Note: This valuation was done just before Q2-2016 numbers came out.
Like Berkshire, Markel has three profit centers: (i) the insurance business, (ii) investment portfolio, and (iii) Ventures.
Unlike Berkshire today, the key source of value is the investment portfolio. It is impossible to value Markel, without forming some opinion on the investment portfolio. This will become apparent in the valuations below. The profitability of the insurance business and Ventures matter much less.
There are two key variables that one needs to consider to project the performance of the investment portfolio:
- the average return of the portfolio
- the investment leverage
To estimate normalized values for these two key variables, it is useful to look back their historical range. We can then “adjust” these numbers by accounting for the different ways in which present times differ from history.
MKL’s overall investment portfolio has averaged between 6-8% historically. The current composition of the investment portfolio is tilted towards bonds (75-80%), rather than equities. Given low bond yields, and the elevated levels of the stock market, it is reasonable to use an average return on the lower side of the historic range.
A conservative (but not over-conservative) normalized number may be something like 5.5%.
MKL’s leverage ratio has varied between 2.25 and 4.5. Currently, we are at the lower end of this range. Due to the soft insurance market, MKL isn’t writing too much new business. This translates to lower amounts of new float, which leads to a smaller investment portfolio, compared to historic norms.
A conservative normalized leverage ratio may be 2.4.
Of course, these numbers (5.5% returns and 2.4 leverage) are just guesses.
Armed with these numbers, we proceed to the valuation stage. I will explore 3 different methods.
1. Simple DCF
This is the simplest and most basic valuation for MKL.
Given the investment return and the leverage, we can estimate the normalized growth in book value per share (BVPS)
BVPS Growth = Investment Return * Leverage
With our numbers, we get BVPS growth rate = 5.5% * 2.4 = 13.2%.
Over its history MKL has compounded BVPS in the high teens. In the recent past, this has decreased to the low teens. Given MKL’s size, a normalized 13.2% seems like a reasonable projection.
Let us assume a holding period of 10 years, and a discount rate of 10%. The target P/B ratio for a company with these cash-flow characteristics is:
Historically, MKL’s P/B ratio has ranged between 1 and 2.5. So this number seems quite reasonable.
On Q12016 (03/31/2016), MKL’s BVPS was $590 (compared to $562 on 12/31/2015).
This implies an intrinsic value of $590 * 1.33 = $786.
2. Two Column Method
If MKL is a mini-Berkshire, perhaps we should use the same method (Buffett’s two-column method) to value it.
The left column is obtained by finding the net investments per share. As of Q12016, MKL had total investments of $18.5B and debt of $2.3 B. The number of shares outstanding was 13.968M.
Thus, the net investments per share is (18.5B – 2.3B)/13.968M = $1,162 (up from $1,142 on 12/31/2015). This is the “left column”.
As of Q12016, MKL Ventures had a TTM EBITDA of $99M (compared to $91M on Q42015). Buffett likes to slap a 10x multiple on pre-tax earnings of quality businesses.
Let us assume that businesses contained in MKL are of high-quality. Since EBITDA does not include depreciation costs, it overstates owner’s earnings. Thus, let us use an 8x multiple on EBITDA, instead.
This is arbitrary – I only know that I should be using a number less than 10x.
At 8x, I get $99M/13.968M * 8 = $57M for the right column (value of Ventures).
Note that currently, the right column is much smaller than the left column, highlighting once again that the principal value of MKL is buried in its investment portfolio.
Using the two-column method, we get an estimate of $1,162 + $57 = $1,219/share.
It is quite likely that this number overestimates MKL’s intrinsic value. At this price, the P/B ratio would be 2.6, which is above the historical range of 1 – 2.5.
3. Three Separate Units
Finally, one can separately consider the three business units: (i) the insurance business, (ii) investment portfolio, and (iii) Ventures. This is fashioned after the analysis by Just Value.
We find earnings of the three business segments, and slap on a reasonable P/E multiple on the combined earnings. I use numbers as of Q12016 everywhere.
(i) Insurance Business: MKL’s insurance business has been consistently profitable on its own. So it might make sense to value it as an independent business. Even BRK has started to do this of late.
The total premiums were $3,838M. We can assume a profit margin of 3%. Thus, insurance earnings = $3,838 * 0.03 = $115.14M.
(ii) Investment Portfolio: The total invested capital ($18,532M) * return (5.5%) = $1,019M earnings from the portfolio.
(iii) Ventures: Slap an 0.8x multiple on Ventures $99M EBITDA to get $79M.
Thus, the total earnings are $115M + $1,019M + $79M = $1,213M. Dividing by the number of shares, I get a comprehensive EPS of $87/share.
We can convert the simple DCF target P/B ratio to a target P/E ratio by multiplying it by (1+ROE)/ROE, where the ROE in this case is the comprehensive ROE, which is equal to the growth in book value (13.2%).
This gives me a target P/E of 1.33 * (1 + 0.132)/(0.132) = 11.42.
Multiplying the comprehensive earnings by this target PE multiple, I get 11.42 * 87 =$992/share.
We got three numbers $786, $992, and $1,219 for the intrinsic value. Which one is correct?
Of course, none of them is correct! But we have a range roughly $800-$1,200, which can be justified based on different assumptions.
At the current price of around $950/share, MKL appears fully valued.
However, if you are into MKL for the long haul, it might not be prudent to sell shares even at these prices. This is true for compounding machines in general, since the intrinsic value rises steadily. Even if you pay a tad too much for such a business, it doesn’t really hurt you, if you hold on to your shares for say 20 years.
In the next post, I will do some sensitivity analysis by using the three valuation techniques considered here, and doing Monte Carlo simulations.