- Most DCF calculations out there are wrong, specifically those that discount “retained earnings” in every single year. When an assumption is made about the earnings/FCF growth rate, that assumption comes with it an implicitly assumed payout ratio (i.e. an assumption on a certain amount of the earnings/FCF to be retained and reinvested). The payout ratio is a driver of future earnings growth rate. Those “retained earnings” cannot be discounted every single year as part of free cash flow because they are not truly free, i.e. they cannot be taken out of the business without impacting the future earnings stream. When you discount the “retained earnings” every single year, you are making the implicit assumption that those money can be taken out without impacting the future earnings stream, which is not the case (See previous post Determining the Intrinsic Value of a Business for more details).
- Even when you base your earnings growth rate assumption on historical earnings growth rates, those historical earnings growth rates were obtained due to certain earnings being retained before (i.e. an implicit payout ratio).
- These “retained earnings” need to be “rolled-over” to the following year, and the year after that, and the year after that, and so on, until a “certain point” where ONE discount can be done.
- The dividends however can be discounted normally every single year as those are truly taken out of the business.
- Given the dividend cash flows every single year, and a lump value at the “certain point”, an IRR can be calculated using the current price.
- The problem then becomes,
If that lump value is the value of the company at that point in time, then you get into a circular argument, coz you are basing the current value today on its value say 10 years later, which is based on its value 20 yeas later, etc. If you can calculate the value of the company 10 years later, you should be able to calculate the value of the company today. How do you calculate the value of a company at any point in time? Actually the questions should be more like “how do you know how much to pay”, “what price will give me a 15% IRR”? Very different from asking “what is the value of the company”.*how do you calculate the lump value at the “certain point”?* **The lump value is what you can sell the company for.**- There are three methods to determine what you can sell for:
__Non-going concern – Liquidation__– You can definitely sell the company at liquidation value (minus expenses and a chunk for the effort incurred), so the liquidation value is a lower bound for the lump value.__Going concern – From Transactions__– LBO and M&A prices provide an indication of what you can sell the company for (to knowledgeable buyers). Christopher Browne collects metrics of P/CF, P/E, P/S, P/B, etc. of such transactions and apply them to similar companies to determine their lump value.__Going concern – From Market Prices__– The price at which you can sell a company for is by definition P/E * E. If earnings are stable and predictable (for a good business), then the volatility of the price is due to the P/E fluctuations. Assuming mean reversion for the P/E (between booms and busts), you can get the price at which you can eventually sell the company. Warren Buffett (according to Buffettology) uses Average P/E * Projected Earnings 10 years later to determine the lump value. This is conservative for two reasons:- Using the projected earnings 10 years later assumes that the price will only be realised 10 years later (i.e. P/E will revert back to its mean by 10 years time), which is a conservative assumption.
- Using the Average P/E, as opposed to historical P/Es that are higher than average, is conservative because we don’t have to bank on irrational exuberance to realise our projected selling price.

- You don’t know when this lump value will materialise, can be 1 year, 5 years, or 10 years. The earlier it materialises the higher your IRR (picture the jump of P/E to the Average P/E, the quicker the jump, the higher your IRR).
- Using Buffett’s logic (according to Buffettology) on Christopher Browne’s method, Browne should have collected many different sets of metrics at different periods of the economic cycle (e.g. boom, bust, etc.). That is similar to the high and low P/Es experienced by a stock in the market. Browne can then use an “average” P/CF, P/E, P/S, P/B metric set to calculate the lump value (the “average” set can be different for each different industry).
- Determining the lump value is essentially the same as determining the
of a DCF calculation. Bruce Wasserstein in*terminal value**Big Deal*highlighted two methods- Determine the multiple at which comparable companies trade or have been acquired and apply the multiple to the company’s earnings in the last year of projections (e.g. 10 times EBIT). This is similar to point 8 above, and also highlights that for point 8(3), Buffett’s method can be applied with other ratios apart from P/E.
- Capitalise the final projections based on some perpetual growth rate. As with my arguments for DCF, I think only projected
*dividends*should be capitalised.

## Discussion

## No comments yet.