The dangers of discounted cash flow

I’ve been re-reading an important book on investing, Value Investing by James Montier, which I highly recommend if you want to separate in your mind the nonsense spouted by “expert investors” from the truly rational.
Today I want to look at what James wrote on discounted cash flow, a method we are all supposed to follow, i.e. estimate cash attributable to shareholders generated by the firm after paying for capital items, extra working capital, etc., to maintain its competitive position and (possibly) allow for growth.
James says, first, there are difficulties forecasting future cash flows
Second, we don’t know what discount rate to use. For example what equity risk premium should be in the final discount rate?
He suggests alternative approaches, that accept complex maths is not required.
Can analysts forecast?
A colleague of James’, Rui Antunes, looked at analysts’ forecasts 24 months before the actual reporting of the profits, then 23 months, then 22 months, and so on (for individual firms).
He found, on average, 24 months before the results came out the size of the error was 93%. Even one year beforehand analysts managed to miss the real profit number by 47% (USA data)
When he looked at European forecasts he found the 24 month forecast error was 95% and the 12 month forecast error 43%.
Montier says “Frankly, forecasts with this scale of error are totally worthless” (p50).
But this is not too surprising when we are dealing with many companies subject to a great deal of change such as a Tesla or Alibaba. Hence Warren Buffett’s great care to only look at companies NOT subject to great change. Anyway back to Montier and Antunes:
They separate European shares into five categories based on “value” characteristics or, at the other extreme, “growth” characteristics.
Those in group 1, the “value” shares had historically low annual growth in earnings per year, around 6%. Whereas those in group 5 grew earnings in the past at over 17% per year. Naturally, most analysts extrapolated with a small nod to the phenomenon of reversion to the mean in growth rates.
So they reckoned that the value shares would remain the slowest growers but would move a bit more to the average in the future, at 9% pa. on average.
For the growth shares in group five they forecast around 16%.
How good were those forecasts? Well, for the value shares they were pretty well spot on (on average) at around 9%.  But they were way off the mark for the growth shares which went on to grow earnings at only 5% on average over the long term. Yes, below the rate for the "value" shares!
“So, it appears than analysts are most often wrong on the things they are most optimistic about!”
The discount rate
According to the university textbooks (I’ve written a few and therefore guiltily admit I’ve inflicted this on countless students) the discount rate consists of,

• The rate of return an investor requires if investing in a completely risk free security. No such thing exists, but we usually take the rate of interest on government securities, bills or bonds (if the government is reputable). There is debate on whether to use the short term rate (three month UK Treasuries currently yield 0.5%, or the 10 year bond rate, currently UK govt bonds yield 1.4%). Most people opt for the long-bond rate, PLUS
• The risk..To read more subscribe to my premium newsletter Deep Value Shares – click here http://newsletters.advfn.com/deepvalueshares/subscribe-1