Tesla (TSLA) again, with fewer numbers plucked out of the air

I got a comment from Russ about my recent TSLA post, in which I concluded that a ton of future growth is already priced into the stock. The gist of his comment is that earnings could be a lot higher than I’ve been assuming. What powers a conventional car is an internal combustion engine–an expensive machine spewed out of large-scale factories that need to operate at close to capacity just to break even. The technology is also mature. In contrast, TSLA uses gigantic batteries, a technology in its infancy, i.e., one where costs are falling.

So I decided to abandon my back-of-the-envelope approach and take a look at TSLA’s 2019 financials. Call this my front-of-the-envelope analysis.

2019 sales and direct expenses

TSLA sold 367,500 cars in 2019, at an average price of $54,341 each. It made a gross profit from auto sales (I’ve ignored leases) of $4.0 billion, or $10,900 per car. Gross profit means after all direct cost of manufacture, including materials, labor and the cost of running the plant (including depreciation).

general expenses

The company spent about $4 billion on R&D + sales, general and administrative expenses (SG&A). This offset basically all the manufacturing profit. TSLA had interest expense + taxes, so it made a loss for the year.

where the operating leverage is

The essential point, which somehow eluded me last week, is that the general expenses laid out in the preceding paragraph have been steady for the past several years. Typically, an analyst would have them accelerate at some (low) trend rate of increase. Still, any increase in this expense number will be dwarfed in short order by the rise in gross income if TSLA continues to grow at the current high rate.

If we thought that TSLA could expand sales at 30%/year for the next half-decade, while retaining the same gross profit, then it would have gross profit in 2025 of about $15 billion. If general expenses increase at 5% yearly, they would be around $5 billion then. Net income would likely be about $7.5 billion, or close to $40/share.

new battery?

What about the battery? Let’s assume–I’m just making numbers up here–that the cost of parts is a third of total gross costs (three kinds of costs: materials, labor, factory) and that the battery is half of the cost of all parts. That would make the cost of today’s battery about $7000 per car. Say new technology cuts that figure in half. That would add $3500 to the gross profit per car for TSLA, assuming (unlikely, I think) it could retain the whole amount. TSLA’s unit sales in 2025 would be about 1.4 million at a 30% cagr, meaning $14/ share more in eps, if so.

As I’m writing this, the TSLA price is about $1700, implying a pe of just over 40x, assuming no changes in unit profits and 30+, on the same assumptions, if all the profits of new battery technology accrue to TSLA.

dealing with Tesla (TSLA)

If the question came up about whether to buy TSLA at $1400 (I have no intention to do so), how would I decide?

If I were making a recommendation to someone else I’d do a detailed spreadsheet in which I’d try to project the level of future earnings, the rate of their growth and how long I thought the superior growth would continue.

Today, I’m going to do a quick, back of the envelope, calculation. My aim is to get an idea of what level of future earnings is already reflected in the price of TSLA.

valuation

Let’s say the current price is $1400 a share. The market cap is $260 billion, implying that about 185 million shares are outstanding. Let’s say that I would be willing to pay 30x future earnings for TSLA, once it starts to earn on a regular basis. That implies earnings at some point of $45 a share, or earnings of about $8.5 billion a year, to justify today’s price.

Let’s assume this all comes from selling electric cars. Let’s put the average selling price at $40,000 and the net profit to TSLA from each at $4,000. If so, how many cars does TSLA need to sell to make $8.5 billion? The answer is 2.2 million units. That’s as many as Mercedes or Kia or BMW do and would put TSLA at the low end of the top ten global auto brands.

If I’ll only pay 20x earnings for an auto company, TSLA has to sell 50% more vehicles, or 3.3 million, to get the $1400 share valuation. That would put the company in the lower middle of the top ten, somewhere around Chevrolet, Hyundai or Nissan.

Is that doable? Well, the American car companies, operating in a heavily protected market, have been by and large pretty sorry companies for most of my lifetime. The Europeans are currently embroiled in a scandal that’s resulted from widespread falsification of emissions testing results for the diesel cars they sell in their home markets. The chaebol and zaibatsu conglomerate structures in Korea and Japan mean profits and innovation are at best secondary considerations. In other words, the competition isn’t particularly stiff.

The world market is about 70 million units annually, so 3.3 million would be around a 5% market share. Again, not impossible–although at the high end of what traditional auto companies have been able to achieve recently.

time

How long will it take for TSLA to be able to manufacture 2 million+ cars a year? The company was making them at a 400,000 unit annual rate at the end of 2019. At a 30% growth rate, it would take close to ten years. At a 50% growth rate it would take four or five.

In other words, today’s stock price is discounting very large growth for TSLA and paying today for earnings that are easily a half-decade in the future. What I think is significant about this is that in my experience the US stock market rarely discounts earnings more than two years ahead. How so? Pre-financial crisis studies (when there were lots more experienced analysts) show that securities analysts aren’t able to make accurate earnings forecasts more than a year ahead. Also, how far in the future the market is willing to discount is also a measure of market bullishness. I’ve rarely seen markets where investors are willing to pay today for estimated earnings three years in the future. Eighteen months is more usual. In bear markets, no one pays for any future earnings! In this case, though, the market is willing to pay for profits much farther ahead than in typical bull markets.

my take

I can imagine a world where TSLA is the leading maker of electric vehicles, with a global market share of, say, 10% and where electric vehicles are the dominant form of ground transportation. That outcome is not in the current TSLA price, in my view. But my guess is that if this happens, it’s also at least another half-decade in the future. Unless/until enough time passes that the market wants to pay for this, my guess is that TSLA will be at best a market performer.

I think the market’s willingness to discount far into the future is primarily a function of super-low interest rates. There’s also the sense that substantial structural changes to global economic life are in the offing and that it’s important to have a portfolio oriented toward companies of the future rather than those of the past. But if fixed income investments were to become more attractive–that is, if interest rates were higher–portfolios would shift toward bonds. At the same time, I think, the discounting mechanism in the stock market would become more conservative/less willing to look five-ten years ahead. That would spell trouble for TSLA …and many other tech-ish names.

my $260 price target for Tesla (TSLA)

A regular reader recently asked how I arrived at $260 as a price target for selling TSLA, which many of you will know I have been trading alongside my younger son for a while.

 

To some degree, I’ve regarded TSLA in the way I would look at any growth stock.  That is to say, I look for signs that the company can grow faster than the market expects and/or would sustain an above-average expansion pace for longer than the market believes.

I then try to use the company’s financial information, plus government, trade association and individual competitors’ data to project an income statement and flow of funds statement for several years into the future.

In TSLA’s case, reliable data are scanty.  In addition, the company’s own projections of its production capacity and sales have not been anything to hang your hat on.

 

Typically, and for a stock I would consider as a core part of my portfolio, I would try to make a projection that I believe to be reasonable, or perhaps mildly aggressive, and would reassess as/when the stock reaches the future target price that projection would lead me to.

In the case of TSLA, however, I did something different.  I decided to make an extremely optimistic projection, one I thought would leave the shares at least temporarily overvalued–and which would serve as a sell signal.

 

This is what I did:

–I decided to make my projection based on 2018 earnings, since I don’t think anyone has the faintest idea of what earnings for years farther in the future will be

–I took the TSLA production forecast, which has proved consistently too optimistic, of 500,000 cars during 2018 as a given

–I decided that most of these cars would be Model 3s, where the average selling price indicated in preorders is about $42,000, as another given.  To account for a small portion of more expensive models, I rounded the average selling price up to $50,000

–I examined other publicly traded auto companies worldwide.  The highest pre-tax margin I found (yes, although this is a dangerous shortcut, I did use margins) was Porsche, at 15%.  I assumed that TSLA would achieve that margin in 2018, even though I think the figure is substantially too high

–I assigned a financial reporting tax rate of 25%, which is probably too low.  That gave me net financial reporting profits of $2.8 billion, and earnings per share of $17.10

–I decided to apply a price earnings multiple of 15x to the earnings number, even though auto companies typically trade at single digit valuations.  I don’t think TSLA should trade at the 20x+ multiple that an IT stock would merit, so 15x seemed about right.  That gave me a valuation, based on 2018 earnings of $257–which I rounded up to $260.

 

As we can clearly see from a TSLA chart, the stock blew through my $260 target in March-April as if it weren’t there.  It peaked, for now, at least, at $386–leaving an embarrassingly large gap between me and it.

 

Where did I go wrong?

The obvious place is in the PE multiple.  The prevailing market view would seem to be that 2019 earnings will be, say, $25/share, and that it’s safe to be factoring anticipated earnings for two years from now into today’s stock price.  Another way of saying the same thing, one which I think must be TSLA bulls’ belief, is that TSLA is this decade’s Amazon (AMZN)

That may well be correct.  Personally, I’m uncomfortable making the bet, though.  One thing experience does tell me, however, is that, barring a total corporate collapse, the stock is highly unlikely to get back to the $180 – $200 range that has marked the lows over the past couple of years.

a bad day for Tesla (TSLA) shares

First, let’s put yesterday’s negative price action for TSLA–down by 7%+–in context.  Prior to yesterday, the stock had risen by 75%+ since the opening bell in January.  So a down day–or even a down few weeks–shouldn’t come as a shock.

What happened yesterday:

–TSLA reported 2Q17 results.  Profits were hurt by another production foulup–a shortage of batteries this time–that prevented the company from churning out cars at a higher rate.  The good news is that the problem was solved intra-quarter and shouldn’t affect results for the second half

–TSLA also said it intends to be churning out 20,000 Model 3s a month by the end of the year and 500,000 in total during 2018

–two negative analyst reports were released, arguing that TSLA is substantially overvalued.  Reasons:  plateauing demand for older models, increasing competition from other auto companies and TSLA’s less-than-perfect production experience.  Goldman Sachs says it now thinks the stock is worth $180 a share (down from $190 previously)

–Volvo announced it intends to become a exclusively a hybrid/electric car company in 2019; Baidu announced it will give its autonomous-driving car technology away for free in return for usage data.  Takers include a bunch of other Chinese carmakers + Ford and Daimler

my take

–I sold my last shares of TSLA at $260, based on the idea that this is the highest price I can reasonably conceive of TSLA trading at during 2018, assuming the company does indeed make and sell 500,000 cars.  I guess that’s my bottom line

–the negative reports are good news in the limited sense that they imply the authors’ firms see no possibility of future investment banking business from TSLA.  Maybe their negative analyst stance in the past has already ruled them out.  But emphatically underlining the fact suggests to me they think TSLA needs no further funding to carry out its production plans

–the possible turn to significant profits being earned in 2018 is a mixed blessing.  On the one hand, say, $5 a share in earnings for the company, with the promise of more to come in 2019 is better than the current situation.  On the other, the emergence of earnings–and of a more easily predictable future–means an end to the “dream” of unparalleled riches that many early-stage-company investors routinely harbor with any of their stocks.  For a certain percentage of “dream” stocks, the minute the earnings begin to arrive marks the peak in the stock price.  A minerals exploration company that owns a single orebody peaking the day the mine opens is the stock example.  Euro Disneyland is another.

Tencent (700:HK) now owns 5% of Tesla (TSLA)

The Chinese internet conglomerate Tencent filed a 13G form with the SEC yesterday, fulfilling its legal requirement to declare 5% ownership of a publicly traded US firm–in this case, TSLA.

Filing a 13rather than the better-known 13D indicates Tencent intends to remain a passive investor rather than seeking a voice in TSLA operations.

According to the filing, Tencent acquired its 8.2 million shares (at a cost of $1.8 billion) both by participating in TSLA’s public offering on March 17th and through market purchases.  Tencent reached the 5% level on March 24th.

When I first heard of the stake, it struck me as peculiar that Tencent would make open market purchases, in which the money goes to third parties, rather than arranging for a private placement of stock from TSLA, in which case all the money would go to fund TSLA.  Looking at the 13G a little more closely, however, I realized that Tencent’s total cost implies an average acquisition price of $219 a share, meaning Tencent has been patiently accumulating shares at lower prices.  Now I’m thinking that Tencent took part in the recent offering to provide some financial support to TSLA–and then rounded its position up to 5% during the following few days in order to file a 13G that publicly declares its backing.

the TSLA offering

The TSLA offering raised about $1.3 billion, through an issue of $400 million in common stock plus $1 billion minus in 2.375% convertible five-year notes.  The conversion price is $327.50, a 25% premium to the stock price at the time of issue.

The notes are convertible, at the option of the holder, but, practically speaking, only if they are trading at a 30% premium to conversion value.  To my mind, though, they represent a much better deal than fixed income investors have gotten in prior TSLA offerings.  This seems to me to imply that these buyers see much greater credit risk with TSLA today than they have in prior years.