Here’s why one hedge fund manager thinks Alibaba could be a big fraud
SEPTEMBER 18, 2015, 5:42 PM EDT
Compared to Amazon and UPS’s figures, Alibaba’s numbers don’t add up, he says.
Alibaba has already had a terrible first year since its IPO: Its shares are down 28%. But one well-known hedge fund manager has a suspicion that, if true, could potentially destroy Alibaba’s stock completely.
In a post on his blog this week, Bronte Capital hedge fund manager John Hempton laid out reasons why the Chinese e-commerce company’s delivery figures seemed fishy. The possibility that Alibaba BABA -1.36% might be a fraud, he wrote, “is a thesis worth testing”—and plans to gather evidence, and potentially short the stock, depending on what he finds.Hempton, an Australia-based investor known recently for making a long bet on Herbalife HLF -1.37% opposite hedge fund manager Bill Ackman’s famous short, said he got the idea from fresh-out-of-school candidates whom he interviewed for an entry-level position:
[Several] suggested that we could not dismiss the idea that Alibaba was faking their numbers. Most people who suggested that knew what an extraordinary (and potentially outrageously profitable) suggestion it was. After all some two-bit reverse merger Chinese company might be a fraud – but you can only make a limited profit from that. The idea that Alibaba – a company with half the market cap of Google GOOG -1.99% might be making its numbers up is – well – extraordinary.
Here’s how they got that idea, by the numbers:
On Alibaba’s “Singles’ Day” shopping event (“think of it as Black Friday in America,” Hempton wrote), in November 2014, the company said it received and shipped 278 million orders.
It’s also “more parcels in a single day than Amazon had users [244 million] in a whole year.”
Alibaba reported about 35,000 total full-time employees as of March 31, 2015. Meanwhile, UPS had more than 12 times as many (435,000) and Amazon had more than four times as many (150,000)—plus robots, Hempton added.
Then Hempton did some math: “To truly deliver at a larger intensity than Amazon, Alibaba and its outsource network would need more staff or capital (or both) than Amazon and UPS combined.”
Alibaba said that Alipay, the online payment processor it divested in 2011, handled 2.85 million transactions per minute at the peak of Singles’ Day.
By comparison, “Visa’s V -1.65% peak transaction volume globally [840,000 transactions per minute, by Hempton’s calculation] is only about 30 percent of Alipay’s peak minute. This suggests a level of shopping in China that puts the US, Europe and most of Asia to shame,” Hempton wrote.
Earlier this week, Alibaba responded publicly to a Barron’s article that expressed a similar skepticism about the “seeming improbability” of some of Alibaba’s reported figures. “Alibaba stands by our reported financials and operating metrics,” the company wrote in its response. Alibaba further explained that it is “flawed” to make conventional assumptions about its customers’ spending habits because they are a relatively exceptional demographic: “Shoppers that come to Alibaba’s platforms are early adopters of technology and tend to be urban and more affluent,” the letter continued. And as for its logistics and delivery staff, Alibaba said that rather than “taking on large headcount increases ourselves,” it instead “partner[s] with other companies to leverage their expertise and scale.”
Meanwhile, Hempton (who acknowledged in his post that he’d read Alibaba’s response) concluded his analysis saying: “At this point I know the numbers are wonky.” But before he’ll begin shorting Alibaba, he needs more evidence. As for how to collect that evidence, he’s welcoming suggestions.
Posted: 15 Sep 2015 11:40 PM PDT
Recently Bronte advertised for an entry level position. We have now hired.
There was one interview question which nobody gave an entirely satisfactory answer to, and many people gave unsatisfactory answers to.
This is not surprising. I don’t have an entirely satisfactory answer either. The question was about Alibaba – the Chinese internet giant – and the only reason I am going public is that Barrons has asked several of the same questions and Alibaba has responded.
As preamble I told the prospective employees what Alibaba was and what Singles Day was. This was graduate recruitment so I could not be sure that everyone would know even that. So I said:
i). Alibaba inside China is the biggest but not only online shopping site with Tabao and TMall. These were essentially flea-markets where lots of people had stalls (shops within shops) and they sold a huge number of items. Think of it as a combination of Amazon and Ebay and you will not be far wrong.
iii). Singles Day is a sort of anti-Valentines Day. On Valentines Day if you are attached you go out with your beloved. On Singles Day you go out and try to find a beloved. However more importantly it was the big day of online sales – think of it as Black Friday in America or Boxing Day in Australia. It is the biggest shopping day of the year.
I then gave the candidates two stories about Alibaba on Singles day. One was sourced from the BBC and one from an Australian website. The BBC reported in US dollars, the Australian site reported in Australian dollars. [Ultimately both are sourced from an Alibaba press release.]
I also indicated that the questions would be along the lines of “how interesting is this” and “how much bigger can it get”. The rule of thumb of course is that if an internet company can grow 500% from here you probably should own the stock. However if it is near the end of its growth period you should not own the stock – because it will derate.
Here were the articles – and I want you to read as if you were doing the job interview. Allow yourself 15 minutes…
The BBC article…
and the Australian article…
Starting question: How much bigger can this get?
Lots of people started by arguing from China’s growth rate – and they got projections 5x, 10x bigger than now. But a few noted that 278 million deliveries on Singles Day was “a lot” and thought that maybe it could double from here but ten times was not likely.
But at this point I wanted to explore the 278 million deliveries number. So the question is “how many internet users in China”. Most people guessed a number that was reasonable – so I pointed them to official numbers. Roughly 650 million would be a good guess. The official number at the end of 2013 was 618 million – and the growth rate has already slowed.
So there was 278 million deliveries for a total prospective 650 million users. In one day! The question is how much bigger can this get and what drives it to be bigger?
I then pointed them to official numbers that said that Amazon (worldwide) had 244 million users. That was 244 million separate accounts made at least one order in the past twelve months. The Amazon numbers are here.
Alibaba delivered more parcels in a single day than Amazon had users in a whole year.
Most people then figured that the users had multiple deliveries which is I guess a possibility. However it worth comparing this to Cyber Monday in the US. This article – which I presume is accurate – gives Amazon orders for Cyber Monday. To quote:
In order to uphold its reputation for fast deliveries, Amazon hired 80,000 seasonal workers in anticipation of Cyber Monday and the holiday shopping season, according to a report released by the company. Last year, Amazon sold about 426 items per second on Cyber Monday, and the online retailer expects to sell even more this year.
426 items per second is almost 37 million items. This is more directly from an Amazon press release. Singles day is 7.5 times bigger. To match Amazon in timeliness of delivery they would need to hire 600 thousand seasonal workers who work at the same efficiency as Amazon workers. Note that Amazon workers work with a lot of robots and other technology to pick and file and deliver items.
The Alibaba 20-F (the SEC filed annual report) does not show anything like this number of employees… to quote:
As of March 31, 2013, 2014 and 2015, we had a total of 20,674, 22,072 and 34,985 full-time employees, respectively. Substantially all of our employees are based in China.
The following table sets out the breakdown of our full-time employees by functions as of March 31, 2015:
(1) The number of employees presented in this table does not include third-party consultants and contractors that we employ, substantially all of whom are based in China. These consultants and contractors primarily performed work related to sales, research, logistical support and customer service.
(2) Our total number of employees increased to 34,985 as of March 31, 2015 from 22,072 as of March 31, 2014. Of the increase in employees, approximately 7,300 was due to the completion of our acquisitions including UCWeb, OneTouch, Alibaba Pictures and AutoNavi, and a majority are engaged in engineering and data analysis.
At the accounting date the total number of employees was under 35 thousand. Amazon – which has much smaller peak loads and far more obvious investment in high efficiency warehouses has more than 150 thousand employees.
However Footnote 1 on the above table notes that this does not include “third-party consultants and contractors that we employ, substantially all of whom are based in China. These consultants and contractors primarily performed work related to sales, research, logistical support and customer service.”
If Alibaba were as efficient as Amazon (meaning just as many computers/robots etc) and had the same delivery schedules as Amazon there would need to be about a million employees at Alibaba and its consultants and another 600 thousand or so seasonal employees around Singles Day.
These numbers are consistent with the annual report.
Again to quote the 20-F.
We believe that orders from transactions generated on our marketplaces represented a significant portion of our delivery partners’ total delivery volumes in the twelve months ended March 31, 2015. According to data provided by them as of March 31, 2015, our top 14 strategic delivery partners employed over 1,400,000 delivery personnel in more than 600 cities and 31 provinces, directly controlled municipalities and autonomous regions in China. Collectively they operated more than 100,000 delivery stations. This network managed the delivery of over 8.6 billion packages from our China retail marketplaces to consumers in the twelve months ended March 31, 2015.
It may be possible that Alibaba has outsourced logistics almost entirely – with a million outsourced employees controlled by 35 thousand in-house employees. The numbers in the annual report are consistent with this. This however is a much more “virtual” view of Alibaba than I had previously held. Amazon wins through logistical superiority. You would not want to compete with them. Alibaba it seems wins through logistical outsourcing of the highest quality.
We should also note the scale of the 8.6 billion packages the network claims to have delivered. UPS delivered 4.6 billion packages per year and had 435 thousand employees, 539 aircraft (including charters), about 100,000 delivery vehicles and almost 2000 operating facilities. [UPS factsheet here.]
To truly deliver at a larger intensity than Amazon Alibaba and its outsource network would need more staff or capital (or both) than Amazon and UPS combined.
The size and scope of Alipay
Only a couple of my candidates (all of whom wound up on the short list) noticed the throw-away line in the Australian article about the size of Alipay. To quote:
At the peak of the event, 2.85 million transactions were processed every minute by Alipay.
Now I asked “how many transactions per day do you think is 2.85 million in the peak minute”?
If they started to multiply by 60 and then 24 this was a problem. Obviously the peak minute is much higher than the volume at 3AM. I mean when have you ever used PayPal at 3AM? The peak day is obviously enough less than 24 hours worth of the peak minute.
A pretty good guess is that the peak day should be between 4 and 9 hours of the peak minute (in other words the peak day is somewhere between 2.85*60*4 million transactions and 2.85*60*9 million transactions). This seems sensible. I have discussed with several people in the payment industry the ratio of peak minute to peak day volumes and got numbers in this range – but I do not have definitive numbers. [If someone from PayPal or Visa wants to come through with something definitive I would like it.]
7 hours of peak minute (which my informal research says is reasonable) is about 1.2 billion transactions in the day. That is two per internet user in China. That is to say a large number.
Now Alipay is roughly the equivalent of PayPal. It might be used more than PayPal because people send remittances on it. It is unlikely to be used as much as Visa. After all Visa is used in the shops all the time. Look at your own life and ask yourself how much you use Visa and how much you use PayPal.
We can do a comparison to Visa. Visa overbuilds capacity – they have built capacity to 56 thousand messages per second. We know from another press release that was about 4 times the actual peak in 2013. So the 2013 peak was 14,000 transactions per second – which is 840 thousand per minute.
Visa’s peak transaction volume globally is only about 30 percent of Alipay’s peak minute. This suggests a level of shopping in China that puts the US, Europe and most of Asia to shame.
So I asked: Are there any other ways of reconciling these numbers?
Some people suggested (a) multiple deliveries for items, (b) that many of the items delivered were virtual items such as emoji and their delivery is not technically problematic and (c) insufficient data handling capacity and hence queuing explaining the almost absurdly large peak minute on Alipay.
So I asked: How would you test this. And I got some interesting answers – some of which might require a little work to pan out. However some answers wound up looking silly. If you think for instance that a substantial proportion of the transactions are low-value emoji then the remaining transactions wind up with implausibly high average value.
But a few – knowing Bronte’s history – suggested that we could not dismiss the idea that Alibaba was faking their numbers.
Most people who suggested that knew what an extraordinary (and potentially outrageously profitable) suggestion it was. After all some two-bit reverse merger Chinese company might be a fraud – but you can only make a limited profit from that. The idea that Alibaba – a company with half the market cap of Google might be making its numbers up is – well – extraordinary.
So I responded – fairly – that
(a) extraordinary claims require extraordinary evidence (known ubiquitously as the Carl Sagan standard), and
(b) extraordinary claims offer extraordinary opportunities for profit (as the market is very wrong) and hence it might be worth gathering the evidence so we can accept or reject the extraordinary claim.
And so here was the key point to the question: how would we go about gathering that evidence and what sort of evidence would you require to put the trade (short Alibaba) on?
I should be clear that I do not have the extraordinary evidence – but I do think it is a thesis worth testing…
This was a tough question – and I did not expect many (if any) good answers. So I am going to leave it to you dear readers for comments. I am perfectly willing to accept answers which demonstrate that the “fake numbers” thesis is wrong. Indeed the thesis is extraordinary and hence likely to be wrong.
However before I go I should look a little at delivery data. The 20-F (quoted above) suggested that the delivery network had 1.4 million employees and over 100,000 delivery stations. This suggests an average of 14 employees per delivery station and there must be a huge number with less than 10 employees including the truck drivers who actually do the delivery.
There was a Wall Street Journal story recently about delivery infrastructure in the provinces. There was a picture of a delivery station in Northeastern China. It was – to put it mildly – crude.
Another image of even more crude sorting technology heads a Bloomberg article:
Ramping something with this lack of automatic sorting technology up for Singles Day would require a lot of staff. Some candidates thought of trying to count them.
With 100,000 delivery stations it is unlikely to be possible to do it on a statistical basis by counting trucks. There are far too many locations and they are too heterogeneous.
Maybe you could do it with official employment numbers. But in China that is almost certainly a dead end. Chinese employment numbers are amongst the most suspect of all Chinese economic data.
Alibaba’s own promotion material contains several videos of their logistic operations. Here is a video interview of a woman describing the seasonal work packing boxes for Singles Day.
The woman states she will normally pack about 200 packages a day but on Singles Day she will do 800-900. If it takes her 50 seconds to pack a package (reasonable looking at the process) and she takes 5 minutes break per hour packing 800 packages will take about 12 hours. I guess that is reasonable. But to do the volume for Singles Day would require 350 thousand such people just in packaging, not even delivering.
The company has videos of its logistics as well. Here is a summary video – with pictures of the delivery network:
(The original can be found here…)
The company also has b-roll material (which they license widely) which has quite extensive video of the conveyor belt systems and the sort.
Strangely much of the material involves people standing around and multi-handling material on conveyor belts rather than doing the real function of a sorting facility (that is making choices therefore sorting).
Similarly there is b-roll material of packaging happening in one of Tabao’s merchant shops. This involves distinctly low-tech (and slow) packaging and sorting for Singles Day. It is however clearly in part aimed at a Western audience (and presumably Western investors) as the whiteboard is in part in English.
We have some data about the size and the amounts of capital deployed in this delivery network. Here is a relatively old fact sheet (stated volume has risen substantially since then).
Here is an article and video about Alibaba/Tabao rural service centers:
There is some evidence in this video about people making multiple orders as in some rural centers there is a single woman who places multiple orders for an entire village. The distribution center however is not heavily endowed with advanced technology. There is no evidence that in any way the capital equipment at Alibaba and its distribution partnership comes close in scale and quality to the infrastructure of (say) UPS. However the delivery volumes are larger than UPS and Fedex combined and the employment levels are similar (in aggregate).
It is worth comparing this to videos of Amazon’s warehouse and packing system.
There is also a Fedex and UPS documentary on Bloomberg which gives some idea of the scale of these behemoth organisations. When this was filmed (2012) the combined volume of UPS and Fedex was well below the volume of Alibaba’s delivery network. Still it is worth comparing how much capital equipment the US giants have to how little is in the Alibaba network:
At this point I know the numbers are wonky – but working out whether this is material or not (or how material) will require some fine measurement techniques.
The two best candidates both had good ideas on this – but again testability is going to be hard. I would accept ideas by email too.