In one sentence: Running Fiat Chrysler's 2017 numbers gives an owner-earnings value far above the share price, but analysts' 25% growth forecast is absurd; a sober 8% estimate still points to a discount, though both admit they don't understand the business.
Key ideas
- Numbers in euros. Copy the statement's own labels and convert at the end. 2017 figures read out: net profit about 3.5B, depreciation and amortization about 5.9B, receivables about −0.2B, payables about +0.1B, tax about 2.7B. [01:00–04:30]
- Tax line caution. The tax figure that matters was in the explanatory notes, larger than the cash flow statement's line. [03:00–04:00]
- Conservative maintenance. Danielle subtracted all PP&E purchases (about 8.7B) as maintenance, which is ultra-conservative: roughly 5B of owner earnings and a 10-cap price near 50B. [03:30–10:00]
- Depreciation as a proxy. Phil suggests depreciation approximates maintenance if assets are depreciated over true lives, and subtracting 5.9B instead gives about 8.5B and an 85B price against 26B market value. When unsure, take the safer number. [05:00–08:00]
- Gap needs an explanation. A price at a third of the value implies fear or something unknown. They guess a leadership change, tariffs, peak-cycle worries and a rotation out of autos (GM and Ford fell too), without confirming. [08:30–15:00]
- Analysts' 25% growth. Applied for ten years it implies a value of $242 to $318 a share against $16. A number that crazy should make you suspect the input. [15:00–19:00]
- Why analysts bias upward. Danielle's caveat: Phil is describing a general conflict, where an analyst's bank wants business from the company, not alleging anything about this firm. Their audience also trades on weeks, not decades. [16:00–23:00]
- The book's growth rule (p. 222). Use the lower of the historical rate or the analyst rate; if both look wrong, use your own "windage" rate. Run several rates. [18:30–19:30]
- Fiat's actual numbers. ROIC about 4%, ten-year growth in the low-to-mid single digits, cash flow about 8%. At 8% growth and a PE of 16: sticker about $23, margin-of-safety price about $11.50, payback-time price about $16 against a $16 market price. [19:00–21:30]
- Predictability test. Chipotle is easy to call bigger in ten years (200 new stores a year, no debt); Fiat is not (electric cars, new leadership). Look for six-inch hurdles, not five-foot ones. [23:00–25:30]
- Both admit it is not a real investment case. It is an example; nothing is on sale for beginners. [08:00–09:00, 25:00]
How it maps to RuleOne
- The screen's growth input follows the same rule: historical or analyst, whichever is lower. Overriding it is a conscious "windage" choice, and /stock/TICKER/ lets you see the effect.
- A value far above the price is a prompt to find what the market knows, not a buy signal.
Buffett, Munger and Graham links
- Skepticism about forecasts: Buffett's 1989 letter on projections; Graham, The Intelligent Investor, chapter 12 on per-share earnings.
- Know-your-limits: the circle of competence in 001, and the "too hard" pile in 185.
- Mohnish Pabrai as cloner and contrarian: see 164.
Words to know
- Windage growth rate: your own estimate when the usual inputs look unreasonable.
- Sticker price: the estimated future value discounted back to today.
- ROIC: return on invested capital; here only about 4%.
Try this
On any stock page in /stocks/, change the growth rate between the historical figure, the analyst figure and one of your own. Note how large the swing in sticker price is and which input you trust.
Check yourself
- Why subtract all PP&E purchases as maintenance?
Answer
It overstates maintenance and so understates owner earnings, the conservative error. - What do you do when analysts forecast absurd growth?
Answer
Use the lower of historical and analyst rates, or your own windage rate, and test several. - Why shouldn't a big gap between value and price be taken at face value?
Answer
It may reflect risks you do not understand, so find out why before acting.
Short quotes
"You want stuff that just is obviously on sale." (Phil Town, ~25:30, auto-transcribed)