Description

This buy-here-pay-here used car seller has been written up 4 times before on VIC. While I agree with most of the basic points of these write-ups, I think they do not adequately explain the underlying economics of the business. The stock has sold off substantially this year as the company has missed expectations. These reasons include lending too aggressively in some of their newer lots in Texas (since reined in) and the impact of the hurricanes (September 2005 was the worst month for new car sales in Arkansas - where over 50% of their lots are located - in 20 years. Car-mart was profitable in September despite a steep drop in purchases). It is my contention that the fundamental competitive dynamics of the business have not changed, and Car-Mart will continue to expand while maintaining historic average margins for a long time. I meant to post this write-up a few days ago, and the stock is up 5% since then - I apologize for not posting sooner. Nevertheless, the fundamental thesis that this company should grow earnings at 20% per year, on average, and reinvest capital at an after-tax IRR above 20% for a long time remains intact.

The simple valuation is: The company had 81 lots as of November 2005, which sold an average of 26 cars per month for an average price near $7,000. 27k cars sold in 2006 with that number growing at 15% annually(trailing 5 year average growth pre-tax income is 16%) multiplied by a post-tax NPV of $700 per car sold equals $18.9 million after-tax normalized income (the company, for GAAP purposes, does not use gain-on-sale accounting).

Historic per-car pre-tax profitability:
1998: $ 877
1999: $ 865
2000: $ 979
2001: $1,046
2002: $1,036
2003: $1,046
2004: $1,020
2005: $1,106
6M06: $ 892 (Fiscal year ends in April, so the very bad month fell in Q2.
On Q2 call management indicated things had recovered to normal.)

At a share price of $18.5, you are paying $209 million, or 13.7x trailing numbers (and 11.1x forward normalized numbers), for a company that has a conservative approach to the subprime market, considerable barriers to entry (based on difficulty, culture, and length of time in business - i.e. institutional learning), and should continue to grow the top and bottom line at 15% - 25% per year for at least another decade, and possibly much longer.

---------- The Business ----------

The first thing to do when deciding on whether a business is a good one is to ask, does it make sense that it exists and has good ROIC? And, in Car-Mart's case, if the business makes sense why hasn't it been done before? I'll address these two questions, then do the detailed per-car valuation, talk some about the culture and a visit to car-mart, and I'll finish by considering some of the common complaints/risks others see with this company.

The used car market is large, fairly stable over multiple years, and highly fragmented. There are a few large companies taking a stab at it (e.g., KMX), but they focus primarily on the higher-end and urban areas. The buy-here-pay-here (BHPH) portion of the market is dominated by mom-and-pops, and is large - the subprime vehicle financing market is estimated at anywhere from $70 to $100 billion annually, and there are at least 30,000 BHPH dealers in the US. The majority of BHPH lots, especially rural ones, are family-owned and small. They eke out a marginally profitable existence and generally don't possess the systems, capital or talent to expand successfully. The elements of success in the business are access to capital, inventory control/mix, customer relations, successful credit underwriting, and good collections. All but the first of these elements are improved with experience, and all but the first two are improved the longer and more integrated a lot is with the community. This is an important element of this business, and part of the reason there are so few chains - unlike the sub-prime market, the sub-subprime credit business is very much a local business. Furthermore, proprietors can either take capital out of the business, or expand the business, but not both (banks are leery of BHPH's seeking leverage) - growing income depends on growing the balance sheet through earnings retention.

Car-Mart has successfully carved out a profitable niche for itself through careful organic growth and by creating a strong culture that gives stability to the employee base within local markets, allowing bonds with customers to form. This is a highly surprising result, considering the number of repos in this industry, but it is the case, and the company has the stats to support their claim: in mature markets (open 10+ years) 40% of their cars are sold to previous customers and an additional 10% - 15% are sold to customers referred to Car-Mart by other customers. In addition, firmwide they have over 3,000 customers who have bought 5 or more cars from Car-Mart. According to the company, statistics are not commonly this high for BHPH lots - Car-mart has an incentive structure that encourages managers to sometimes forego near-term profits for long-term customer retention and value.

Why is this a good business? Car-mart's culture allows it to match/beat the local used car dealer economics, and Car-mart's size and institutional memory/experience allow it to push margins higher through incremental improvements. Car-mart has a capital advantage, which allows it to sell cars with smaller downpayments due to their large balance sheet and write credit more easily on an as-needed basis. Car-mart has a scale advantage, which yields inventory acquired more cheaply (you can spread a buyer across multiple lots) and better optimized. Car-mart has a technology advantage, which allows it to see trends that are affecting its business more quickly. And Car-mart has a management/communication advantage - best practices can flow around the institution and employees can be trained properly prior to going onto the lot.

Over time one would expect, in aggregate, for small used car dealerships to enter or leave the market based on the short-term profitability of the business environment, but there should be a general equilibrium of profitability for those companies over multi-year cycles. If Car-Mart successfully drives significantly higher profitability through their business infrastructure, then they should be able to expand quickly and profitably. This is what they have been doing for the last 21 years (every year they have been profitable), and what they intend to continue doing indefinitely (they are far less than 1% of the market, so they have plenty of room).

Three main hurdles people have to get over before owning this company: 1) Why hasn't it been done before? Answer: It is a very difficult business that is not a good roll-up and is not quickly scalable. Getting people excited about working for a company where you are constantly collecting from sub-subprime credits is not easy. 2) How can it be profitable with such a poor customer base? Answer: Everyone agrees that at some mark-up from cost, this business is profitable. They achieve that mark-up. 3) Isn't their current profitability partly explained by the fact that they take massive losses once every 10 years or so when the economy goes through a bad cycle (i.e. black swan risk)? Answer: No. It is a good business even considering this. One, they are substantially under-levered. Two, the additional defaults that occur in a downturn are not so bad as to damage the business significantly, and the decreased competition resulting from a downturn means loans underwritten then are more profitable coming out of the cycle. Three, they have been profitable every year since inception. Four, their customers are poor (the majority have almost no assets, work close to minimum wage, and have no bank accounts) - times are always hard for poor people.

There are two levels at which one could build a valuation model: per car or per dealership. In this write-up I will approach it from the per car level, but before my detailed analysis of per-car economics, it is worth noting that dealerships become much more valuable over time. Cars sold per year, and profitability per car, both grow steadily at existing dealerships, and continue to do so even once they are mature. According to the company, even the most mature dealerships continue to improve, on average, year-in and year-out. I will build this fact into the multiple analysis after discussing the per car valuation.

---------- Detailed per-car value analysis ----------

First off, my analysis is different than that you will get via looking at the filed financials. I built a per-car default-based model below, whereas the 10-Qs and 10-K deal in aggregates and pull the NPV of the defaults into loss reserves. I chose the method below because I believe i) it better illustrates what actually happens and ii) it gives a better sense of the sensitivities of their business to changes in the assumptions.

Car-Mart's fundamental business is trading cars for loans on subprime credits. Using the latest 10-K as indicative of the business economics going forward, the economics are:

Average car price(TTM): $7,300

COGS per car: $3,924
SG&A per car: $1,341
Interest per car: $ 47
D&A per car: $ 16

Current credit statistics:

Risk-free rate: 5%
Average loan length: 26 months
Average default rate: 38% (assume this is a normal distribution of defaults peaking at month 10)
Average interest rate: 12.5%
Assumed recovery on repossessed cars: 35% of COGS (more likely to be broken and more used)

For the average default rate, the observed average default rate over the past few years has been 35%. I added 3% to the average default rate to take into account that rate jumping to 50% once every 5 years. The reason I don't have it jumping higher is: 1) 40% of the company's defaults are from the car breaking down. There is no reason to suspect this is correlated to the economy in general. The cars sold by the company are on-average 6+ years old and 70k+ miles. 2) The remaining 60% of defaults are somewhat correlated to the economy. 60% * the 35% overall default rate yields 21% of defaults resulting from job loss, high medical bills, etc., over a two year period. This means 10.5% of these defaults occur annually. Assuming that 2/3rd of those defaults are correlated to the economy and that the rate triples in a really bad economy, a really bad downturn adds 14% to the default rate (yielding a 24.5% default rate in one year due to reasons other than the car breaking down). Tripling is more conservative than the actual loss experience of subprime credit portfolios I spot-checked. For instance, two different Capital One Auto Finance portfolios I looked at had net annual credit losses ranging from 4.5% in a good year to 9% in a bad year. By assuming that the tripling occurs once every 5 years, I believe I am being conservative. Actual observed provisions for credit losses bear out my thinking - in fact, Car-mart did surprisingly well in 2001 and 2002 when other used-car finance companies were in serious trouble. More on this later in the write-up.

Using these credit statistics, and valuing based on discounted probabilistic cashflows (a risk-free default-based model), I get a NPV of loan cashflows worth $6,400 (the cashflows tend to be slightly more front-end weighted due to various incentives the company offers for early payments).

NPV calculation: For defaults, I have assumed a normal distribution peaking during the 10th month (per management's guidance regarding typical default patterns) and eventually reaching 38%. A probabilistic model of the cashflows would yield the following.

p_i = I used excel's NORMDIST function to create a cumulative normal distribution and then multiplied times 38%
r_i = recovery value = COGS * 35% * remaining principal to pay = $1050 and declining
monthly payment assumes constant principal amortization + interest
month 1 payment: p_1 * r_1 + (1 - p_1) * monthly payment
month 2 payment: p_2 * r_2 + (1 - p_1 - p_2) * monthly payment
month 3 payment: p_3 * r_3 + (1 - p_1 - p_2 - p_3) * monthly payment
etc...

To get to the $6,400 value, you discount each monthly payment by the risk-free rate.

With a $6,400 NPV for the company's loans (taking into account defaults and recovery of collateral), it is straightforward to determine the value per car sold:

NPV: $6,400
COGS per car: -$3,924
SG&A per car: -$1,341
Interest per car: -$ 47
D&A per car: -$ 16

Pre-tax NPV: $ 1072
Taxes (35%): -$ 375

NPV per car sold: $ 697

Note on sensitivities:

Default rate - I find the NPV falls $100 per 4% you increase the default rate - i.e. taking it from 38% to 42% yields after-tax NPV falling from $697 to $596. I am fairly confident that the true multi-year average lies between 35% and 42%, thus yielding an NPV between $775 and $600.

Interest rate - I find the NPV falls $100 per 3% you increase the risk-free rate - i.e. taking it from 5% to 8% yields after-tax NPV falling from $697 to $597. This is prior to taking into account the likely higher interest rate they would charge on loans written in a higher interest rate environment (which would clearly occur). If you take into account higher interest rates on loans, the effect disappears.

Recovery - I find the NPV falls $100 per 20% you decrease the recovery as a percentage of COGS - i.e. taking it from 35% to 15% yields after-tax NPV falling from $697 to $597.

---------- Management, Culture, and Company Visit ----------

I had the pleasure of visiting Car-Mart in Bentonville recently. In order to believe the company will reinvest cashflow intelligently, grow margins, grow sales, and successfully stave off future competition, I believe it is helpful to get a close view of how the company operates.

Driving over from Tulsa, Oklahoma (nearest airport that Southwest Airlines flies into), Highway 412 probably typified many of Car-mart's markets. The highway was a series of large stretches of empty highway broken up by lower-end strip malls. In the town sections there were numerous BHPH car lots (buy-here-pay-here) - Car-mart is not without significant competition in all of its markets. However, for a company that is better at what it does than other companies, competition is not a burden, it is a proof of concept.

Judging management and culture is difficult and necessarily anecdotal, so take my observations as you will. I believe the management team is quite impressive. All of the people in charge of operations (this excludes the CEO and CFO) have many years (15+) with the company, and they have been the architects of 21+ years of profitable double digit growth. This growth has not been the result of the company being static - it is constantly evolving. As the president put it, it is a constant cycle of experimentation and tweaking. The vast majority of the employees and management come from modest backgrounds, but the company appears to have a very effective system of pulling talent up from within. For example, the store manager I met with seemed to fit the stereotype of a Southerner of modest background (at first blush, at least). He joined Car-mart at minimum wage after finishing his enlisted period in the military. As he describes it, he needed a job and they were willing to pay him. He says the thing that changed his sense of the possibilities for him at Car-Mart was when he heard his manager complaining about all the taxes on his $50k bonus check. He worked every role on the lot, and stood out in all the roles. In particular, though he hated it, he was good at collections (the highest turnover role they have). He made assistant lot manager after a little more than a year, and then got his own lot shortly thereafter. He has moved all over the company, including a stint in Texas as an underperforming lot fix-it-up man (i.e. the manager who comes in and improves a lot that has slipped some). One comment he kept repeating was that the management "took care of its employees... they will drop everything to help a worker out". In his case, this involved a nasty divorce while he was in Texas after which his family and kids moved back to Northern Arkansas. According to him, he was a wreck and expressed his need to be close to his kids to the president of the company. The management team made room for him in Northern Arkansas and it is apparent that things like this have earned his loyalty. That, and the fact that someone of his background is probably earning $100k - $150k/year now (though I don't know the exact amount).

His comments on the CEO, president and COO were interesting as well. As he describes them, they are just normal gregarious guys who are abnormally sharp (this was my impression as well). The store manager was surprised, for instance, when they had a problem with their computer systems and the president came in and altered the code (because he had written all of it). They now have a CTO who they hired out of Walmart, and so technology is no longer the president's issue. Nevertheless, I think the story goes to show the bootstrapped nature of the operation combined with what I believe is likely virtousity among management.

Other things I liked:
1) There was a fat book of best practices for every role in Car-mart that is updated quarterly.
2) Managers are constantly in contact with each other (as opposed to a top notch enterprise like Home Depot, where my (once again anecdotal) evidence is that they don't talk to each other as often, though they do have training seminars regularly). They have monthly get togethers where they discuss ideas.
3) District and regional managers are constantly on the lots helping keep an eye on ways to improve things.
4) Top management is constantly experimenting and improving. They have no shortage of answers to the question "how can you get better at what you do?". It sounds like they have lists a mile long for every executive on things they can do to improve margins, grow sales, and reduce risk.
5) They believe something that sets them apart is their focus on keeping the customer. They are much better at coming up with alternative payment plans for people who can't make their payment, or finding a job for people who lose them, than their competitors. This builds customer loyalty and lifetime customer value.

One comment the President made stuck out in mind as worthy of additional focus. Paraphrasing him, he said "as the company gets bigger, we've realized that there are times when we have been too much of a burden on our store managers' backs. We are very cognizant of our duty here at headquarters to help our store managers, and so we work very hard to always focus on simplifying while increasing their productivity". There are two aspects of this I really like: 1) Management realizes its job is to help the lot managers make more money, not get in their way for unnecessary reasons. 2) It shows the self-reflective nature of top management in their desire to make bigger better, not more cumbersome. I think both of these aspects are necessary to truly drive efficiencies through the company and market.

I don't know how often a company like Car-mart is the intermediate step in the evolution of an industry. I imagine there is often a trailblazing company in the early stages of most industries that shows that there are available economies of scale and information. Car-mart appears to be that company in BHPH lots. There are a few other semi-large BHPH lots, but it appears they have nowhere near the ability to generate high IRR on every single lot that Car-mart does. There is no escaping that Car-mart is in the business of selling cars to poor people, but it is important to realize they do it very well. I wonder whether, when Wal-mart was small, people had similar worries about their customers' exposure to things like rising oil prices, or just their generally poor clientelle.

---------- Company Valuation ----------

Last year the company sold 25,399 cars and is on pace to sell 26,800 in the year ending April 2006. The company believes that if it stopped opening new locations the current base of stores would probably sell 35,000 cars within 5 years. Furthermore, these 35,000 cars would be sold more profitably due to improved credit statistics in markets that have been open longer. This result makes sense in the context of likely lot-by-lot aging dynamics and chain-wide averages.

Management intends to open new lots at the rate of 8% - 13% growth per year. This will likely add another 10,000 cars sold 5 years from now. Looking at this company on an ongoing basis 5 years from now, the NPV per car sold * 45,000 cars yields NPV created of: $31.4 million

What multiple to place on a company that is growing annual NPV created at 5% - 10% prior to expansion? In addition, the multiple needs to be expanded based upon the company's ability to grow indefinitely at 10% - 20%/year. I will assume this justifies the company trading at 20x, very conservatively.

20x * 27,000 cars * $700/car = Value of $378 million.

Value: $378 million

Fully-diluted shares outstanding: 12.0 million

Value per share: $31.5, or up 70% from today's price of $18.5

Alternatively, one could consider what the steady-state valuation of the company is (assuming they stop expanding) and then add a premium on for the high ROIC opportunities they have of opening new lots:

Using my steady state 35k cars in five years, assuming some expansion in NPV per car to $800 by that point (due to more seasoned lots) and assuming that net income from that point grows at inflation, we would be seeing net income of $28 million. Applying a 15x multiple:

15x * 35,000 cars * $800/car = Value of $420 million.

Next, discounting back five years to the present at 10%, we would get a value of:

Value: $261 million

Shares outstanding: 12.0 million

Value per share: $21.7/share.

This is still a 17% premium to the current stock price, and this is prior to giving the company credit for the ability to reinvest cashflows almost indefinitely at greater than 20% post-tax ROIC.

Balance sheet:
I haven't mentioned the balance sheet yet, except in subtracting out interest expenses from the per-car value. The majority of the left-side of the balance sheet is finance receivables because the company turns inventory 15x/year. The company had $135 million net finance receivables (net of provision for losses and not included expected interest payments) out of $160 million in total assets. The company had $36 million drawn under its revolving credit facility. This yields a debt:receivables ratio of 0.27x and equity:receivables ratio of 0.83x. For comparison, Americredit (whose portfolio defaults at 30% - 50% of the rate of Car-mart's, excluding cars breaking down) has debt:receivables ratio of 0.91x and equity:receivables of 0.2x (note: I used a value of $10 billion for Americredit's receivables due to including restricted cash and investments in Trust receivables). It appears Car-mart could stand to lever itself more without running any reasonable risk of a liquidity crisis.

Stock buyback:
After the second quarter 2006 earnings call the company upped its share repurchase authorization to 1 million shares (out of 12 million outstanding). I expect to hear that they purchased a significant number over the last three months, as it is clear from speaking with them that they believe their stock is significantly undervalued, and has perhaps never traded at its fair value.

---------- Summary ----------

Car-Mart currently trades at $18.5/share. I think this is a business that, prior to considering excellent ROIC opportunities, is worth $21.7/share conservatively. After considering the ROIC opportunities, the company deserves to trade conservatively for at least $31.5/share and possibly much higher. The business is relatively unique, difficult to emulate, has enormous expansion room and is unlikely to face institutional competition for a long time.

The "risks" section will be the first comment.

Catalyst

- Operating profit per car trends back to historical averages as Katrina market dislocations fade.
- At 14x trailing twelve months earnings (and 11x forward) with 20% growth and a stable business, a catalyst is less important.