Claims And Issues

"PREDICTIVE ANALYTICS" AND INSURANCE CLAIMS.

Actuaries have all the fun.  It seems that way sometimes.  But that may be because sometimes I am not sure what they are saying.

"Predictive analytics" is a formula used to model claims, in basic and simple terms.  This was the subject of a fine presentation at the Casualty Actuarial Society Annual Meeting in Orlando.

The CAS Annual Meeting continues with more fine presentations today, Wednesday, November 16, 2016.

One of the presenters said that his company did not have any negative reaction to report from regulators in his company's modelling of claims.  That drove me to think about our own experience in this State when predictions were introduced into rate filings to replace or supplant actual results from the real world, so to speak.

In Florida, we have a State-run computer model to predict catastrophe claims based on the data of actual experience.  Our statute was recently amended to allow insurance companies to use their own computer models including models based on predictions and assumptions, but not necessarily based on any real-world results.  The Florida experience is pretty well summarized in this April 16, 2012 article, posted here.

"Predictive analytics," like any tool, may be fine in one area such as predicting payouts on claims and predicting which claims are likely to hit bigger than the claims department might have thought at first, but not so accurate in other areas such as predicting premium rates.  These trends have been building for years.  See, for example, these issues addressed in these articles from 2012, about automobile liability insurance premiums here, and about "evidence-based funding criteria" especially in setting premiums for all lines of insurance here.

The actuarial presenters of CAS in Orlando made a compelling case for learning more about "predictive analytics" in modelling claims.

Please Read The Disclaimer.  ©2016 by Dennis J. Wall.  All Rights Reserved.

COMPUTER MODELS NEED "BIBO": "BETTER INFO IN, BETTER INFO OUT."

Image courtesy of NASA.

Based on computer models filled with facts predating current climate change facts, the National Oceanic and Atmospheric Administration ("NOAA") reports a stunning repudiation of existing modelling.  The recent flooding in Louisiana is the eighth (8th) time since May, 2015 that so much rain fell in so short a time that it exceeded the NOAA computer predictions of events that would take place once every 500 years.

To put it another way, 8 times in the last 15 months rainfall exceeded the amount that NOAA's computer would have predicted had a 0.2% chance of happening.  Jonah Engel Bromwich, "Flooding in the South Looks a Lot Like Climate Change" (New York Times Online, Tuesday, August 16, 2016).

Time to update the computer models to model present-day realities with "Better Information In, Better Information Out."  That prescription goes for computer models used to predict hurricanes and to justify insurance premium rate requests as well.  See, for example, "ROCKY MOUNTAIN GANG PROBABLY PREDICTS NO HURRICANES IN ROCKY MOUNTAINS ... Not So Many in Other Areas of the Country Where Hurricanes Have Actually Been Seen in the Past," and sources cited in this article, which was posted here on August 12, 2014; and see this Postscript to it, "P.S. NOAA's CLIMATE PREDICTION CENTER," an article posted here on August 18, 2014.

Please Read The Disclaimer.  ©by Dennis J. Wall.  All Rights Reserved. 

MODELS IN INSURANCE.

An interesting take on modelling in insurance is provided by corporations and apparent academics collaborating in the U.K.  They have issued several 'white papers' on the topic.  Their umbrella appears to be the "Amlin" organization.

Here is a link to their 'white paper' titled, "Did Your Model Tell You All Models Are Wrong?" (undated but apparently circa November 4, 2015).

Please Read The Disclaimer.  ©2016 by Dennis J. Wall, author of "Insurance Claims and Issues" (forthcoming Thomson Reuters 2016).  All rights reserved.

"INSURANCE CLAIMS AND ISSUES," THE BOOK: COMING IN 2016!

 

 Storm Clouds lined up over the Gulf of Mexico.  Photograph and text copyright 2015-2016 by Dennis J. Wall.

I am hard at work on an anthology of the best blogs from "Insurance Claims and Issues."  The anthology I submit to my publisher will keep all hyperlinks intact, including links to the Insurance Claims and Bad Faith Law blog.  When articles posted on Insurance Claims and Bad Faith Law blog are important to the anthology of "Insurance Claims and Issues," I am including them in the anthology too.

As each previously posted article is updated, the posted articles will remain just as available here as they always have been.  New articles will benefit from me being better informed and better equipped to discuss future insurance claims and issues.

I have contracted with Thomson Reuters (West Publishing Company to those fortunate to have the experience) to publish "Insurance Claims and Issues," the Book, in the late Spring or early Summer of 2016.  Best wishes to everyone for a joyous and Happy New Year!

Please Read The Disclaimer.  ©2016 by Dennis J. Wall, author of "Insurance Claims and Issues" (forthcoming Thomson Reuters 2016).  All rights reserved.

Related articles

Convincing Anyone Yet? Coverage, Contraception, and the Court.

CYBERINSURANCE:  MANY DATA-LESS RISKS TO UNDERWRITE & PREMIUMS TO CHARGE.

THE ALI RESTATEMENT PROJECT OF THE LAW OF LIABILITY INSURANCE.

Majority Holds No Negligence Cause of Action by Insured Against I.A.

Computer Models, Catastrophe Predictions, Premiums, and Claims.

    Continuing the trend toward Federal involvement in Insurance issues, the Government Accountability Office ("GAO") was asked by several U.S. Senate and U.S. House Members to prepare a report on what would be involved for private insurance companies to take on the additional risks of Flood Insurance.  (Private insurance companies avoid Flood risks; actually, private insurance companies refuse to cover them.  That is why the National Flood Insurance Program was enacted.)

    The GAO reported back that private insurance companies still do not want to wade into Floods, so to speak.  In general, if private insurance companies were going to take on risks in addition to those which they already cover, then private insurance companies would of course have to evaluate those risks, the GAO reported.  Here is the GAO report:  "Flood Insurance / Strategies for Increasing Private Sector Involvement" (released under date of January, 2014).

   At the same time as the GAO released the report, private insurance companies announced that they are establishing their own computer modeling platforms for evaluating catastrophic risks including Hurricanes and other 'weather events' which cause catastrophic damage including from floods.  Charles E. Boyle, "21 Insurers, Brokers Launch Oasis 'Open Source' Cat Model," posted online in "Carrier Managment", January 29, 2014.  Parenthetically, "Carrier Management" is a relatively new newsletter which apparently belongs to the same publishing group as the more widely known "Insurance Journal" and "Claims Journal" publications. 

© 2014 by Dennis J. Wall.  All rights reserved.  No claim to original U.S. Government works.

Please Read The Disclaimer.

Climate Change: The Insurance Companies Know. Don't You?

   The United Nations is in the midst of one of the many controversies swirling around climate change.  See Justin Gillis, "By Degrees / A Climate Alarm, Too Muted for Some" p. D1, col. 1 (New York Times Nat'l ed., "ScienceTimes," Tuesday, September 10, 2013).

   To property and casualty insurance companies and their reinsurers, there are no controversies.  There are only increasing risks and rising payouts as a result of climate change.  See, in particular, the Category "Climate Risk" on this blog.

Please Read The Disclaimer.

Related articles

Policyholder Attorney's Fees Award in a Florida "Confirmed Sinkhole Case".

Pull it Off the Computer, Print it, we're done.

Canadians Wade into Sandy Losses With Their Own Computer Models, and the Insurance Information Institute (III) Estimates Too.

    Aon Benfield, based in Canada, estimates that as of May, 2013 the known and predicted losses from Sandy include "$72 Billion in total economic loss".  "Aon Benfield's cat modeling arm" estimates that total "insured losses" included in that amount "are estimated to be $30 billion.  The $30 Billion estimate of the Cat Modelers includes roughly $7.2 billion" in Flood Insurance payouts.  "$72 billion in total economic loss from Hurricane Sandy:  Aon Benfield" in "Daily News" posted online on May 14, 2013 by Canadian Underwriter.ca.

    The industry publication reports from Canada on a confusing and apparently competing array of facts.  The apparent conclusion is that Sandy may not trigger Hurricane Deductibles.  On the other hand, the computer modelers are reportedly convinced that (excluded) flood and surge caused much more damage than (covered) wind perils.

    The Insurance Information Institute reported in "The Topic" online in July, 2013 regarding "Catastrophes:  Insurance Issues".  The III posting announced a much lower estimate of insurance claims at that time as a result of Sandy, in the amount of $18.8 Billion, nearly half of the estimated Canadian Underwriter amount in May, 2013, cited above.  Total Sandy damage was also reported by the III in its post as of July, 2013 at $50 Billion, a figure clearly less than the estimated $72 Billion in total economic loss, above.

    Like the Canadian source discussed above, however, the III too concludes that most of the claimed Sandy damage is excluded and otherwise not covered under homeowner's policies or commercial policies purchased by small businesses.  Insurance Information Institute, "Catastrophes:  Insurance Issues / THE TOPIC JULY 2013," posted online at http://www.iii.org/issues_updates/catastrophes-insurance-issues.html#at_pco=cfd-1.0.

 Please Read The Disclaimer.

Canadians Wade into Sandy Losses With Their Own Computer Models.

    Canadian Catastrophe Computer Modelers are wading into Sandy losses.  Aon Benfield estimates "$72 Billion in total economic loss" resulting from Sandy.  "Aon Benfield's cat modeling arm" estimates that of that amount,  total "insured losses" caused by Sandy "are estimated to be $30 billion.  The $30 Billion estimate of the Cat Modelers includes roughly $7.2 billion" in Flood Insurance payouts.   "$72 billion in total economic loss from Hurricane Sandy:  Aon Benfield" in "Daily News" posted online on May 14, 2013 by Canadian Underwriter.ca.

    The industry publication reports confusing and apparently competing facts but apparently concludes that Sandy may not trigger Hurricane Deductibles.  On the other hand, their Computer Modelers are convinced that (excluded) flood and surge caused much more damage than (covered) wind perils.  Id.

Please Read The Disclaimer. 

Related articles

After Sandy, Underwriters, CAT Models View Northeast Like Southeast: NAPCO

For All Models, Reality Requires Evidence.

This is an update to the November 7, 2012 article posted here, "Computer Models Doing Good After Hurricane Sandy?  Not so Much," and to the articles linked in it.

    In whatever form models may come, as program models or as computer models, one commentator recently observed in the context of social programs that "evidence-based funding criteria" is "a powerful new incentive for the development, rigorous evaluation and -- if effective -- dissemination of new program strategies and models.  It would catalyze evidence-driven improvements in an otherwise moribund system."  Jon Baron, "Applying Evidence to Social Programs" published on the Economix Blog on The New York Times website, posted November 29, 2012, http://economix.blogs.nytimes.com/2012/11/29.

    "Evidence-driven improvements" are exactly what certain computer models are designed in the Insurance context to replace.  They are not a substitute for measuring known losses backed by solid documentation, whether to predict weather-related catastrophic losses before they occur, based on years of records, or to quantify the known losses after they occur, by examining and evaluating the actual destruction.

Please Read The Disclaimer.

Computer Models Doing Good After Hurricane Sandy? Not so Much.

    Computer models have been used by Property and Homeowner's Insurance Companies to justify higher prices.  See the Category on this Blog, "Computer Models to Predict Premium Rates".

    Now, it appears that computer models are being used to 'predict' insured losses for claims not yet made following the destruction wrought by Hurricane Sandy.  E.g., Beth Pinsker Gladstone, "Insurance Divides Haves From Have-Nots After Sandy" (Reuters, Saturday, November 3, 2012); "Too Much Computer Modeling Makes Fools of Us All," article posted here on October 31, 2012.

    Far from justifying higher prices as yet, these computer 'predictions' of Hurricane Sandy Damages Claims are bad news for Homeowner's and other Property Insurance Companies in the Northeast:

    Insurers, the group that will feel the storm's effects most acutely, plunged as analysts warned that the storm will eat into their income.

"Rising Estimates of Costs of Storm Weigh on Wall St.," a "Stocks & Bond" copyrighted report of the Associated Press, published by the New York Times, p. B7, col. 6 (New York Times Nat'l ed. Saturday, November 3, 2012).

    Computer models.  What good do they do when losses can actually be determined, not merely 'predicted'?

Please Read The Disclaimer.

 

Too Much Computer Modeling Makes Fools of Us All.

    Hurricane Sandy is a big event.  It is focused on "nearly one-quarter of U.S. Homeowner's Insurance Market," wrote employees of A.M. Best's insurance news service on Monday, October 29, 2012 (subscription required).

    Hurricane Sandy is a fearful event.  The Hurricane already caused damage on the ground in Jamaica and the Bahamas, damage evidenced by things that can be seen and inventoried.

    Incredibly, computer models are being offered to "predict" the known and knowable damage in those unfortunate places.  And the information is offered for a price.  "Models Show Jamaica, Bahamas Damage From Sandy at Nearly $400 Million," Monday, October 29, 2012 (subscription required).  A slightly longer version of this "content requires a subscription" is offered by AIA's "First Read" EMail aggregation of insurance news published by the AIA on Tuesday, October 30, 2012.  With luck, you will be able to access it through the AIA offering.

    When computer models provide Insurance Companies' estimates of insured losses to take the place of verifiable damages, insurance claims provided by Policyholders will not be far behind.  Whether Judges and Juries, and lawyers and parties, will also treat estimates as reliable substitutes for verifiable inventories and proofs of loss, remains to be seen.

    But if that happens, it will not be the Judges and the Juries who are responsible.  The first step has been taken by someone else, not by them.

 Please Read The Disclaimer.  

 

HAPPY HALLOWEEN!

 

 

COMPUTER CLAIM MODELS USED TO INCREASE AUTO INSURANCE PREMIUMS ...

... FOR ACCIDENTS THAT HAVE NOT HAPPENED.

   The use of computer models to justify Premium increases is rampant.  The increasing use of computer models to justify Catastrophe Premium Increases for Homeowner's Insurance and Property Insurance has been commented on for years.  See, for example, the articles posted here on April 16, 2012, December 7, 2011, and June 29, 2011, and the many links in them to other sources as well.

    The use of computer models to justify Automobile Liability Insurance Policy Premiums has begun.  As in the case of Hurricane Damage and other Catastrophes, the database does not include -- as in not any longer -- evidence of actual occurrences and events.  Instead the computer claims model is premised on assumptions.  When such computer claims models are used in the context of increased Premiums for Homeowner's Insurance in Florida, for example, the "Garbage In, Garbage Out" ("GIGO") database consists of predicted future Hurricanes with the intensity of Hurricane Katrina and does not include actual payouts for actual covered storm damage caused by a Hurricane Isaac, if any.

    In the case of Automobile Liability Insurance, the computer database does not include accidents that have actually happened, but includes instead predictions about accidents that have not happened.  See David Sirota, "Big Brother in Your Car" (posted Friday, September 7, 2012)(©2012 Creators.com).

     Automobile Liability Insurance includes an added factor in the Premium equation.  Automobile Insurance is mandated by State statutes throughout the United States.   Auto Liability Insurance Companies have a built-in competitive advantage over many other Insurance Companies offering other lines or types of Insurance, thanks to State Governments.  In that view, then, it is appropriate for the same State Governments which mandate Auto Liability Insurance, to refuse to allow Auto Liability Insurance Companies to base their Premiums on losses that never occurred and on accidents that have not happened.

    At the same time, there is a larger question reflected by this particular development in one field of Insurance Premiums.  To say again,  the use of computer models to justify Premunium increases is rampant.  Computer models are increasingly used to justify any and all lines or types of Insurance:  Catastrophe Coverage including Coverage for losses caused by Hurricanes in Florida and by Earthquakes in California,  Homeowner's Insurance losses in Texas and Property Insurance losses in New York, and now Automobile Liability Insurance anywhere in the U.S.

    That is the real story, I think.  Practitioners, purchasers and underwriters of Insurance will either deal effectively with computer claims models used to support increased Premiums, and which are no longer based on actual, actuarial evidence but which are instead based on assumptions, or they will confront higher Premiums for all lines of Insurance.  No other choice appears.

Please Read The Disclaimer.

Florida Homeowners Insurance Premium Catastrophe: Imaginative Premium Base Spiked Premiums.

   Florida Homeowners are facing a Catastrophe without a named storm.  Premiums for Homeowners Insurance have spiked sharply, even skyrocketed, even over the Premiums that were charge for the same Homeowners Insurance last year.

 One reason for higher Homeowners Premiums is the higher cost of Reinsurance for Homeowners Insurers.  See Scott Maxwell, "Taking Names / Why Homeowner Insurance Rates Are Skyrocketing" p. B1, col. 1 (Orlando Sentinel, Sunday April 15, 2012).  But they are not just passing along costs as the market would ordinarily dictate.

 Instead they are pushing a false model, some say, which is based not on evidence but on opinions about how major storms -- if they strike Florida -- might sometimes result in Property Damage claims.  See "Current Computer Models Predict Premium Increases, Not Hurricanes," posted here on December 7, 2011 with links to prior posts.

 Computer models matter.  That is why many Homeowner's Insurance Companies are putting money in the bank, and why many Homeowners are not.

 Please Read The Disclaimer.

Current Computer Models Predict Premium Increases, Not Hurricanes.

   Hurricanes are not the only Catastrophes.  Today is Pearl Harbor Day.  Remember Pearl Harbor and remember, too, that everyone's life is entitled to respect if only because all our lives, already too short, are every one subject to unexpected Catastrophes and losses.

    The role that computer models play in current Insurance Premium Rate Requests is again being looked at.  This is a recurring, if sporadic, occurrence.  See, for example, the post here on June 29, 2011, "More Computer Models Print Out In Put:  This Time, Coverage?"

   The Massachusetts Attorney General's Office sent a letter to the Massachusetts insurance rating bureau on November 17, 2011.  In it, the A.G. pointed out her position that "faulty computer-generated hurricane models have contributed to unnecessarily high home insurance rates for property owners across [Massachusetts]."  Mark Alan Lovewell, "A.G. Faults Hurricane Models For Inflating Homeowner Insurance Rates" (Vineyard Gazette, published online on Friday, December 2, 2011).

   Although the A.G.'s letter itself does not seem to be on the A.G.'s website, a Press Release dated November 17, 2011 is accessible there.  In it, the Massachusetts Attorney General makes clear both that transparency does not exist in making these computer models, and that accumulated actual evidence is ignored in the input into these models.  These are real problems for homeowners who must pay higher Premiums:

Most Massachusetts insurers use private prediction models to estimate expected hurricane losses when rating and underwriting a policy.  AG Coakley noted in her letter that the insurers are not appropriately detailing how the companies are using these models, and no regulatory body has reviewed most of these models to ensure that they are suitable for use in Massachusetts.  Not only are models not being reviewed, the Commissioner is allowing the use of models that have been rejected in other states, such as Florida.  Historically, these rejected models have inflated expected losses well above actual experience.

 Massachusetts Attorney General's Press Release of November 17, 2011 concerning her letter to the Commissioner of the Massachusetts Insurance Rating Bureau, entitled "AG Coakley Urges Hearing on Excessive Homeowner Insurance Rates / Estimates Possible Overcharges of Half a Billion Dollars Due to Untested Hurricane Models".  Both of these points are made in the linked Vineyard newspaper report as well.   Parenthetically, so-called 'information asymmetry' or unbalanced access to information used in the writing of Insurance and in the handling of Insurance Claims is also a frequent subject of current commentary.  See, e.g., "Asymmetrical Information Overload" posted on Insurance Claims and Bad Faith Law Blog on November 22, 2011. 

   It is reasonable to expect an even greater increase in the use of allegedly faulty computer-generated models of catastrophe risk, along with the leverage that comes from information asymmetry.  Through the first nine months of 2011, Property and Casualty Insurance Companies in the United States experienced "catastrophe-related losses" that doubled all of the catastrophe claim payouts of 2010, or $38,600,000,000.00 ($38.6 Billion) in the first nine months of this year.  See "U.S. P/C Insurers' Catastrophe Losses Top $38 Billion" (Insurance Journal, published online on Monday, December 5, 2011).

The author of this post is also Co-Author of "CATClaims:  Insurance Coverage For Natural And Man-Made Disasters" (Thomson Reuters West).

Please Read The Disclaimer. 

More Computer Models Print Out In Put: This Time, Coverage?

   The U.S. Senate Banking, Housing and Urban Affairs Committee is considering whether Computer Models can be used to predict how much of a Hurricane Loss is due to Wind vs. how much is due to Flood.  See Lydia Mulvaney of McClatchy Newspapers, "Models Can Determine Wind and Water Damage, Panel Told" posted in political blog on MiamiHerald.com on Thursday, June 23, 2011.

   There is a major drawback:  Bad Faith.  Although it has been held by the Courts that it is not Bad Faith simply to use Computer Models as tools to determine Coverage for all or part of a Loss, Courts have also held that Insurance Companies expose themselves to liability beyond their Policy Limits if they abdicate to Computers their responsibility to determine whether all or part, if any, of a Loss is Covered.  See the post on Sunday, June 26, 2011 on Insurance Claims and Bad Faith Law Blog where this issue is addressed, and see the related post on Tuesday, June 28, 2011.

   There is a second drawback:  These Models are not necessarily as accurate as they are advertised to be.  There is no such thing as "Garbage In, Gospel Out".  See the post on Insurance Claims and Bad Faith Law Blog on Tuesday, June 28, 2011.  The people selling these Computer Models have a right to sell us this software, if we let them, but they have no right to call their Computer Models "the Gospel Truth".

   There is a third strike:  The way in which these Models are being sold is suspicious, although it may be sincere.  Predictions by people who profit from the predictions are inherently suspect predictions, whether the predictions are made by people who will profit by selling the software or by trying to sell a lower payment.  Advertising of these particular computer programs seems often to be based on the same marketing playbook that has been used recently in financial transactions and natural gas schemes:

"They want to bend light to hide the truth."

  Anonymous retired Unocal geologist writing in a February, 2011 E-Mail "about other companies invested in shale gas," quoted by Ian Urbina, "Insiders Sound an Alarm Amid a Natural Gas Rush" p. 1, col. 5 (New York Times Nat'l ed., Sunday, June 26, 2011).

Please Read The Disclaimer.

Florida Questions Catastrophe Computer Models.

   Florida newspapers slowly continue to publish reports in which Insurers' Catastrophe Computer Models are questioned, in a subject addressed here in posts that began "on or about" December 3, 2006.

   First up in the current newspapers is Paige St. John, "Insurers' Computer Models Deeply Flawed" (Sarasota Herald Tribune Online Monday, November 15, 2010), part of a series of reports over a year in the making in Florida's Sarasota Herald Tribune.

   More to come, and the more responses and answers to these many questions, the better.

 Insurance Questions including how to handle Catastrophe Claims Investigation and Evaluation, have in general terms been addressed by many Courts, many Insurance Companies, and lots of Insureds, in many decided Cases.  Dennis Wall is Co-Author of the leading book on Insurance Coverage for Catastrophe Claims, "CATClaims:  Insurance Coverage for Natural and Man-Made Disasters" (West Publishing Company 2008; 2010 Supplement in process).  An experienced Expert Witness in Insurance Cases himself, Mr. Wall will speak on a panel at the American Conference Institute's Bad Faith Litigation Conference in Orlando, Florida on November 30, 2010 on "Dealing With Catastrophic Disasters:  How to Properly Investigate and Handle Overwhelming Claims".  The author has been advised by the American Conference Institute that the ACI is continuing to offer its discount to readers of this Blog:  Download ACI Advises Readers of the Insurance Claims and Issues and Insurance Claims and Bad Faith Law Blog are entitled to a discount.

Please Read The Disclaimer.

 

Computer Models Calculate Risks of Buying Your Life Insurance Policy.

   Not content with the limitations of securitization and bundling so amply revealed in late 2008 and early 2009, many investors are turning to Computer Models to calculate the risks of buying Life Insurance Policies -- from you and your clients.  The prospects held out by this approach include bigger returns on investments for Wall Street and maybe most investors, but not for you.

   Wall Street investors are reportedly gathering steam as they grow excited about the returns they could receive on investments in other peoples' Life Insurance Policy payouts.  Investors' risks in "life settlements" and "viatical settlements" include the Insured outliving her or his life expectancy.  "The sooner the seller [Insured] dies, the more money the investment makes."  Michelle Singletary, "Playing Investment Games With Life and Death" (Washington Post Online, Thursday, September 10, 2009).

   This particular "investor's risk" carries with it the specter of potential fraud and potential peril to the Insured seller.  If investors receive a sufficiently greater return on their investment if you or your client die sooner rather than later, i.e., if the purchaser of your Life Insurance Policy payout gets more money the sooner you die, there is obviously a range of incentives built in to that investment, ranging from an incentive to perpetrate mayhem to  at least acquiesce in all sorts of life-threatening policies, practices, and procedures.  Talk about "Moral Hazard"!  This type of 'investment' gives a whole new meaning to that phrase.

Please Read The Disclaimer.

No Storms Proof to Some Catastrophes Are On the Way.

    There have been no named storms so far this year in Florida.  Knock on wood (you cannot see me right now, but I am tapping my forehead).

    This is good news for most of us.  Even the Computer Models used in the Rocky Mountains to predict Hurricanes in Florida (under an arrangement made by the previous administration) and the National Weather Service have downgraded the number of likely Hurricanes and damaging storms for 2009.  (Isn't it an oxymoron to label individuals in Colorado as Hurricane Experts?  It seems a lot like saying that people who work in Florida are the perfect candidates to predict rockslides in the Rockies.)

    The absence of storm damage and the lowered predictions of Catclaims are especially good news for most of us that live here in Florida.  See generally Emily Nipps, "It's Quiet Now, But Keep the Canned Goods Handy" (St. Petersburg Times Online, Thursday, August 6, 2009), republished online on August 7, 2009 under a totally different headline..

    Please Read The Disclaimer.

    

Hope.

    Dr. King's Birthday will be celebrated on January 19 in 2009.  Dr. King is being honored with a National Holiday to be observed on the occasion of his birth because he was a great American.  His life gave evidence to us all that perfection is not required in order to be a good person, let alone, in his case, a great one.

    Hope is a singular quality he provided.  And which his memory continues to provide.  I call often to my mind something that Dr. King said about a person who had lost hope.  This is what he said:

The cessation of breathing in his life was merely the belated announcement of an earlier death of the spirit.  He died when he lost hope.

[Emphasis in original.]

    Hope can be found for most Americans in the Inauguration of a new President.  See, e.g., Maya Angelou, "At Long Last, Ready For The Real Deal" (Washington Post.com, Sunday, January 18, 2009).

    That is much.  More even than that, and in addition to that, hope can be found in other things previously despaired of.  Such as courage and competence under pressure.  The pilot and all the members of the crew of Flight 1549 provided that kind of hope in the week before this post.  Their names are Chesley B. Sullenberger III, the pilot; co-pliot Jeffrey B. Skiles; and Flight Attendants Doreen Welsh, Sheila Dail, and Donna Dent.  See, e.g.,  Ray Rivera, "In a Split Second, a Pilot Becames a Hero Years in the Making" p. A19, col. 1 (New York Times Nat'l Ed., Saturday, January 17, 2009); Ray Rivera and Liz Robbins, "For Flight 1549, With Age Came Wisdom," The Lede web log, January 17, 2009 post.

    Hope.  That Insurance Premiums will be forced to be based on actual actuarial information and not Computer Model GIGO 'projections'.

    And that among other things Insurance Companies' market performance will be a result of their strong Insurance operations rather than CDSs and CDOs and other false profits.

    Hope.  This is a great message sent by Dr. King's Birthday.  His Birthday is observed as Washington's Birthday is observed together with Lincoln's Birthday.  The commemoration of Dr. King's Birthday was not enacted by the U.S. Congress as a day they gave Dr. King as his "Day," which they can take away.  To the contrary.  An annual observance of Dr. King's Birthday was enacted into law because of the accomplishments which Dr. King had already conferred upon the Country.

Please Read The Disclaimer.

Computer Models: The New Astrological Charts.

"Their credibility suffered for it last year."   Louis Uchitelle, "Forecasters See Fast Recovery; Others Doubt Their Eyesight"  p. A1, col. 5 (New York Times Nat'l Ed., Sat., Jan. 3, 2009).

    Their credibility suffered because they failed.  They could not predict what they said they could predict. 

    They used Computer Models.  The Computers were filled with information.  The Models digested that information and displayed the directions reflected by that mass of information.  The Computer Models did not work.

    Computer Models of Catastrophes?  In a sense.  The Catastrophes in this illustration are financial.  But financial Catastrophes only for the many.  A few including the self-proclaimed predictors made a lot of money.

    Even though they failed.  Even though their Catastrophe Models were wrong.

    Now they are at it again, the predictors of things that they do not have the capacity to predict.  For some reason, they are still speaking, still demanding to be listened to, still touting their failed Computer Model printouts as Astrologers touted (and continue to tout) their Star Charts.  See generally Louis Uchitelle, supra.

    Property Insurance Companies use their own Computer Models too.   They too jealously guard the secrets of their Computer Models.  They do not generally allow others to test the information they put into their Computer Models, either.  They too tout untested or just plain false Computer Model printouts, and for the same purpose:  Financial Astrology.  In their case, higher Premiums for Property Insurance.

    Otherwise, why would they hide the information that goes into the Computer Models, why would they make their Computer Models so mysterious, and why in short would they pretend that these Computer Models are able to predict the future when so clearly they cannot?

Please Read The Disclaimer.

 

Catastrophe Model Companies Estimate Damages, Provide Basis to Evaluate Themselves.

    Companies marketing Computer Models of Catastrophes have now opened a new line of potential profit for themselves.  They have become estimators.  Their estimates of Catastrophe Claims (not including Uninsured Damages) from Hurricane Ike in Texas vary as widely as the scoring in Little League Teeball games:

                    1.  Eqecat, Inc. (mentioned in a post here on September 7, 2008 that implicated CatClaim Models) reportedly estimates $8,000,000,000.00 to $18,000,000,000.00 or $8 Billion to $18 Billion in Insured CatClaim Damages.  Eqecat is based on Oakland, California.

                    2.  AIR Worldwide, located on the other Coast, in Boston, Massachusetts, estimaes $8 Billion to $12 Billion.

                    3.  Risk Management Solutions Inc., based in Newark, California, reportedly estimates $6 Billion to $16 Billion.  See Erik Holm, "Hurricane Ike's Toll in Texas May Cost Insurers Up to $18 Billion" (Bloomberg.com, Sunday, September 14, 2008).

    The next potential profit line that the Computer Modelmakers may venture into, may be  to provide estimates of paid claims on which Premium Rate Filings will be based -- rather than on the actual payments of actual CatClaims.

Please See The Disclaimer.

Federal Disasters Declared, Catastrophe Models Unspecified But Described As Consistent.

     During his two terms as president, Mr. Bush has issued 422 declarations of disaster.  That is described as more than President Bill Clinton and far more than President Ronald Reagan, also two-term presidents.  The largest number of Mr. Bush's disaster declarations in any given year came in 2004.  Austin Bogues, "Bush Declared 422 Major Disasters" (New York Times Online, Sunday, August 10, 2008).

    Mr. Bush's government committed $40 Billion in disaster relief from the Federal Government after Hurricane Katrina and Hurricane Rita in 2005.  Since they took office, they have committed an additional $47 Billion, for a total of $87 Billion in Federal Taxpayer Funds used for disaster relief, so far.  With or without the relief following Hurricanes Katrina and Rita, Mr. Bush has assigned more Federal money to disaster relief than his predecessor by some 2 and 1/2 times after adjusting for inflation, it is reported in the linked newspaper article.  The likely cause is, ironically, global warming.

    Before concluding this post, it is worth noting that a curious quote in the linked report is attributed to a person with the Environmental Defense Fund.  He is quoted as observing that "'weather events'" are "'consistent with what the models predict'".  However, the linked article does not provide a description of whose Computer Catastrophe Models they have in mind.

Please Read The Disclaimer.

Florida O.I.R. Lifted Suspension of New Allstate Policies...

... The Same Day Florida's First District Court of Appeal Denied Rehearing of its Opinion Upholding O.I.R.'s Order Suspending New Allstate Policies.

     Press reports published on Saturday, May 17, 2008 made it appear that the Florida Office of Insurance Regulation (O.I.R.) lifted its suspension on new Allstate Insurance Policies on Florida on the day before, or Friday, May 16, 2008.  As you will recall, the O.I.R. issued its Suspension Order in the first place because Allstate failed to comply with O.I.R. subpoenas for various groups of documents including unauthorized Catastrophe Computer Models purchased by Allstate in connection with its Premium Rate Increase requests.

     According to the O.I.R., it announced a stay of its Suspension Order on Wednesday, May 14, 2008, immediately after the Florida appellate court denied rehearing of its Opinion upholding the O.I.R. Order.  The reason given by the O.I.R. was its receipt of a sworn affidavit from Allstate "certifying that it has complied with Florida law by freely providing all documents requested" by the O.I.R.  See Florida Office of Insurance Regulation Press Release, "Florida Insurance Commissioner McCarty Stays Allstate Suspension, Continuing Compliance Required," Wed., May 14, 2008.

     Dennis Wall, Chair of the Insurance Law Committee, will speak at the Orange County Bar Association Center in Orlando, Florida about Preparing for Hurricane Insurance, on Wednesday, May 21, 2008 from Noon to 1:00 P.M.  The Orange County Bar Association Center is located at 880 North Orange Avenue in Orlando, Florida.

    You are cordially invited to attend this free public event.  Hope to see you there. 

Please Read The Disclaimer.

Balderdash to Insurance Regulation by The Feds ....

.... Whether or Not These People Say Katrina Was So Good, They Want to Do It Again.

     Ms. Leigh Ann Pusey, identified as "the senior vice president for government affairs" at an Insurance trade group identified with the current administration, calling itself the American Insurance Association, is quoted as endorsing the current Federal government's offer to take over Insurance.  The current Federal government's proposal was placed without publicity among the 200+ pages of a document it recently issued in response to unfair financial practices.  Their offer is to take over Insurance Premium Regulation by doing away with it.

     See this quote by Bloomberg in a news report published on April 16, 2008:

"With all due respect to many of the leaders at the NAIC [National Association of Insurance Commissioners], I think they are structurally and functionally incapable".

Ms. Pusey was apparently talking about regulating Insurance Premium Rate Hike Requests.  She is also quoted in the news article as saying that Federal regulation of Insurance will "happen sooner than people think."

     Thinking like that is mistaken.  On several grounds.  After looking at quotes like these, investing a little effort into research pays dividends in a reality check.  One easily understood fact is that, to the contrary of what anyone may think, the NAIC does not regulate Insurance Premium Rate Hike Requests.

    Insurance Departments in the 50 States regulate Insurance Premium Rate Hike Requests.  They have done that job for 150 years.  All of a sudden we are to believe that 50+ Departments of Insurance have been rendered incapable of doing that job?  What changed?

     The news article does not even try to address what if anything has changed.  Neither do the Federal government supporters quoted in it.  And that would have been a revelation.

     Nothing is known to sentient beings that changed at all except that the people in charge of the current Federal government, before they cease to be in charge of the current Federal government, want to provide Insurance Companies with no oversight for any Insurance Premium Rate Hikes the Insurance Companies may want.

     It is of interest, particularly to businesses and people in Florida at this time, that besides of course the people currently in charge of the Federal government who made this proposal (surprise, they endorse their own proposal, duh), their proposal is "endorsed" by Allstate Corporation according to this and other news articles.

     Allstate is currently appealing a Court Order that affirmed the Florida Office of Insurance Regulation's suspension of Allstate's authority to write new Insurance Policies in Florida until Allstate provides documents it has withheld that were subpoenaed by the O.I.R. months ago -- which include documents and unauthorized Computer Models that Allstate used to endorse its own Florida Insurance Premium Rate Hikes.

Please Read The Disclaimer. 

Allstate and the State: Subpoenas, Catclaim Models, Reinsurance ... and Premiums.

     The Florida Office of Insurance Regulation (Insurance Department) has filed an administrative Complaint against Allstate.  The object is to force Allstate to provide documents previously demanded by subpoenas served on Allstate by the State.  The subpoenas demand documents from Allstate that describe alternative CatClaim Models favored by Allstate in computing its Premium Rate Increase Requests in Florida, Allstate's reinsurance treaties and agreements, and other matters that do or may factor into requesting higher Premiums in Florida.

    This administrative proceeding will be heard by an administrative law Judge from the Department of Administrative Hearings.  One available sanction if the State prevails on its Complaint is to yank Allstate's license to engage in the business of Insurance in Florida.

    In a separate and independent proceeding in the interim, the appeal is continuing that Allstate filed after the D.O.I.'s attempt to issue an order directly to Allstate that would halt Allstate's issuance of new Insurance Policies in Florida.  Florida's First District Court of Appeal has stayed that order while that appeal is pending, meaning that Allstate may continue its business of selling new Insurance Policies while the appellate proceeding goes forward.

    These developments are reported recently by Anika Myers Palm, "State Seeks to Block Allstate/It's the Latest Step to Obtain Documents From the Company Related to Hurricane-Risk Models" p. C1, col. 5 (Orlando Sentinel Central Florida Business Section, Wed., Feb. 20, 2008); and by Tom Zucco, "State Goes After Allstate on Code Violations/A Broader Pursuit Alleges [sic] Also Seeks to Block the Insurer From Writing New Policies" (St. Petersburg Times Online, Wed., Feb. 20, 2008).

Please Read The Disclaimer.
 

Computer Models Authorized and Unauthorized Used for Premium Rate Increases.

   Some Homeowner's and other Property Insurance Companies have requested rate hikes in Florida using unauthorized computer models to project and predict huge Hurricane losses in the next few years.  These requested rate hikes, based on unauthorized, unknown computer models have all been denied to date.

   Other Homeowner's and Property Insurance Companies have complied with Florida law and have not found any problem doing so.  They have reduced Premiums for the benefit of Florida Homeowners and other Property Insurance Policyholders because they have lowered their profit margins.

   The most important law governing authorized vs. unauthorized computer models to predict future Hurricane losses is Florida Statute Section 627.0628.  It has been around since 1995 and it was enacted partly in response to Hurricane  Andrew.   Here is a searchable link to the Florida Insurance Statutes and other Statutes provided by the Florida Legislature's web site.

    In Section 627.0628, the Florida Commission on Hurricane Loss Projection Methodology was born.  Here is a link to the web site of the Florida Commission on Hurricane Loss Projection Methodology.

    Subsection (3)(c) of Section 627.0628 tells what kinds of catastrophe models can be used by insurers in their rate filings, or requests for premium increases, which are governed by Florida Statute Section 627.062, "Rate Standards".  The law governing authorized vs. unauthorized models, Section 627.0628, says that Insurers making such a rate filing requesting authority to charge higher Premiums may use models "found by the commission to be accurate or reliable to determine hurricane loss factors for use in a rate filing".

   Parenthetically, Florida is perhaps the only State that also provides "a public model for hurricane loss projections."  By law, this public model is required to be accurate and reliable "pursuant to the same process and standards as the commission uses for the review of other hurricane loss projection models."  Florida Statute Section 627.06281.

   There does not appear to be any doubt about the meaning of the words in Section 627.0628 in any past year since the words were first written.  If insurers want to make a rate filing or request a premium rate increase, then the kinds of models they use are models found by the commission to be accurate or reliable to determine hurricane loss factors for use in a rate filing.

    EVEN IF authorized models are used in making the rate filing, the resulting "findings and factors are admissible and relevant in consideration of a rate filing" ONLY IF IN ADDITION the Florida Office of Insurance Regulation and the Florida Consumer Advocate are allowed to see all of the factors that went into the insurer's computer models (among other things) AND the insurer does not try to keep them secret.  What happens if, as appears to be the case this year, an insurer makes a rate filing using a computer model which has NOT been found by the commission to be "accurate or reliable to determine hurricane loss factors" for use in the rate filing AND which the insurer tries to keep  secret?  The statute provides in what clearly seems to be express language that hurricane loss factors are  NOT  "admissible  and relevant" WITHOUT providing the access AND DROPPING the secrecy.

    Result of using unauthorized computer models WITHOUT providing the access AND WITHHOLDING secrets about the models?  Rate Filing DENIED.  That is the result that the Florida Statute clearly and reasonably seems to require.  That is the apparent result also in every rate filing in which the insurer used such computer models to date.  (This reality is, sadly, contrary to some opinions, some of which in turn are arguments for an absurd interpretation of the Florida Statute.  Everybody has an opinion, and some opinions are expressed here in a newspaper report suggesting that since a few but by no means all Insurance Company representatives argued for such an interpretation of the Florida Statute, then maybe an absurd interpretation is the same thing as a reasonable interpretation.  Not in the Florida Courts.)

    There does not appear to be any "loophole" actually in the Florida law, not on the face of the Florida Statute nor in the practical interpretation provided by the agency applying it.

Please Read The Disclaimer.

Florida to Allstate: No Doc Production, No Auto Insurance Production.

     Allstate is currently the 2nd largest writer of Auto Insurance in the State of Florida.  However, in the last 4 years, Allstate has dropped about 400,000 Homeowner's Insurance Policies and raised rates for the remainder.  The State of Florida served subpoenas 60 days ago for a 2-day Hearing scheduled to start on January 15, 2008.  Florida wants to know what the effects of reinsurance are on Allstate's Homeowner's Insurance performance.  Florida and many other States also want to know much more about an unauthorized CatClaims Computer Model used by Allstate and other Homeowner's Insurance Companies to justify their past rate increases and other industry conduct concerning Homeowner's Insurance, such as policy cancellations and nonrenewals.

    On January 15, 2008, rather than provide the documents, Allstate withdrew its pending request for a future rate increase and filed 51 pages of objections to the Florida subpoena.

    On January 16, 2008, the Florida Office of Insurance Regulation suspended Allstate's certificate of authority to write new Auto Insurance business and therefore prevented Allstate from collecting any new Auto Insurance Policy Premiums in Florida.  It is reported that the suspension will not affect  Policyholders under previously issued Allstate Auto Insurance policies in Florida.  See Anika Myers Palm, "Florida Cracks Down on Allstate/State Regulators Won't Let the Insurer Write New Auto Policies", p. A1, col. 5 (Orlando Sentinel, Thurs., Jan. 17, 2008).

Please Read The Disclaimer.

Catastrophe Models, Catastrophe Prices ....

   They're Not Just Your Parents' Florida Premium Payment Any More.

     Computer models for Catastrophes have recently come under closer scrutiny everywhere in the United States.  No more so anywhere else than Florida.  The focus is not so much on the Computer screen as on what information is selectively input into the Computer that ends up on the screen -- and in some Insurance Companies' Rate Hike Requests:

    

State insurance regulators have questioned why the models used by insurers only look at the past five hurricane seasons -- years that were unusually active and destructive -- instead of a broader period when determining how to set their rates.
                               
                                   * * *

     And the National Association of Insurance Commissioners is investigating the models and the way insurers use them to come up with conclusions about risk.  Florida Insurance Commissioner Kevin McCarty, who is chairman of the association's property and casualty insurance committee, has promised to give the models special scrutiny as part of his effort to lower homeowners-insurance costs for Floridians.

Anika Myers Palm, "State, Insurers to Clash Over Pricing Model," Orlando Sentinel Central Florida Business p. 30, col. 6  (Mon., January 7, 2008).

More to come.

Please Read The Disclaimer.

Profits Up, Payments Down: Hurricane Winds Still Blow -- Part 2.

     The second reason for a large increase in profits for Property Insurance Companies, besides the reductions in Insurance Coverage discussed in Part 1 of this post, is an increase in Premium charges.

     Increased Premiums are abetted by Computer Models that project Catastrophic Losses on a more frequent basis than they occur.  As posted here previously, these Computer Models are not authorized by State government, and Florida is an example, in part because they are often wrong.  The Computer Models are, in addition, predicated more on opinions and projections than on demonstrated facts.   Nonetheless, these Computer Models favor the imposition of high Premiums because they are used to predict more Catastrophes which trigger CatClaims in greater numbers and for more dollars.

     Again, for a comprehensive newspaper report which summarizes trends in Insurance Coverage Reductions and Insurance Premium Increases, an excellent report is by Joseph B. Treaster, "Insurers Shift Cost Burdens to Homeowners" p. C1, col. 2 (New York Times Business Day Nat'l Ed., Fri., Nov. 23, 2007).

     Premium Increase trends may be leveling off if predictions are accurate for 2008.  This is the conclusion of Dr. Robert P. Hartwig, CPCU, President of the Insurance Information Institute, in his "Special Report: Earlybird Forecast 2008" posted online by the III on December 17, 2007.  Dr. Hartwig averaged various forecasts and arrived at an average among them of a 0.3 percent decline in Property/Casualty Insurance Premiums.  Newspaper reports have reprinted this average projection.  See, e.g., Lavonne Kuykendall, writing for Dow Jones/Associated Press on Dec. 18, 2007, published online by South Florida Sun-Sentinel.com.  Dr. Hartwig's conclusion is that premium growth is "Stuck in Neutral," if these projections are accurate.

     "Neutral" is of course a relative term.  As Dr. Hartwig points out in his post, in 2007 Property and Casualty Insurance Companies are on track to post one of their best underwriting results since about the time of the Stock Market Crash year of 1929.  If industry predictions reflected in the post are accurate, Dr. Hartwig also points out that 2008 Premiums would decline for the first time since the middle year of America's involvement in World War II, or since 1943.

     The "Earlybird Forecast 2008" includes a table listing 10 analysts and their projections.  The average of a 0.3 percent decline in 2008 P/C Premiums does not account for the weight of their projections.  For example, Tillinghast and Fitch forecast increased Premiums, at 1.1 and 0.1 percent respectively.  Merrill Lynch and Standard & Poor's are shown in the table as forecasting a continuation of the level of Premiums from 2007 to 2008 without projecting increases at this time.

     Time will tell. 

Please Read The Disclaimer. 

Catastrophe Claims, Catastrophe Models

     Karen Clark has been engaged in making, implementing, and selling catastrophe models since the last Century.  Okay, for decades.  On her online site, www.karenclarkandco.com, this observation is written from that experience:

The catastrophe models have another important requirement, which is detailed data about the exposure at each location that could be affected by a catastrophe. For residential and commercial properties, this includes the replacement value, construction type, occupancy and other building specific details. Without this information or if this information is inaccurate, the models will not provide reliable results.

For more observations born of the experience of Karen Clark and Company, click here which will bring you to the first page of their web site, and click there on "catastrophe models" for the full posting.

     A new computer model to predict the probability, consequences and necessary preventive measures for wildfires illustrates the need for reliable factual information in computer models in this report by Kirk Johnson, "Enlisting Computers to Help Fight Wildfires" p. A14, col. 1 (New York Times Nat'l Ed., Saturday, December 15, 2007).  The computer model was caused by 8.9 million acres that burned in the United States this year.  That is the second highest total -- and the year is not finished yet -- since 1960.   

     The reported wildfire computer model relies on soil moisture data  provided by the United States Forest Service, images from Civil Air Patrol training flights, and analyses of ash from California wildfires.   Those concerned with wildfires and their effects on people have made the wildfire computer model a priority for their own time and effort: