I recently came across a post by Greg Speicher from 2011 that described Tom Gayner’s method for valuing Berkshire Hathaway (h/t Value Seeker @Find_Me_Value for the link). The post referenced a 2011 Value Investor Insight interview in which Gayner discusses his valuation method for Berkshire. It’s really a great interview with a ton of insight from Gayner. He’s always fairly open with his thinking, but I found this interview particularly informative. I might have to do another quick post on it.
Anyway, Greg did a great job summing up Gayner’s method by creating a corresponding valuation spreadsheet including Gayner’s assumptions. I thought it might be interesting to update the spreadsheet and see how Gayner’s valuation method looks today with shares at seemingly cheap levels. (quick note: Greg’s original post was in July 2011; a very good time to buy BRK shares, which were around $75 per B share, 1.1x -1.2x book value and have since produced 12.4% annualized. The price actually dropped to $65 per B share later that September, producing 15% annualized.)
Gayner’s method for valuing Berkshire:
TG: “I break the business into three parts, insurance, investments and the other operating companies. For the insurance business, which is currently relatively soft but is well-positioned and well-run, I look at the level of insurance premiums running through the business and assume at the low end that it breaks even on an underwriting basis, in the middle range makes 4-5 points of underwriting profit and at the high end makes 8-9 points. For the investment portfolio, I assume it earns from 3% at the low end up to 10-12% at the high end. For the operating businesses, I look at the actual cash flow produced over the past three years and overlay my expectations over the next few years to get a range of possible normalized earnings. On all of that, I apply 10x, 14x and 18x earnings multiples to arrive at a fair-value range for the company overall.”
The TL;DR version:
- underwriting profits
- investment returns
- normalized operating cash flow
Gayner then takes the sum of all three and applies a multiple of 10x, 14x and 18x.
It’s pretty simple. What’s interesting about this method (to me) is that Gayner’s not just some hedge fund investor valuing a business. He is essentially valuing his business, which is as close to Berkshire as it gets.
I adjusted the spreadsheet a bit from Greg’s by doing five scenarios instead of three, and used slightly different multiples. For insurance premiums I used Q3 ttm and assumed a range of zero to 8% underwriting profits with 4% as the base case. I estimated the investment portfolio at $196.7B, which includes all investments as of Q3: equity portfolio, fixed income, “other”, the Kraft-Heinz investment and $25B cash. I don’t recall how much cash the PCP deal involves, but you can adjust the cash level as you see fit. I used Gayner’s range of 3% investment returns on the low end, to 10% on the high end, with a 7% base case. For operating earnings base case I estimated future growth of 10% annually for three years off of $12.47B ttm, and took the average earnings. Then I took the total net profit and applied multiples of 10, 14, 15, 16 and 18.
The base case has a FV of $181 per B share or 39.8% upside from Friday’s close of $129.53 per B share (1.79x book value). Not to shabby. Yes, it’s very simplistic but I think this gives a good idea of what possible. Can Berkshire achieve 4% underwriting profit? I think so. 7% investment returns? Probably. And a 15x multiple on that is not unreasonable.
At $12.08 EPS Berkshire trades at ($129.53 per B share):
- 10.7x comprehensive earnings
- 12% ROE
The main thing I took from Gayner’s method was the visual of just how big the operating businesses are becoming for Berkshire and its intrinsic value. Growth here should lead to a declining importance of book value multiples, which we sometimes tend to anchor to. Gayner alluded to this in the Markel 2014 Annual Letter stating that growth in book value CAGR is the best way to measure growth in intrinsic value opposed to using absolute book value multiples.
Anyway, this is just one method of many for valuing Berkshire, but one I’ll probably continue to use if for no other reason than to model various return scenarios from the three segments.
For a good discussion on BRK valuation take a look at Value Seeker’s (@Find_Me_Value) recent tweets. Some great detail, and interestingly a similar FV despite the very different approach.
I attached the spreadsheet below so you can adjust/play with the assumptions yourself. I’d love to hear what people think about this so feel free to comment.