Any real estate company that owns property in states exposed to hurricanes or
earthquakes is well-aware of how expensive it can be to insure those properties. Florida
alone accounted for $65.5 billion out of $427.8 billion of U.S. insured catastrophe losses
from 1982 to 2011.1 California’s Northridge Earthquake in 1994 resulted in $19 billion
in insured losses,2 and computerized models indicate that a repeat of the 1906 San
Francisco earthquake could produce insured losses of as much as $95 billion.3 Those costs
are factored into insurance premiums and are borne by policyholders.
Catastrophes include natural events such as earthquakes, hurricanes, tsunamis and
tornadoes, as well as man-made disasters such as large-scale terrorist attacks. Insuring
property in catastrophe-exposed regions always will be expensive, but real estate
companies should not be passive about their insurance costs. By understanding the
factors that contribute to the price of insurance, they can take steps to lower their
exposure to catastrophe losses and proactively manage their insurance spend.
The elements of property insurance pricing
The elements that comprise an insurance premium are similar for all types of insurance. In
general, premiums consist of “expected loss” – the amount allotted to pay claims – plus
various costs incurred by the insurer and an amount allocated for the insurer’s profit. For
some types of insurance, such as personal automobile insurance, a very large number of
comparatively small claims over time make it relatively simple for actuaries to determine
expected loss with a great deal of precision. Much the same can be said for calculating
the fire loss portion of expected loss for commercial property insurance. Natural
catastrophes, on the other hand, occur with far less frequency than do automobile
accidents and building fires, and the losses caused by a single event can range in the tens
of billions of dollars. Consequently, standard actuarial techniques, which rely on the law
of large numbers, do not work as well for estimating expected loss from hurricanes and
earthquakes. Insurance premiums for catastrophe-exposed properties therefore reflect
various costs associated with the uncertainty of insuring these risks.
For property insurance in catastrophe-exposed regions, three elements are of particular
significance: (1) the catastrophe component of expected loss; (2) the cost of reinsurance;
and (3) the cost of capital set aside to absorb large catastrophe losses.
Expected loss
Calculating the catastrophe component of expected loss is far more challenging than
calculating most other types of expected loss. Traditional actuarial techniques are
not very effective, so the expected loss component is often calculated using complex
computerized catastrophe models that mathematically simulate large catastrophes and
estimate the damage to properties in their reach.
The complexity and lack of transparency of most catastrophe models has led some critics
to term them “black boxes.” Although the inner workings of catastrophe models can be
a mystery, it is clear that various factors affect the outputs. Many of these factors can be
managed. For example, properties that are clustered in an exposed area are more likely
to have a higher total expected loss than a similar portfolio of properties spread over a
broader geographic area. Construction characteristics play a large role in how susceptible
buildings are to damage.
Calculating the catastrophe
component of expected loss
is far more challenging than
calculating most other
types of expected loss.
Based on PCS data adjused for inflation.
Cited in “Florida Hurricane Insurance: Fact
File,” Insuring Florida www.insuringflorida.org
Property Casualty Insurers Association of
America, www.pciaa.net
The 1906 San Francisco Earthquake and Fire:
Perspectives on a Modern Super Cat, RMS
www.rms.com
For hurricane exposed buildings, features such as roof anchors, engineered shutters
and how the roof sheathing is attached can have a big influence on how much damage
is expected to be sustained in a large storm. As a consequence, of the elements that
most influence property insurance premiums in catastrophe-exposed regions, real estate
companies have the most control over expected losses.
Catastrophe reinsurance
Among the costs that are factored into the price of insurance is the cost of reinsurance.
When reinsurance costs rise, as they often do following a large catastrophe, the price of
insurance is likely to rise as well.
Few insurance companies retain all the risk assumed under the policies they underwrite.
Even the largest insurance companies find it necessary – or at least financially desirable –
to transfer some portion of their risk to the reinsurance market. Reinsurance is a way for
insurance companies to guarantee that worse-than-expected claims will not financially
impair the company. An important role of reinsurance is to absorb property insurance
losses from large natural catastrophes.
For the most part, the net cost of reinsurance – the difference between the premiums
insurance companies pay for protection and the losses they recover from reinsurers over
time – is quite low relative to the total premium, and year-to-year fluctuations have little
or no impact on premiums paid by policyholders. However, when reinsurance costs shoot
up sharply, which sometimes happens after one or more very large catastrophes, it can
have a material impact on insurance premiums.
Individual real estate companies have no way to influence the cost of reinsurance and its
impact on their insurance premiums. However, they can make insurance buying decisions
based in part on how much reinsurance an insurer buys and, consequently, how much
influence reinsurance costs can have on property insurance premiums. In general, large
globally diversified insurance companies have less need for reinsurance, and have better
control over their reinsurance costs.
Volatility and risk capital
Another factor built into insurance pricing is the cost of the capital insurers set aside to
pay claims in case losses are much worse than expected. The amount of capital insurers
set aside varies by type of insurance, and is a function of the volatility of the losses for
each insurance type. The more volatile the claims experience, the more capital is required.
Because natural catastrophes represent one of the most volatile insurance exposures,
insurers that write property insurance in catastrophe-exposed areas must hold more cash
to cover potential claims than they do for property insurance in less exposed regions.
In order to get an equivalent return on the capital committed to catastrophe-exposed
business, insurers must charge higher premiums. The impact on property insurance
pricing will necessarily be much greater in states like Florida, Mississippi and California
than, for example, in Illinois, Minnesota or Nebraska.
As is the case with reinsurance, real estate companies have little control over this cost.
However, also like reinsurance, insurance buyers can make decisions that minimize the
impact of this factor on their insurance premiums. Once again, it is larger, more globally
diversified insurers that tend to fare better. Smaller companies with heavy concentrations
of properties in catastrophe-exposed areas typically must maintain proportionately more
cash reserves to provide the necessary buffer.
Catastrophe models
Computerized models that estimate the impact of natural disasters on properties
play important roles in all three elements of catastrophe premiums discussed above:
expected loss, reinsurance premiums and the cost of capital to absorb claims volatility.
Understanding how these models work provides important insights into the property
insurance pricing process. Additionally, real estate companies – especially those with large
property portfolios – can benefit from cat models to assess the catastrophe exposure to
their properties, to evaluate risk management options, and to better understand how to
lower their insurance costs.
All the major catastrophe modeling firms have models for both hurricane and earthquake
risk in the United States. Hurricane models mathematically simulate large numbers of
storms of various intensities, and calculate the aggregate damage, in dollars, they would
cause to properties in their paths. Key variables include concentrations of property values,
how close properties are to shorelines, and construction characteristics. Similar processes
are used to model earthquakes and the damage they can cause.
The computerized catastrophe modeling industry took off like a rocket in the aftermath
of Hurricane Andrew in 1992. Andrew landed a direct blow to Florida’s Atlantic coast
below Miami, causing $15.5 billion in insured losses and more than $25 billion in total
economic losses.4 The hit left the insurance industry dazed and panic-stricken. Prior to
Andrew, insurers widely agreed that a hurricane could result in insured losses of no more
than about $8 billion. The nearly $16 billion in insured loss from Hurricane Andrew was
virtually incomprehensible, invalidating insurers’ forecasting techniques and prompting an
urgent reassessment of catastrophe risks.
Computerized catastrophe modeling was still in its infancy, but it held out the promise
of a scientific, disciplined approach for pricing and managing catastrophe risk. One of
the first catastrophe modeling firms was Applied Insurance Research, now known as AIR
Worldwide, which was founded in 1987. After Hurricane Andrew, other firms followed in
its footsteps. Catastrophe modelers initially focused on US hurricanes, but subsequently
developed models for all types of catastrophes, including terrorism and pandemics, for
regions throughout the world.
Hurricane Katrina, which caused more than $40 billion in insured losses in 2005, was the
single most costly natural catastrophe in U.S. history.5 All the major catastrophe models
badly underestimated the damage caused by the storm, leading catastrophe modeling
firms to reconsider their assumptions about the amount of damage that can be caused
by a hurricane. One modeling company, Risk Management Solutions (RMS), revised its
models to better account for what it termed “loss amplification,” a “cascade of far more
damaging consequences” that can follow in the immediate wake of a major catastrophe.
Other modeling companies made similar adjustments.
Hurricane Ike in 2008 again challenged the models when the storm, which made landfall
in southern Texas, combined inland with another storm system and wreaked havoc as
far north as Ohio. Ike caused $12.5 billion in insured losses, making it the fourth most
costly storm in U.S. history. In 2011, RMS released version 11 of its model, based in part
on lessons learned from Ike. RMS 11 substantially raised certain estimates of potential
hurricane losses.
The nearly $16 billion in
insured loss from Hurricane
Andrew was virtually
incomprehensible,
invalidating insurers’
forecasting techniques and
prompting an urgent
reassessment of
catastrophe risks.
“Hurricane Andrew changed the worldwide
reinsurance market,” Business Insurance
www.businessinsurance.com
“Hurricane Katrina: Insurance Losses and
National Capacities for Financing Disaster
Risks” www.congressionalresearch.com
Catastrophe modelers have made huge strides in understanding and accounting for a
vast array of variables, but the models continue to develop. Hurricane Irene in 2011 and
Superstorm Sandy in 2012, for example, were comparatively weak storms that inflicted
enormous damage. Hurricane Irene raked the Caribbean and the eastern U.S. before
making landfall in the New York City area as a tropical storm. Irene continued inland,
causing wind damage and extensive flooding as far north as Vermont. Sandy, which
resulted in an estimated $25 billion in insured losses, proved especially challenging for the
models. Like Irene, Sandy had fallen below hurricane strength when it made landfall near
Brigantine, New Jersey. Nonetheless, it still managed to be one of the most destructive
storms in U.S. history due to its enormous size, the concentration of values in its path,
and high tides that amplified its storm surge. Undoubtedly model revisions are in the
works that will reflect new information gained from these events. Future events will
almost certainly compel modelers to again reassess their assumptions, perhaps resulting
in yet higher estimates of expected losses.
Catastrophe models take into account the growing concentration of property values in
areas highly vulnerable to catastrophes such as southern Florida and southern California.
A recent report from catastrophe modeling consulting firm Karen Clark & Co. estimated
insured building values in the U.S. at more than $40 trillion, with increasing concentrations
in areas subject to earthquakes and hurricanes. Nearly $15 trillion of insured property is in
the first tier of Gulf and Atlantic coastal counties. One reason Sandy was so devastating
was that the states in Sandy’s path account for 23 percent of U.S. GDP.9
Catastrophe modeling firm AIR Worldwide estimates that catastrophe losses will
double every decade due to this growing residential and commercial density. Higher
damage estimates from growing concentrations of property values are likely to result
in higher reinsurance costs over time, and may require larger commitments of capital
as buffers against worse-than-expected losses. These are costs that will be passed on
to policyholders.
For real estate companies – especially those with large numbers of properties in
catastrophe exposed regions – catastrophe models are useful for understanding how
much is at risk in a major storm, and how altering certain variables can influence loss
estimates. Catastrophe models can also arm insurance buyers with information about
their exposures than can be very useful when evaluating insurance options.
Future events will almost
certainly compel modelers
to again reassess their
assumptions, perhaps
resulting in yet higher
estimates of expected
losses.
“Natural Catastrophe Response.” NAIC
www.naic.org
“Natural catastrophe statistics of 2012
dominated by weather extremes in the USA,”
Munich Re www.munichre.com
Sandy’s economic hit may be softened by
cleanup, rebuild and insurance payments, NBC
News Business www.nbcnews.com
Conclusions
Globally, 2011 produced $107 billion in insured losses, the second highest on record
according to reinsurance intermediary Aon Benfield. At an estimated $25 billion in
insured losses, 2012’s Superstorm Sandy trails only Hurricane Katrina for insured losses
from a U.S. hurricane. Munich Re statistics strongly imply that the number of U.S. natural
catastrophes is on the rise. Higher insurance premiums for real estate in catastrophe-
exposed regions is inevitable, but real estate owners can take steps to control the
increases. Buildings can be made more resistant to wind and earthquakes, which may
result in lower premiums. Additionally, they can carefully choose their insurance partners.
Large, globally diversified companies are less reliant on reinsurance and have the benefit
of a broader spread of risk. Over the long run, these factors can contribute to more
stability and greater security.
Zurich
1400 American Lane, Schaumburg, Illinois 60196-1056
800 382 2150 www.zurichna.com
The information in this publication was compiled from sources believed to be reliable for informational purposes only.
All sample policies and procedures herein should serve as a guideline, which you can use to create your own policies and
procedures. We trust that you will customize these samples to reflect your own operations and believe that these samples may
serve as a helpful platform for this endeavor. Any and all information contained herein is not intended to constitute legal advice
and accordingly, you should consult with your own attorneys when developing programs and policies. We do not guarantee the
accuracy of this information or any results and further assume no liability in connection with this publication and sample policies
and procedures, including any information, methods or safety suggestions contained herein. Moreover, Zurich reminds you that
this cannot be assumed to contain every acceptable safety and compliance procedure or that additional procedures might not be
appropriate under the circumstances.
The subject matter of this publication is not tied to any specific insurance product nor will adopting these policies and
procedures ensure coverage under any insurance policy.
©2013 Zurich American Insurance Corporation
A1-112001577-A (06/13) 112001577
