Overview

Katrina is the largest insured loss ever seen in today's dollars. Even when measured as a percent of industry resources, it is the largest loss current professionals have seen, surpassing hurricane Camille in 1969 (Katrina at 10%+ of US P&C industry surplus vs. Camille at 8.5%). Neither of these events approach the impact of the 1906 San Francisco Earthquake. Katrina ranks among the top ten deadliest US events. While it is a big one, it is not necessarily "the" big one. We have conclusions and observations about the impact of Katrina on the market, but please note that we are writing this with Rita strengthening in the Gulf, and the situation could change dramatically.

There have been a variety of releases and estimates from modelers, rating agencies, reinsurers and other analysts variously estimating Katrina as costing from below $15Bn to over $150Bn, including various categories of insured and uninsured losses. This is a difficult loss to measure, because it is widespread and involves literally hundreds of insurers and reinsurers. It is a challenge to adjust, because transportation and other infrastructure problems are still keeping adjusters far from some affected areas. This is also the first major loss under the Sarbanes-Oxley disclosure rules, and less insurer information is available than in prior events. Loss reports may also be "spiraling" as the same gross loss is in part reported by both insurers and their reinsurers. (Although worldwide insurers' published net, pre-tax losses, already exceed $18Bn. (See Exhibit 5.)

Holborn estimates that the direct insured Property and Marine market loss due to Katrina will be between $45Bn and $65Bn, assuming that Flood exclusions and sub-limits are interpreted as intended. Eliminating Flood exclusions could add tens of billions more exposure, but that is outside of our current estimates. We consider the direct loss and LAE that will be covered by the worldwide Non- Marine and Marine insurance markets, including state pools and funds that have assessment authority, but excluding the U.S. Government's National Flood Insurance Program (NFIP), which is funded outside of the industry's resources. Only reinsurers' direct losses are included in our assessment of industry totals. The high fatalities will cause losses to Life, WC and Prem-ops coverages. Agents' E&O may also have losses. But it is too early to make meaningful estimates for these coverages.

Total Insured Property and Marine Loss and LAE:
Non-Marine Wind - South Florida Landfall	$1Bn to $2.25Bn
Non-Marine Wind - Gulf Coast Landfall	$25Bn to $32.5Bn
Marine Wind	$5Bn to $10Bn
Market Insured Flood	$5Bn to $15Bn
Other Insurable Perils	$2.5Bn to $7.5Bn
Flood Coverage Disputes	Loss of $0 (current assumption) and LAE of $1Bn to $3Bn
Insured Loss and LAE	$45Bn to $65Bn

In nominal dollar terms, Katrina surpasses even the September 11th, 2001 attacks, which we currently estimate to be slightly below $40Bn (although litigation continues), and the combined 2004 hurricane season, at approximately $35Bn. (Our estimate for 9/11 includes WC and LAE, and Marine, Aviation, Foreign and Non-Admitted Carriers, and is thus significantly higher than PCS's estimate for the admitted market.) Katrina is multiples of any previous individual natural catastrophe event, such as Andrew, Northridge, Hugo, Japan's Kobe and Mireille losses and the Lothar and Martìn European winter storms.

As a percentage of industry resources, Katrina is comparable to, but even greater than, Andrew (at less than 8% of U.S. industry surplus). However, we expect fewer insolvencies (although some are probable) due to a better understanding of hurricane peril and improved risk management at many of the affected insurers. There will be many downgrades among the twenty companies Best's put "under review" on September 16th. The impact on the reinsurance market will be significant, although we do not expect as many reinsurer impairments and withdrawals in 2006 as resulted from Andrew during 1993 renewals. We expect critical re-examinations of capacity offered, pricing and underwriting approach by every Property reinsurer in the market.

Losses by Peril

Non-Marine Wind Losses:

Katrina had total non-marine Wind losses of $26Bn to $35Bn. Katrina is the largest ever Non-Marine windstorm loss, although relative to surplus, Betsy, Camille and Andrew were greater on-shore industry wind-only events (6% - 7% vs. 8%+).

Katrina's first landfall was near North Miami Beach, Florida, with 80 mph winds, at 7PM EDT on Thursday, August 25th. Various model estimates of a $1Bn - $2Bn loss seem reasonable. Southeastern Florida was largely spared by the 2004 losses, although we expect demand surge to be above what these model loss estimates anticipate. At these levels, the FHCF should be only minimally involved.

	Gross Losses ($Billions)
National Personal Companies	$.30 – $.60
Florida HO Specialists	.15+ – .35+
Regional and Other Multi-Line	.20 – .40
Citizens Insurance	.30+ – .60+
Primary Market – Incl. LAE	$1.1 – $2.25

There is much wider uncertainty about the wind losses caused by Katrina's landfall on the Gulf Coast, crossing over Plaquemines Parish, Louisiana at 7AM local time on Monday, August 29th, and continuing on-shore again near Gulfport, Mississippi by 10AM. The modelers' most recent estimates for Non-Marine Wind losses on this landfall are:

• EQECat $14Bn - $22Bn


• AIR $17Bn - $25Bn


• RMS $25Bn - $35Bn

This averages to a range of $19Bn - $27Bn. We believe that these estimates may prove to be low for two reasons. First, as a direct comparison to hurricane Andrew in 1992, while the wind speed is slightly lower, the affected population area is much larger (See map in Exhibit 1). In both Andrew and Katrina, most Weather Service reporting stations were destroyed by the extreme winds. Based on offshore readings, and the degree of damage seen on-shore, current estimates are that Andrew had maximum sustained winds over land of 165 miles per hour, while Katrina was moderately lower at 145 miles per hour, although even higher at sea, at 185. Either storm's peak on-shore wind speed is sufficient to destroy most frame structures. Although, there may be moderately less damage from Katrina to well-built commercial structures due to the difference in wind speed.

Katrina had a deeper central pressure than Andrew, at 26.64 (902 mB) inches vs. 27.23 inches (922 mB) and by this measure is the most extreme storm observed in US waters since this level of data has been captured. Only hurricane Gilbert that hit Mexico in 1988 was more intense at 26.22 inches (888 mB).

However, the population affected by Katrina is over 25,000,000 people. Katrina was a notably wide storm (Hurricane winds out to 125 miles, the widest measured Category 4), while Andrew was a very focused "shotgun blast," and largely missed the core of the Miami-Dade metropolitan area. Also, the interior regions of Florida moving west of Homestead are entirely everglades, and there was little or no property value exposed to destruction inland. The affected population for Katrina was significantly greater than that for Andrew, at least one-third higher. Tragically, we note that the toll of lives in Mississippi alone is multiples of, perhaps ten times, the 26 persons who perished in Andrew.

The destruction of the weather instruments means that the modelers' estimates are essentially based on aerial photography taken immediately after landfall. At the time of this writing, we are concerned that the modelers have not been able to provide more detailed information about specific local damage levels. In insurance claims practice, no news is rarely good news. We do not believe that the amount of damage caused by Katrina is any less than that caused by Andrew, in physical terms, and that it would at least meet the $21.4Bn of loss that Andrew cost, when restated in 2005 dollars (See Exhibit 3). This includes Business Interruption (BI) and other time element coverages.

Demand surge will be a very challenging issue for the Gulf Coast landfall. The Mississippi and Alabama coasts and Florida Panhandle are still recovering from Category 4 hurricane Ivan on September 16th, 2004 and Category 3 hurricane Dennis on July 10th, 2005. (See Map in Exhibit 2.) The entire southeastern construction industry is still reflecting the very significant stress from the 2004 rebuilding efforts, which are not complete at this date. Moreover, the economic impact of the current insured and uninsured Flood losses in both Mississippi and the New Orleans basin are competing for scarce labor and material, and further bid up the cost of repairing wind damage. Holborn and others estimate that the cost of repairs will be closer to 40% above normal cost levels in this area. This is significantly higher than what prevailed during the post-Andrew period and also higher than the 20% - 30% estimated for Katrina by the modelers' wind-only perspectives.

The industry's recognition of mold is another factor which is different from the time of Andrew. While mold is excluded as a peril from most Property insurance policies, mold that results from water damage following an insured peril is generally covered, although perhaps subject to a sub-limit. In addition, whenever water damage is a cause of loss in a climate where mold is likely, insurers' standard of care in reconstruction work on partial losses is now much higher than it was following Andrew. Several high-profile verdicts against insureds and insurers have raised the bar. Alabama, Louisiana and Mississippi have climates that certainly promote mold and, unfortunately, also promote insurance litigation. We believe that the recognition of mold, either as an insured loss, or as higher standards for repair work, could add up to 10% to losses, when compared to the costs after Andrew.

We estimate that the industry Non-Marine loss and LAE due to the Wind peril, and the insured parts of coastal storm surge, and other related perils, such as glass breakage and debris removal, will total $25Bn - $32.5Bn for the Gulf Coast landfall, considered alone. This is in addition to the estimated $1Bn - $2.25Bn loss from the first landfall.

Marine Wind:

We estimate that Katrina's Marine loss due to wind and wave action is $7Bn - $10Bn of loss and loss adjustment expenses. Katrina is also the largest ever Marine loss, although 1998's Piper Alpha explosion in the North Sea (a $2.7Bn loss) may have been a greater share of Marine insurance industry income.

Our starting point for analyzing this exposure is hurricane Ivan, which brushed the off-shore oil and gas fields south of the Louisiana and Mississippi coastlines. The most recent estimate for the off-shore loss on Ivan is $2.6Bn in insured Marine property damages and related coverages. Katrina had comparable wave forces to Ivan, although with stronger winds, but the storm moved through the center of the producing area, as opposed to having brushed the side. We estimate off-shore damage at 50% to 150% above Ivan's. In addition, the off-shore loss is subject to the same demand surge as the on-shore losses, and one year of price inflation in the already over-heated energy sector. There is also a significant amount of on-shore damage to piers and marinas, as well as what is likely the largest-ever yacht loss across this very wide recreational area.

Flood Losses:

We estimate $5Bn to $15Bn of insured Flood loss and LAE. Katrina was the largest-ever market insured Flood loss, exceeding the 2003 Czech floods, which cost the industry Euro 4Bn ($5Bn).

At the afternoon high tide at 5PM on August 29th, it appeared that Lake Pontchartrain had reached the top of some levees, but that the levees had survived. Some water came over, not through, the levees. Later, on the 30th, it was clear that there were three major breaches, and the low-lying areas of New Orleans were inundated by water up to lake and river levels. 80% of this city of 500,000 was beneath as much as 20 feet of rapidly contaminating water.

On September 8th, RMS released a press release citing a combined insured loss of $40Bn - $60Bn for Katrina and "the great New Orleans flood," which they view as two separate events, suggesting two reinsurance retentions, and also two limits. The announcements left their wind loss unchanged at $25Bn - $35Bn of Marine and Non-Marine loss. Their estimates of insured New Orleans Flood losses are $15Bn - $25Bn. The insured loss estimate includes a very rough sum of $10Bn for losses insured by the NFIP. Thus, RMS estimates that the insured New Orleans Flood losses covered by the private insurance market are between $5Bn - $15Bn. We feel this is reasonable based on a review of the exposures and coverages which we believe apply.

RMS based this estimate on a 40% take-up rate for the NFIP program in Orleans Parish, applied to all coverages, and thus implies a total (insured and uninsured) New Orleans Flood loss of $37.5Bn to $62.5Bn. We believe that 40% is also a reasonable estimate of the Commercial values covered for Flood, considering 25% - 30% of Property Policy limits and 100% of Auto, IM and Equipment coverage.

Assuming that residential values are two-thirds of the total, RMS's estimate suggests a cost for lost residences of $25Bn to $42Bn. That is consistent with our calculation of 170,000 destroyed homes in New Orleans, at an implied cost of $150,000 to $225,000 per unit, including contents and ALE. Considering the neighborhoods involved, many of the homes are rented apartments. While they may have Flood coverage (especially if written through the national stock companies), it is often sublimited to $1Mn or $2Mn per policy, with time element coverages, such as BI, included in the sublimit. On a typical 5-to-25 units BOP policy, $1Mn - $2Mn is a meaningful limit. Many Tenants policies also include Flood, as do all Personal Articles policies. Mobilehome policies written on Automobile or similar forms include Flood. We do not believe that many off-premises or contingent BI forms cover Flood.

We assume that the insured values are one-third each for private homes, apartments, and retail, and that implies $25Bn to $40Bn of Commercial Property exposure in New Orleans. That develops $10Bn - $16Bn of Commercial Flood loss in New Orleans alone, assuming full limits are placed. Sub-limits on Commercial policies will reduce this. But there is also loss outside of New Orleans, on other sides of Lake Pontchartrain and in Mississippi, and on Auto and IM policies. We agree with the RMS range of $5Bn to $15Bn, but for the whole region. The individual risks are likely almost all total losses, so demand surge will not be as much of a factor. (Guaranteed replacement cost coverage is unusual for Commercial Flood.) We do need to consider loss adjustment at 5% - 10%, but with the width of this range, that is not significant.

Other Insurable Perils:

We estimate that the perils other than Wind and Flood will have insured loss and LAE of between $2.5Bn and $7.5Bn. Considering a total Flood loss of $37.5Bn - $62.5Bn, and an insured part of $15Bn - $25Bn (including NFIP), we consider what other insured perils could respond to the remaining $12.5Bn - $37.5Bn of damage. In many cases there is Wind loss, however this is already included in our wind estimates above. Also, there are many properties that are damaged by fire and looting, and smaller coverages, such as glass breakage, fidelity, and mysterious disappearance of valuable items are covered as well.

There are two other areas that are less clear, but we believe will likely prove to be insurable. Intentional acts by an insured are excluded from almost all Property policies, because they are not fortuitous and insurable. Intentional acts of others are generally covered. We have all seen the extensive search and rescue efforts undertaken while the flood left many residents trapped. This involves forced entry by rescue teams, who broke doors, windows and in some cases roofs to look for survivors. While the actual damage from entry is small, for homes left unoccupied, there may be continuing damage and mold. In a similar way, we understand the Superdome (at $400Mn) and the Convention Center are both total losses after their intentional, but unexpected, damage by refugees.

Another peril for concern is Collapse. Under some policy forms, total collapse of the structure is covered as a named peril. Arguably, this would be true even if the collapse was caused by flood erosion or wave actions, which are otherwise excluded perils. We estimate that in the New Orleans area, up to 20% of the property loss that is not insured as Wind or Flood will fall into these other coverages, creating an insured loss of $2.5Bn - $7.5Bn.

Potential Disputes:

We estimate $1Bn to $3Bn of LAE on coverage disputes. Our estimates assume that courts uphold the Flood exclusions, and these potential losses are outside of our estimates. There are several areas where coverage disputes seem likely:

• "When in doubt pay it as wind." - Mississippi Insurance Director Dale


• Mississippi A.G. Hood's suit: "That wasn't a Flood.that was wind-driven water." Is Flood exclusion "unconscionable"?


• Louisiana: "That wasn't a Flood.that was a Barge-Levee Collision"


• Florida's Mierzwa decision: "Pay all total losses as if they were all wind".


• Separating Time Element into Wind (covered), Flood (sometimes), Public Safety evacuation (not covered)


• Separating Flood from Looting, Fire, Collapse, and Intentional Acts, such as Search and Rescue

The Mierzwa Flood vs. Wind precedent could be applied both in other states, and on Florida policies that renewed before new forms were filed. The legal theory here is that the "valued policy" law in Florida applies to all total losses, even if only part of the loss is due to covered perils. Louisiana has a similar law, but the Florida precedent doesn't automatically apply.

On standing structures that are uninhabitable, separating living expense and other time element coverages between Wind (covered), Flood (only if purchased) and government orders (not covered), will cause disputes. Although a significant Federal program of grants and other aid could lessen the pressure on insurers, as seen after the 9/11 attacks.

Our current estimates assume that Flood exclusions and sub-limits are, in the end, upheld. Win or lose, there will be disputes. Reviewing and resisting disputes will cost the industry a great deal in loss adjustment. LAE of $1Bn to $3Bn seems reasonable if many people will "take a shot" to recover a significant part of their $10Bn to $30Bn of uninsured loss, and LAE costs 5% - 10% of disputed amounts. The legal climate is adverse to insurers in all of these states.

Fairness Rankings of State Legal Systems
50th	Mississippi
48th	Alabama
47th	Louisiana
42nd	Florida
Source: U.S. Chamber of Commerce, March, 2005

Direct Loss by Market Segment:

Combining these estimates, we expect $45Bn to $65Bn in Property and Marine market insured loss and LAE. This makes Katrina the market's largest dollar-amount loss ever, and the greatest loss since 1906 when measured as a percent of industry capital. (These estimates assume that policy language is interpreted as intended and exclude Life, WC and Liability exposures.) Based on clients' losses in prior catastrophes, we estimate the loss will split among major products as follows:

Personal Property	$16Bn – $23Bn
Commercial Property	$12Bn – $18Bn
Insured Flood	$5Bn – $15Bn
Automobile	$2.0Bn – $3.2Bn
Marine	$7Bn – $10Bn
All Insured Lines	$45Bn – $65Bn

Based on market shares by line and state, we expect the gross losses to split among major industry segments in these categories:

National Commercial and Multiple-Line Cos. (20)	$$16Bn — $23Bn
National Personal Lines Companies (7)	$12Bn — $18Bnn
One-Region Companies (8)	$4Bn — $6Bn
Multi-Regional Companies (14)	$3Bn — $5Bn
Direct by Admitted Reinsurers (7)	$1Bn — $2Bn
Marine Specialists (5)	$0.6Bn — $1.0Bn
Citizens Insurance (Florida)	$0.3Bn — $0.5Bn+
Florida HO Specialists (25)	$0.2Bn — $0.4B
All Others, including E&S and Non-admitted Foreign, such as Lloyds’	$8Bn — $12Bn
Market Direct Loss and LAE	$45Bn — $65Bn

Net Losses by Market Segment

There are several issues that need to be considered in reviewing how reinsurance protections will apply to such a widespread and continuing loss. First, in 2005, most contracts have limited hours provisions in the definition of loss occurrence. The industry standard for windstorm had been 72 hours, although alternative quotes have been provided for 96 hours, or even on the basis of "any single atmospheric disturbance" (Flood is covered for 168 hours). It is not clear how much if any 2005 wind reinsurance protections have hours clauses greater than 72 hours. The first and second landfalls of Katrina do not fall inside a single 72-hour period. Another common treaty provision is that losses from a single windstorm cannot be separated into two separate 72-hour periods, and ceding companies may only choose their "best" 72 hours. Thus, in many cases, no reinstatement would be allowed. We believe most national companies will have coverage for the Gulf Coast landfall only.

In the New Orleans area, Wind, Flood, Fire and Looting extended over more than 72 hours. A few insurers could attempt to define these as two occurrences: Windstorm and Flood or Windstorm and Civil Commotion.

The structure of reinsurance protection that companies buy is a key consideration. The Florida Hurricane Catastrophe Fund (FHCF) protects Florida residential-only property and will treat the entire Florida loss as a single occurrence. Because of the FHCF and regulatory issues in Florida, insurers often take a different approach to covering Florida exposures. Many national insurers only buy protection above the FHCF with respect to losses in Florida, or even countrywide. Other insurers have purchased Florida-only programs because of the need to support Florida Homeowners rate filings.

Companies with significant Marine exposures often purchase separate Marine programs, with separate retentions.

In the retrocessional market a great number of contracts are purchased on an industry loss warranty basis (ILW). While the dollar amounts required to trigger ILW contracts are likely to have been exceeded by a loss at $45Bn or more, it is important to note that many Gulf coast or Eastern US specific contracts are often limited to only cover admitted market Wind losses, without LAE. We believe that there are several ILW contracts attaching at $25Bn of Wind loss and that it may be difficult to find an objective source to demonstrate a covered loss.

The following estimates of net and ceded loss are based on market share allocations of the industry losses by line and state, our knowledge of catastrophe reinsurance terms, and simple assumptions on the effect of Risk, Aggregate and facultative coverages. We do not have an opinion on the losses of individual companies, which would require specifics on exposures and protections. Thus, there is a wide range of uncertainty around these industry estimates.

Range of Net Insured Loss and LAE ($Bn)
Market Segment	45	45	50	55	60	65	70
Reinsurers (incl. direct operations)	15.8	17.6	19.5	21.4	22.9	24.3	26.1
National Personal Cos. (7)	10.5	12.1	13.9	13.7	15.1	16.4	17.6
National Commercial and M-L Cos. (20)	6.7	7.9	9.2	10.6	12.0	13.3	14.7
One Region Cos. (8)	2.6	3.0	3.4	3.8	4.2	4.7	5.1
Multi-Regional Companies (14)	1.0	1.0	1.2	1.3	1.4	1.6	1.8
Citizens Insurance (Florida)	0.3	0.3	0.4	0.4	0.5	0.5	0.6
Florida HO Specialists (25)	0.2	0.2	0.2	0.3	0.3	0.4	0.5
Marine Specialists (5)	0.1	0.1	0.2	0.3	0.3	0.4	0.4
All Others	2.7	2.8	3.0	3.2	3.3	3.5	3.6

The net losses do not spread evenly across the market when measured by either surplus or premiums. Residential companies' shares vary based on geographic spread; commercial companies' vary on the classes written, with HPR, Energy risks, chemical plants, universities, casinos and churches particularly exposed. Several large Personal Lines companies bought limited reinsurance protection for these states as a conscious strategy. Some Commercial companies bought only to their modeled 1- in-100 year loss and have exhausted their coverages. While there will be several notable charges this quarter, we do not see any major, national companies becoming impaired due to Katrina alone.

Regional companies face more severe concentration problems, particularly for one-region companies, but also for some of the multi-regional insurers. Several appear to have exhausted their reinsurance protections, especially at the highest loss estimates. We believe that some of these companies will not be able to trade forward without replenishing their capital levels.

In addition to insurers' catastrophe protections, we believe the majority of U.S. property retrocessional contracts are now total losses.

In Florida, this loss is the first storm to really test the new carriers formed to write "de-population" business, largely in the Miami-Dade area and surrounding counties. The first landfall should not have exhausted any of these companies' FHCF protections, but we suspect that some will have recoveries from the FHCF, and will have absorbed significant, but not ruinous, net losses. These companies also have "take out bonuses" available to help fund one or more losses. Many new companies have exhausted their potential bonuses on this event and will require more reinsurance going forward.

Catastrophe reinsurers manage their capacity by zone. Losses of this magnitude in a single zone should not impair a well-run reinsurer. Unexpected combinations of Marine, Non-Marine and Retrocessional losses may cause some surprises. We believe the combined capital of the U.S., Bermuda, London and Continent reinsurers is over $100Bn, so $20Bn - $25Bn of reinsured loss from Katrina is painful, but survivable. (A second major loss in 2005 would cause insolvencies.) We believe that nearly one-half of the reinsured Katrina loss will ultimately fall to "super Cat" retrocessions, written by three or four highly-rated reinsurers. We do not expect many smaller reinsurers to fail solely as a result of Katrina, although we do expect many to have significant changes in their appetites. We also expect reinsurer downgrades.

Gross and Net Losses in Context

Katrina losses represent 35 to 50 loss ratio points on the admitted industry's countrywide Property premiums. Taken together, 2004 and 2005 have cost insurers $80 to $100 billion in gross U.S. catastrophe losses, excluding any loss on Rita or subsequent storms.

Beside 9/11 at about $35Bn, some other major market events (in nominal dollars) have been Andrew at $15.5Bn, Northridge at $12.5Bn, Kobe at $5Bn (insured only, uninsured was many multiples of this), Hugo at $4Bn and Lothar and Martin at $10Bn, combined. Two more significant events for the market were hurricanes Betsy (1965) and Camille (1969) which are described in more detail in Exhibit 4.

Another significant "storm" loss has been the reinsurance industry under-reserving in 1998 through 2000. The US on-shore reinsurance industry has already acknowledged over $15Bn of loss reserve shortfall on these accident years. The worldwide market's shortfall is certainly over $20Bn, making the reserve deficiency one of the market's major events. After Katrina and 9/11, the reinsurance industry's largest "hit" in dollar terms has been this self-inflicted one.

Although Katrina was a notably strong and wide storm, it was only a Category 4 over land, not a 5. While reinsurers are still absorbing the magnitude of this loss, we are already hearing some comments that at least the windstorm portion of it could have been worse, even much worse.

The impact of Katrina and previous major losses are shown below.

Market Observations:

Previous major market events have led to a market swing, coming at the end of a soft market. 2005's loss follows a period of rising prices. Considering 1998 - 2000, 2001, 2004 and 2005, the worldwide reinsurance market in total has no profits to show for the last decade or more of investments. Most Property-focused reinsurers now have losses in three of the last five years, 2001, 2004, 2005. It is not clear how reinsurers and buyers will react. Our intelligence after meeting with markets in London, Bermuda and domestically follows:

• The market is currently quoting both third-event covers for the remainder of insurers' contract term, and also "Live Cat" covers for individual named storms. Prices are rising, terms are limited, and quotations have short time limits. Third-event covers for the last quarter of January 1 anniversary covers are priced at well over 50% of the annual premiums.


• Retrocessional capacity is down sharply, prices are up sharply, and there will be coverage limitations. One move is to exclude Per Risk coverage from Retro Cat protections.


• Lloyd's syndicates are re-examining their 2006 business plans and are now expecting to grow capacity, rather than to reduce it as planned earlier. We expect other reinsurers to take similar steps, although their business plan information is not as public as the information on Lloyd's syndicates' stamp capacity.


• The demand for Catastrophe reinsurance will increase at upcoming renewals. Also, more buyers will request or require ECO/XPL, Ex Gratia or follow-the-fortunes language in their Cat covers to prevent disputes on gray-areas claims, such as Mierzwa.


• Some insurers, not necessarily exposed to Katrina, are already testing rate increases in the primary market. How much increase will stick - and where - is not clear.


• Some Cat bonds will be hit by Katrina. Many hedge funds sustained direct losses on retrocessions. Some hedge funds have also taken simultaneous losses on energy speculation.


• Clearly, reinsurers' profit margins in good years have not funded the bad years. Reinsurers overall either need to charge more or to cover less. Market negotiations will decide the specifics of how this happens.

How Will the Industry Rebuild Capital?

• How many open shelf offerings are there that could follow Montpelier's example? Is it too late?


• Will capacity come from new capital at new reinsurers? Is there a credible talent base to start new companies?


• Will hedge funds withdraw, or will they make larger commitments?

Impact on the Economy

Insured and Non-Insured Loss:

We are troubled by the many estimates of "economic loss," because the definitions are so vague. Indirect costs resulting from lost oil production and hotel revenues may make some sense to include. But some published estimates include secondary economic effects driven by supply chain delays and price movements. We believe these effects cannot meaningfully be separated from the overall economic picture and so are not useful. Instead, we consider only "market loss and uninsured property damage." Our $45Bn - $65Bn range for the insured market is combined with an estimated $10Bn for NFIP, at least $20Bn - $30Bn of uninsured private Flood damage and S&P's $25Bn estimate for government losses, such as the I-10 bridges. We therefore estimate $100Bn - $130Bn of damages, excluding any uninsured off-premises and indirect losses.

Economic Observations:

• Major hits resulted to oil, shipping, airlines, steel, auto, agriculture, hotels, municipal bonds.


• DJIA has stabilized since then (see chart in Exhibit 7).


• The 1906 San Francisco and 1995 Kobe earthquakes were larger percentages of GNP. Both events involved a massive uninsured loss. Both events caused significant stock market and interest rate disruptions.


• The Gulf Coast has a relatively small economy, so the relative impact will be more focused than the national trends will show. The impact of Federal assistance will also be greater.

Impact on Catastrophe Models and Analysis

• The majority of this loss was not modeled under current technology ($19Bn - $27Bn vs. $45Bn - $65Bn). This is true of many prior events, as well. Reinsurers are learning that models are only one tool to assess risk. They will rely more on deterministic scenarios to establish their capacity, and on traditional underwriting techniques to evaluate risk and pricing.


• Cat Models have now called the losses in 1999, 2001, 2004 and 2005 as 1-in-50+ year scenarios. In the last 48 months there has been over $150Bn in insured loss on fifteen events of over $1Bn each. Would 2001's models have even recognized that as a possibility? The extreme probability events within the models are limited by the imagination of the model engineers. Nature has more than we can imagine.


• Catastrophe risk clearly isn't "Zero Beta." Is there an implication for investors' interest in Cat bonds? For the FHCF's Bonding capacity?


• Hurricane models are not meaningful if they don't reflect year-by-year variations in storm activity (such as Dr. Gray's forecasts).


• Demand surge cannot be calculated on an individual event, or even individual year, basis.

Fatalities

Worst U.S. Natural Disasters in Human Lives:

1. Galveston, Texas, Hurricane, 1900 - estimated 8,000 deaths

2. September 11, 2001 Attacks - 2,752 in NYC, plus approximately 200 in D.C. and Pennsylvania

3. Great Okeechobee Hurricane in Florida, 1928 - estimated 2,500-plus

4. Johnstown, Pa., Flood, 1889 - estimated 2,200-plus

5. Louisiana Hurricane, 1893 - 2,000-plus

6. South Carolina-Georgia Hurricane, 1893 - 1,000 - 2,000

7. Great New England Hurricane, 1938 - 720

8. San Francisco Earthquake, 1906 - 700

9. Georgia-South Carolina Hurricane, 1881 - 700

10. Tri-State Tornado in Missouri, Illinois and Indiana, 1925 - 695

11. Texas City Explosion, 1947 - approximately 600

12. Labor Day Hurricane in the Florida Keys, 1935 - 405

Katrina: 975+ confirmed, with 2,500 unaccounted for and areas of New Orleans still unexplored, is number 6 or 7 on the list.

Attachments

On the Cover: NOAA satellite image of Katrina on August 28th, 2005

Exhibit 1    Population Densities: Katrina 2005 vs. Andrew 1992

Exhibit 2    Recent Gulf Coast Landfalls

Exhibit 3    Katrina vs. Major Market Losses, in 2005 dollars

Exhibit 4    Hurricanes Camille, Betsy, Andrew and Katrina

Exhibit 5    Insurers' Loss Disclosures through September 23rd

Exhibit 6    Market Share Data

Exhibit 7    Dow Jones Average Performance through Katrina

Exhibit 1: Population Densities: Katrina 2005 vs. Andrew 1992

Exhibit 2: Recent Gulf Coast Landfalls

Exhibit 3: Katrina vs. Major Market Losses, in 2005 dollars

Exhibit 4: Hurricane Betsy

Date of (most significant) U.S. Landfall: September 9, 1965 - Grand Isle, Louisiana

Maximum Sustained Winds: 155 mph

Lowest Recorded Pressure: 941 Mb mb/27.76 inches

Storm Surge: 8 - 10 feet above mean tide

Fatalities: 76

• One of the most intense, deadly, and costly storms to make landfall in the United States.


• Discovered on August 27th about 560 km east-southeast of Barbados.


• By midday on the 27th, the central pressure was 1007 mb and by evening the depression was named Betsy.


• The depression did not intensify until two days later when it passed through the Lesser Antilles.


• By the afternoon on the 29th, Betsy intensified to hurricane status with sustained winds in excess of 74 mph. Betsy remained a mature hurricane through September 10th.


• On September 5th, Betsy made an unusual shift southwestward. Prior to this point, the hurricane appeared to be headed towards the Carolinas, but a high pressure system draped over the northeast forced the storm southwest.


• By the 6th, Betsy was just off Great Abaco Island. The hurricane continued through the northern Bahamas with the eye passing just to the north of Nassau. Only one fatality occurred in Nassau.


• Betsy then moved on a westerly course and passed over the Florida Keys. Damage from winds and storm surge were limited to the area south of Fort Lauderdale.


• The hurricane then turned toward the northwest upon entering the Gulf of Mexico. Betsy intensified while moving through the Gulf. Just hours prior to landfall, the storm was nearly at Category 5 status, with maximum sustained winds of 155 mph.


• The storm weakened to a Category Three shortly after making landfall at Grand Isle, LA, just west of New Orleans. This path sent a 10 ft. storm surge up the Mississippi River into Lake Pontchartrain, generating widespread flooding.


• New Orleans experienced winds of approximately 125 mph. The levees thankfully held, limiting the amount of flooding in the city.


• Damage was most significant along the Gulf Coast from Grand Isle, LA to Mobile, AL. The greatest devastation resulted from the storm surge of 8-10ft.


• Total damage was $1.42Bn in 1965 dollars (over $8.5Bn in today's dollars). It was the first hurricane to ever produce in excess of $1Bn in damages, earning it the nickname "Billion Dollar Betsy."


• The storm is most noted for its large eye of 40 miles in diameter and a storm track that included two complete loops.

Exhibit 4: Hurricane Camille

Date of U.S. Landfall: August 17, 1969 - near mouth of the Mississippi

Maximum Sustained Winds: 190 mph

Lowest Recorded Pressure: 909 mb/26.84 inches

Storm Surge: 22-25 ft. above mean tide

Fatalities: 255 (additional 8,900 injured)

• Hurricane Camille is the most intense storm to strike U.S. mainland in modern history.


• Storm discovered by satellite as a tropical disturbance moving west in the Caribbean Sea on the 14th of August.


• Became a strengthening hurricane while located off the western tip of Cuba.


• Crossed over Cuba on the 15th, entering the Southern Gulf of Mexico with wind speeds of 100 mph.


• By midday on the 16th, wind speeds were up to 115 mph.


• The hurricane slowed nearly to a stand-still as night fell, rapidly intensifying during the process to wind speeds near 150 mph.


• On the night of the 16th, Camille was predicted to strike near the mouth of the Mississippi River.


• By the afternoon of the 17th, Camille sat about 100 miles south of the Mississippi coast and had intensified to nearly unprecedented levels.


• Barometric pressure was at 905 Mb. Sustained wind speeds were up to 190mph.


• Landfall was now predicted to occur at midnight on the 17th along the coast of Mississippi.


• As Camille made landfall, civilization from near Ansley to Biloxi was erased. The resulting damage resembled that caused by an atomic bomb.


• Nearly 100 years of growth and progress along the Mississippi coast was completely destroyed in three hours time.


• Most of the structures within ½ mile from the ocean vanished. (Plumbing systems had even been uprooted.)


• Gusts were estimated in the 210-220 mph range.


• At that time, Camille produced the highest hurricane tidal surge ever recorded in the United States. No pacific Coast tidal wave or Atlantic Coast storm (hurricane or winter storm) had ever submerged so much land to such a depth.


• After moving inland, Camille weakened significantly, but not before triggering catastrophic flash flooding and landslides over the mountains of the Southeastern United States.


• A single location in the Blue Ridge Mountains in Virginia measured total rainfall of 27 inches.


• More than 100 were killed in Virginia and Tennessee alone.


• 15,000 people were left homeless and there was no water, food, or fuel.


• All means of communication were wiped out.


• Roads, bridges, airports, and even railways were impassable or destroyed.


• A vermin control problem added to the devastated landscape.


• 1,000 federal troops were sent into the area by Nixon, and Governor John Williams of Mississippi declared martial law.


• Thousands of dead farm animals, pets, and wildlife posed a huge problem. The dead animals attracted insects and rodents. Heavy equipment was used to bury thousand of cows and other livestock.


• Rattlesnakes and fire ants also claimed dozens of victims in the aftermath of the hurricane.


• The state and federal government supplied thousands of bulldozers and dump trucks to cart away the twisted wreckage. Debris removal efforts continued for the rest of 1969.


• Nearly 14,000 housing units were damaged, and 6,000 others were totally destroyed.


• The total damage was $1.42Bn (in 1969 dollars).

Exhibit 4: Hurricane Andrew

Date of U.S. Landfall: August 24, 1992 - Homestead, Florida

Maximum Sustained Winds: 165 mph

Lowest Recorded Pressure: 922 mb/27.23 inches

Storm Surge: 17 feet above mean tide

Fatalities: 26

• Started as a tropical wave that emerged from the west coast of Africa on August 14th.


• Over the next few days, Andrew strengthened first to a tropical depression and then to a tropical storm.


• On August 20th, Andrew appeared as if it might completely dissipate due to vertical wind shear.


• Continued west-northwestward movement brought Andrew into warmer, more favorable stormdeveloping waters between Bermuda and Puerto Rico by the 21st.


• Andrew experienced rapid strengthening on the 22nd and 23rd, reaching Category Four status by nightfall of the 23rd.


• Andrew weakened when it passed over the western portion of the Great Bahama Bank and the pressure rose to 941 mb. However, the hurricane rapidly reintensified during the last few hours preceding landfall in extreme southern Florida, reaching Category Five status on the 24th.


• After crossing the southern tip of Florida on the 24th, Andrew continued westward into the Gulf of Mexico where it gradually turned northward.


• Andrew made its second U.S. landfall on the central Louisiana coast on the 26th as a Category Three hurricane.


• The hurricane weakened significantly after landfall and eventually merged with a frontal system over the Mid-Atlantic States on August 28th.


• Total economic damage was estimated at $25Bn. The damage was most extreme in southern Dade county, from the Kendall district southward through Homestead and Florida city.


• Andrew completely destroyed 25,524 homes and significantly damaged 101,241 others. 99% of mobile homes in Homestead were completely destroyed.


• Andrew was a relatively compact storm; the radius of maximum wind was about 11 nautical miles. Had it been slightly larger, or made landfall just a few miles to the north, the devastation would have been greatly magnified. The area just to the north includes downtown Miami, Miami Beach, Key Biscayne, and Fort Lauderdale.

Exhibit 4: Hurricane Katrina

Date of (most significant) U.S. Landfalls: August 29, 2005 - Plaquemines, Louisiana and near Gulfport, Mississippi

Maximum Sustained Winds: 184 mph

Lowest Recorded Pressure: 902 mb/26.64 inches

Storm Surge: 30 feet above mean tide

Fatalities: 943 confirmed, over 2,500 still missing (as of 9/19)

• Started as a tropical depression over the southeastern Bahamas.


• The system was upgraded to tropical storm status on the morning of the 24th and then to hurricane status midday on the 25th prior to making its first U.S. landfall at 6:30 p.m. between Hallandale Beach and Aventura, Florida.


• Katrina weakened to tropical storm status while over Florida on the 26th, but then reintensified to Category Two hurricane status with sustained winds of 100 mph.


• While over the Gulf of Mexico on the 27th and 28th, Katrina rapidly intensified, reaching Category Five status by late morning of the 28th.


• The hurricane made its second landfall early morning on the 29th near Buras-Triumph, Louisiana as a Category Four.


• A third landfall occurred near the Louisiana/Mississippi border with 125 mph Category Three winds. Individual reports, however, suggest that Katrina was much more intense. The Mayor of Slidell, Louisiana, for example, reported maximum sustained winds of 176 mph (Category Five).


• High and damaging winds and storm surges slammed the Mississippi coast, including Biloxi and Gulport. Significant damage was also witnessed on the Alabama coast.


• The storm surged breached the levees that protected New Orleans from Lake Pontchartrain, flooding over 80% of the city and some surrounding areas.


• Katrina weakened after landfall, losing hurricane strength approximately 100 miles inland. It was downgraded to a tropical depression near Clarksville, Tennessee.


• The storm continued northward, dumping large amounts of rain over a large portion of eastern North America.


• Federal disaster declarations blanketed 90,000 square miles of the U.S. This area is only slightly smaller than all of the United Kingdom.


• An estimated five million people were left without power, and over a million people were displaced from their homes.


• Total economic damages are currently estimated in the $200Bn range, making Hurricane Katrina the most expensive disaster (both natural and man-made) in United States history.


• The hurricane has resulted in a humanitarian crisis on a scale unseen in the U.S. since the Great Depression.

Exhibit 5: Hurricane Katrina Loss Estimates (000's)

Exhibit 6: 2004 Personal Lines (HO and FO) Market Share
Louisiana, Mississippi, and Alabama

Company Name	Market Share (%)	DWP ($000)
State Farm Mutual Group (01767)	31.30%	768,448
Allstate Insurance Co Group (00086)	14.16%	347,689
Alfa Insurance Group (00051)	9.91%	243,247
Farmers Insurance Group (00698)	4.93%	121,051
Mississippi Farm Bureau Mut Ins Co (14605)	4.37%	107,201
St Paul Travelers Companies and Affi (00070)	3.44%	84,488
United Services Automobile Asn Group (02003)	3.12%	76,519
Nationwide Group (01406)	2.90%	71,312
Louisiana Farm Bureau Mutual Ins Co (14427)	2.28%	55,957
Metropolitan P&C Ins Co & Affiliates (02410)	2.13%	52,414
Liberty Mutual Group (01112)	1.99%	48,979
American International Group (00124)	1.58%	38,918
Amer Natl Prop & Cas Co & Affiliates (04081)	1.37%	33,758
Shelter Insurance Cos Group (01236)	1.36%	33,355
Combined Federal Ins Co & Affiliates (00388)	1.20%	29,477
Combined Safeco Ins Co Grp (01635)	1.16%	28,472
Auto-Owners Insurance Group (02801)	1.14%	28,070
Hartford Fire Group (00914)	0.94%	23,159
Cotton States Mutual Group (00493)	0.86%	21,149
Continental Casualty Group (02186)	0.81%	19,996
Hanover Insurance Co Group (00884)	0.75%	18,365
Cincinnati Insurance Group (02445)	0.74%	18,219
So.Farm bureau cas ins co (04839)	0.67%	16,572
Assurant Group (00191)	0.64%	15,754
Firemans Fund Insurance Group (00760)	0.55%	13,464
National Security Fire & Casualty Co (03166)	0.53%	13,129
Unitrin Prop & Cas Ins Grp (00629)	0.51%	12,566
Horace Mann Group (03000)	0.43%	10,459
Auto Club Family Insurance Co (27235)	0.43%	10,440
State Automobile Mutual Group (01759)	0.32%	7,972
Employers Mut Co of Des Moines (00620)	0.32%	7,736
Nc Farm Bureau Insurance Group (03240)	0.31%	7,573
United Fire & Casualty Group (02488)	0.30%	7,477
Winterthur Us Holding Group (04006)	0.30%	7,251
Republic Companies Inc (00063)	0.29%	7,088
All Other Companies (45)	1.94%	47,690
State Aggregates	100.00%	2,455,414

Exhibit 6: 2004 Commercial Lines Market Share
Louisiana, Mississippi, and Alabama

Company Name	Market Share (%)	DWP ($000)
St Paul Travelers Companies and Affi (00070)	8.83%	170,205
Zurich Insurance Co Group (02127)	7.13%	137,443
American International Group (00124)	6.37%	122,742
Continental Casualty Group (02186)	5.21%	100,332
State Farm Mutual Group (01767)	5.19%	99,997
Fm Global (00655)	4.15%	79,896
Alleghany Grp (05010)	2.69%	51,809
Assurant Group (00191)	2.32%	44,744
Ace American Ins Co (00626)	2.32%	44,709
Nationwide Group (01406)	2.24%	43,074
Combined Federal Ins Co & Affiliates (00388)	1.97%	38,018
Allstate Insurance Co Group (00086)	1.92%	37,008
Alfa Insurance Group (00051)	1.91%	36,745
Mississippi Farm Bureau Mut Ins Co (14605)	1.74%	33,607
Firemans Fund Insurance Group (00760)	1.71%	32,900
Liberty Mutual Group (01112)	1.68%	32,450
Hartford Fire Group (00914)	1.68%	32,327
Auto-Owners Insurance Group (02801)	1.54%	29,651
Great American Group (00841)	1.42%	27,372
Unitrin Prop & Cas Ins Grp (00629)	1.38%	26,511
United Fire & Casualty Group (02488)	1.34%	25,875
Combined Safeco Ins Co Grp (01635)	1.30%	24,968
American Modern Home Group Inc (01279)	1.11%	21,432
Cincinnati Insurance Group (02445)	1.09%	20,964
Hanover Insurance Co Group (00884)	1.05%	20,276
Guideone Mutual & Guideone Spec Mut (03034)	1.00%	19,234
Assctd Int' Markel Ins Cos (07854)	0.96%	18,561
Axis Specialty Ins Co & Affiliate (00056)	0.92%	17,807
Employers Mut Co of Des Moines (00620)	0.88%	17,055
Allianz Insurance Company Group (07617)	0.87%	16,797
Louisiana Farm Bureau Mutual Ins Co (14427)	0.86%	16,669
Argonaut Insurance Co Group (02151)	0.85%	16,396
Qbe the Americas (07960)	0.84%	16,251
National Security Fire & Casualty Co (03166)	0.76%	14,702
Church Mutual Insurance Co (18767)	0.76%	14,635
All Other Companies (166)	22.00%	423,975
State Aggregates	100.00%	1,927,137

Exhibit 6: 2004 Ocean Marine Market Share Countrywide

Company Name	Market Share (%)	DWP ($000)
American International Group (00124)	12.78%	414,800
Ace American Ins Co (00626)	9.49%	308,196
St Paul Travelers Companies and Affi (00070)	8.96%	290,951
Continental Casualty Group (02186)	8.82%	286,516
Xl Reins America Inc & Affiliates (00743)	5.82%	188,957
Firemans Fund Insurance Group (00760)	5.71%	185,242
Combined Federal Ins Co & Affiliates (00388)	5.30%	172,040
American Steamship Owners Mut Prot (13366)	5.06%	164,246
Navigators Insurance Group Inc (05100)	4.91%	159,303
Onebeacon Ins Grp (00442)	4.62%	150,108
Zurich Insurance Co Group (02127)	3.07%	99,758
New York Marine & General Ins Co (02569)	2.96%	96,030
Hcc Ins Holdings Grp (09849)	2.42%	78,474
Great American Group (00841)	2.20%	71,413
Assctd Int' Markel Ins Cos (07854)	1.87%	60,850
Millea Holdings Inc (09504)	1.48%	48,053
Hartford Fire Group (00914)	0.97%	31,406
Farmers Insurance Group (00698)	0.92%	29,953
Arch Capital Grp Inc (00727)	0.87%	28,235
Liberty Mutual Group (01112)	0.84%	27,169
Employers Reinsurance Corp Group (03271)	0.82%	26,585
Swiss Re America Corp Group (01813)	0.78%	25,473
Nationwide Group (01406)	0.78%	25,469
Seaworthy Insurance Co (37923)	0.76%	24,722
Fm Global (00655)	0.72%	23,284
United Services Automobile Asn Group (02003)	0.64%	20,838
State Farm Mutual Group (01767)	0.64%	20,695
American Modern Home Group Inc (01279)	0.56%	18,276
Sompo Japan Ins Co of Amer (00066)	0.41%	13,168
Mitsui Sumitomo Ins Grp (00026)	0.38%	12,475
Hanover Insurance Co Group (00884)	0.36%	11,634
Samsung Fire & Marine Ins Co Ltd (38300)	0.34%	11,188
Seven Seas Insurance Co Inc (37672)	0.33%	10,686
Atlantic Mutual Group (00248)	0.28%	9,133
Sunderland Marine Mutual Co Ltd (10838)	0.27%	8,810
All Other Companies (59)	2.85%	92,649
State Aggregates	100.00%	3,246,785

Exhibit 6: 2004 APD (Pers and Comm) Market Share
Louisiana, Mississippi, and Alabama

Company Name	Market Share (%)	DWP ($000)
State Farm Mutual Group (01767)	26.62%	809,989
Allstate Insurance Co Group (00086)	10.03%	305,288
Progressive Casualty Group (01554)	9.39%	285,853
Alfa Insurance Group (00051)	6.83%	207,986
So.Farm bureau cas ins co (04839)	5.15%	156,769
National Indemnity Co Group (00310)	3.84%	116,930
United Services Automobile Asn Group (02003)	3.09%	94,074
Nationwide Group (01406)	2.72%	82,921
American Modern Home Group Inc (01279)	1.91%	58,213
Usagencies Cas Ins Co Inc (10295)	1.70%	51,762
American International Group (00124)	1.66%	50,580
St Paul Travelers Companies and Affi (00070)	1.60%	48,784
Safeway Insurance Group (02577)	1.58%	48,042
Liberty Mutual Group (01112)	1.47%	44,701
Farmers Insurance Group (00698)	1.38%	41,943
Combined Safeco Ins Co Grp (01635)	1.19%	36,072
Metropolitan P&C Ins Co & Affiliates (02410)	1.18%	35,940
Assurant Group (00191)	1.07%	32,571
Auto-Owners Insurance Group (02801)	1.01%	30,656
Hartford Fire Group (00914)	0.96%	29,269
Motors Insurance Corp Group (00795)	0.75%	22,685
Shelter Insurance Cos Group (01236)	0.72%	21,960
Zurich Insurance Co Group (02127)	0.70%	21,270
Direct General Group (00720)	0.69%	21,048
Amer Natl Prop & Cas Co & Affiliates (04081)	0.65%	19,759
Louisiana Farm Bureau Mutual Ins Co (14427)	0.57%	17,201
Unitrin Prop & Cas Ins Grp (00629)	0.54%	16,281
Cotton States Mutual Group (00493)	0.53%	16,139
Continental Casualty Group (02186)	0.48%	14,541
Clarendon National Ins Company (04847)	0.48%	14,534
Canal Insurance Co Group (02623)	0.45%	13,569
Mississippi Farm Bureau Mut Ins Co (14605)	0.37%	11,335
Vesta Fire Insurance Corporation (09580)	0.37%	11,300
Guideone Mutual & Guideone Spec Mut (03034)	0.37%	11,185
The Usa Insurance Company (10313)	0.34%	10,388
All Other Companies (137)	7.61%	231,538
State Aggregates	100.00%	3,043,076

Exhibit 7: Dow Jones Industrial Average as of September 16, 2005