Just for fun, what might a dividend discount model estimate for IV? Assume that Berkshire could pay out an annual dividend equal to 3% of BV, or about 2% of market cap, indefinitely, while still growing BV and the dividend by 6%/year. This would give an annualized total return of 6% growth +2% dividend = 8%, versus the 9.2% annualized return since Dec 1999. (In Dec 1999 Berkshire’s P/B was

1.48, as compared to 1.40 today, so multiple contraction has had a negligible impact on the total return, only about 0.25 percentage points.) The appropriate discount rate would be equal to the total return, or 8%. The present value of this perpetual annuity, or the IV, would be

PV = dividend/(discount rate - growth rate)

PV = 3%*$314,090/(0.08-0.06) = $471K, or 1.50 times the current BV.

Dividend discount models, like all discount models, are notoriously error prone. Changes to the growth rate and especially to the discount rate, change the present value dramatically. The discount rate must equal the annualized total return, but one is then really starting with the answer, the annualized total return, in doing the calculation. With the assumptions used above, the estimated IV is pretty reasonable, but with different assumptions the calculated IV changes quite a bit. I seriously doubt that Warren uses discounted cash flow or any other discount model.