I recently wrote an article describing the key errors in Ackman's Herbalife (HLF) presentation which lead him to wrongly conclude that the company is a pyramid scheme. Most of these were a result of Ackman and his analyst Shane Dineen's misinterpretation of Peter Vander Nat's definition of pyramid scheme. Bill Ackman's analyst responded with some questions. This article is my response to their questions and aims to help them better understand the errors they made in their work. For new readers, I strongly recommend reading my previous article (first link above) before reading this one.
Vander Nat & Keep offer a theoretical model to analyze MLM compensation schemes. However, MLMs are inherently complex enterprises, whose GAAP reported metrics do not fit neatly into the equations provided in their paper.
The main question is not how well GAAP reported metrics fit in the model. The important thing here is you misunderstood the basic concept of retail sales commissions (direct and indirect). This resulted in you ignoring indirect retail commission or Advance Retail Commission ((ARC)) completely in your presentation and wrongly classifying $1,740 million as earning with no retail basis-- thus, supporting your flawed argument that Herbalife is a primarily recruitment based organization.
Let me make it more clear. In your December 20th presentation you said that total recruitment rewards for 2011 were $1,740 million, which is greater than retail earnings which distributors generate from sales, thus Herbalife is a pyramid scheme. The fact is, most of this $1,740 million is getting financed by retail sales to the end customers. As long as distribution rewards in an MLM organization are primarily getting financed by retail sales, it is a legitimate MLM.
Put simply, retail commission consists of two things:
a) Direct Retail Commissions
b) Indirect or Advanced Retail Commissions
You missed Advanced Retail Commissions part entirely in your presentation and came to a wrong conclusion.
I have tried to explain this concept with an example on another website here. Since you have already commented on it, I am sure you read my article there. I suggest you to again take a deeper look. All I did there was design an MLM with Advance Retail Commissions covering distribution rewards; and distribution rewards intentionally kept higher than direct retail commissions made by the participants. The organization thus designed caused no harm to the participants and was not a pyramid scheme. You can argue that it is a simple model, and I welcome you to make it more complex. However, as long as ARC is primarily covering distributor expense, it will be a genuine MLM.
I hope you now understand that you have missed the Advance Retail Commissions part completely, which resulted in you arriving at the wrong conclusion. In case you still have some questions, I am happy to answer them.
Let's have a look at your questions now. I will start with the last one first.
5. Finally, you say that Ackman misinterpreted the term "primarily from recruitment" in his analysis based on the Vander Nat & Keep economic model. However, their paper explicitly leaves the definition of this phrase to policymakers ("any definition of the term 'primarily' implies a policy decision.") Therefore, why do you rely on the economic model in the second half of Vander Nat & Keep to define this phrase instead of case law that incorporates definitions of a pyramid scheme that are similar to that found in the first half of Vander Nat & Keep's paper (e.g., FTC v. Five Star)?
You and Ackman didn't misinterpret the term "primarily". You misinterpreted the term "primarily from recruitment." As explained in my previous article, Effective Recruitment Rewards (ERR) or rewards which have no basis in retail are calculated as
ERR = Distribution Rewards - ARC
When one is talking about MLM organization supported "primarily from recruitment", he is saying that Effective Recruitment Rewards (ERR) or rewards with no basis in retail form more than half of the total distribution rewards, i.e.,
ERR > 1/2 (Distribution Rewards), or
Distribution Rewards - ARC > ½ (Distribution Rewards), or
ARC<1/2 (Distribution Rewards)
i.e, Advanced Retail commission cannot even cover half of the distribution rewards. Now coming to the use of "primarily" which you say is a policy decision. I / Vander Nat have defined "primarily" as majority or more than half in the above derivation.
ERR > ½ (Distribution Rewards)
The term "primarily" is a policy decision in the borderline cases where ARC is covering approximately half of the recruitment rewards. However, in the case of Herbalife as I derived in my previous article, 70% retail sales (your assumption) will mean Distribution Rewards are getting completely funded by retail sales. It leaves no ambiguity for policy decision.
To sum up, ERR in relation to Distribution rewards is the determinant of whether an organization is primarily dependent on recruitment or not. Your use of distribution rewards in isolation was incorrect. Also, the use of word primarily in policy decisions is relevant for the borderline cases, which clearly Herbalife is not.
One of the interesting things to note from this question is how smartly you played with the word primary to deviate the discussion and call it a policy issue. This has been your tactic since the December presentation to confuse and misguide the investment community, and present an inaccurate picture based on wrong interpretations of key arguments.
This brings me to your second question.
2. In Vander Nat's Declaration in FTC v. Trek Alliance, he asks the following question, "Are the revenues from the sales of goods and services to general consumers sufficient to cover the following costs and expenses: immediate production costs, the various marketing expenses, general overhead, and the promised rewards for enrolling new members?" In light of Vander Nat's Declaration, why have you excluded SG&A from your calculation of "full production costs" or 'f'?
Again you have tried to force your interpretation and misguide investment community. This question is about overall profitability of MLM operations. This is not a question asked in the context of determining what should be used in calculation f. Peter Vander Nat has provided a link to his paper for more detailed explanation. Also, this is just one of the questions asked in economic assessment of an MLM scheme.
Snapshot from Vander Nat's declaration
Let me show you with the help of some examples, what your (mis)interpretation and making it a must meet/ mandatory criteria will lead to.
Here's Peter Vander Nat's statement:
Are the revenues from the sales of goods and services to general consumers sufficient to cover the following costs and expenses: immediate production costs, the various marketing expenses, general overhead, and the promised rewards for enrolling new members?"
i.e., r(Net Sales) - COGS - SG&A - Royalty Overides (or Distribution Rewards) > 0 or not
Let's start with companies following the "Sell and Earn" model: Avon Products(AVP) and Tupperware (TUP). Assume 90% of sales are outside network, or r = 90%. In 2011, Avon Products posted total revenues of 11,291.6 million. Cost of Sales was 4148.6 and S,G&A expenses were 6025.4. Assume Royalty Overrides / Distribution Rewards are completely captured in SG&A. The wholesale commission would make things worse, but even without considering it Avon will be classified as a pyramid scheme according to your (mis)interpretation. Here's the calculation.
0.9 x (11,291.6) - 4148.6 - 6025.4 = -11.56 <0
Worse still, if any MLM company makes operating losses, it will be classified as a ponzi scheme. Here are financials of Tupperware for the South American Market for the Years 2007, 2008 and 2009.
So, according to your interpretation Tupperware operated a ponzi scheme in South America in the year 2007 and 2008, and even if 90% of sales in 2009 sales in South America were outside its network, its operations would still qualify as a ponzi scheme. And this is for companies with the "Sell and Earn" model. "Save and Consume" operations like Nuskin (NUS) and USANA (USNA) will be by default ponzi operations, because you can take any obscenely low number while assuming r.
Clearly, your assertion that Peter Vander Nat declared that one should include SG&A in f is you putting your words/faulty assumptions in Peter Vander Nat's mouth. That being said, your remaining questions deal with how to make some adjustments while applying Peter Vander Nat's model to the companies. I will focus on that in Part 2 of my article.
Additional disclosure: The article was written with substantial inputs from an investor who has a long position in “in the money” call options of Herbalife. He may size up / exit / re-enter his position at any time.