Graham or Growth? Part 4 - The Car Industry

May 2019
James Anderson

The value of any investment and any income from it, can fall as well as rise, meaning that investors may not get back the amount invested.

When we first attempted to explore and explain our enthusiasm for Growth Investing 15 years ago it was natural to try to learn from our predecessors. The problem was that there was very little literature to guide us. The only text in the canon of investment that espoused Growth Investing was Phillip Fisher’s ‘Common Stocks and Uncommon Profits’. It dates from 1958. What follows is James’s musings after he reread Ben Graham’s value bible, ‘The Intelligent Investor’.

There are few, if any, industries where the decibels of disagreement about stock selection are as rancorous as in the car industry. Why this should be is something of a mystery to me as I have no great enthusiasm for the product. In the last year this has reached fever pitch as amidst more conventional debates the rise of EVs and the very specific controversies surrounding Tesla have reached new levels of hysteria. Amidst the volume of noise the Graham comparison technique seems to offer room for gentler reflection. In this industry we are better off considering a wider set to capture the full range of companies from value to brand to disruptive potential. So here are five examples running the full gamut of the industry and the stock market:


The first incarnation of General Motors created a large amount of value, despite bankruptcy in 2009. The marvellous Bessembinder study suggests that GM ranked as the eighth largest value creator in US markets since 1926. The new GM has already taken on many of the hues of the old – apart from its repute for managerial incompetence which has been impressively dispelled by Mary Barra’s team. This has included a vigorous attempt to innovate in both the EV segment via the Bolt and autonomous vehicles through Cruise that have been the content underpinning a new slogan ‘Zero crashes, Zero Emissions, Zero Congestion’.

For 2018 GM has announced automotive free cash flow of $4.4 billion on a volume of 8.38 million vehicles on earnings per share of $5.72. According to the Graham formula used previously, GM is discounting moderately shrinking returns in the future (though it may not be a surprise that analysts forecast 8.5 per cent annual growth). So GM appears to have a value case if you believe that it will contain its vicissitudes to “ups and downs”.


BMW probably stirs more favourable emotion than any other German company or car company. Drivers usually love the product and analysts are just as prone to superlatives about the calm and collected approach to data and finances. Yet despite this BMW meets the Graham standards for an out of favour blue chip. The shares have persistently fallen from a high of €120 in 2015 to a level of €70–75 in recent months. Perceived virtue has not been its own reward.

BMW generated €4.46 billion (approx. $5.1 billion) in automotive free cash in 2017 on unit sales of a little under 2.5 million. After three quarters in 2018 cash flow was down by 25 per cent. Although the car industry demands respect for cash more than earnings BMW currently trades at 6.5x likely 2018 earnings. These earnings have been hurt by multiple factors from diesel regulation to loss of Californian market share to weak market conditions in China. The multiple assigned to the earnings and free cash has fallen even more sharply. This translates into the Graham equation predicting a state of affairs similar to GM in the next five years: BMW too discounts annual earnings shrinkage in the low single digits. Plainly this is more of a break with the past decades than for GM.


For those who believe that the car industry is dependent on large volumes and is bereft of profitability above the cost of capital Ferrari is the exception that proves the rule. In 2018 Ferrari made €405 million ($463 million) of industrial free cash which was an improvement of 23.5 per cent on sales of just 9,251 cars. So this translates into free cash of €43,779 per vehicle or over 10 per cent of GM’s absolute dollars on just over a thousandth of the production volume. The Graham formula indicates that at a share price of €110 Ferrari is still thought capable of compounding its earnings at around 12.5 per cent in the next 7–10 years. There are no analyst forecasts that go out this far but this is below the five year forecasts that do exist.

Many observers of Ferrari opt to assess it as more of a luxury brand than a car company. But chairman and Agnelli family leader John Elkann disagrees with this. He believes that unless Ferrari is at the cutting edge of automotive technology that brand, let alone nostalgia, will not keep it thriving. What is clear is that if the returns are sustainable then this is a fabulous business. Selling fabulous businesses is something we should be wary of doing on a regular basis. Or as Philip Fisher wrote the time to sell is “almost never”.


The Graham approach with its reliance on simple financial data should be still more valuable in assessing Tesla. Emotions in capital markets are seldom higher than with Tesla. That comes as a persistent given between the founder, the short positions and their associated personalities and the significance of the project to both backers and critics.

So what do the basic numbers say? Tesla ended 2018 with free cash flow of -$2.377 million. This red zero is the amalgamation of two contrasting halves: an outflow of $1,793.5 million for the first six months followed by an inflow of $1,791 million in the last two quarters. It’s plain that the finally productive ramp of the Model 3 was the cause of the stark difference between the two periods but there is still room to argue about the future pattern. Of course we’re happy to argue that the second half is the lead indicator but that is unfair to the Graham reticence to predict especially when optimism is involved. What is clear is that to judge Tesla we need some forward-looking contentions. I’ll return to this after a last example.


For all the controversies surrounding Tesla it’s actually a very basic investment story. It’s completely clear that it dominates the EV competitive landscape in America and Europe. In fact as a pure play it has basically zero competitors. In terms of time frame it may be 6–7 years ahead of anybody else in its markets. If you gain conviction in the appeal and economics of EV technology then the case for Tesla is simple. But for NIO the challenge is very different.

China appears to have close to 500 EV manufacturers. The chaotic and exuberant scene is much like that of the nascent US automotive market before the transforming rationalisations of first Ford and then GM. NIO is far from being the current leader with BYD the historic leader. NIO has few claims to technological leadership and indeed most of its operations are outsourced. NIO has no imminent prospect of turning cash flow positive. It’s uncontroversial to suggest that Graham would not have owned NIO.

Nio car


If we try to put this all together, how should either a fair minded Value or Growth investor feel about these five car companies? However hard I try I find it hard to see a plausible mean reversion explanation of the past or a convincing hypothesis based on it to guide an assessment of the future. This applies both to the business fundamentals and to equity outcomes.

Although it’s probably not what mathematicians would identify I’d hazard that the commonest day-to-day practical incarnation of mean reversion in the car industry is the common view that says that it’s the profit on every vehicle produced that must mean revert. That’s why I mentioned vehicle sales numbers for several of these examples. This is most frequently aimed at Tesla on the basis that its 245,000 sales in 2018 make it inconceivable that it can be worth as much as GM (or Ford or potentially even Toyota). Oddly though, very highly reputed and very self-confident auto analysts find it impossible to imagine that the historic premium returns and premium market rating of BMW are anything other than a permanent and sacrosanct feature. Ferrari meantime is seen as such an extreme outlier that, as we’ve discussed, the standard approach is to refuse to see it as a car company at all.

But surely the dominant characteristic of this industry and its stocks isn’t mean reversion. It’s been deep cyclical uncertainty and it’s now sudden change. It’s very much analogous to the story of the turkey. Every day it sees the farmer approach and every day without variability it receives a decent supply of food. It looks like the perfect stock – constant dividends with no volatility! That’s the very definition of low risk.

But then the farmer kills the turkey. This is not far away from GM. From September 1908 to June 2009 come wealth, crisis or war GM survived and mostly prospered enough to mean that even eventual bankruptcy could not remove its proud status as one of the best lifetime investments in US history. So if you could have the owned the shares for 101 years your institution was very happy. If you bought in 2008 you were less so. But of the mean in performance there is little to be seen.

Is there any way through these thickets? I’d suggest that there is but that it requires re-thinking most of the traditional verities. It seems clear that the automotive industry is subject to such wild lurches that picking one outcome is remarkably foolhardy. What we can do to ameliorate the situation is to acknowledge existential doubt. The notion of one forecast of the future, of one expected growth rate and one associated discounted cash flow analysis is just too simplistic. This issue arises in comparing Coke and Facebook’s upside but here it’s much more serious and dramatic. We need multiple versions of the conceivable futures stretching from the scarcely imaginable best case to the end of the turkey’s life.

With considerable humility it’s then probably helpful to probability weight these different scenarios and update as the path develops. In general it’s imperative to push the boundaries of the extreme cases much further than analytical caution usually permits. This isn’t about a conventional bull and bear case. On the upside indeed creativity rather than analysis has to be the focus. We’re back to Charlie Munger and Atlanta in 1884, as discussed in the last article. In the opposite direction it’s better to assume bankruptcy is an ever present possibility for all stocks unless argued explicitly otherwise. That’s salutary. It doesn’t leave much room for a Margin of Safety. We cannot be safe and investors. But it may leave the possibility of significant upside.

If we adopt such a process for the automotive stocks then I think we come to a more realistic perspective and one that is considerably more likely to pay off especially when put in the context of an overall portfolio rather than left in isolation. To illustrate this we will take a closer look at the return spectrum for Tesla and NIO.

What persistently surprises me is how straightforward it is to construct a roadmap for Tesla being worth many times its current market value. It also requires less imagination than in most of the investments we make. This is probably because Tesla ‘only’ needs to capture currently existing markets to have dramatic potential whereas many of the internet platforms have had to create a new world. On the other hand, whilst it’s unlikely that Tesla goes bankrupt (this was a plausible outcome even six months ago) it’s still quite possible that the share price could fall 75 per cent. But isn’t this the type of skewed, asymmetric and to many frightening sets of potential pay-offs that we should welcome. One way or another it will be growth at an unreasonable price.

"But surely the dominant characteristic of this industry and its stocks isn’t mean reversion. It’s been deep cyclical uncertainty and it’s now sudden change."

Let’s first simply consider Model 3 economics on their own. Unit sales could easily reach 1.5 million a year even considering just the current markets for the luxury producers before any beneficial demand shock (akin to the iPhone compared to Nokia). We know that aggregate pricing points will exceed the $35,000 base ambition by some way. So revenues of $75 billion are conceivable. Current operating margins are 5.8 per cent but the consistent long-run target has been double digit (and indeed up to 30 per cent at the gross level). The unparalleled control of the battery supply, its technology and data make the competitive advantages plausibly sustainable.

Remembering that it isn’t a central case we’re discussing but an exploration of the possible upside, let’s posit operating margins of 20 per cent (Ferrari makes 25 per cent) and net margins of 16 per cent. That would give earnings from the three of $12 billion per year with free cash of similar dimensions as capital expenditures would principally be of a maintenance type. Ferrari sells on a free cash flow yield of 2.5 per cent so let’s say 3 per cent for this lesser comparator. That gives us a putative equity value of $400 billion in five years’ time. What are the chances of such a scenario happening? We are dealing with an unlikely but far from outrageous scenario. The competitive moats seem secure over such a time frame. Customer demand and satisfaction seem supportive. So let’s be conservative and say there’s a 20 per cent chance of such an outcome for the Model 3 project.

But, of course, the Model 3 is most unlikely to be the sole contributor. The crossover Y series soon to be unveiled attacks a market at least as large and with even greater pricing indulgence. Then there is the Tesla Truck. Then there is Tesla Energy which is gradually showing its potential under more vigorous leadership. There’s the underpinning emphasis on software and software driven upgrades but all this is before the most unlikely but potentially most rewarding opportunity of all in the form of autonomous vehicles. That Tesla has an unusual approach to this challenge should surprise no one. The attractions are that its chances of success seem to be inching upwards from the highly improbable to the merely unlikely and that the size of the prize is hard to estimate but large in the extreme. Adding these possibilities together make it seem that Elon Musk’s comment to us six years ago that: “There is a small but growing possibility that Tesla will be the largest company in the world” represents a now highly realistic scenario. None of this means that it is a certainty.

For NIO the situation is substantially more unpredictable. It’s quite clear that the range of outcomes we consider must include a substantial possibility that the shares prove to be worthless. I don’t want to inflict too much detail on the reader but our parameters go from a 30 per cent chance of zero or its functional equivalent to a 5 per cent chance of making a 65x return. This is a world of deep uncertainty and few, if any, anchors. In Nicholas Taleb’s terms it’s Extremistan. I find the idea that there can be a Margin of Safety here extremely odd and a mean to revert towards is once again a meaningless nirvana. What matters is the path dependent but attractive potential asymmetric pay-off. Naturally NIO needs to be a small portion of an adequately diversified portfolio until it has negotiated the terrifying chasms that it has to cross. But we think NIO’s leadership and strategic bravery offer that chance.

The dominant point that I’d like to establish is that whilst for both Tesla and NIO there are severe challenges, many outcomes and no certainty about the future path there is undeniable and asymmetric upside. Probability adjusted upside is substantial. For BMW and GM we find this a great deal harder to establish. If the former proves more resilient to the challenge of EVs then it might go back to prior valuations but that would not multiply the value of any holding. For GM it seems to us that upside is dependent on the moves to electric and autonomy succeeding. But we know transformation of an established company is very tough.

The views expressed in this article are those of the author. Its express purpose is to highlight areas of intellectual thought and debate which inform the investment philosophy that underpins Scottish Mortgage, in the hope that they may be of wider interest. The author(s) therefore make(s) no suggestion that this article constitutes independent investment research and it is not subject to the protections afforded to such.

Further, it is not intended to be considered as advice or a recommendation to buy, sell or hold any particular investment. As referenced above, private companies may be more difficult to buy or sell, meaning short-term changes in their prices may be greater. Those considering investing in any of the areas highlighted in the article should undertake their own research and seek advice if unsure. For those looking for information on Scottish Mortgage specifically, please visit