Tuesday, May 27, 2014

What Florida's SB 542 teaches us about catastrophe insurance in the US...

Recently, Florida passed SB 542 setting up the regulation of a private flood insurance market.  I think the wisdom of this can be debated.  Regardless, the wording of the bill highlights the the dynamics at play in determining the cost of catastrophic property insurance for flood and hurricane.  (The earthquake insurance market may be a little more like a traveling circus than like the more patterned dynamics of flood and hurricane).

Let's start with the introductory wording of the bill.  I have highlighted a few things.  Then, I'll say a few words about why the wording is interesting and what it demonstrates.

627.715 Flood Insurance.—An insurer may issue an insurance
199  policy, contract, or endorsement providing coverage for the
200  peril of flood on any structure or the contents of personal
201  property contained therein, subject to this section.
202         (1) The Legislature finds that:
203         (a) The National Flood Insurance Program (NFIP) is a
204  federal program that enables property owners in participating
205  communities to purchase flood insurance. A community
206  participates in the federal program by adopting and enforcing
207  floodplain management regulations that meet or exceed federal
208  floodplain management criteria designed to reduce future flood
209  risk to new construction in floodplains. The program was created
210  by Congress in 1968 because insurance covering the peril of
211  flood was often unavailable in the private insurance market and
212  was intended to reduce the amount of financial aid paid by the
213  Federal Government in the aftermath of flood-related disasters.
214  Since the creation of the NFIP, generally flood insurance
215  coverage has been unavailable for purchase from private market
216  insurance companies.
217         (b) The Biggert-Waters Flood Insurance Reform Act of 2012
218  reauthorized and revised the NFIP. The act increases flood
219  insurance premiums purchased through the program for second 
220  homes, business properties, severe repetitive loss properties,
221  and substantially improved damaged properties by requiring
222  premium increases of 25 percent per year until premiums meet the
223  full actuarial cost. Primary residences lose their subsidized
224  rates if the property is sold, the policy lapses, repeated and
225  severe flood losses occur, or a new policy is purchased.
226  Policyholders whose communities adopt a new, updated Flood
227  Insurance Rate Map (FIRM) that results in higher rates will
228  experience a 5-year phase-in of rate increases to achieve
229  required rate levels.
230         (c) The Biggert-Waters Flood Insurance Reform Act of 2012
231  also encourages the use and acceptance of private-market flood
232  insurance. The Legislature finds, however, that there has been a
233  long-term inadequacy of private-market flood insurance available
234  in this state. Such inadequacy suggests that the private market
235  in this state is unlikely to expand unless the Legislature
236  provides multiple options for the regulation of flood insurance.
237  In addition, the consumers of this state will be protected from
238  excessive premiums by the continued oversight of insurance rates
239  by the Office of Insurance Regulation and the continued
240  availability of flood insurance from the NFIP.
241         (d) The NFIP, as amended by the Biggert-Waters Flood
242  Insurance Reform Act of 2012, will prevent many property owners
243  from obtaining affordable flood insurance coverage in this
244  state. The absence of affordable flood insurance threatens the
245  public health, safety, and welfare and the economic health of
246  this state. Therefore, the state has a compelling public purpose
247  and interest in providing alternatives to coverage from NFIP by
248  promoting the availability of flood insurance from private
249  market insurers at potentially lower premium rates so as to
250  facilitate the remediation, reconstruction, and replacement of
251  damaged or destroyed property in order to reduce or avoid harm
252  to the public health, safety, and welfare, to the economy of
253  this state, and to the revenues of state and local governments
254  which are needed to provide for the public welfare.

Flood (and hurricane) insurance is not really about the flood and hurricane risk
What risk does this legislation seek to manage?  A gut reaction to this question is likely, "Flood risk, obviously".  But this is not really so.

The strong emphasis of this bill is on NFIP and what Biggert-Waters did to NFIP rates.  So strong is this emphasis that the bill begins with an entire paragraph on the history of the federally run NFIP.  The risk being managed by this bill is uncertainty surrounding federal decisions about NFIP rates and the potential effect these decisions may have on Florida's real estate heavy economy.

Certainly, much of Biggert-Waters was recently repealed by Obama and the passing of the Homeowner Flood Insurance Affordability Act.  But what about the next flood event that has Congress scrambling to appease the public by again scheduling to increase rates on flood properties?

To the extent that Florida can successfully encourage the private market to pick up flood risk while also controlling the rates charged then the state can buffer against Congressional fiddling with flood risk.
Thanks to ramseur78 on Flickr
By creating the goal of "affordable property insurance" for flood, Florida establishes a means to enable the public to participate in decision making about the size of flood risk they wish to insure against.  The public's perception of flood risk may be different from Congressional or market determinants of flood risk, but, then, that is precisely why this legislation was created.    

The state's residual market for hurricane risk (primarily), Citizens Property Insurance Corporation, shares the same wording of "affordable property insurance."  Legislators worded it this way for similar reasons- market judgements of hurricane risk were far greater than the public cared to manage with insurance.

Solutions to US catastrophe insurance woes do not necessarily relate to causes
Paragraph c, lines 232-236 establish a weird cause and effect conclusion.  True, private market insurance has traditionally been unavailable because, traditionally, insurers found flood insurance difficult to measure.  But, that this "suggests" the need for state rate regulation is a straw man.

Some would have argued that it "suggests" the need for a federal backstop.  It could also "suggest" the warrant of having a federal flood insurance program.  It could suggest power politics between the federal government and the insurance industry (sort of a "Yeah?! well make us provide flood coverage" mentality).   It could also "suggest" a need to revise land management and real estate policies.

That is, a lack of private flood insurance could suggest many things and the legislature's leap is not intuitive.  Leaps in logic, when imbedded into public policy, can result in policy that is unhelpful and possibly harmful.

Perhaps more interesting than why private market insurers have not historically provided flood insurance in the US is why they are suddenly excited to do so now.  In recent years, modeling methodology has warmed insurers up to the idea of covering flood risk.  With a growing comfort around flood modeling and a high demand for big risks by the capital market the insurance industry has been chopping at the bit for the US to release some of its flood risk to the private market.  Florida, lead by (unsettling) insurance savvy, Rick Scott, has done just that.  

Threats to the state and national economy are bipartisan.
After regulatory changes to Citizens rates in 2007 and the introduction of "affordable property insurance," the then governor, Charlie Crist, became notorious in the insurance industry.  He was, along with his orange glow, symbolic of all that was wrong with "political" ratemaking in Florida. In general, it was felt that the democrats had it out the for insurance industry.

(05/30/14: Well, it was and still is but with my reversal of Crist's party affiliations it may not really have much to do with Crist's actions...)

Yet, Rick Scott, heralded as an insurance industry mascot, has demanded affordable property insurance, as well.  This is made complete with legislative fiddling with regulatory requirements to get there.

The difference between the two governors is that Citizens is a public insurance facility whereas this state flood insurance legislation seeks to encourage private insurers to write flood.

However, I wonder how material the difference actually is.  Insolvent insurers, whether a public residual market or private market companies, are still a public problem and money to resolve the problem comes from the same public.

The point is that when the cost of insurance threatens the economy, whoever is in charge and whatever their partisan flavor, they give in to economic concerns.  Likely, rightfully so.

But it also demonstrates that ratemaking is a political process: from choosing science to go into a model, to choosing how to implement models, to choosing what rates to charge regardless of models and negotiating those decisions with the public, insurers, reinsurers and capital markets.

Charlie Crist: The (ex) Democrat that enabled
affordable property insurance for hurricane
5/30/14: Sorry!
The (ex) Republican, then Independent now Democrat
Rick Scott: The Republican that enabled
affordable property insurance for flood

Insurance is intended to provide for the general welfare 
Finally, line 254 states what nearly all insurance legislation that I have ever read states: "...provide for the general welfare."

In many ways, insurance is a form of privatized government and provides for the economic stability and safety nets that we put government in charge of.  Jobs, the creation of capital, spreading risk over larger populations, etc are all benefits of using insurance instead of relying wholly on government.    

The successful implementation of an insurance regime is judged largely on how well general welfare goals are being met- detected mostly in the mood of the public.  If those goals are not being met (e.g. people can't hold onto their mortgages because insurance costs are too high and they are upset) then government steps in to provide an alternative.  

Therefore, the insurance industry as a whole has a responsibility to  and interest in meeting the needs of the public.  When it doesn't it faces great challenges in access to policyholders and limits on freedom to judge the characterization of risk for itself.   

Thursday, May 22, 2014

The politics of catastrophe models

The above infographic is taken from here- a article providing an interview with Karen Clark, a foremother of catastrophe modeling.

The image gives a general schematic of decision points for creating a catastrophe model.  At each point the model builder must make a choice of input or how to construct an input (such as how to construct a catalog of hypothetical events).  These choices can and do produce different outputs or rather, different estimates of risk.

Is one choice of an input better than another?  Well, it depends on what you are trying to accomplish.

The goodness of model input is justified by the desirability of its output.  Economist and Philosopher Marcel Boumans has likened the model building process to "baking a cake without a recipe,"
The ingredients are theoretical ideas, policy views, mathematisations of the cycle, metaphors and empirical facts… However, a recipe is not unique in the sense that it is the one and only way to integrate a certain set of ingredients.  Thus a new recipe is a manual for a successful integration of a new set of ingredients [emphasis in original].  
The recipe creation process ends when the baker (modeler) creates a cake (output) to his liking.

Ultimately, the risk we choose to model, accept and insure against are fully malleable.  This provides ample political opportunity to influence choice in the risk estimation and risk acceptability process.

Think about risk estimates as highly technical propaganda.  They represent specific political positions about what the future could look like given a certain set of assumptions.

This does not mean they are not useful for planning.  Certainly they are.   Which is where they derive power to influence decision making.

However, their specificity should be taken with a grain of salt.  While the model may foretell what the future may look like by a given perspective and his/her values and expert judgement, it certainly cannot tell you what to do about it.

Thursday, May 15, 2014

The rise of risk

I created the graph above (which you can also get here) using Google and its seemingly endless capabilities.

Risk is a specific type of uncertainty.  Generally, risk is the possibility of something that represents a threat.  For the purposes of insurance, risk is a measurable uncertainty.

The graph suggests that society has long thought and written about risk and uncertainty.  Seems reasonable as these two concepts were central to early statistical and probability theory.  While we have remained consistently uncertain we have grown remarkably concerned about risk since about 1970.

So, if we are equally as uncertain today as yesterday how could we be more risky?

I'll leave this open ended for now, but I have my theories =)