"The report of my death was an exaggeration."
-Mark Twain, June 2, 1897
Tesla Motors (NASDAQ:TSLA) is a manufacturer of expensive electric cars. Tesla's stock is expensive, too. Tesla's stock price has quadrupled in the last year and, despite a recent pullback, still trades for 40 times sales, 126 times book value, and an undefined multiple of profits, since the company has never had a profitable year.
It is hard to justify Tesla's current stock price. Even Tesla's CEO, Elon Musk, recently said that the stock price "is more than we have any right to deserve." Musk went on to say that the current price might be justified by profits far in the future, but even that case is hard to support. NYU finance professor Aswath Damodaran considered a pipe-dream scenario in which Tesla grows 50-fold over the next 10 years and ends up with the revenues of Audi and the profit margins of Porsche. Even in that pipe dream Damodaran could not justify paying more than half the stock's current price.
What pin could prick the bubble of Tesla's stock price? One possibility would be a decline in the demand for Tesla's cars. SeekingAlpha contributor Paulo Santos recently reported that he had seen early signs of such a decline. I have very high regard for Santos' work and briefly considered taking a short position on Tesla after I read his article. After probing the data more deeply, however, I have concluded that in this case Santos overstepped. The data do not support the case that Tesla's sales have hit the skids.
Santos' article was based on data collected by Craig Froehle, a business professor at the University of Cincinnati. Since June 2013 Froehle has been encouraging new Tesla buyers to provide him the last five digits of their Vehicle Number (VIN5). The VIN5 is a serial number, so the trend of VIN5 values over time can offer a rough portrait of the rate of Tesla purchases over time, even if only a fraction of buyers actually report their VIN5s. For example, when Froehle's thread started in early June, new buyers were reporting VIN5 values between 14500 and 15100; more recently, in early November, new buyers have reported VIN5s between 27000 and 27600. This implies that about 12,500 Tesla cars have been purchased in the past five months (12,500=27,000-14,500; also 12,500=27,600-15,100), for an average rate of about 2,500 cars per month, or 555 per week. This figure is compatible with Tesla's latest shareholder letter, which reports that the company is now producing 550 cars per week.
What caught Santos' eye was the two-month trend in Froehle's data, which is reproduced below. The blue points and blue line plot VIN5 values against the date on which they were issued. The points have recently gone from above the linear trend to below it. The red points give the implied rate of purchases over a sliding 2-month window. These sliding windows also suggest a slowdown; two-month windows ending in September or October imply the purchase of 650-725 cars per week, but more recent two-month windows, ending in November, imply a slower purchase rate of less than 580-650 cars per week.
Santos finds the two-month trend concerning, and I would, too, if the two-month trend were the only data available. Elsewhere, however, Froehle provides the following five-month trend, which puts the recent numbers in a larger, more reassuring context.
(click to enlarge)
The five-month trend sticks close to a straight line, implying that purchases have been pretty stable over the whole period. The data points sometimes go above or below the line for a few weeks, but these brief departures have not been predictive of anything more sustained. In particular, there have been two previous occasions-from 7/13 to 8/10, and again from about 8/20 to 8/25-when the data points fell as far below the linear trend as they are today. After both dips, VIN5 issuance recovered and in fact went above the linear trend for much of September, before dipping again in October. It would be instructive to check the series again in the spring and see if the October dip predicts a longer-term flattening of the sales trend. I suspect it does not.
While the five-month trend does not suggest that sales are slowing, it doesn't suggest that they are accelerating, either. That may seem odd for a company whose story depends on growth, but it's actually compatible with statements by Tesla's management. In its most recent shareholder letter, Tesla management repeated the claim that "demand exceeds supply." According to management, Tesla is making cars as fast as it can; consumers are buying the cars as quickly as they are produced; and sales will rise as soon as Tesla can increase production. Bulls may believe this, while bears argue that the latent demand for Tesla cars is much smaller than management's statements imply. It will be hard to settle the argument until Tesla starts to crank out more cars.
I do believe that Tesla's stock will fall eventually, and I expect that Tesla's sales will plateau well before the company grows as large as Audi. That said, I believe it is premature to look at the October VIN5 issue rate and conclude that the end is nigh.
The situation reminds me of the 1897 reports of the death of Mark Twain, which also originated with valid but misinterpreted data. (Twain's cousin was deathly ill.) Twain sent a famous correction to the New York Journal, noting that "The report of my death was an exaggeration." Twain did die eventually, but it took another 13 years. I am not sure how long it will take for Tesla's stock to fall in line with the company's potential earnings.
Acknowledgment: I thank Craig Froehle for permission to use his graphs.