IPCC Risk & Uncertainty

I haven’t had time to go into the latest WGIII IPCC report in detail (indeed I haven’t had much time recently to spend on blogging) but I had a quick browse through the report and there is an excellent chapter on “Integrated Risk and Uncertainty Assessment of Climate Change Response Policies” which goes through many of the key elements of current risk management theory and practise and how they can be applied to climate change.

A previous post highlighted the difficulties of making predictions given the uncertainties involved. The report highlights the “large number of uncertainties in scientific understanding of the physical sensitivity of the climate to the build‐up of GHGs” and acknowledges that these “physical uncertainties are multiplied by the many socioeconomic uncertainties that affect how societies would respond to emission control policies”. The report calls these socioeconomic uncertainties “profound” and lists examples as the development and deployment of technologies, prices for major primary energy sources, average rates of economic growth and the distribution of benefits and costs within societies, emission patterns, and a wide array of institutional factors such as whether and how countries cooperate effectively at the international level.

The IPCC gives a medium rating (50% probability) to the statement that the “current trajectory of global annual and cumulative emissions of GHGs is inconsistent with widely discussed goals of limiting global warming at 1.5 to 2 degrees Celsius above the preindustrial level”.

Included in Chapter 2 are the graphs below. According to the report “the representative concentration pathways (RCPs) are constructed by the IPCC on the bases of plausible storylines while insuring (1) they are based on a representative set of peer reviewed scientific publications by independent groups, (2) they provide climate and atmospheric models as inputs, (3) they are harmonized to agree on a common base year, and (4) they extend to the year 2100”. The 3 scenarios (A2, A1B and B1) are multi-model global averages of surface warming (relative to 1980–1999) shown as continuations of the 20th century simulations. Shading is the plus/minus one standard deviation range of individual model annual averages and the orange line is where concentrations were held constant at year 2000 values. Time permitting; it demonstrates that the conclusions and scenarios presented in the latest report are worth finding out more about.

click to enlargeIPCC global surface temperature scenarios from RCPs

Although each scenario is likely at the mercy of the uncertainties highlighted above, the open and thoughtful way the report is presented, including highlighting the underlying weaknesses, doesn’t mean that they (or the report) can be ignored. Indeed, the recent output from the IPCC will hopefully provide the basis for informed thinking on the subject in the coming years.

The report includes a reference to Kahneman-Tversky’s certainty effect where people overweight outcomes they consider certain, relative to outcomes that are merely probable. That implies that a 50% probability of the temperature blowing through 2 degrees celsius may not be enough to force real action. Unfortunately the underlying scientific and socioeconomic uncertainties inherent in making forecasts on temperature change over the next 30 to 50 years may mean that the required level of certainty cannot ever be achieved (until of course it’s too late).

Pricing Pressures & Risk Profiles

There have been some interesting developments in the insurance market this week. Today, it was announced that Richard Brindle would retire from Lancashire at the end of the month. The news is not altogether unexpected as Brindle was never a CEO with his ego caught up in the business. His take it or leave it approach to underwriting and disciplined capital management are engrained in Lancashire’s DNA and given the less important role of personalities in the market today, I don’t see the sell-off of 5% today as justified. LRE is now back at Q3 2011 levels and is 25% off its peak approximately a year ago. As per a previous post, the smaller players in the specialty business face considerable challenges in this market although LRE should be better placed than most. A recent report from Willis on the energy market illustrates how over-capacity is spreading across specialist lines. Some graphs from the report are reproduced below.

click to enlargeEnergy Insurance Market Willis 2013 Review

One market character who hasn’t previously had an ego check issue is John Charman and this week he revealed a hostile take-over of Aspen at a 116% of book value by his new firm Endurance Specialty. The bid was quickly rejected by Aspen with some disparaging comments about Endurance and Charman. Aspen’s management undoubtedly does not relish the prospect of having Charman as a boss. Consolidation is needed amongst the tier 2 (mainly Bermudian) players to counter over-capacity and compete in a market that is clustering around tier 1 global full service players. Although each of the tier 2 players has a different focus, there is considerable overlap in business lines like reinsurance so M&A will not be a case of one and one equalling two. To be fair to Charman the price looks reasonable at a 15% premium to Aspen’s high, particularly given the current market. It will be fascinating to see if any other bidders emerge.

After going ex-dividend, Swiss Re also took a dive of 9% this week and it too is at levels last seen a year ago. The dive was unusually deep due to the CHF7 dividend (CHF3.85 regular and CHF4.15 special). Swiss Re’s increasingly shareholder friendly policy makes it potentially attractive at its current 112% of book value. It is however not immune from the current market pricing pressures.

After doing some work recently on the impact of reducing premium rates, I built a very simple model of a portfolio of 10,000 homogeneous risks with a loss probability of 1%. Assuming perfect burning cost rating (i.e. base rate set at actual portfolio mean), the model varied the risk margin charged. I ran the portfolio through 10,000 simulations to get the resulting distributions. As the graph below shows, a decreasing risk margin not only shifts the distribution but also changes the shape of the distribution.

click to enlargeRisk Premium Reductions & Insurance Portfolio Risk Profile

This illustrates that as premium rates decline the volatility of the portfolio also increases as there is less of a buffer to counter variability. In essence, as the market continues to soften, even with no change in loss profile, the overall portfolio risk increases. And that is why I remain cautious on buying back into the sector even with the reduced valuations of firms like Lancashire and Swiss Re.

Deflating Bubbles

The market is currently full of overdue anguish, with the air coming out of some of the frothier areas of the market notably in the biotech and internet sectors. To get an idea of the movements, I had a quick look at the S&P500 against a number of other indices such as the Powershares S&P high beta, S&P low volatility, & Nasdaq Internet ETFs plus the Nasdaq Biotech Index (SPHB, SPLV, PNQI & NBI respectively) as per the graph below.

click to enlargeS&P500 vrs SPHB SPLV NBI PNQI

One of the more amusing bubbles in the recent run-up has been that surrounding the creeping legalization of cannabis in the US. Penny stocks in the sector, as if straight out of “The Wolf of Wall Street”, have been rocketing. Some of the more dubious firms have jumped on the bandwagon by coming up with fanciful plans on exploiting cannabis markets after having tried their luck as software, oil exploration or even tanning companies! Firms such as CannaVest (CANV) and Vape Holdings (VAPE) have shown classic pump & dump penny stock rises and falls in recent months.

One stock that has rode both the biotech and the cannabis buddle is a UK firm called GW Pharmaceuticals (GWP.L) founded in 1998 to develop cannabinoid prescription medicines to meet patient needs under medical supervision. Their main product, Sativex, a treatment for moderate to severe spasticity is approved or near approval in a number of countries such as Norway, Israel, and Austria. Bulls point to approval in the US of Sativex and the potential for other cannabinoid products in areas such as cancer and diabetes to justify the current valuation of multiples of revenue for this loss making firm. GWP has risen from 50p last year to a high (forgive the pun!) of 400p in March with a fall back to 250p recently. Cannabis stocks offer the ultimate high for aging stoners, add in some biotech hype for GWP and the sky is the limit to a happy ever after fantasy……….will people ever learn!

To me, the deflating of sector bubbles is a very healthy sign of a rational market.  Whether an outbreak of rationality will last is another matter.

Smart money heading for the exits?

Private equity is rushing to the exits in London with such sterling businesses as Poundland and Pets at Home coming to the market. PE has exited insurance investments, following the successful DirectLine float, for names like Esure, Just Retirement, and Partnership. It was therefore interesting to see Apollo and CVC refloat 25% of BRIT Insurance last week after taking them off the market just 3 short years ago.

The private equity guys made out pretty good. They bought BRIT in 2011 for £890 million, restructured the business & sold the UK retail business and other renewal rights, took £550 million of dividends, and have now floating 25% of the business at a value of £960 million. To give them their due, they are now committing to a 6 month lock-up and BRIT have indicated a shareholder friendly dividend of £75 million plus a special dividend if results in 2014 are good.

I don’t really know BRIT that well since they have been given the once over by Apollo/CVC. Their portfolio looks like fairly standard Lloyds of London business. Although they highlight that they lead 50% of their business, I suspect that BRIT will come under pressure as the trend towards the bigger established London insurers continues. Below is a graph of the tangible book value multiples, based off today’s price, against the average three year calendar year combined ratio.

click to enlargeLondon Specialty Insurers NTA multiples March 2014

Munich’s Underwriting Cycle

Munich Re had a good set of results last week with a 12.5% return on equity on a profit of €3.3 billion (with the reinsurance business contributing €2.8 billion of the profit). A €1 billion share buyback was also announced contributing to the ongoing shareholder friendly actions by industry players. Munich is targeting €3 billion for 2014 but warned of challenges ahead including “the lingering low-interest-rate environment, increasing competition in reinsurance, and changes in demand from clients in primary insurance”.

Torsten Jeworrek, Munich Re’s Reinsurance CEO, cited tailor-made solutions as a strength for Munich highlighting “multi-year treaties (occasionally incorporating cross-line and cross-regional covers), retroactive reinsurance solutions, transactions for capital relief, comprehensive consultation on capital management, and the insurance of complex liability, credit and large industrial risks”.

Whilst looking through the 2013 report, I noticed historical calendar year combined ratios (COR) for the P&C business (reinsurance & primary) including and excluding catastrophes. I dug up these figures going back to 1991 as per the graph below. A small amount of adjustment was needed, particularly in relation to the 24.3% and 17.1% of deterioration for 2001 and 2002 relating to 9/11 losses (which I included as catastrophes in the CaT ratio for those years). As with a previous post on underwriting cycles, I then “normalised” the COR excluding catastrophes for the changes in interest rates using a crude discount measure based upon the US risk free rate for each calendar year plus 150 bps over 2.5 years. That may be conservative, particularly for the 1990s where equities were a bigger part of European’s asset portfolio. I then added the (undiscounted) CaT ratio to the discounted figures to give an idea of the historical underwriting cycle.

click to enlargeMunich Underwriting Cycle

The “normalised” average discounted COR (excluding CaT) since 1991 is 87% and the average over the past 10 years is 83%. The standard deviation for the series since 1991 is 6% and for the last 10 years 4% indicating a less volatile period in recent years in core ratios excluding catastrophes.

The average CaT ratio since 1991 is 7% versus 9% over the past 10 years. The standard deviation for the CaT ratio since 1991 is 8% and for the last 10 years 9% indicating a more volatile period in recent years in CaT ratios.

Adding the discounted CORs and the (undiscounted) CaT ratios, the average since 1991 and over the past 10 years is 95% and 92% respectively (with standard deviation of 11% and 9% respectively).

As Munich is the largest global reinsurer, the ratios (reinsurance & primary split approx 80%:20%) above represent a reasonable cross section of industry and give an average operating return of 5% to 8% depending upon the time period selected. Assuming a 0.5% risk free return today, that translates into a rough risk adjusted return as per the Sharpe ratio of 0.44 and 0.80 for the period to 1991 and over the past 10 years respectively. Although the analysis is crude and only considers operating results, these figures are not exactly earth-shattering (even if you think the future will be more like the last 10 years rather than the longer term averages!).

Such results perhaps explain the growing trend of hedge funds using reinsurance vehicles as “float” generators. If the return on assets over risk free is increased from the 150 bps assumed to 300 bps in the analysis above, the Sharpe ratios increase to more acceptable 0.73 and 1.13 respectively. And that ignores the tax benefits amongst other items!

As an aside, I again (as per this post) compared the underlying discounted COR (excluding catastrophes) from Munich against a credit index of global corporate defaults (by originating year as a percentage of the 1991 to 2013 average) in the graph below. As a proxy for the economic & business cycles, it illustrates an obvious connection.

click to enlargeMunich Underwriting & Credit Cycle