I have been closely following the debate between Shweta Dubey of Fusion Research, Shane Dinneen of Pershing Square (William Ackman's primary analyst on Herbalife), and William Keep who co-authored an important paper with on identifying illegitimate pyramid schemes regarding the legitimacy of Herbalife (HLF).
If you have not been following it, the debate started with a posting by Shweta wherein he argued that using the methodology outlined in Dr. Vander Nat's paper he could demonstrate that Herbalife operates a legitimate MLM and that Shane Dinneen and William Ackman had misapplied the methodology in their analysis of Herbalife. In a separate posting Shweta tried to provide an example of a legitimate MLM in order to illustrate his point. William Keep and Shane Dinneen both posted comments to this posting including a series of questions from Shane Dinneen. Shweta has provided responses to the questions in two separate postings.
As has been widely reported, Bill Ackman and Shane Dinneen of Pershing Square have alleged in a lengthy presentation that Herbalife operates a pyramid scheme and urged the FTC to take action against the company. Herbalife has denied the charge including providing a presentation of its own attempting to refute the Pershing analysis. As Dr. Vander Nat is an economist employed by the FTC who has served as the Commissions expert witness in a number of pyramid scheme matters understanding how he might analyze Herbalife's operation is important to assessing a potential investment in Herbalife.
While the debate has been vigorous and thoughtful, I believe that all the participants thus far have missed an important aspect of the issue, which I think I've fleshed out below. I look forward to your thoughts based on my analysis.
Review of Dr. Vander Nat's Paper
Dr. Vander Nat begins his inquiry by reviewing relevant case law. He determines that based on applicable law an MLM is considered a pyramid scheme if "participants obtain their monetary benefits primarily from recruitment rather than the sale of goods or services." In an effort to convert this legal guidance into a quantifiable question, Dr. Vander Nat reinterprets this legal guidance into a thought experiment that he calls the "retail question," which is "if the MLM parent company were required to use actual gross retail sales as the basis for paying all the direct and indirect retail commissions (as well as production costs and related expenses), to what extent could it do so."
This interpretation is the key insight of the paper. Dr. Vander Nat conceives of an MLM as an entity which generates cash to fund commission payments through a combination of retail sales (net of costs) and recruitment. He conjectures that by looking at the mix of these two sources an analyst can determine which source is the "primary" source of monetary rewards to participants and thus answer the court's question. Using this framework, MLM's can be divided into three groups:
- Companies where the proceeds from actual gross retail sales less production and operating expense exceed commission expense, (i.e., the business would be profitable if the distributors were consolidated in the income statement). For these businesses all commission expenses can ultimately be funded entirely by retail sales and thus these businesses are definitively not pyramid schemes.
- Companies where the proceeds from actual gross retail sales less production and operating expense are insufficient to cover any of the commission expense (as would be the case if actual retail sales were zero for example). In this case, all the commissions are ultimately due to recruiting so these businesses are definitively pyramid schemes
- Companies where the proceeds from actual gross retail sales less production and operating expense cover a portion of the commission expense. In these instances Dr. Vander Nat would defer to the court's interpretation of "primarily" and the specific facts and circumstances, but posits that if "primarily" were to be interpreted as more than half, that this framework can provide a specific answer.
Dr. Vander Nat continues his paper by assuming a variety of characteristics about an MLM and computing solutions to the "retail question" subject to his assumptions. However a number of these assumptions are not consistent with the specifics of the Herbalife MLM. Specifically, Dr. Vander Nat assumes that participants "purchase the product primarily for resale" while not explicitly stated the implied assumption embedded in his mathematics is that any product purchased by a participant and not sold is ultimately discarded. In addition, Dr. Vander Nat assumes that any product which is not discarded is sold to the public at the retail price R with a specified retail margin. There is no mechanism in Dr. Vander Nat's specific model for a participant to sell or return product at a discount to recoup part of his/ her investment or to consume the value of that product as a retail customer and thus internalize the retail sale. These are important factual distinctions between the MLM theorized by Dr. Vander Nat and Herbalife. As such, simply plugging into Dr. Vander Nat's equations as both Shane and Shweta have attempted to do does not provide meaningful insights into whether or not Herbalife is a pyramid scheme.
Answering the Retail Question for Herbalife
In order to determine if Herbalife is a pyramid scheme it is necessary to answer the retail question based on Herbalife's specific circumstances. Specifically, we need to know two things:
- What are Herbalife's retail sales less as production and related expenses?
- What are Herbalife's commissions?
Herbalife's retail sales less production and related expenses:
Herbalife has very little visibility into the actual retail sales of its products. It sells to distributors then those distributors either consume the product (effectively retailing it to themselves) or retail the product. Pershing's data regarding online sales, however, can help to range bound an estimate. The data laid out in their presentation shows that any distributor can readily sell his/ her inventory for 65% of MSRP on EBay or other websites. So if some distributors sell their product for more than 65% of MSRP and the remainder sells their product for 65% of MSRP we can confidently say that 65% of MSRP represents a floor to actual retail sales.
A bit more on internal consumption since it is such a source of controversy. There is no reason why in answering Dr. Vander Nat's retail question a participant cannot both be a distributor and an end user provided that the participant would genuinely purchase the product even if he/ she were not participating in the MLM. Since we know that any distributor can readily sell his/ her product on EBay for 65% of MSRP if a distributor voluntarily elects not to avail his or herself of this option, but does not sell the product to a third party outside the network, it is safe to conclude that he/ she derives greater utility by consuming the product himself effectively acting as a retail customer.
This view on internal consumption is consistent with FTC's January 14, 2014 staff advisory opinion which clearly indicated that internal consumption, in and of itself, is not of importance when evaluating an MLM. What matters is if the product sale is "incidental" to the purchase of the business opportunity.
In 2012 Herbalife's reported retail sales, (a company reported metric which assume that all the product that was shipped was sold at 100% of MSRP) totaled $6,404.3 million. Assuming that distributors were only able to retail the product for the theoretical minimum of 65% of MSRP then actual retail sales would have totaled $4,162.8 million. Subtracting all other non-commission costs of operating Herbalife's business (SG&A and CoGs) of $2,072.3 reveals that total retail based proceeds to pay commissions totaled $2,090.5 million.
It is worth noting this $2,090.3 million represents the minimum retail based proceeds available to pay commissions. If, for example the average retail price was 80% of MSRP vs. 65% of MSRP, retail proceeds to pay commissions would be $3,051.2 million.
Herbalife reported paying $1,338.6 million of royalty overrides in 2012 in addition to that, assuming retail sales of $4,162.8 million (65% of MSRP) and subtracting Herbalife's revenue (all of which was paid to it by distributors) of $4,072.3 million reveals that distributors as a group made $90.5 million of margin buying wholesale and selling retail. Thus, aggregate compensation to distributors including both royalty overrides and retail margins was $1,429.1 million.
Distributors purchase product at between 25% and 50% of MSRP. When a distributor purchases product at less than a 50% discount to MSRP, Herbalife pays the difference between the actual discount given (25-42%) and 50% to other distributors higher in the organization. Herbalife accounts for these payments by netting them into revenue. Pershing Square has argued that these payments ought to be considered royalty overrides. Assuming Pershing's argument is valid, the difference between Pershing's method for computing royalty overrides and Herbalife's method is, at most, 25% of $6,404.3 million or $1,601.1 million, though it is probably significantly less than this as not all sales are made to distributors who buy with the minimum discount.
If we use the Pershing methodology for computing Herbalife's commissions, then Herbalife would have paid the $1,338.6 million of royalty overrides that it reported plus an additional $1,601.1 million for a total of $2,939.7. This increase would be entirely offset by a reduction in distributors' wholesale margin so aggregate compensation to distributors would be unchanged at $1,429.1 million.
One could argue, however, that when computing commission for the purpose of computing compliance with applicable legal standard, that gross payments to distributors who net receive money from the company without offset by losses incurred by distributors who incur losses is the buying product and selling it at a loss should be considered. Without taking a view as to the validity of this argument, if we assume that Pershing's view is accurate and that what matters is gross compensation to distributors (vs net compensation) then total commissions to distributors could not exceed $1,429.1 million of compensation under Herbalife's methodology plus $1,601.1 million theoretical adjustments to reflect Pershing's views on a gross basis, resulting in total compensation to distributors of $2,939.7 million.
Putting it all together: Is Herbalife a Pyramid Scheme?
Given that Herbalife's retail sales less production and related expenses are at least $2,090.5 million (likely substantially more) and its total compensation to its distributors is $1,429.1 million, we can show that Herbalife pays all of its compensation to its distributors out of the proceeds from retail sales. Thus, we can make a definitive determination that Herbalife is not a pyramid scheme.
If we further assume that actual compensation to distributors should be computed using Pershing's methodology and computed on a gross basis, then total commission to distributors would be less than $2939.7 million. This would imply that under the most draconian assumptions, 71% (2090.5/2939.7) of Herbalife's gross payments are ultimately due to retail sales. While this is a grey area (type 3 from above), since 71% > 50% this likely would satisfy the "primarily" standard identified by Dr. Vander Nat. What's more, the $2090.5 million retail profits used to compute the 71% is the absolute theoretical minimum. If Herbalife could show that actual retail sales were made on average at 78% of MSRP or more (vs the 65% of MSRP available on EBay) then we could make a definitive determination that it is not a pyramid scheme.
Based on this analysis, I don't believe that Dr. Vander Nat could reasonably defend the contention that Herbalife is a pyramid scheme, and thus I don't think that the FTC is likely to pursue the company on these grounds or ultimately prevail if they chose such a course of action.
This analysis does not mean that none of Herbalife's distributors fail/lose money. The fact is many or most do. This fact however is not relevant to the law, nor should it be. The reality is in many endeavors in our society ranging from trying a new product to starting a business, failure/ an unsatisfactory experience is commonplace and necessary in order for people to engage in the experimentation which is central to progress. In the case of Herbalife the cost of failure seems to be quite low. A distributor who decides he or she likes the product and might be able to sell some can purchase a few month's supply and if it isn't sold, consume it him/herself, return it to the company, or sell it on EBay. In each of these scenarios the consumer may be slightly worse off for the effort, but not horribly. For the small percentage of distributors who do succeed, Herbalife can offer a unique opportunity to sell a product that they believe in and earn a little extra money or in rare instances a good living.
I subscribe to the view that investing in businesses which do harm to society is generally a poor idea because these sorts of franchises inevitably get their comeuppance to the detriment of shareholders and because I believe that it is important and ultimately in everyone's best interest to profit honorably. With this in mind, I've given some consideration to the whether or not Herbalife is good for aggregate social welfare. I think the answer clearly yes. Herbalife contributes to social welfare in a variety of ways. Herbalife's retail consumers (including its distributors who consume what they also sell) enjoy gains from trade derived by from consuming Herbalife products versus their next best alternative. Some of these consumers use the product as a springboard to lose weight/ live a healthy active lifestyle benefiting themselves, their families and the community. Herbalife's 3 million distributors share (unevenly) in the $1.4 billion of royalty overrides; and of course Herbalife's suppliers, 6,200 employees and shareholders all benefit as well. This is not to say that there aren't bad actors in the Herbalife network, the law of large numbers means that there most certainly are, but on balance, I'm proud to be a shareholder.