Every year reinsurance actuaries and other professionals gather to collaborate and discuss a wide range of issues facing the reinsurance industry.
The event is sponsored by the Casualty Actuarial Society (CAS), the oldest actuarial society in the United States and the only one focused on property and casualty.
This year it was held on June 6th and 7th in beautiful Bermuda and I am very happy to have attended. One thing I learned over there is that shorts (known as Bermuda shorts) and dress socks are considered formal business wear on the island.
But it was far from just me over there. Eight other Willis associates attended as either presenters or moderators. I am pleased to report that Willis’s contribution to this year’s conference was exemplary.
Prasad Gunturi, the tornado catastrophe modeling practice leader, presented the Willis eXTREME Tornado product.
As many of us are aware from the recent events in Oklahoma, summer storms can be devastating – both in terms of lives lost and destroyed property. Insurers in tornado prone areas must understand their risks.
Prasad presented on the challenges of tornado modeling generally and on Willis eXTREME Tornado specifically, which can take a client portfolio and estimate losses from high severity events.
Unlike big vendor models, the event set used in eXTREME Tornado is built specifically for the client portfolio as opposed to industry exposure.
Prasad also spoke on the importance of data quality in catastrophe modeling data. With improved data quality our clients can potentially reduce their expected losses through better underwriting guidelines.
As some of you may know, Willis has a long history as a global leader in marine insurance. In fact, Willis was the insurance broker for the ill fated Titanic. Jason continued the Willis tradition on marine leadership with a presentation on the modeling of cargo accumulation at ports.
Transportation by ships is a major form of transportation for goods around the world. The cargo accumulates at ports where they can be exposed to catastrophic wind, fire and water, as well as theft and mishandling.
The exposure at ports from an insurance perspective is poorly understood, and Jason presented techniques to improve our understanding of a particular client’s exposure, such as mapping the client’s mix of goods to the ports that actually handle those goods. Jason also presented a summary of the damage at the port of Newark due to Hurricane Sandy.
Medical professional liability (MPL) dates back to the time of Hammurabi, when doctors could receive severe physical punishment for medical errors.
Today MPL claims can cost doctors and hospitals millions of dollars, a liability which can be insured and reinsured in modern markets.
Brian Ingle, chief actuary of Willis Re North America, presented on the state of the MPL re/insurance market in the United States, particularly in light of the health care reform bill (PPACA), signed into law by President Barack Obama in 2010.
According to Brian, MPL insurers in the United States have made record profits in recent years, leading to record levels of capital and a therefore continuing trend towards a reduction in reinsurance premiums.
However, the PPACA may increase the demand for MPL reinsurance. While greater industry consolidation might reduce the need, reforms in the system will likely lead to greater frequency (more people receiving health care than before) and may lead to more mass torts.
In the last couple decades we have seen the use of catastrophe modeling in the re/insurance industry become nearly ubiquitous. Insurance companies seeking to understand their exposure to perils such as hurricanes and earthquakes may turn to variety of cat model vendors.
Each vendor may have multiple versions of the same model and multiple methodologies (e.g. long term vs. near term hurricane), which can lead to an overwhelming situation for many of our clients!
Given this situation, many Willis clients may want to simplify their view of their own risk by blending several models into a single model. Some of this is being driven by Solvency II, which requires a single model.
Ian Cook, Chief Actuary for Willis Re International and Specialty, presented cutting edge applied research from Willis regarding the proper ways to blend models.
One thing I found interesting was that it would be correct to average probabilities for a given size catastrophic loss, but it would be incorrect to average catastrophic loss sizes for a given probability.
A typical insurance contract, called an indemnity contract, pays out when the insured has an actual monetary loss as per the contract – a fire loss, a theft loss, etc. However, a major innovation in the insurance industry has been the use of indexes.
Under an index based contract the payout happens not when the actual loss happens but when the index is triggered. For example, an agriculture contract might pay out not when crops are actually lost to drought, but rather when a specified number of inches of rain have failed to fall.
Bill Dubinsky, Head of Insurance-Linked Securities at Willis Capital Markets & Advisory, presented on the pros and cons of using indexed based contracts versus indemnity contracts in reinsurance.
On the pro side, index based reinsurance is more transparent, less complex (less associated expenses) and more appealing to capital market investors.
On the con side, index based reinsurance introduces a risk called basis risk – which is the risk that the index will differ significantly from the actual loss. However, this risk can be quantified and mitigated for prospective Willis clients.
As we all know, the insurance industry goes through good times and bad times. It’s a cyclical pattern known as the underwriting cycle and understanding and predicting it can have significant impact for Willis clients.
Raj Bohra, head of Financial Services North America, presented his analysis of the underwriting cycle for the primary casualty market in the United States – workers compensation and general liability. In the good part of the cycle, premiums relative to the exposure are high or increasing (hardening) and loss ratios are low or decreasing.
But eventually premium rates start to decline (softening) and then the loss ratios get worse, leading to the bad part of the cycle.
According to Raj the workers compensation market has seen clear signs of hardening in 2011 and 2012. The general liability side has also seen some hardening, however the sustainability of its decreasing loss ratio trend is unclear.
Understanding market cycles can help companies time their reinsurance purchases and new entrants pick the best time to enter. Raj also moderated a forum on hurricanes in the Northeast region of the United States and, along with Brett Shroyer, helped to organize the conference.
Neil Bodoff is one of the top researchers in the actuarial community. He was written papers on topics as diverse as economic capital allocation and catastrophe bond pricing (the latter which won him, along with fellow Willis colleague Yunbo Gan, VP the 2009 Prize in Reinsurance).
At this conference he presented a paper entitled “An Actuarial Model of Excess Policy Limits”.
Sometimes a court of law may force an insurer to pay an amount greater than the limit of the insurance policy. This is called an excess policy limit (EPL) loss.
It is often not contemplated in the pricing of the original insurance policy and often comes as a shock to the insurer. EPL losses have important implications in reinsurance pricing as well.
Neil presented an elegant framework for analyzing EPL loss contribution to excess of loss (XOL) reinsurance contracts.
He starts by assuming that there is a probability p that a policy will not have an EPL loss. p is a high number, such as 99%. Thus there is a small probability (e.g. 1%) that there might be an EPL loss.
He then shows how to use actuarial formulas and simulation methods to derive the possible EPL contribution to the XOL contracts.
And so that leaves me! I’m the actuary for the Miami office of Willis Re, which focuses on placing reinsurance for our clients in Latin America and the Caribbean.
Many clients in this region are geographically concentrated, with high exposure to hurricane and earthquake.
Accurate pricing for their catastrophe excess of loss reinsurance (Cat XL) is a very important budgeting concern for these clients, some of which can spend nearly half of their gross premium on their Cat XL programs.
I presented a paper entitled “Pricing Catastrophe Excess of Loss Reinsurance using Market Curves”.
The idea is to price a Cat XL program not with the typical actuarial methodology of using catastrophe models and risk loads (bottom up), but rather by adjusting an actual Cat XL program that has been placed in the market (top down).
For example, we could fit a power curve through last year’s Cat XL program, and after adjusting for the growth in exposure for the company, estimate costs for varying Cat XL layers.
This is a method that has been around in the London market since the early 1990s, and is in fact widely used in Willis. I explored this method and further refined it by using a spline curve instead of a power curve. My paper was awarded the 2013 Prize in Reinsurance.
What Can we Conclude From all This?
Willis’s contributions to this reinsurance conference and to reinsurance in general is wide and deep. We had nine colleagues from Willis serve the conference in various capacities – presenters, moderators, organizers and paper writers.
We are on a roll with the Reinsurance Prize, winning two of the last three prizes. Perhaps we aren’t the world’s biggest broker, but clearly we have the best talent and we should all be proud to work here!
David Morel is a consulting actuary for the Miami office of Willis Re. He provides dynamic reinsurance analysis, pricing and other actuarial services to clients throughout Latin America and the Caribbean. He joined Willis and achieved his associateship in the Casualty Actuarial Society in 2009 (ACAS).