I work on various financial portfolios, two of which focus on US equities. In both of them, we held Tesla at a zero weight in the beginning of this year. One of the reasons we held Tesla at zero weighting, even before recent public scandals, was that it performed poorly on our forensic accounting analysis. I've written before about the topic of forensic accounting, which, when applied to investing, is the science which examines a company's public financial statements in order to ascertain the risk that those statements have distorted the actual financial position of the company. Typically, this means that companies engage in 'earnings management' in order to make the most recent period look better than the actual reality. This can be done by exaggerating revenues or by under-stating expenses. Forensic accounting is not designed only to detect actual fraud (though it can be used that way), because most distortion isn’t fraud. When we flag a company for forensic accounting problems, we are not accusing them of illegal fraud, and we’re not necessarily even accusing a company of legal forms of financial distortion: it’s more like we're identifying a higher risk of shenanigans (or is it rigmarole?). We don't have to be sure, because with so many thousands of companies to invest it, we can afford to be on the safe side and exclude a company which shows significantly greater risk of not quite playing straight with investors, and buy one which is at low risk.
When it comes to Tesla, it was not too tough a choice. Out of 13 tests we apply, Tesla had significantly worse than average scores on ten. Three were above average, but only one of those was a very high score, namely payment delays.
Without going through each category in great detail, we'll look at overall patterns. Two really jump out to me. First, this sure looks like a hype stock. We'll see that more clearly when we look at valuation, but even before then, forensics paints a picture of a company which has an encore problem: meaning a company with past revenue growth so high that it will have trouble keeping up with expectations. This creates incentives for distorting earnings and it also creates incentives to rush production in order to keep the momentum going. Both are big warning signs, and they were flashing before Tesla's production delays were exposed in the press.
The other pattern which really jumps out at us is that this a company which is so capital-intensive that it is going to be very difficult to get a decent return on assets. Tesla has a great deal of capital equipment, very expensive equipment. My colleague, Jim Huntzinger, who is a manufacturing engineer, has written about this. There is such a thing as too much automation, and a company with a public persona of being ultra-high tech is exactly the sort of company which would be tempted to fall into that trap. The principle of human productivity certainly is not hostile to automation in and of itself: machines are extensions of human beings and magnifiers of our productivity. But there still are jobs that humans can do better (and cheaper) than machines can. The forensic accounting analysis above is certainly consistent with a picture of a company which has over-invested in expensive machinery to a degree which makes it very difficult to sustain a good return on assets.
That's not the only reason that we held it at zero, there's also the issue of valuation. Even a company with numerous problems can eventually get cheap enough that investors can have a decent probability of meeting reasonable return goals. At the beginning of this year, was Tesla pricey or on sale? We'll look at that next time.