Someone who reads my blog emailed me this question:
“How do you value insurance companies? Berkshire, Markel, Progressive...
There are many different ways to do it. E.g. with Berkshire, people like to do e.g. 1) investments plus multiple on earnings 2) pre-tax earnings less insurance plus investment portfolio (haircut or not) 3) book value based on buyback level. There are also float-based methods and so on.”
Honestly, I have always thought the same way about insurance companies, closed-end mutual funds, holding companies, etc. What matters is the annual return you get. That annual return can either be in the form of some payout to you (like a dividend) or it can be “growth” in the sense of an increase in the economic earning power of the business from year to year. I know others will use a price-to-book approach – or something similar. I do not think that make sense.
I will use Bancinsurance (no longer a public company) as an example. I wrote a couple letters to the board of that company. That was an unusual thing for me to do, so you can be sure that what I put in those letters was something I thought was important. My argument was that book value might be a relevant metric for some insurers – even most insurers – but it was not a relevant metric for Bancinsurance. The company usually had a combined ratio below 100. Basically, it did not have a positive cost of float in almost any year. So it had “free money”. As long as the company’s premiums did not decrease by a lot from year to year, it had a stable source of essentially 0% (or less) money. I do not see any reason why a permanent accounting liability (like float) should be treated as an economic liability at all. If you were going to liquidate the business, it would make sense to net out that liability. But you are not going to liquidate the business, so it does not matter what you see on the balance sheet. What matters is the normal earnings power of that float.
You can do a check to prove what I’m saying. In the very long run, the S&P 500 has maybe returned close to 9% a year. Any insurer that had a 9% return on equity in each and every year should not be valued below book value. If the company had a 6% return on equity each and every year, then it should trade at a discount to book value. If the company had a 12% return on equity each and every year, then it should trade above book value. My point here is that book value does not matter. What matters is the annual return in the business and the annual return in the S&P 500. The “fair value” of a stock should be the price that equalizes the forward return on that stock with the forward return on the S&P 500. So, if Bancinsurance was earning a 10% return on equity, and the S&P 500 was priced to return 8% a year – I would argue that Bancinsurance needed to trade above book value (not below it) to equalize the forward return on its shares with the forward return on the S&P 500.
For an insurer, you can break the earnings situation down into the underwriting profit and the investment profit. For any one year, this is not really that helpful. Both underwriting results and investment results are cyclical. For example, Progressive has made good enough underwriting profits since 2008, but it hasn’t made much of an investment profit at all. That is because Progressive (PGR, Financial) invests in high-quality short-term bonds. These are tied closely to the Fed Funds Rate. So if the Fed raises rates, Progressive’s investment profits will rise. If the Fed cuts rates, Progressive will earn less on its investments. I do not think the intrinsic value of Progressive changes from year to year based on the Fed Funds Rate. It changes based on the amount of float the company has and the cost of that float. This is the same approach I use when valuing Frost (CFR, Financial). I have used the example before that from 2008 to 2016, Frost’s deposits roughly doubled. However, EPS did not grow. To me, the company doubled its intrinsic value from 2008 to 2016. Other investors may disagree with this. They think a company’s intrinsic value must follow its actual earnings. That makes no sense to me. The retention rate on deposits is high. At Frost, it is probably just over 90%. That is very similar to the renewal rate on auto insurance policies at some place like GEICO. It is even higher than Progressive’s renewal rate. Progressive tends to have younger customers who have “unbundled” policies. It probably has more single people than some insurers. Those kinds of policyholders renew at lower rates than older married people. Old married people almost never switch insurers. I do not see any difference between Progressive’s “float” and Frost’s deposits. They are both stable sources of funding. Their cost can be estimated fairly easily. Progressive’s combined ratio – like GEICO’s – is cyclical, but it’s not impossible to estimate. Both Progressive and GEICO have so much lower costs than other auto insurers that they are not likely to have large losses for long due to competitive pressure. It’s possible. There was a year where even GEICO and Progressive had combined ratio problems because of intense rivalry with another big insurer, but it’s rare. There is no pressure like that in banking. Frost cannot be forced to pay a lot more on deposits to stop them from flowing out of the bank and going elsewhere. Most bank customers do not really chase higher interest rates on their money. The customers who do are not the kind of customers you want. So I look at Frost and Progressive as being very similar.
My newsletter co-writer (Quan) and I picked both Progressive and Frost. We wrote issues on each stock. I like Frost a lot more than Progressive. That is not because of differences in insurance versus banking. It’s for unrelated reasons. One, I think driverless cars will eventually – decades from now – obsolete the business of both Progressive and GEICO. So, there is a limited lifespan left on those companies. Banking will be around a hundred years from now. I expect bank branches to mostly disappear over time. But other than disappearing branches – which are potentially positive for bank stocks – I do not expect much change in the bank industry.
The math is really the same for Berkshire, Progressive  or Frost. Pretend we wanted to find the right price-to-book ratio for each of these stocks. How much do they pay out in dividends as a percent of book value? How much stock does each company buy back each year? And then how much does the company’s shareholder’s equity grow each year? If a company has 5% growth in shareholder’s equity, a 2% share buyback rate and a 2% dividend yield – it’s priced to return about 9% a year. It’s fairly valued. Maybe a little cheap compared to where most stocks trade at today, but it’s not clearly undervalued by a lot or overvalued by a lot. This assumes that today’s earnings are normal. Often, they are not. In the case of both Frost and Progressive, we picked those stocks specifically because we thought earnings were at a cyclical low. Investment profits were lower than they would be in a higher interest rate economy. So, you always must adjust earnings to a “normal” level. That is true for all stocks.
Let’s look at Berkshire Hathaway (BRK.A, Financial)(BRK.B, Financial). Berkshire is easy. It does not pay dividends and rarely buys back stock. It also does not issue much stock. So, for the most part, we can just focus on the change in book value for the entire company. Say you believe – as I do – that the S&P 500 is not going to return much more than 7% a year from now on. Since the financial crisis, Berkshire has grown per-share book value by something like 12% a year. I just eyeballed the non-compounded (simple) numbers from the annual letter. That is not a very accurate average I just gave you, but it lets us know what kind of number we are talking about. The stock market has performed very well since the financial crisis. Berkshire is a lot bigger now than it was in 2009. So I think we cannot possibly assume a growth rate in per-share book value as high as Berkshire had since the crisis. I will use 10%. I think 10% is a fair estimate for what Berkshire will compound book value at in the future. This is especially true if Berkshire starts buying back stock. Stock buybacks could help sop up funds when the shares are undervalued. It may hurt the chances of very rapid book value growth, but it will also lower the range of possible outcomes for the stock. I think share buybacks will make it more likely the stock will compound growth at something like 10% a year. So let’s see what that means for book value.
Again, my estimate is that the S&P 500 will not do better than 7% a year. I’m also estimating that Berkshire will not do worse than 10% a year. These are the two critical assumptions. Everything about what I’m going to say rests on these two assumptions. Other investors will pick different numbers for both of these. If we go with my numbers – the “fair value” formula at Berkshire is 10% / 7% = 1.43. If you believe Berkshire definitely will grow per-share book value by at least 10% a year and you believe the S&P 500 definitely will not return more than 7% a year – you should sell your S&P 500 index fund and buy Berkshire Hathaway stock whenever Berkshire is priced below 1.43 times book value. I think it makes sense to buy Berkshire whenever it goes for 1.5 times book value or less. When Berkshire trades at a much higher price – like two times book value – I would be hesitant to recommend the stock over holding cash. I have no problem with anyone putting whatever cash they have lying around into Berkshire whenever the stock is at 1.5 times book value or less. Notice, however, that this would change depending on the level of the S&P 500. The higher the price on the S&P 500, the more comfortable I am with the idea of owning Berkshire. That is because my idea of “fair value” is a relative value approach. A stock is fairly valued when it’s priced such that the forward return on an alternative is similar to the forward return on the stock.
Let’s use Bank of Hawaii (BOH, Financial) as an example of a bank instead of an insurer. The dividend yield on that stock is 2.3%. The share buyback rate is usually no lower than the dividend. We can assume a combined dividend and share buyback rate of about 2.3%. That gets us to 4.6%. You can use this figure in one of two ways. One, you can estimate the growth rate needed for BOH to match the S&P 500. Assuming the S&P 500 will return no more than 7% in the future – as I do – you get a required growth rate of 7% minus 4.6% equals 2.4%. Bank of Hawaii needs to grow deposits at about 2.4% a year. In reality, I do not even think it needs to do that to match the S&P 500. I think normal earnings at BOH are higher than what it is earning now. That is why I picked the stock for the newsletter. Let’s look now at the actual growth rate I expect for Bank of Hawaii. I feel certain the bank will grow deposits more than 3% a year but less than 6% a year. Let’s use the midpoint (4.5%) as the most likely figure. A 2.3% dividend plus a 2.3% buyback rate plus a 4.5% deposit growth rate gets you to a 9.1% annual return at today’s price. You can divide 9.1% by 7% and get a 1.3 multiplier. This means BOH stock should be trading 30% higher to equalize its forward return with the S&P 500. BOH stock is now trading at $85 a share. If you do $85 times 1.3, you get about $111 a share. So the stock should be priced at something like $110 a share to have the same forward return expectation as the S&P 500. Again, this does not consider the lack of normal earnings today. So BOH should be priced at about $110 a share if you expect interest rates to stay where they are forever. I expect them to rise. So I expect the stock should be worth more than $110 a share. Again, this is a relative approach. For the newsletter, Quan and I took an “absolute” approach. We got a value of $97 a share for BOH. That is the correct value if what you mean is that you need an 8% to 10% annual return in the stock. This figure is lower than the fair value of over $110 a share I gave – and a lot over $110 a share in a higher interest rate environment. Again, this is because I expect the S&P 500 to return 7% a year or less not 8% to 10% a year (like it did in the past).
I think it’s most useful for an individual stock picker to define “fair value” as the price at which a stock has the same expected future return as the S&P 500. For Bank of Hawaii, I am confident that fair value – defined that way – is no less than $110 a share. And it might be more.
Disclosure: Long CFR.
Start a free 7-day trial of Premium Membership to GuruFocus.
