Graham & Doddsville Spring Letter is out! The issue features…
A. Rama Krishna, CFA of ARGA Investment Management
Clifford Sosin of CAS Investment Partners
Chris Begg of East Coast Asset Management
It also includes photos from from the 20th annual CSIMA Conference and the 10th annual Pershing Square Investing and Philanthropy Challenge.
Michael Gelband’s Exodus Point launched in 2018 with $8.5 billion in assets. Expectations were high that the former Millennium Management executive would be able to take the skills he had learned at Izzy Englander’s hedge fund and replicate its performance, after a decade of running its fixed income business. The fund looks to be proving Read More
Excerpt and full PDF below – Note Clifford Sosin of CAS has a huge long in Herbalife and another long in World Acceptance Corp. (WRLD), both are big Whitney Tilson shorts.
Q1 2017 Letters
2016 Fund Letters
G&D: We saw, in prior letters, that Herbalife is one of your biggest long positions. Can you walk us through the research process with that idea?
CS: It was always on my radar. If you did a “Magic Formula” screen, it always showed up cheap. It was this company that always grew and had great economics, but I had never bothered to grab the 10-K. I knew the name [Herbalife], and watched the Ackman presentation a week or two after it had been done for roughly the same reason people stop to stare at car accidents. I originally thought the Ackman presentation was great and I even wrote to one of my friends that I thought Herbalife was a pyramid scheme but a very profitable pyramid scheme.
However, there was something that bothered me. Ackman spent a lot of time talking about the product and the $100 MSRP. He showed that on eBay, the same product trades for 65% of MSRP. That sounds bad. He used that to show how there’s no significant retail profits. The weird thing is that if you’re a distributor and you have any volume, you buy at 50% of MSRP. So that seemed to imply that distributors were making money even selling on eBay.
knew the name [Herbalife], and watched the Ackman presentation a week or two after it had been done for roughly the same reason people stop to stare at car accidents.” If I told you that there was a loan and nobody wanted it and it was worth virtually nothing, what would the market clearing price be? The answer isn’t 65 cents on the dollar. The answer is closer to two cents on the dollar. In the case of Herbalife, you can compare it to furniture. Used furniture trades at a much bigger discount at MSRP than Herbalife product. Herbalife’s a food. When you’re buying food on eBay, it’s a little weird. There probably should be some discount. This bugged me. It was inconsistent.
You can’t have a situation where large numbers of people are buying the product they don’t want because they want to participate in a money transfer scheme. It doesn’t companies must sell to ultimate users.
Dr. Peter Vander Nat tried to put some math to the legal standard. It’s a very sensible paper. He says, let’s imagine that we consolidated the economics of the distributors with the multilevel marketer. There’d be a certain amount of gross profit, there’d be a certain amount of overhead, and then there’d be a certain amount of sales and recruiting commissions. If the gross profits less overhead cover the sales commissions then the organization is clearly not a pyramid scheme – after all, it could work as a consolidated entity. Conversely, he posits that if the gross profits don’t even cover the overheads then all the commissions paid are essentially wealth transfer among the sales people so it is a pyramid scheme. Somewhere in between, there’d be some percentage of the amount paid out to distributors that comes from gross profit and some percentage that comes from new distributors coming in and going out. He said 50% would be an interesting tipping point.
This is the way that Dr. Vander Nat tried to put some math to the legal standard, and there is an equation that falls out. In Shane Dineen’s part of the Pershing Square presentation, he tries to fit Herbalife into this equation. Ackman’s team used a bunch of assumptions and shows Herbalife’s a negative number, ergo a pyramid scheme.
The problem is that the original paper starts with a bunch of implicit assumptions. Among the implicit assumptions is that the product is either sold at the retail price or not sold at all. The problem is, we know for Herbalife that anyone can sell it on eBay for 65 cents on the dollar. That is different than the embedded assumptions in the model. If you were to just make that change you can’t come up with any way that Herbalife is a pyramid scheme in the Vander Nat model.
Also, let’s look at this in a Bayesian sense. Just ignore everything you know about Herbalife. Imagine some company that’s lasted for 37 years, is publicly traded, exists in 90 countries, and has been in regulated markets all around the world. Is it or is it not a fraud? The answer is it may be a fraud but the prior probability that it’s a fraud is quite low.
When I finished going through the Vander Nat paper, I added that to the mosaic of things and decided I was reasonably confident that it wasn’t a pyramid scheme. I essentially looked at it in a Bayesian sense. The prior probability is low. Ackman presents evidence it is a pyramid scheme but we have determined that much of that analysis is faulty (and there were a lot of other issues beside the ones I mentioned). Then you layer the availability of returns, the high price on eBay and the perception of this as a class act among industry veterans, and you come away thinking it is vanishingly likely that Ackman is right.
At the time, I definitely did not understand why it was a good business, but I was willing to take the trend of 30 years of performance at face value.
work if people try to sell, can’t sell, become stuck with product and then eventually throw it in the garbage when, at the same time, there is a secondary market where the stuff trades above the wholesale price. People are not that ignorant. It’s not like one dollar of the stuff trades on eBay; millions of dollars’ worth is traded on eBay. Herbalife sells billions so it’s this teeny little piece, but it’s millions of dollars in a secondary market.
That was the string in the sweater that I started pulling at so to speak. I was also fortunate to have a friend who had done some work on it and believed it to be a good business, so he steered me the right direction. John Hempton also put out his somewhat famous blog post. He wrote about going to an Herbalife club and guess what? It was filled with customers. They’re drinking shakes. I started to put it together.
I spoke to people who are experts in the space (lawyers and such) who say Herbalife’s not just a legitimate company, it’s the white gleaming example of multilevel marketers in the industry. They call it the gold standard. Herbalife’s turnover is the lowest in the industry. People love it. It’s been around forever. I started noticing all these things.
Finally, I sat down and revisited the section of Pershing Square’s presentation where they were quoting this paper from the SEC’s former economic consultant. I read the paper. The legal precedent from the Koscot case for multilevel marketers is that the
G&D: Could you walk us through the WRLD thesis?
CS: WRLD gets a bad rap. Subprime lending, in general, gets a bad rap. I think that people tend to confuse their desire not to have a society with any desperate people with the fact that once people are desperate, WRLD is a “lender of last resort.” I think the role of “lender of last resort” is extremely important. It gives people with nowhere else to turn an opportunity to borrow money to solve urgent needs. In performing on those loans, individuals can improve their credit scores, which will ultimately improve creditworthiness in the future.
It’s an incredibly difficult, risky, and thankless business, but providing this ladder, from the very bottom to a notch or two above the very bottom, is incredibly important for social mobility. Compare the role of these lenders in society to the importance of chewing gum manufacturers—I think subprime lending is a far more important business. “It’s an incredibly difficult, risky, and thankless business, but providing this ladder, from the very bottom to a notch or two above the very bottom, is incredibly important for social mobility.” Some people don’t like the industry because the interest rates are high. But as an industry, there are not reams of profits to be made. If you look at a typical installment lender, they are not making money hand-over-fist. The prices are covering their costs and their losses and create a modest profit. These loans are expensive to originate and service, and they have a lot of embedded loss due to the risk. They’re also small and have short duration. To make a reasonable dollar profit, you need to have a high implied rate.
Another way to think about WRLD is to consider 200 people, all of whom have large problems. They need to repair their water heater, or fix their car or, less practical but emotionally important, they can’t buy Christmas gifts or travel to a friend’s funeral, etc. All come to WRLD’s office and WRLD sends half of them packing. Those are the people who really suffer; nobody lends to them.
Then, WRLD makes a loan to the remaining 100 of them. Those people get money and solve their problem. Sure, they pay for it with an interest rate of 60% on average. But that is a lot better than the alternative of not having a car to go to work. Over time, some people will renew that loan. The average person renews twice and is in debt for 24 months. Almost 80% of them will eventually exit the repayment door as opposed to the charge-off door and their credit is improved.
Who are the victims here? When the customer repays, WRLD makes a healthy profit but the customer got the cash he or she needed and his or her credit improved. Tough to argue that those borrowers are victims.
When customers default, they experience the discomfort of having debt collectors call them and further degradation to their credit score. But otherwise, they are better off. WRLD gave them more money than WRLD received back. This is different than many payday transactions where the lender can often profit even when the borrower defaults. It’s hard to argue that the borrowers were victimized. And even if you could, you can’t make the other 80 loans without experiencing the twenty who don’t repay although they sure do try.
I’ll add one more piece. WRLD is an installment lender, which is fundamentally different than a payday lender. Payday lenders charge very high interest rates, typically 400%, for very short-term loans, two weeks on average. Payday lenders have an ability to reach into someone’s bank account and pull money out. With these two features, payday lenders can and often do make money on loans where the borrower defaults. WRLD is an installment lender. They give people longer-term loans with fixed payments. A typical loan might be $900 payable in twelve $100 installments. WRLD has no ability to enforce repayment if the person doesn’t voluntarily repay.
G&D: How does WRLD collect from customers?
CS: The most common method is paying in cash in person. WRLD has expanded the payment options, and there are people who pay with check. The customers are not all under-banked. There are people who pay by phone. There’s debit card. But WRLD has no ability to take money out of people’s accounts.
In the vast majority of charge-offs, WRLD loses money. The incentives are well aligned. WRLD wants to make loans that people can repay. The only way WRLD can get people to repay them is if the borrower’s income less expenses is enough to service the debt.
G&D: How long are the loans on average?
CS: The average loans are twelve-month, but they’re monthly installments. The average duration of a twelve-month loan is six months. The average duration of a portfolio of twelve-month, linear amortized loans is three months.
G&D: What do you think about increased political scrutiny and regulation of the industry?
CS: A variety of politicians will paint these guys as evil. It’s not hard to find someone who had some really bad experience. Yet WRLD’s net promoter score is 68%, which is amazing. WRLD is very popular with its customers.
The academic work on payday loans is mixed. There is a wonderful piece by John Caskey that summarized all the academic work and makes this point nicely.
On the negative side, there is work that shows that the career performance of Air Force members stationed at places with access to payday loans is worse than ones without access. Of course, that’s bad. There is also evidence that different disclosures about sustained use of loans by borrowers can importantly reduce their propensity to borrow. So those would indicate that perhaps payday loans are bad.
But there is also research that shows that counties in California with access to payday lending have reduced rates of suicide and robbery after earthquakes than other counties which have banned the product. Similarly, when Oregon put in place a ban on payday loans, economists used the occasion to study how the change impacted people right on the border of Oregon and Washington. They found that indicators of financial suffering, including phone disconnections and job loss, were higher after payday and installment lending were removed on the Oregon side of the border than just a few miles away on the Washington side of the border where payday loans were available. So these studies support the idea that payday loans are good for society.
Evidence supports both sides.
What I think is clear from the research is that whether payday lending is good or bad, it is not very good or very bad. Economists have studied this too closely and had too many conflicting findings for the impact to be very strong one way or another.
Installment lending is far more user-friendly than payday lending. So if payday lending is at worst a little bad for social welfare, I think it’s highly probable that installment lending is very good for society. All those people who need cash are served by these businesses.
Still, obviously, despite the logic and evidence this is an industry that is under a lot of scrutiny. The CFPB is clearly of the view that high-cost short-term consumer loans are probably bad for consumers. There is a very lengthy legal discussion we could have about all this but it is too involved for this interview. I think though it would leave you thinking the risk isn’t as big as it might seem. But it is a big risk.
G&D: What evidence would indicate that your investment is wrong? CS: Not an easy answer. Let’s start with what the economic mechanism is.
Our hypothesis is that the business has what we call “loyalty effect economics.” It’s an economic phenomenon in The Loyalty Effect by Fred Reichheld. It describes how in some industries a business in the upper-right of a 2×2 matrix of customer retention and employee retention is the most lucrative. If you can establish an organization with long-tenured employees and long-standing customers, you will be much more profitable than your competitors.
Now let’s look at the installment lenders. WRLD has been very successful at this. WRLD’s average branch employee has five years’ experience. This is tremendous. Way higher than most other front-line employees in most companies or industries. These aren’t big communities. They know who is who in town. They know your mother. They go to church with you. People come in and if they’re good at it, WRLD’s people have some reasonable ability to underwrite your loan and
“Installment lending is far more user-friendly than payday lending. So if payday lending is at worst a little bad for social welfare, I think it’s highly probable that installment lending is very good for society. ”
Full letter below