Author's Note: This article has been revised to reflect the comments and corrections noted by the Seeking Alpha membership who read it as it was first published. I tried to acknowledge each of you for your contribution in the text and I apologize to whomever I may have missed.
It is a much more useful work as a cooperative venture than it could ever have been being written by me alone. Thank you all. I am learning a lot.
Retired Securities Attorney
The Potential Impact of Quasi-Lease Repurchase Obligations on Tesla's Future: Part 2
As some of you requested, and as I promised, I have revised the ranges of my input data to reflect the superb comments and criticism I got from board members. In the process of doing so, I noticed that my view of the relative risk/reward valuation in the stock was changing direction.
So first, let me thank all of you for your input. Getting that kind of investing support for free is a wonderful thing. Thank you Seeking Alpha for making it possible.
This is an extended, updated and revised article on the potential impact of the quasi-lease repurchase obligations on Tesla's (TSLA) future. It changes many of the assumptions that I built into the first article in response to good reasoning and input from the board. It also corrects errors in logic and calculations from the original of this article published this morning as a result of member input.
The results have changed. For those of you who want to save time, I'll give you the results upfront. The potential losses from the purchase obligations on the quasi-leases do not now seem to be material on the amended range of assumptions and inputs I used. You can skip to item 4 below where I start doing Tesla six-year projects on various assumptions.
This article contains some research and calculations on some key numbers and metrics in the analysis like: what is the average selling price of a model S; what is the probable resale price, which Tesla can get for the turned-in vehicles, etc. It also projects Tesla Sales and Income out six years based on a range of assumptions and metrics. The projections themselves are somewhat simplistic as of yet but they do serve to quantify the lease repurchase risk. The input assumptions and metrics range from a very pessimistic scenario (which does not need a spreadsheet calculation) through a moderate scenario to very optimistic scenario. As before, I have uploaded a configurable spreadsheet, which you can download and use to pick your own scenario. If you change a key number, all the other numbers in the spreadsheet will automatically update. I have uploaded the spreadsheet here.
This analysis deals only with income streams from the sale of automobiles. No other potential sources of revenue are factored in. For starting sales numbers and sales projections, I used Tesla's non-GAAP pro forma accounting, which does not book quasi-lease repurchase obligations as liabilities on the balance sheet but rather treats them as completed sale. I did not take any account of the potential costs of Tesla's warranty which, on June 30, 2013, had a booked GAAP liability of $28 million, or about 7% of the then-current quarter's sales.
Key assumptions are the sales growth rate; the rate at which the sales growth rate increase or decreases over time; and the net profit margin Tesla will achieve on those sales.
Meanwhile, here are the assumptions I made for this article. Please feel free to disagree with some or all of them.
1. I calculated the Average Selling Price for a model S sold (not leased) in the second quarter 2013 and came up with an incorrect number. I have now corrected it thanks to input from Logical Thought and CoverDrive. The Average Selling Price of a Tesla model S in the second quarter 2013 was $90,408. I derived this price by dividing the total revenue from GAAP sales by the total GAAP number of vehicles sold. I got the revenue number from the company's 10-Q as filed with the SEC, adjusted for the non-auto sales numbers pointed out by Logical Thought and CoverDrive. I got the number of vehicles sold from the company's press release. I got the percentage represented by quasi-leases from interviews Elon Musk gave to reporters. I ignored operating lease GAAP revenues booked as automobile sales because it was immaterial at $1.9 million, less than a rounding error.
This is a little bit less than my original assumption, which was $96,070. I got that $96,070 number by pricing a middle-of-the-road model S on the Tesla website, not a particularly sophisticated way to do it. I want to thank contributor 4121 who suggested that I redo the numbers using the best criticisms from the board and do another with new variables. This is a valuable learning experience for me.
2. On the lease repurchase guarantee of 50% of the base price and 43% of the option prices, the repurchase blended average repurchase price for this vehicle is 48.43%. or $43,787.
I tacked on re-marketing costs of $8,650 for a total of repurchase cost to Tesla of $62,151. I cut the prepping costs by 80% thanks to the board's input; cut advertising in half and left the commission since I still think it makes sense to pay one in order to move the turned-in vehicles. Many members think that a commission will not be necessary but I still disagree. You can change all the numbers in the spreadsheet to match your own viewpoint and the numbers will all update across the spreadsheet.
3. I estimated the resale price as a percentage of the original cost using data from three-year-old Tesla roadsters and roadster sports advertised for sale online and on the Tesla website for the original cost and list prices. I want to thank stephenpace and Rik for pointing me to data on Roadster original cost and resale prices. These are all very low mileage vehicles and average 8,473 miles each. I expect a turned-in Tesla model S will have 45,000 miles or more on it and so will be priced as a vehicle with that mileage. Since lower mileage cars sell for more than higher mileage cars, I went looking for a metric to adjust the projected selling price. I used two.
First, to get an adjustment for the difference in the cost to buy a certified vehicle from a dealer with 8,473 miles on it (the average mileage of the listed model S's) and the cost to buy a certified vehicle from a dealer with 45,000 miles on it (which is what I expect the turned-in model S's will have on them), I used Kelley's blue book values for a three year old certified used Mercedes-Benz CLS550 Coupe, which is about the same price as the model S. I thank Jack Rice for pointing out that the buy-back benchmark was based on the S class Mercedes because it has a steep depreciation rate. It's not perfect since the two cars won't wear the same way but it'll give us an idea. I priced the Mercedes identically equipped on Kelly's Blue Book both with 8,473 miles and with 45,000. I then compared the two prices and found that the higher mileage vehicle costs 11.5% less. I used the 11.5% to lower the list prices of the listed model S's to account for their unusually low mileages. I also tried deducting 15 cents a mile to see what that did. The percentage deduction produced a lower resale value than the mileage adjustment. At 23 cents a mile, they're about equal.
I assumed that the seller would take 12% less than the advertised price to get the vehicle sold. I have now tracked the seven online listings for eight days and none have been sold. Haggling is a way of life in the used car business. Like everything else, the 12% is configurable in the spreadsheet.
The vehicles listed on the Tesla website produce a presumed resale value of 51.8% of original cost with these parameters. The vehicles listed online produce a presumed resale value of 59% of original cost with these parameters.
The resale calculation is a complicated worksheet in the spreadsheet with a lot of variables. It's too wide to post a full picture here so I hid many of the columns.
I wasn't satisfied with this result. I want to know two more things.
1. How much change does it take to produce losses which are potentially substantial and in what variables? I wanted to know how delicate the numbers are; how sensitive the results are to small variations; and which variables have the more significant impact on the end result.
I did the work and I found that the results make the potential impact of the quasi-lease obligations on Tesla's future no longer so much of a cliffhanger. Change one or more input parameters for the resale price a little bit for the better or a little bit for the worse, and the impact on Tesla is no longer material. For example, doubling the discount the buyer of a used vehicle gets off list hardly moves the needle at all. You have to change key inputs a lot to move the needle.
2. How serious is any realistic potential negative impact? What if the input parameters do move a lot. Is Mr. Musk betting the farm on the resale value of model S's? It's all right to bet the farm, especially if you have another farm or three in the next state. I've done it more than once. But I always knew that I was betting the farm when I did.
So I had to go back and do more work.
I needed a number of years projection of Tesla's sales, income, and the quasi-lease aggregate liability. I decided on six years since I'd already roughed that out before. This projection can be in annual results since I need only rough numbers for a sensitivity analysis.
4. So I went out six years using the prediction by Elon musk for 2013, 2014 and 2016 vehicle sales. I don't remember him saying anything about 2015. I just filled it in the middle between the 2014 and 2016 numbers. I used the ASP I derived and 20,000 vehicles for 2013, 40,000 for 2014, and 100,000 for 2016. I put 2015 2017 and 2018 at reasonable numbers. I came up with a projection that had some logical errors in it, which Michaelgeis was kind enough to point out. This is the corrected version:
This is a simplified projection for Tesla's results six years out using an assumed capitalized annual growth rate of 100% for the first year and slowing it down at increasing deceleration rates starting at 29% to a sales growth in the sixth year of 51% over the prior year. The net profit margin remains consistent at 15%, which I think is reasonable. I grew the share count at 25% a year to take account of capital needs and option exercises. It is non-GAAP.
The two key configurable rates in this spreadsheet, expressed as percentages, are those two rates. One is the rate at which sales grows in each quarter over the previous quarter. It can be changed to suit your own estimate; held flat by entering a 0; or made into decreasing revenues by entering a negative percentage. The other is the rate by which that sales growth decelerates each quarter, expressed as a percentage that the first rate slows each quarter. Normally, the super high growth of disruptive technology companies have slower growth rates as they mature. One reason is that they start to run out of first-adopters for whom price is not significant; competitors start copying their technology or producing their own. But this may not be a "normal" situation.
The moderate scenario results in a sixth year projected EPS of $15.57.
I know some of you think that Tesla's growth rate is going to accelerate over the future. I think it is a possibility, which cannot be ignored. What if, for example, Tesla's lead in battery technology and already-there superchargers results in more major auto companies buying Tesla drive trains. Mercedes and Toyota have signed on. What are Ford, GM, Honda, etc. going to do? What if Tesla actually can build a good $35,000 to $50,000 mass-market model E EV? I don't know how likely either or both of these scenarios are but they are certainly within the realm of the possible and should be considered.
Here is a projection that considers the scenario of a successfully developed $35,000 mass market model E EV with a fully built out supercharging network in the US; a partially built out supercharging network in Europe; and some build out in the rest of the world.
I grew sales for the model S the same way as under the moderate scenario. I started model E sales in 2015 at 20,000 vehicles and grew sales from there at 100% accelerating at a capitalized annual growth rate of 25% as more capacity comes online. This huge annual sales growth and acceleration assumes that the model E is so popular that it is continuously on back order and that Tesla has access to virtually unlimited capital since it looks to the market like Tesla will come to dominate the world auto market. I am not saying that this will happen – just that it is within the realm of the possible and should be taken into account. I grew shares outstanding at a quite large 30% annual growth rate to take account of a massive need for additional capital and the exercise of outstanding stock options. Thanks to Irfrexuberance for pointing out that my share count wasn't growing. I corrected the formula error. I eliminated accounting for quasi-lease losses since they will not be material under this scenario. I used an average selling price for the model E of $35,000 although I personally believe that it will be higher than that.
The much better scenario results in a sixth year projected EPS of $58.98. At a $50,000 ASP for the model E it would be $73.42.
A couple of questions occurred to me:
1. What is the worldwide addressable market for model S vehicles at about $100,000 each. Assuming that the typical buyer would not want to spend more than 5% of his/her net worth on a vehicle limits the addressable market to multi-millionaires. I searched for an answer online and got this. It says that there are 100,000 multi-millionaires in the world. But the same site says that there are 1,000,000 multi millionaires in the US. Oops. So we don't know. But I'm going to call it at .0005% of world population or 3 million individuals. Half of that is enough to keep sales going for a long time. Super luxury car sales combined however, for June 2013 in North America were 5,081 or about 60,000 annualized. I read somewhere that North America is less than half of the worldwide sales but I can't find it now. That's a big enough market for the model S for awhile but puts a cap on growth in the fifth year and later years unless the market expands dramatically. So I held production steady after the fifth year.
2. What is the worldwide addressable market for the model E. There were 44,435,330 autos produced worldwide at 11:15:13 am on August 31, 2013 PST. The site continuously updates. Toyota sells about 64,000 Camrys a month worldwide.
Anyway, it's plenty large enough.
Lastly, just for fun, I went out another three years to give the model E a full six years to impact sales. The ninth year EPS at an unrealizable growth rate of 244% from the sixth through ninth years reaches another astounding level.
On the other hand, what if Ford finds a better battery technology and patents it? What if some 18-year-old young lady invents a battery that can fully charge a mobile phone in 25 seconds and someone patents a way to use it for EVs? Scenarios like these are also within the realm of possibility and prudence dictates that we should take account of them also.
Some of the pessimistic scenarios are easy and do not require any calculations. If Ford or someone else obsoletes Tesla's battery technology and this happens after Tesla has borrowed a lot of capital to expand its operations, Tesla goes bankrupt. This result is true of mostly all scenarios where Tesla expands; borrows to do so; books unfunded liabilities in the process like the lease repurchase obligations; and then has the bottom kicked out from under it for one reason or another. It goes under. This is a risk for all new businesses and Tesla is no different. You gotta take the risks to reap the rewards.
There is one somewhat worst-case scenario that could be quantified. That is the one where Tesla cannot develop a lower priced EV and competition limits its addressable market for the model S. That will possibly put an end to its ability to grow sales and they will begin to shrink as customers who can afford a model S either get one and are still driving it ten years later or opt for a competitors offering. The cap that the sales figure reaches; when it reaches it; and how much sales shrink on a compounded annual negative growth rate basis can be projected with those inputs and rates estimated. But those inputs and rates are so speculative at this point that I would be randomly guessing to pick any. So I'm not going to try at this point and just say the result may be that Tesla stock proves to be overvalued at its current level.
I look forward to hearing the board's take on the assumptions I used and the results I reached.
I'm also looking forward to hearing about the logic or calculation mistakes I made. I'm sure I made some. I always do.
Additional disclosure: I may put on a volatility spread by going long out of the money calls and puts at the same time within the next 72 hours.