Insurance

In preparation for Q4 earnings and the outlook for 2012, we have attempted to organize the available pricing data into an “expectations” analysis.  This is difficult given the quality of the various surveys and company commentary, but we do think this can be useful for thinking about stock performance when combined with harder data

The main outcome of this analysis is that we are changing our Short-Term Qualitative View on Reinsurance to Neutral from Positive; the rest of our View remains unchanged.  We would play any pricing cycle improvements through P&C Specialty in particular, along with P&C Multiline and P&C Commercial (both Short-Term Positive).  P&C Specialty has the potential to benefit from both insurance and reinsurance pricing improvements Read the rest of this entry »

Market risk perception was one of the best determinants of insurance stock performance in 2011 (Exhibit 1).  The change in market risk premium by insurance subsector during the financial crisis did a better job of forecasting 2011 stock performance than fundamental growth, valuation, or capital management (e.g. buybacks)

We largely expect this behavior to continue short-term into 2012, except that we think the probability of an upside surprise on Reinsurance pricing is increasing.  Otherwise, we expect Brokers, P&C Multiline, P&C Specialty, and P&C Commercial to remain the best sources of long ideas

We were not negative enough on the “riskier” subsectors in 2011 (Life Investment, Reinsurance, Multiline, Guaranty), but given the relative multiple compression that has already occurred, we do think underperformance in these “riskier” subsectors eases in 2012, absent a more general market downturn

Unlike longer timeframes, where fundamental growth can trump valuation, changes in valuation (specifically price/book) were the biggest determinant of 2011 performance.  While risk perception captured this well, no typical fundamental metric did.  For example, the long-run positive correlation of ROE to price/book has no relationship in the short term

Perhaps more notably, share buybacks had no significant short-term impact on performance.  There was some slight improvement in performance as buybacks increased, but it was mostly owing to less underperformance

“Deep value” did not work well, either: among stocks trading below book, higher valuations slightly outperformed lower valuations.  For stocks trading above book, neither buybacks nor value investing were significant

We have resisted making market assessments of risk perception, but we attempt such an analysis for the first time here.  Our conclusion is that the bond market seems to be positioning itself for a negative outcome in equities.  We see this in both longer-term corporate spreads, as well as Treasury spreads over inflation, both of which have some predictive correlation to the S&P.  We do not claim this is causation, but there is a resemblance to current spreads and past “risk-off” events

If one believes in this bearish market outcome, then our Long-Term Qualitative View would be our guide to how to invest, which is mostly Negative across the board.  In fact, a major part of our 2012 research will be in determining when to make the shift from Short to Long-Term, whether fundamental or market-driven

But conversely, if one is betting on a recovery from risk aversion and a return to pre-crisis “normalcy,” insurance may not be the best way to do this.  Banks seems to have the most obvious post-crisis re-rating, whereas the apparent re-rating of insurance largely was relative to the market.  Thus, banks offer more potential “bang for the buck” in the sense of re-rating potential

This remains true even if one looks at the “riskiest” end of insurance, like Life Investment.  While banks were clearly and sharply re-rating downward versus history, life insurers on average are trading at the same relative valuation they were trading in the early 2000s.  Thus, one could plausibly conclude that 2004-06 life insurance valuations were an anomaly, and that life insurers would only relative recover in a “risk-on” market

• Our current research has focused on many of the complex cross-currents impacting risk management overall and Insurance specifically.  In today’s note, we take a different tack, and look at the industry from the standpoint of a generalist portfolio manager, emphasizing where to look for constructing a “core” portfolio
• We conclude that PMs may want to look to the least “hard core” Insurance subsectors: Brokers (e.g. MMC, AON, WSH), Life Other (e.g. AFL, AMP), and P&C Personal (e.g. PGR, ALL).  If one is more concerned about owning the best average performance most of the time, and is not focused upon timing the cyclicality of insurance, taking less risk seems to have been the better approach, particularly (but not just) when the financial crisis (2008-?) is taken into account
• Other conclusions of this analysis are summarized as follows:

1)      Insurance is intrinsically a “longer-term” industry: Insurance is more likely to outperform over the longer-term (we use 3 years in this note) than the shorter-term (6 months in this note).  This is true pre- and post-crisis

2)      Pre-crisis, a risk “barbell” was effective…: Prior to 2008, investors were most likely to outperform either by taking a great deal of risk (e.g. Life Investment or Reinsurance), or by taking the least possible balance sheet risk (e.g. Brokers or P&C Personal)

3)      …But the crisis wiped out all gains from risk-taking…:  However, if we add in the performance from 2008 on, all the outperformance from Life Investment and Reinsurance vanishes.  We think this is very big deal for Life Investment in particular.  Reinsurance has always been a hard sell for many investors, given its specialist nature and opacity.  But Life Insurance generally used to be considered a good business, whereas now annuity-focused life insurance (what we call Life Investment) appears thoroughly “broken”

4)      …yet left many other subsectors unscathed:  Brokers and P&C Personal did worse during the crisis, but not so much that it ruined the long-term track record.  In addition, moderately risky subsectors like Life Other (e.g. less annuity exposure) and P&C Specialty (e.g. complex underwriting) produced overall good results

5)      Insurance can be a decent indexing source: Because most insurance is tied to overall economic activity, it should not be surprising that insurers market-perform most of the time.  But this is not necessarily a problem if it comes with more limited downside.  In addition to the 4 subsectors noted in the previous bullet, the large-cap heavy P&C Multiline subsector can be defensive long-term, though is more likely to underperform shorter-term.  As a corollary, most insurance subsectors strongly market-performed during the crisis, so they were relatively but not absolutely defensive

6)      There are some opportunities for speculation: Subsectors like Guaranty, and especially Services, are less likely to market-perform, and have more symmetric distributions of out- and under-performance, particularly over the shorter-term.  These subsectors may be less appropriate for the average PM, which is important to know given that Services, for example, superficially resembles the low capital intensiveness of Brokers, but behaves very differently

7)      Value creation matters most: Book value growth mostly offsets multiple contraction over the longer-term, so picking the highest earners, unsurprisingly, is a good strategy.  But Life Investment is a major exception here: 20% of its pre-crisis performance was multiple expansion, and this all vanished post-crisis.  At present, no other subsector looks particularly at risk for (more) downward re-rating, but in general high multiple expansion should be distrusted over time

• The biggest risk we face from this analysis is the possibility that the 2008-2011 “post-crisis” period is only “temporary”, in which case formerly high performance but risky subsectors like Life Investment (e.g. MET, PRU) and Reinsurance (e.g. RNR, PRE) may be the areas that deliver the best performance in any recovery.  However, the body of our research does not lead us to this conclusion.  We think the changes in risk aversion crystallized by the financial crisis have been a long time in coming (i.e. per our thesis of risk quantumization) and are therefore “permanent.”  Even if this is incorrect, calling a recovery in financials seems to us mostly a market and global macro call, where we have no research opinion

 Print This Post

There has been no appreciable change in the valuation discount that insurance underwriters have suffered in the wake of the 2008 financial crisis. Average price/book ratios for “capital intensive” insurance subsectors, like reinsurers and annuity-exposed life insurers, are 35% below the long-term average prior to 2008, putting such subsectors at risk for trading chronically below book value

While many investors are therefore attracted to insurance as a once in a lifetime value opportunity, it may be more prudent to consider the possibility that this revaluation is “permanent”. Or more specifically, that the revaluation is mostly a market-driven change in risk preference, and thus there may be nothing insurance underwriters can do short of raising ROEs to unrealistic levels

Using several risk-adjusted ways of examining the valuation discount, it is hard to avoid the conclusion that the market has been internally consistent in the sense of assigning a uniformly higher risk premium in the years post-2007

Insurance subsectors that are arguably not capital intensive (e.g. brokers, personal lines), have also been revalued down 35%, literally the same amount as for capital intensive subsectors. But the difference is that these subsectors are still trading above book value on average

The market has revised price/book ratios downward in a way that is statistically consistent with the volatility of price/book, or the beta of stock prices. That is, price/book ratios both pre- and post-crisis lie along the same continuum of price/book volatility and stock beta, both of which are inversely related to price/book

Most interestingly, price/book ratios as related to the ROE Sharpe ratio show only a change in the average level of price/book, whereas the relativity of price/book across ROE Sharpe ratio is nearly unchanged pre- and post-crisis. In other words, the valuation discount is almost totally independent of individual company risk: all companies have been tarred with the same brush

The ROE Sharpe ratio analysis reveals startling details. For example, the average price/book ratio for a 0 ROE Sharpe ratio (i.e. no expected future return above risk-free) has fallen from 1.25 pre-crisis to 0.80 post-crisis. Investors familiar with the life insurance concept of embedded value—e.g. the economic value of all business booked to date, with no future business—may realize that this result is consistent with the market saying that insurance underwriters have moved from a perceived 25% credit to accounting book to a perceived 20% hole

Herein lies a possible way forward for insurance underwriters, although it is such a potentially drastic way forward that it seems unlikely to occur on a wide scale. It seems all but impossible that the average underwriter can raise its sustainable ROE by the 15+ points implied by ROE Sharpe ratios to get back to pre-crisis valuations. But what certain companies might be able to do is ring-fence past liabilities. We are reaching a point where the possibility of running off past business may be the best future course for many companies, even if this is very costly

A qualitative way to think about this is the recent increase in insurance interest by private entities, be these hedge funds or true PE groups. The idea that private investors would need to be the predominant source of intensive capital is fully consistent with our thesis of risk quantunization, and now we are seeing the beginning signs that investors understand this as well

If this is correct, incumbent underwriters need to consider the possibility that investors wanting exposure to future insurance writings, be these a pricing cycle turn in non-life insurance or a new product cycle in life insurance, will bypass companies with possible legacy liabilities in favor of “clean” capital that will not lose 20% of its value immediately once it is injected into a public company. This may be true even for historically strong companies who think they can demonstrate that their liabilities are “manageable”

If you think we are being too pessimistic, such extreme views on past liabilities are not at all unprecedented. Lloyd’s of London was forced to mutualize all of its pre-1993 business in 1996 in order to end litigation and relieve investors of the capital strain from “open years”. While Lloyd’s stated off badly in the subsequent soft market of 1998-2000, in hindsight Reconstruction & Renewal as it was called is largely deemed a success, with the Lloyd’s franchise now greatly strengthened and more disciplined, and restructuring vehicle Equitas now owned by Berkshire Hathaway

It does not seem at all outlandish that many investors may consider old assets and liabilities to be “toxic” post-2007, or at a minimum something they do not want to bother with if there are alternatives. A private investment in a startup vehicle is one such alternative. Although current insurance managements may not see it this way, many investors might put legacy annuity products or long-tailed non-life contracts, along with various well-known asset classes, into the “do not want” bucket. Getting rid of old business perceived to be a capital drag is rewarded in the market, as was recently observed when Genworth sold off an Australian subsidiary

Of all the insurance subsectors, reinsurers may be best able to execute an organized runoff of some past business. The businesses of runoff and commutation are well-established in the reinsurance industry, and for the right price financial runoff might be a desirable investment for PE, as the founder of JC Flowers recently articulated. It is difficult to forecast exactly who might attempt runoff, but general possibility of such an occurrence needs to be on investor’s radar