Using technology to manage seismic risk in your property portfolio

Perspective on current seismic risk

The “out of sight, out of mind” approach is often that taken with respect to managing seismic risk in many regions of the United States. Although there are horrific reminders of seismic risk such as the recent devastating earthquake in Chengdu, China, or more gentle reminders, such as the recent magnitude 5.2 earthquake in Southern Illinois on April 18 earlier this year, seismic risk is typically a low priority on the list of issues addressed during the planning, design, construction and financing of new and existing facilities. Even in California and the Pacific Northwest, which have both experienced several damaging earthquakes during the last twenty years, the memory of seismic damage (which impacts the preparedness for future seismic events) tends to fade quickly after damage from recent events has been repaired.

A recently updated earthquake forecast for California was released by the Working Group on California Earthquake Probabilities [1] (WGCEP). A summary of the information is contained in the USGS fact sheets 2008-3027. The WGCEP determined that over the next 30 years, there is a 99% chance of having one or more magnitude 6.7 or larger earthquake in California. This magnitude of earthquake is similar to that of the 1989 Loma Prieta earthquake (M6.9) and the 1994 Northridge earthquake (M6.7). The probability of an even larger, potentially more damaging earthquake in the next thirty years in California is estimated at 50%.

Figure 1   San Francisco Bay Area Earthquake Probabilities [2]

The probabilities noted above can be broken down further by region, with the probability of a magnitude 6.7 earthquake in the San Francisco and Los Angeles regions in the next thirty years estimated to be approximately 65%.

The 750 mile long Cascadea subduction zone, which extends north from Northern California to Vancouver, Canada, continues to pose a significant risk to the Pacific Northwest with a USGS estimate of a 10% probability of a magnitude 8 to 9 earthquake along this zone in the next thirty years.

In the seismically active regions of the Central and Eastern United States, the absence of a recent damaging earthquake has led to questions and fierce debate about the potential for damaging seismic events in the region. In 1811 and 1812, the Midwest experienced the largest earthquakes ever recorded in the continental United States. The USGS estimates that the probability of an earthquake of a magnitude similar to that of the 1811–1812 earthquakes (magnitude 7.5–8.0) is approximately 7–10%, while the probability of a magnitude 6.0 or larger is estimated at 25–40% over the next 50 years. Due to the nature of the attenuation of the seismic waves in the Central US, an earthquake on the New Madrid Seismic Zone of a given magnitude will be felt over a much wider region as compared with an earthquake of similar magnitude on the West Coast.

In 1755, the Boston area was struck by an earthquake of magnitude 6 and in 1886,  Charleston, South Carolina, was devastated by a magnitude 6.7 earthquake. Since building codes in these areas of the United States have typically trailed those in the West Coast in requiring seismic design, there is significant potential for a high level of damage and financial loss if earthquakes of similar magnitude were to occur today.

Why worry about seismic risk in the portfolio?

A recent probabilistic study [2] by FEMA utilizing HAZUS to compare earthquake risk across the United States found that over 75% of potential average annual earthquake loss ($5.3 billion) in the US is situated on the West Coast and 66% of the total potential is concentrated within the State of California. However, $1.1 billion or 23% of annualized earthquake risk is distributed throughout the remainder of the US; some in areas that are not considered as having earthquake risk by investors and the financial markets.

In 2006, EQECAT conducted a study focusing on potential industry damage and loss resulting from a cascading earthquake rupture along the southern portion of the San Andreas fault and concluded that such an event could result in insured losses in excess of $27 billion on average and up to $50 billion if the Cholame/Coachella segment were to rupture. If such an event occurred, this would affect population and assets extending roughly from the City of San Bernardino in the north to Imperial County (California) in the south. Total societal expected damage and loss estimates for this event were $200 billion, with an upper range of up to $300 billion. [3]

Figure 2   EQECAT, Inc. Estimated Damage Areas, Simulated Rupture on San Bernardino – Coachella Fault Segment

Earthquake risk can have a significant impact on the financial performance of portfolios of residential and commercial building assets if not clearly identified and managed. Many owners of properties exposed to natural disasters such as earthquakes, carry inadequate levels of traditional earthquake insurance due to cost, insufficient capacity or other market conditions. Portfolio owners who are providing financial support to some or all of these properties may believe, inaccurately, that they have effectively managed or limited the risk to their portfolios through underlying or forced placed insurance coverage, loan restrictions, covenants or other means.

Changing financial market conditions such as reduced market values or increasing LTV (loan to value) ratios, could increase the risk of impact to portfolios from earthquake events otherwise considered to be immune from financial impact. Over time, new properties are typically added or removed from portfolios as mortgages are refinanced or properties are sold, potentially changing the risk profile of the portfolio given certain unique characteristics of the properties including location, age, construction, in addition to the wide variety of seismic events and their characteristics.

In a recent earthquake risk analysis of a nationwide residential and commercial mortgage portfolio, EQECAT found that the determined earthquake risk was quite sensitive to a decline in property values. An assumed 10% decline in property value within the studied portfolio resulted in a mortgage loss increment increase by approximately 2.5 times in the higher seismic risk areas for single-family books, and 1.3 times for multi-family property books. This study underscores an increase in the expectation of damage to a portfolio should a major seismic event occur simultaneously with a downturn in market property values in concentrated asset areas.

Technical standards to estimate damage from seismic events

Probable Maximum Loss (PML) studies are a standard measurement in the insurance, real estate, and financial industries for quantifying risk due to extreme hazards and the resulting financial losses.

The event of choice for the standard earthquake PML has historically been an event with a 475-year return period, which can also be expressed as an event that has a 10 percent chance of exceedance in 50 years. There are many theories as to the origin of the classic 475-year return period event; however, this return period had been used to define the Design Basis Earthquake in several of the primary building codes in the United States prior to the creation of the new International Building Code (IBC).

ASTM International recently released an update to the Standard Guide for theSeismic Risk Assessment of Buildings (ASTM Designation E2026-07). The 2007 update is similar to the 1999 standard, as much of the Guide concerns standardization of seismic risk assessments. The current ASTM provides definitions of terms related to both deterministic (scenario) studies and probabilistic studies that have become increasing popular because they provide  an evaluation that considers a statistical combination of all the events that may occur, but have not necessarily been observed at the site.

Although one of the objectives of the ASTM Standard is “to establish guidelines on what reasonably can be expected of and delivered by a Provider in conducting the seismic risk assessment of building”, facility owners are still faced with many technical choices with respect to definition of the seismic hazard and the actual methodology utilized to perform the evaluation.

Catastrophe models developed for the insurance market are used to support risk assessment from natural hazard events such as earthquakes. The EQECAT USQuakeTM model provides a view of a site or portfolio’s PML, in accordance with the ASTM Standard through a more detailed risk assessment given key characteristics of assets such as location, construction, configuration, age, height, and other factors. ABS Consulting’s new Property RiskTM website embedded with USQuakeTM , identifies damage potential for individual properties at varying return periods (expected intervals of recurrence), enabling users to “screen” properties for inclusion in a portfolio based upon earthquake risk. Inclusion or exclusion can then be subject to rules based upon damage contribution and correlation with the balance of sites within an existing portfolio.

Managing seismic risk through risk screening technology
Site asset characteristics can play an important role in the development of quantified damage to a property portfolio. The insurance market routinely measures its potential financial loss from claims arising from natural hazard events in its underwriting and accumulation management activities. This process depends upon catastrophe models such as USQuakeTM, which, given key characteristics of sites within insurance policies, can provide a realistic assessment of the average annual damage and loss (AAL) and probabilistic damage and loss at key return periods to support pricing and capacity management. Property portfolio owners can utilize this same technology to identify, mitigate and manage their embedded earthquake risk within their internal risk management program, enabling increased options and improved decisions on managing risk.

A property portfolio consisting of 170 sites across the US was analyzed using USQuakeTM to identify the 500-year return period PML for earthquake risk. This metric provides a damage threshold that has a 0.2% probability of being exceeded in any given year from one or more earthquakes. Conversely, there is a 99.8% probability that the damage from earthquakes in any given year would be below this threshold value.

In this study, certain key building characteristics were captured from the portfolio schedule including street address, construction type, age, number of stories and replacement value, and imported into USQuakeTM. The model produced a 500-year PML mean damage for this portfolio of $ 15.9 million (3.7% of replacement cost) with a 90th percentile damage of $222.5 million (5.3% of replacement cost) for the portfolio. Once this baseline had been established, the portfolio owners began using the web-based Property RiskTM tool to evaluate future inclusion of subsequent properties at the same level of return period based upon modeled percent damage. Their rule base allowed automatic inclusion of properties with a modeled percent damage value lower than 20%; required a more detailed on-site engineering analysis for those with modeled damages between 20% and 35%, and prohibited inclusion for those with modeled damage values exceeding 35%.

An example of the possible use follows: a client was considering two similar opportunities in the San Francisco Bay area for inclusion into its portfolio, and wanted to know if there were any differences from an earthquake risk perspective, since both properties were in the same earthquake zone. The underwriter used Property RiskTM for both locations and determined that Site A ($75 million replacement cost) incurred a 500-year PML of $14.3 million with a percent damage of 19.1%. Site B ($60 million replacement cost) developed a 500-year PML of $7.6 million but with a damage percentage of only 12.8%. In both cases, the PML estimates fell below the damage threshold for automatic consideration.

Figure 3   USQuake™ Generated 500-Year Portfolio Damage by Zip Code for Bay Area

In comparison with the original portfolio analysis, it was determined that Site A, when added to the existing portfolio, elevated the 500-year damage to $224.2 million, a PML increase of about $1.7 million versus an increase of only $500,000 for Site B. Since the firm was managing its portfolio aggregates to $250 million at the 500-year return level, Site B had a more beneficial impact to the overall portfolio while limiting the earthquake risk aggregation and so was selected.

Conclusion
Seismic risk is just one of several components that can affect the financial health or profitability of a property portfolio, but it is one that can be quantified and managed. Recent studies indicate that earthquake risk is increasing in certain areas of the country and this may have a compounded impact upon concentrated portfolios if an actual earthquake were to align with a downturn in property values or marked change in portfolio loan to value ratios.

Relying only upon traditional risk transfer (insurance) protection can be problematic, especially when insurance market capacity and pricing is uncertain and volatile, creating a “gap” in necessary insurance coverage. Any comprehensive risk management program for seismic risk should contain a five-step process that should include a comprehensive risk screening process for all property asset portfolios, regardless of the perceived level of risk from natural hazard events:

• Seismic Risk Quantification of Existing Portfolio
• Creation of Seismic “Underwriting” Procedures
• Development of Risk Screening Process for New Properties
• Optimization of Risk Transfer or Risk Financing Program
• Monitoring of Portfolio Against Changing Market Conditions

Risk screening technology through Property RiskTM provides portfolio owners with the ability to “weed” out certain properties with higher seismic risk prior to their being admitted into the portfolio, so that risk reduction options may be limited to expensive risk transfer. Screening for seismic risk can be used also to support a cost benefit approach to selecting those most likely to reduce overall portfolio risk from earthquake risk mitigation. The use of new risk screening technology may reduce the overall cost of managing seismic risk through limiting the amount of on-site risk assessments required and focusing efforts where the benefits are greatest.

• References:
  2007 Working Group on California Earthquake Probabilities, 2008, Forecasting California’s earthquakes; what can we expect in the next 30 years?:  U.S. Geological Survey, Fact Sheet 2008-3027
• FEMA 500 C Street SW, Washington, D.C. 20472, Annualized Earthquake Losses (FEMA 366), 2006
• Time-Dependent Earthquake Ruptures and Southern California Earthquake Risk, EQECAT, Inc., June 2006