Another Way of Looking at Risk in Financials

Asset/Liability Correlation and Volatility - one way of looking at the risks a financial company faces is to consider the volatility and correlation of its assets and liabilities. As an example, an insurance company like Chubb (CB), whose liabilities arise from providing coverage in the volatile areas of surety, professional, and general liability should (and does) maintain its assets in government or other highly rated bonds. The point is that adverse developments on both sides of the balance sheet can have a multiplier effect on financial risk. An example of correlation risk would be Fannie Mae (FNM): it has substantial liabilities from guaranteeing mortgages it securitized, and substantial assets in mortgages and MBS. The deterioration in mortgage default rates impacted both assets and liabilities simultaneously and unfavorably, with the results that we have seen. While conforming mortgages were not generally considered volatile before the current difficulties, the asset/liability correlation risk was not well thought out.

Allstate's (ALL) primary business operation is personal lines property and casualty, which is volatile with respect to catastrophe losses such as hurricanes, but otherwise presents a stable flow of premium and losses. Assuming that catastrophe losses are held within bounds by reinsurance, as an investor I would have no objections to the company maintaining a substantial part of its investments in MBS or other bonds that have higher yield/risk characteristics than would be appropriate for Chubb. Before the mortgage crisis, Allstate did just that: however, as the situation has deteriorated they have realized substantial losses in de-risking their investment portfolio, and are still under some pressure from analyst/investor opinion to continue the process.

My take is that Allstate at 20 per share is a good buy and holds value proposition considering that its property and casualty operations can generate about 4 per share of earnings: anything they get from investments is gravy. As far as the rating agencies go, P&C companies have a tendency to get reduced ratings as the market weakens, policyholders remain secure, and the ratings should be taken with a grain of salt.

Hartford Financial Group (HIG) is more complex in terms of its asset/liability characteristics. As the situation has developed, the company has volatile and potentially substantial liabilities arising from guaranteed benefits on variable annuities linked to the S&P 500. It has hedged these exposures, so it holds assets to offset the guarantee liabilities: the catch is that correlation needs to be good; if the hedges don't behave like the index, trouble could follow.

Again, on the asset side, Hartford has a substantial amount of MBS which has been creating issues in terms of their willingness and ability to hold to maturity. The behavior of these assets under rating agency stress tests or other worst case scenarios is suspect, creating pressure on the company to de-risk its investment portfolio at the bottom of the market.

In a deep recession, most assets become more correlated than they would normally be, so the S&P 500 and the market values of MBS can be expected to move in tandem for the near term, but long-term that might not be the case.

For the investor who is optimistic about the trajectory of the S&P 500 and the ultimate value of MBS, I regard HIG as an interesting speculative investment, but one with a potential for a very wide range of results due to the volatility and correlation risk on both sides of the balance sheet.

MBIA (MBI), a bond insurer, has potentially volatile liabilities due to the insurance it is providing on 2^nd lien mortgage securitizations (CES and HELOCs), and CDOs and CDO^2. On the asset side it is holding cash, Government bonds, and other highly rated fixed-income securities. So, as the portfolio is currently invested, MBI passes the test in that the volatility is only on one side of the balance sheet.

Before the 2^nd to last round of downgrades, MBIA held substantial MBS backed securities as assets in its Asset/Liability Management (ALM) business. These were volatile, and the downgrades forced the company to realize economic loss in liquidating the assets at a time of depressed values. With the wisdom of hindsight, I now question the ALM business as an appropriate secondary business for a bond insurer, the point being that by its nature it involves holding potentially volatile assets, not a good idea when the bond insurance side of the business has volatile liabilities. At this point, the whole question is moot as further downgrades will not precipitate further calls for redemptions or collateral.

It should be noted that the very similar Ambac (ABK) took a big hit on volatile assets in its own Asset Management business.

Following MBIA's transformation, whereby it split its insurance subsidiary into two legally separate entities, one for muni-only and the other for SF (structured finance) and International, I expect that both will continue to hold primarily very stable and highly rated assets, for the fact that S&P assumes that anything rated less than an A is worthless under the stress case scenario.

But when I think about the SF insurer, the “oldco,” I almost wish they would be willing or able to hold lower rated assets. The most unpredictable liabilities are in the insured structured finance portfolio. CDOs and CDO^2 have the property that they resist mild to medium stress very successfully but go haywire under extreme stress. So a realistic stress case scenario for liabilities of this type is severe. If oldco holds US government or triple A securities, the long term return will be modest at today's “flight to safety” interest rates. Because the return on the assets is predictably low, the results of the liabilities controls the outcome.

But I am willing to bet that there are A or BBB+ assets that would behave more predictably than the CDO and CDO^2 liabilities and over the time frames involved would do better than more highly-rated investments. That would be particularly true today when risk spreads are wide compared to historical norms. So riskier investments in this case might possibly improve the long term outcome under any scenario, to include the worst case.

According to the company's last conference call presentation, payments for losses on the CDOs and CDO^2 would not commence until 2017 and would extend for more than a decade into the future. Defining the task at hand as accumulating investment earnings sufficient to pay stress case losses as they become due in the distant future, a few percentage points on investment earnings is huge.

Mr. Market's Opinion – to get at the market's opinion on relative values here, I have compared the current share price to the latest GAAP Book Value as well as each company's preferred nonGAAP book value metric, from their quarterly supplements:

click to enlarge

Investment Implications - under the line of thinking I am exploring here, Chubb and Allstate are less risky investments in financial companies because asset/liability correlation and volatility are structured in ways that have worked in the past. Hartford is extremely questionable because there is volatility on both sides of the balance sheet, complicated by the necessity of successful hedging, potential rating agency downgrades, and questions about correlation. MBIA, which trades at the most extreme discount to adjusted book value of any of them, is less risky than HIG because its volatility is concentrated on the liability side - the assets are rock solid. In addition, there is limited correlation between the RMBS and CDO type insured liabilities and the highly-rated fixed-income investment assets.

Most analysts would tell you that Chubb is the best of the lot. But here's a thought for you – if MBIA were able to do economic commutations on any sizable portion of the insured CDO and CDO^2 book, it is then similar to Chubb in that the assets are highly-rated, there is is limited asset/liability correlation, and all the volatility is confined to one side of the balance sheet.

Disclosure – Long ALL, CB, HIG and MBI; no position in FNM or ABK. Planning to reduce or eliminate HIG.

Comments

[Ishortyou]

nice ideas thank you very much.

2009 Mar 02 01:35 PM

[James Wilson]

Funny the Underwriters Insurance for those who approved these Subprime Loans never has to pay up .

Or did all the Insurance Company Funds get used for Bonuses or Stock Options ????

2009 Mar 02 04:04 PM

[texalope]

Hi Tom. Have you thought about MBIA exposure to Pension Obligation Bonds and the comments from Warren Buffet .He concludes that insured municipals may be an accident waiting to happen. It will be large enough to wipe out years of profits.

2009 Mar 03 12:58 PM

[Tom Armistead]

texalope,

There was Warren's letter and in the WSJ I think the same day there was a discussion of pension accounting for municipalities, the pension bonds, something I was not aware of, the practice of borrowing money in order to invest it in the pension funds.

My impression, based on my observations of municipalities around where I live, is that town pensions are pretty generous and for example the police or fire departments usually give those approaching retirement a promotion and a raise so their last year sets them up for good payments. Accounting for pensions has always relied to some extent on averaging values over the years because the equity markets go up and down. I studied it some in college although the rules have changed since then. Add to that the politics, elected officials don't have to answer when pensions go bad 20 or 30 years after the deals are made, and the result could be similar to what is happening to corporations about their "legacy" obligations.

I agree that there will be problems, but expect that they will be manageable. The private sector has converted to 401ks and the public sector will do likewise, with older employees grandfathered on their pensions, or defined benefit plans capped as of a certain year, so on and so forth. With stocks down as bad as they are many plans will be underfunded with protracted games of political football. Eventually taxpayers are on the hook for it, the result will be property taxes at the local level, pushing down property values. In effect, the town retirees own part of everybody's house. At the state level it will be income taxes, fees, sales taxes, etc., with more political football. The taxpayers will take the view that private sector has to make do with 401ks so the public sector can do the same.

Just a year or so ago everybody was complaining there were so few losses on municipal bonds that the insurance was a rip-off, California was among the examples cited. There will be losses paid, probably enought to create demand for the insurance and support premium levels, not enough to endanger MBIA's muni operation.

2009 Mar 04 08:06 PM