I am buying common shares of Symetra Financial Corporation (SYA) at $10.83 because it is currently available at a discount of at least 25% from a conservative estimate of its net worth based on statutory book values. As is common in investment analysis, I do not know what the exact value of the company is, but it seems fair to assume that it is higher than 15 dollars a share and maybe more. Various evaluations of value are given below.
Origins Of The Company: In 2004, Safeco decided to jettison its Life Insurance business. The buyer group was led by White Mountains Insurance (WTM), an insurance company with a reputation for good capital management, and Berkshire Hathaway (BRK.B) Each bought a stake of 24% on a fully diluted basis (they received warrants in addition to their common stock purchase). There were other private buyers also involved, most notably, Fairholme Capital Management run by Bruce Berkowitz. The final purchase price was $1.38 billion, comprising $1.065 billion in equity and $315 million in debt. The company's new name, Symetra Financial Corp., is the new name it carries today, although the predecessor company, Safeco, has been around since 1957. (It seems that Berkshire accounts for SYA on an equity method basis with no fair-value option elected. Consequently, it doesn't show up in their Form 13.)
Fast forward to 2007, the buyers tried to take the company public in the fall of 2007. The S-1 filed then shows that they were expecting about $18-$20 per share. Given the onset of the financial crisis, that did not come to pass. The offering was cancelled and a new offering was attempted in January 2010. This time the expected price was $12-$14, and the list of sellers no longer included White Mountains or Berkshire Hathaway. The IPO was priced at the low end at $12. Today these shares trade a tad below that price at $10.83, even though the book value of the company has more than doubled since the IPO.
Evaluation #1 (with training wheels): This approach comes with training wheels, courtesy of White Mountains Insurance. White Mountains accounts for SYA on its books using the equity method of accounting. A fair value election has not been made. However, because the public market valuation has been lower than WTM's carrying value for some time, an OTTI was taken and the value of SYA was written down to $15 per share based on three valuation techniques that led them to a value of between $15 and $30, well above the current price of $10.83. The explanation provided by WTM is so good that it cannot be made better or more concise. (Please see pages 88 and 89 in WTM's most recent annual report.)
Evaluation #2 (on my own this time): This approach relies on book values. This is necessarily an imprecise valuation because there are two levels of limitations: the limitations of GAAP accounting (for e.g., liabilities are not marked-to-market but assets are) and the huge amount of estimation that goes into insurance accounting (liabilities, DPAC amortization, etc.) But there is some hope. While GAAP accounting is little more liberal, statutory accounting (SAP) is very conservative, disallowing items such as capitalized selling costs (DPAC), non-admissible assets (e.g. furniture, computers), and other items. This is because SAP is meant for regulators to assess whether there is enough surplus and capital to allow policyholders claims to be satisfied beyond any reserves. Thus, though accounting has its limitations, SAP accounting may serve as a lower bound on the book-value based valuation as it does represent a kind of liquidation value more so than does GAAP. When I recently had a chance to meet the CFO of a large insurance company who had just made an acquisition, I asked his opinion on statutory book values as a valuation device. He said statutory book value represented kind of a "cash value" in liquidation.
In White Mountains' discourse on insurance valuation referenced above, it is mentioned that the liquidation value as estimated by Statutory Book Value further excludes the future economic gains of already written business and the future economic gains from business to be written in the future. So we can accept SAP book value as a reasonably lower bound on our valuation. SAP book values for SYA and other life insurers, along with their market values, are provided below. For reference, GAAP book values are also provided (as adjusted for FAS 115, which is the industry's standard way of doing it).
Click to enlarge.
From the table above, SYA appears to trade somewhat cheaply - compared to other companies and also on an absolute basis.
Evaluation #3 (also on my own): There has been only one comparable transaction in the Annuities/LI business since the onset of the Great Recession -- the acquisition of Delphi Financial Group by Tokio Marine. There were some other life acquisitions in the post crisis periods, notably ALICO-MetLife and AIA-Prudential (later abandoned), but most of those aren't comparable to SYA because they are either global companies or have products that differ very much from Symetra's products. Global life insurance companies have recently tended to trade and be acquired at much higher premiums than US-only life companies.
Delphi had a business somewhat, but not entirely, comparable to Symetra. Most notably, Delphi included a P&C component which SYA doesn't have. Delphi's merger-related proxy statement also talks about the lack of recent transactions. This is also noted by White Mountains (Appendix A). Delphi was sold at a Price to Book of 1.4x, a Price to Adjusted Book of 1.5x and a Price to Statutory Book of 1.67x.
These multiples are calculated and applied to SYA as below:
A few qualitative considerations: All book values are not created equal. In insurance managements have enough latitude to cause any particular year's numbers what they want them to be, given the amount of estimates involved in valuations of liabilities and, to a lesser but still appreciable extent, assets. So it is important to form a judgment on the quality of the book value number and the quality of the management. Further, it is necessary to consider the impact interest rates (the principal reason why life insurance companies are selling cheap today) on life insurers. We consider these issues in turn.
The Management: There is no question the management of SYA is one of the best in the business. I use the "Ben Franklin test" of judging managements by what they do rather than what they say. The clearest indication of the high quality of the management is provided by the fact that Symetra stayed largely out of the variable annuity guarantees during the 2003-2007 period, when other insurance companies made such guarantees to increase sales. Today, many competitors are reeling from those older promises. The Hartford, MetLife, and others have pulled out of the VA business.
Symetra is now expanding in the market place selling a variable annuity product in 2012 with no onerous guarantees. This defense played prospectively by the company in managing liabilities is what separates winning insurance companies from the losing ones. In fact, based purely on numbers, a few other insurers may be cheaper than SYA. But, in those cases, often the management has proved incompetent. In insurance, a cheap price cannot offset bad management because that same management cannot be relied upon to preserve/grow equity value. Warren Buffett said about the insurance industry, "It tends to magnify, to an unusual degree, human managerial talent - or the lack of it."
The quality of liabilities (liquidity and cost): The quality of liabilities at SYA is quite high with 84% of liabilities considered "illiquid" or "somewhat liquid" (See MD&A in AR2011). The other 16% of the liabilities have surrender charges of 5% or less. However, the liquid assets are about $22 billion. Thus, the risk of a "run on the company" is quite low even in a situation where rising rates might cause some policyholders to remove funds from the company.
The quality of investing/asset reporting: Looking at the current rating agency ratings on an investment portfolio is not very useful even though this information is provided. Rating agencies can be "behind the curve" as they turned out to be during the most recent credit crisis. What does help is to judge the quality of the past performance, especially during tough times. From that, we can form a judgment about how well the investment management side has done. Additionally, a "negative assessment" of the current portfolio must be made similar to that done on the liabilities, i.e. we must be on the look for any suspicious looking investments. Given the benefit of hindsight, looking at the performance of the portfolio during the credit crisis will also give a good idea of the competence of the management. I didn't find any disclosure giving the performance of the investment funds over time. So I focused on the crisis years to see how SYA fared.
Looking back to the crisis years, the AOCI balance developed as follows (in millions):
- 2006... ($0.5)
- 2007... ($12.5)
- 2008... ($1052.6)
- 2009... ($49.7)
- 2010... $432.5
- 2011... $1013.5
From the sales and maturities of fixed income instruments shown, we can judge over the years that the company does hold a lot of its debt investments to maturity, as it claims to do. Consequently, the AOCI number loses some of its significance (it must eventually go to 0 for each security that matures) in book-value valuation. This is why the industry standard adjusted BV also excludes AOCI. However, though excluded there, it is helpful here to track its development to see how the losses and gains occurred so that a judgment on the management can be made. In 2008, there was a contribution of -$1021 from unrealized losses on investments whereas in 2009 there was a net addition of $1019, almost undoing all of the unrealized losses.
The reader may argue that the apparent reversal may be because of earnings/gains on a higher invested assets figure. But this doesn't appear to be the case. From the statements, the following can be constructed for the fixed income assets:
- Amortized cost at 2008: 16,528
- Accretion/Amortization of invested assets during 2009: 25.2
- OTTI in AOCI: 105
- Purchases: 4015
- Maturities: 1313
- Sales: 660
- Amortized cost at 2009: 18,554
(I acknowledge a bit of apples-and-oranges here related to the sales number. But there is no way around it. Amortized cost of investments sold is not provided.)
The OTTI experience during 2008, 2009, and 2010 - $86, $191, and $53, respectively - also appears to be quite benign in light of what the global markets went through. Net realized losses were $158 in 2008, $29 in 2009 and a gain of $40 in 2010.
Thus, it appears that there wasn't any large permanent damage caused to the investment portfolio by the recession. This is a great sign because such investing requires very good prospective thinking which the management has displayed. The reader may draw a parallel to such prospective defense played by the management on the liability side as mentioned above.
In looking at the investment portfolio, I didn't come across any particular investments that would fail the "negative assessment", i.e. set off warnings. There aren't any illiquid LP investments and the expansion in commercial mortgages appears well controlled.
Overall, I feel good about the quality of the investments. What is necessary in the analysis here is to ascertain that the valuation of the investments and their selection is sound enough so that the book values used in reporting can be relied upon. I believe SYA passes this test. The valuation/accounting for investments does not appear aggressive.
Future fluctuations in interest rates: The principal reason insurance companies trade cheaply today has to do with low investment returns. Prognosticating interest rates is not my forte nor is such prediction necessary to sufficiently make our case. Rates will move up and down. Presently, the low rate environment makes the annuity and life insurance businesses look bad. In other periods higher rates have made these businesses look terrific. However, barring any major errors on the liability side (such as onerous guarantees on variable annuities) or the asset side (for e.g. exotic illiquid investments), life insurance companies have shown fantastic longevities. Thus we expect that even if the present low rate environment lasts for a long time, most insurance and annuities businesses will not be adversely affected. They simply continue to write business where implied liability rates are low.
It is, nevertheless, important to consider the various rate scenarios that might play out and how the annuity business may respond to it. Fortunately, from where we stand today, the future of rates has only two possibly directions (up or steady) rather than the usual three (up, down, steady). The recent auction of a U.S. TIPS note at a negative real yield is a testament to that. Could there be more of a bottom in real rates than a little less than zero that was observed in that auction? Even if rates hold steady where they are, we do not expect that the annuity business will be in a difficult position. The rate guarantees made today on business being written are in the range of 1-1.5%, yielding asset-liability interest rate spreads of about 1.8%. While a 1.8% on net assets may not ideal, it represents a decent number to ride out the low interest cycle. To demonstrate this, last year despite the low rates, SYA earned $280 mm before taxes.
Historically, interest rates have tended to move in generation-length cycles. In the US, rates fell from 1861 to 1899, then rose from 1900 to 1920, fell from 1920 to 1946. They then rose to a high of 15% on the 30-year bond in 1981 and have now fallen such that they are "bumping along the bottom". Because annuity liabilities are illiquid, a rise in rates will provide a nice earnings upside for the business. Of course, rising rates would also hurt the value of investments fixed maturity already made. But because SYA holds most of its investments to maturity (and also because we ignore the market value adjustment in AOCI in calculating adjusted book values), these market value fluctuations should not present a problem as long as the liabilities are sufficiently illiquid. Any large additions or deficiencies that develop in AOCI will be forced to 0 when the bonds/loans mature.
Low intangible assets:
The company noted in its IPO prospectus, "In conjunction with PGAAP for the Acquisition, we were required to adjust our consolidated balance sheet to fair value and reset our existing deferred policy acquisition costs, goodwill and intangible asset balances at August 2, 2004 to zero." Consequently, the intangible assets are much lower than comparable companies with DAC of $215mm and goodwill of $30mm.
Disclosure: I am long SYA.