That's sort of a big miss though, factoring in Forex. You can come up with excuses for every quarter, which
I've done in a previous post.
Bottom line, gigantic miss.
If Wall Street was off by that much we'd all be ridiculing them for how dumb/corrupt they are, so need to hold it both ways. While it may not be you explicitly who bashes traditional analyst estimates (too lazy to go through individual post history like some vindicitive wackos here), I see a lot of unfair slander thrown at a lot of these analysts.
...
At the end of the day we're all trying to make projections (short or long) to the best of our abilities, and the biases that we may have does not necessarily mean that one person's word is superior to another.
So I hope, moving forward, some respect is shown to everyone (except JPM and Gordon Johnson), about their thoughts. This should be a humbling experience for you, Rob, The Accountant, James Stephenson, etc. who had ridiculously divergent price targets.
The shade thrown at institutional analysts (with the exception of a few such as Alex Potter and Pierre Ferragu) is completely fair, because their estimation methodology completely sucks. Personally, I will explicitly bash their estimates. They have been very wrong in general and will continue to be very wrong.
Over and over again, institutional analysts have underestimated Tesla's performance, especially over the long haul but also on a quarterly level, as shown in the following table and charts.
The quarterly data uses institutional analyst consensus EPS expectations compiled by Zack's for each quarter since 2019. The average quarterly EPS prediction error was
-25% of the actual EPS, with a cumulative error of $1.49 since 2019. This track record is especially egregious considering that all of these estimates were published immediately before each earnings release,
after the production and delivery numbers were known, which (if we give analysts the benefit of the doubt regarding their good intentions) can only indicate that analysts have been consistently failing to understand the factors affecting Tesla's revenue and cost.
The annual forward-looking estimates for both vehicle production and earnings are far worse. I tried to find compiled data taken in December for each subsequent year, but couldn't find it. Instead, I resorted to looking up a few reports published by individual analysts.
If analysts had a 50/50 chance of guessing quarterly numbers too high or too low, the odds of guessing low on at least 12 of the last 13 quarters due to bad luck is 1/binom.dist(1,13,true) --> 585 to 1. Moreover, for the single instance in Q4 2020 when they estimated too high, it was only by 6% and that was mainly due to a $100M interest expense penalty that Tesla voluntarily paid to settle their convertible notes early. Without this unexpected one-time expense, the analyst consensus would have once again been 20% too low. If we want to count that as another miss, then that's 13 consecutive quarters of underestimation which has a 1 in 8,192 chance of happening randomly.
These results, along with the context of all the misunderstandings that apparently led analysts to their erroneous estimates, conclusively and unambiguously indicate systematic bias.
I would bet my life savings that their estimates for 2023, 2024 and beyond also are much too low just like all the previous estimates. Oh wait...I
have bet my life savings on that prediction.
| Quarter | Non-GAAP Earnings per Share | Estimated | Error | % Error |
| Q1 2019 | -$0.19 | -$0.08 | $0.11 | 58% |
| Q2 2019 | -$0.07 | -$0.04 | $0.04 | 52% |
| Q3 2019 | $0.12 | -$0.01 | -$0.13 | -108% |
| Q4 2019 | $0.14 | $0.11 | -$0.03 | -24% |
| Q1 2020 | $0.08 | -$0.01 | -$0.10 | -118% |
| Q2 2020 | $0.15 | -$0.04 | -$0.19 | -129% |
| Q3 2020 | $0.25 | $0.18 | -$0.07 | -29% |
| Q4 2020 | $0.27 | $0.28 | $0.02 | 6% |
| Q1 2021 | $0.31 | $0.26 | -$0.05 | -15% |
| Q2 2021 | $0.48 | $0.30 | -$0.18 | -38% |
| Q3 2021 | $0.62 | $0.45 | -$0.17 | -28% |
| Q4 2021 | $0.85 | $0.67 | -$0.17 | -20% |
| Q1 2022 | $1.08 | $0.72 | -$0.36 | -33% |
| Q2 2022 | $0.76 | $0.64 | -$0.12 | -16% |
| Q3 2022 | $1.05 | $0.97 | -$0.08 | -8% |
| Totals / % Error Average | $5.89 | $4.40 | -$1.49 | -25% |
There were 1.123 billion outstanding shares (diluted) at the time, so this translates to non-GAAP net income estimates of $10.0B and $11.8B in '22 and '23 respectively. '22 actual is already $9.3B just in the first 3 quarters and it's likely to be around $15-16B after Q4 gets added. This is tracking for a 50-60% beat
despite Shanghai shutdowns and a highly unanticipated war in Europe. $11.8B for '23 is just...I have no words. Even Fremont and Shanghai by themselves will probably earn double or triple that in '23.
For another example, in May 2019 Joseph Spak from RBC Capital provided a price target equivalent to a ~$50B market cap and:
He expects Tesla to post a full-year 2019 loss of $6.58 a share [$1.1B with 173M outstanding shares as of Q1 '19], worse than his previous expectation for a loss of $5.71 a share, before that narrows in 2020 and reaches profitability in 2021. (
Source)
Actual non-GAAP net income results:
'19: $0.04B
'20: $2.46B
'21: $7.64B
Let's bear in mind this estimate was published before anyone had heard of COVID-19, so nobody was anticipating factory lockdowns and major supply chain chaos.
There are only two respectable ways for someone to respond to inaccurate estimation:
- Acknowledge the problem, take responsibility and learn how to fix the model
- (Note: this should be happening every time because all of our estimates are wrong and all we can hope for is to make our models less wrong over time)
- Cease publication of authoritative-sounding "expert" estimates until such time that #1 has been accomplished
I can't help but notice that institutional analysts have selected neither of these options, which means that either they are incompetent at this particular task or are corrupt liars, or both.
www.thestreet.com
