's say that. He himself admits that, so he didn't do very well. He waited another year, took it again, and still didn't do that well. He took it again in the third year, and it finally worked: He got into Peking University, which was one of the top 2 universities.
However, it was one of the less competitive and easier departments—the language department—which is not as popular as physics, you know, those types of things.
Yeah, physics—all the girls, all the glory, all the hype.
Right. So he is someone who is very down-to-earth, and he actually did not want to take this company public. He had 2 other partners who really wanted to get rich. They ultimately left the company and formed a venture arm, in which Michael also holds a stake.
It's called ZhenFund, which is a very famous venture capital fund in China that specifically tries to help New Oriental send students abroad. When students go back to China, ZhenFund, the venture capital fund, helps them launch businesses. It's kind of like a Y Combinator spin-off over there.
Yes. So it's a sort of a close group, right? That's awesome. Michael is from a rural area. He does not want to take the company public. He doesn't care about money.
But he was born in 1962, if I remember correctly, so he's 63 years old. We are trying to change his mind, and he has already changed his mind quite a bit with all the buybacks and dividends. But things take time, and we are trying to gently and in a friendly way nudge him. Hopefully, he will get there.
His mind is actually changing—it's changing more and more in a capitalist-oriented direction, especially in terms of capital return. I mean, it's still going to grow. The EPS is going to keep growing at mid-double-digit rates, so it's not a no-growth, slow-growing company. It's a very cheap, mid-teens EPS-growth company, and the capital return should ultimately increase over time. I believe—I have faith in Michael.
Let me ask one last question here. In edtech, you mentioned the core business is tutoring students, basically, training them for these tests and everything. Anyone who's been following the stock market knows AI is here. It's coming, and edtech has been the area where, in the public markets, you've seen the most disruption.
I'd point to Chegg, the cheat-sheet service for college students, and that business has been crushed. I think it's really impacted education, and there are a lot of reasons it's been impacted. First, there's great data online about education. Second, education users tend to be youngsters who tend to be early adopters. I just think education is at the forefront of AI, in my opinion. I might be slightly overstating that, but I don't think I'm saying anything too crazy.
It strikes me that you have an education company. They're advertising, but it is expensive. They're offering a bespoke product. If you can improve test scores, people will pay anything to improve their children's test scores. But it strikes me that this is a place where education is very vulnerable to AI.
So I just want to ask all of that to you: Is AI a risk here? Or you could come to me and say, “With AI, where you used to need 1 tutor for every 5 students, now you need 1 tutor for every 20 students. So you get a little bit more, you get better results, you get more pricing power, you need fewer tutors, and boom—profit margins through the roof.”
You've got proprietary data because you've got 30 years' worth of test scores and training data. So I could actually see either route. I'd love to just ask you: Is AI a risk or an opportunity here?
Yeah, I believe the fascinating part is this: In China, primary school is from 1st grade to 6th grade, junior high is from 7th to 9th, and senior high is from 10th to 12th. It's slightly different here.
For junior high, because of K–9, the Double Reduction policy is primarily focused on K–9, so you can't teach the kids subject-oriented courses anymore. For junior high, how do you prepare for the so-called Zhongkao? This is the entrance exam to get into the best senior high schools. So it's very involuted, if you will.
Since you can't teach the subjects for junior high, you have to rely on artificial intelligence. You can't have the teachers anymore. What they do is sell devices. In these devices, there are words and phrases already written, as well as recorded videos from teachers. The students buy the device, take it home, and watch the videos. There are also teachers on the back end, so if you have any questions, you can ask them.
In addition, they have decades of data on student behavior. They know what type of student it is, and based on the student's learning curve, they provide the corresponding types of questions that help the student climb the learning curve most effectively. They first prompt you with easy questions, then slightly more sophisticated ones, then cross-disciplinary questions, and finally the most difficult problems on the exam.
Because they are doing this, they no longer need to have learning centers. They don't have to pay the landlords. They don't need schools, and they don't have to lease them. They don't have to have classrooms or teachers. The scalability of the business makes the operating margin of that segment higher than 30%.
Can I pull on that for a second? What you just described is a panacea for a business, right? It's awesome. You get rid of all of that.
But my only counter to that would be: All of this is AI on the internet. You and I tomorrow could go release Andrew Shu's incredible tutorial with AI and all this knowledge and stuff. Doesn't that expose you? Doesn't that really lower the barriers to entry? Yes, you've removed the cost, but now the barriers to entry are gone, and 1,000 products can flood the market?
And maybe these guys are the best, but maybe you and I could launch a product that's 95% as good and gets up to 98% as good as it generates more data, and we launch it at 20% of the cost—30%, I don't know. I'm just speculating. But it does seem to me like there's a risk. It always sounds great when you take all the margin out, but you take all the margin out and, all of a sudden, hey, unlimited demand—newspapers, right?
We're going to lose all the fixed costs; we don't need to print everything anymore. We lose all this, and then, oh, yeah, but now there's unlimited distribution and a thousand places, and all the newspapers are bankrupt. That's a loose analogy, but I could see how that holds here. I just want to ask you that question.
Yeah. So, Andrew, if we were to start a business like that, we would first need to have a really star-level teacher who is willing to do it for us. The best teachers want the highest commissions, right? So, they probably want to go to a bigger platform that can give them scale. That's the first one.
Secondly, we need to collect a lot of questions. Thirdly, we need a lot of data to really know when to give the student the best type of question or prompt to help them learn best. Lastly, we need the resources to develop not just the hardware, but also the software, and to cover marketing and sales expenses.
You went to MIT. I'm relying on you, man. I'm from the South. I can barely read.
Oh, come on. So, I guess the best teachers want the most commission, right? They probably want to go to a bigger platform that can give them that scale. That's the first one. Secondly, we need to collect a lot of questions. Thirdly, we need a lot of data to really know when to give the student the best type of question or prompt to help them learn best. Lastly, we need the resources that we put into developing not just the hardware, but the software, and the marketing and sales expenses.
No, that's a great answer. Can I ask you the question in one more way? Just because, as you say this, it strikes me—and I do remember that China has much more superstar teachers than the United States. I say this from a literal 3,000-mile radius, but I do remember hearing a lot about the superstar teachers.
I do wonder. New Oriental does have advantages here: they have the data and the built-in infrastructure. But in the United States, you've seen MrBeast, Instagram influencers, and all this sort of stuff really supplant the legacy networks because distribution became free. MrBeast is launching chocolates, right? He's trying to disrupt Hershey's. 10 years ago, I would have told you no one could launch chocolates and disrupt Hershey's. Will he be successful? Yes? No? I don't know.
You're seeing people build the brands into themselves and capture all the economics. I would point to Joe Rogan. I'd point to MrBeast. We could point to any number of examples here. And I do wonder if I said, “Hey, with all this AI tooling, maybe the celebrity—the celebrity teachers that you're mentioning—maybe they partner with New Oriental, but they get much higher revenue shares because the celebrity is what drives the signups. The celebrity is what...” Or maybe they say, “Hey, I don't even need New Oriental.”
As we've seen so many—I mean, Joe Rogan 20 years ago would be on SiriusXM. Howard Stern, I think that's obviously a different generation, but his contract with SiriusXM is running out. I wouldn't be surprised if he's launching a podcast. Now, Stephen Colbert, he's getting fired from CBS. I bet you he will make more running a podcast than he was at CBS.
So, I just wonder if these superstar teachers that you mentioned actually drive the signups. Maybe they take more, or maybe they just launch their own network. That will probably be my last question, but I'd love to ask that because I find this fascinating.
It's a great question. Actually, what you described—the star teacher leaving the company to set up their own—has happened throughout the existence of the business. The reason for that is, if you are a teacher there, after all, you are getting a sort of fixed salary, maybe some incentives, maybe some shares. But if you form your own studio, then you can capture all the profit. Yeah.
But this market is, firstly, a very fragmented market. Even in the most concentrated parts, like Beijing, where TAL and EDU are both located, they have a combined market share of only about 15%. So, there are a lot of studios out there competing with them. Competition is probably something familiar to them.
Secondly, you have a constant replenishment of new talent that needs to go to EDU and TAL. Once they have a system that cultivates these teachers, every year the fresh graduates will go there and work a couple of years. Maybe they leave, but the system is still the strongest at EDU.
And, of course, what you mentioned—for the hardware and the scalability question, can we have something like that? I think the answer is probably yes, especially for junior high. I'm not sure whether it will be the case for senior high, because the learning quality is just not as good if you are not face-to-face with a teacher, especially for senior high, where you have a lot of difficult questions like math, physics, and biology. Then you want to have someone see exactly how you think and walk you through that logic, that train of logic.
So, I think—and you can see that in retention—their senior-high-school retention rate is something like 80%. But for junior high, despite the high operating margin, the retention rate is only 60% to 70%. So, I believe senior high is less of a threat. Junior high, yes, they can be threatened.
It will probably be something that, if you—and I'm already a shareholder, and if you end up being a shareholder—it's probably something that we need to keep a close eye on. However, it's so fragmented. I feel like there are so many mom-and-pop studios that will be weeded out in this long slope with a lot of snow.
This was great. Well, I don't know if you heard, but my phone was just blowing up, so I'm sure something terrible or awesome is going on in the portfolio. I'm probably going to have to call this. But this was great. JingshuWrites.com/blog is what I call it.
But you know what we're going to have to do? You've got to write more. I don't think there's been a post up there since late 2024. After this podcast, people are going to demand more writing, man. We've got to get some written word up there.
Yeah. So, I have a Substack. I update that, and at the end of every year I will republish all the letters up there. That Substack is also, I think, just called Guinea Value. I have a Substack that I replenish almost on a bimonthly basis.
Perfect. Perfect.
Yeah.
Well, hey, this has been a ton of fun. This was a really fascinating idea, so I really appreciate you hopping on and walking through all of this. And look, we're going to have to have you back on at some point to talk about something else. You do a lot more than just Chinese, so maybe it's something else, but I'd love to have you back on and we'll go from there.
Sounds good. Thank you so much for having me, Andrew. I've always been a fan. It's so nice talking to you in person finally.