Tropical storm Ingrid should not pose a wind threat to Texas

Tropical Storm Ingrid took shape today in the Gulf of Mexico, but if forecasters are correct, it should not pose a major wind danger to Texas or the Texas Windstorm Insurance Association. The GFS ensemble, shown here, all shows landfall in Mexico. So too do the other major hurricane models.  Of equal importance, the SHIPs intensity model, which attempts the dark art of windspeed prediction, does not show Ingrid getting much above its current 50 knot tropical storm intensity.  So, although I have raised alarm bells for some time about the resilience of the Texas Windstorm Insurance Association to a significant storm, Ingrid does not look like the one that will test the largest insurer along the Texas coast or, for that matter, other insurers.

Ingrid GFS Ensemble

Ingrid GFS Ensemble

Of course, weather forecasters haven’t been having the best year making predictions.  If Ingrid takes a more northerly track, all bets are off.

Invest 92L not showing signs of strength

Good news. The 8 p.m. runs are done and they show growing agreement on what the future holds for Invest 92L.  It looks like Houston and the northern coast of Texas are unlikely to be hit. From an insurance perspective, this is important because this is where most of TWIA’s exposure is situated. Landfall of something ranging from a tropical depression to a high end tropical storm is more likely closer to the Texas-Mexico border. None of the forecasts I have studied are indicating any sort of major hurricane or event that should cause major wind damage. Indeed, the probability of this thing ever becoming getting the “Fernand” name are now down to 50%.

The HWRF model for August 16, 2013 at about 8 p.m.

The HWRF model for August 16, 2013 at about 8 p.m.

The HWRF model points to a a strong tropical storm or category 1 hurricane hitting the Brownsville/Matamoros area on Monday. The GFDL model shows a weak tropical storm hitting a lightly populated area south of Corpus Christi late Monday night. The LBAR model shows something — for some reason the wind speed data is gone — hitting around Port Aransas and Rockport but sooner that the other models suggest, perhaps late Sunday night or early Monday morning. The NAM model shows a hit on Corpus Christi but a storm not even windy enough to be a tropical storm.

Now, all of this could change. And the Brownsville/Harlingen area needs to stay on alert. But I don’t think anyone needs to be making runs to Home Depot for plywood tonight.

Keep your eyes on Invest 92L (Fernand)

The 2 p.m. Computer Model Hurricane Forecasts are in and virtually all of them show what is now “Invest 92L” but what may become “Fernand” hitting the Texas coast in the next few days. One of the projections (GFDL) shows it hitting Corpus Christi. Another model (LBAR) shows it hitting the most populated part of the Texas Coast around Galveston.  It’s too early to say whether Invest 92L will develop into a tropical cyclone — forecasters have been giving it about an 60% chance of doing so — or, if it does, how it will intensify over the world’s largest bathtub, the the Gulf of Mexico, but it would not be a bad time for those at risk from tropical storms, particularly those who have windstorm insurance provided by the undercapitalized (negative surplus) Texas Windstorm Insurance Association, to start thinking about preparations.

2 p.m. August 16 predictions for Invest 92L

The best information I’ve seen on intensity comes from They’ve got three forecasts that show intensity.  The GFS model predicts a very weak depression hitting Corpus Christi on Tuesday, nothing to be concerned about. The HWRF model, a newer but possibly less reliable version of GFS, shows a category 1 Hurricane Fernand going into Mexico.  The GFDL model shows a tropical storm Fernand meandering into Corpus Christi on Tuesday. There is no intensity information that I can find for the LBAR model that implicates the highly populated Galveston/Matagorda area, backed up by the Houston megalopolis.

So, no, Katy need not evacuate yet — nor need anyone else. There isn’t enough to go on yet. This is not the long-track hurricane like Ike or Rita that has been more common recently along the Texas coast and that has, at the end, been a bit easier to predict.  This “thing” is a little closer to what became Hurricane Alicia in 1972. It began life in a similar area but intensified rapidly as it made a fairly quick path towards the Texas coast. On the other hand, there are certainly lots of “things” out in the Gulf that have looked life 92L but have fizzled or gone places other than Texas.  Catrisk will keep you posted.

Experts still predicting active hurricane season

Most mornings I wake up and, after checking the fate of certain baseball teams and world peace, neither of which have had good months,  wander over to the Weather Underground and see if anything is going on in the tropics. It’s been very quiet recently. Does this mean that the experts have been wrong and Texas, among others, is going to escape 2013 without significant storm damage?  Does this mean TWIA will be able to make it at least one more year without running out of money? Has the failure of Texas to address windstorm insurance reform during the past regular legislative session or over the course of three special sessions been, in fact, a brilliant allocation of time?

Unfortunately, the experts say the absence of much tropical activity thus far this summer doesn’t matter that much.  The National Weather Service and Climate Prediction Center says we most likely have 9-15 named storms ahead of us in the Atlantic, 6-9 hurricanes and 3-5 major (Category 3-5) hurricanes. The August 2 forecast from Colorado State University, which for reasons unknown to me is a leader in this area notwithstanding the rarity of tropical moisture affecting Fort Collins, predicts 14 named storms, 8 hurricanes and 3 major hurricanes remaining. CSU sees a 43% chance for at least one hurricane hitting Texas in the remainder of the hurricane season and a 16% chance of a major hurricane impact. Tropical Storm Risk, a private organization affiliated with University College London and with close ties to the insurance and reinsurance industries, reports in its August forecast a a 50% probability that this hurricane season will be the top “tercile,” which means that it will be worse than 2/3 of other hurricane seasons in terms of “Accumulated Cyclone Energy,” a proxy for the destructive power of hurricanes when they hit land.

Now, all of this would not matter if predictions made in August had little value.  After all, although it is seldom reported, predictions made in winter or spring about the forthcoming hurricane season are little better — and sometimes worse — than a blind prediction that the season will be average. Unfortunately, by the time August rolls around, and as shown in the graph below, there is enough climate data for experts to make pretty reliable predictions. It’s these predictions on which the popular press should be focusing. So, the fact that these experts — and others — are persisting in their prediction of a bad hurricane season are worrying.  14 storms over the 113 days left in “hurricane season” means one could see about a storm a week in the near future.

Credits, by the way, to Dr. Jeff Masters, of the Weather Underground for writing an excellent blog entry from which much of the material here has been adapted.

Thanks, Andrea, for heading East while we dither

So, the season’s first tropical cyclone has formed, Andrea.  Fortunately for Texas, the computer models appear in agreement that this one is heading to poor old Florida.  And a good thing too for Texas.  Because, while Governor Perry considers whether to add windstorm reform to the agenda for a special session of the Texas legislature and its members possibly see if they could get close enough to an agreement to make such a session fruitful, Mother Nature is not waiting.  So, could we treat Andrea as a shot over the bow? This year’s tropical storm season has gotten off a bright and early start.

Weather Underground Map of Andrea

Tropical Storm Andrea: Apparently not headed for Texas



Troubling news: TWIA loses $500 million in anticipated funding

The short term finances of the already shaky largest property insurer on the Texas coast took an unanticipated and significant turn for the worse Monday.  Outgoing Texas Insurance Commissioner Eleanor Kitzman rejected Monday plans of the Texas Windstorm Insurance Association to borrow $500 million via a “Bond Anticipation Note” to help pay claims this hurricane season.  The Commissioner did not reject a plan to issue post-event bonds in the event of a significant storm this season.  As a practical matter, however, it may be difficult to persuade the market to loan money to TWIA after a storm due to peculiarities in the existing law that were not ironed out during the regular session of the Texas legislature.

The refusal to permit TWIA to borrow at this time, coupled with the announced $135 million settlement earlier this week of most of the remaining lawsuits against TWIA arising out of Hurricane Ike, probably cuts in half the amount of cash TWIA would have immediately available to pay claims in the event of a storm this summer without having to rely on untested, legally questionable and potentially slow efforts at “post-event” borrowings.  The action leaves both the cash position and the long run finances of the troubled insurer in question.

My best guess is that without the Bond Anticipation Note (BAN), and including its Catastrophe Reserve Trust Fund (CRTF), TWIA probably has between $400 to $700 million in cash with which to pay claims.  That’s not much when your direct exposure is over $75 billion, your total exposure is over $80 billion and a Category 2 or 3 hit at a bad spot on the Texas coast could easily cause losses of over $2 billion. The Bond Anticipation Note would have doubled the amount of cash available to pay claims.

As it stands, and as set forth below, I now believe it is not unduly pessimistic to set the odds of a TWIA insolvency this summer at 10%. If we consider two summers until the next regular legislative session, this risk roughly doubles. Given the grave effects of a TWIA insolvency on the entire Texas economy, this is way, way too high a risk.

Cash position

To understand this, take a look a TWIA’s 2012 Annual Statement. TWIA ended 2013 with about $430 million in cash (Assets, line 5; column 1) and total admitted assets (including the cash) of about the same amount, $430 million. (Assets, line 28, column 3) It has agreed to pay about $135 million in cash to settle the bulk of the Ike lawsuits. How much that will reduce the $323 million in loss reserves (Liabilities, Surplus and Other Funds, line 1, column 1) is unclear.  Because lawsuits remain, it is unlikely to reduce those reserves down to zero.  It will, however, likely reduce TWIA’s cash position by the full $135 million in relatively short order, depending on the details of the settlement. That would leave TWIA with just $295 million in cash.

Of course, it’s a little more complicated.  I don’t have access to TWIA’s financial statements for the first quarter of 2013 or thereafter. TWIA has likely earned some cash since January 1, 2013. It has been earning and collecting premiums, although it has had to pay off about $50 million on a thunderstorm in Hitchcock.  So, let’s be generous and credit TWIA with about $120 million more in new cash. This brings a guesstimate of its cash levels back up to around $415 million.

The problem is that not all of this cash is available to pay policyholder claims.  Some of it will be used to pay for operations, for commissions, and for other matters, including the Ike claims not resolved earlier this week.  So, I would be surprised if someone were to audit TWIA today and found it had more than $400 million in cash available to pay claims before resort to the CRTF. I would not be surprised if the number actually came out in the $300 million range.  And both of these figures will be reduced by $100 million or so less if TWIA succeeds in its plan to purchase reinsurance.

So, without the hoped-for borrowings, TWIA might have had $300 million to pay claims out of operating funds and another $180 million out of its CRTF.  TWIA might have had a total of $500 million.  (If the settlement came out of the CRTF rather than operations, the total would stay the same).  If the BAN had been approved, at least in the short run before TWIA had to pay the loan back, TWIA might have had $1 billion.  Both sums are, of course, grossly inadequate to deal with the $80 plus billion in TWIA exposure. Nonetheless, $1 billion in cash would have left TWIA in a better short run position.

Long run finances

Perhaps the greater impact, however, of the BAN ban is on the ability of TWIA to sell post-event bonds following a storm.  We’ve been through this matter before on this blog, but it is worth repeating because it is so very important.  The short version is, however, that there is a significant risk that very little in post-event bonds will actually be able to be sold.  And, thus, TWIA may very well have less than $1 billion with which to pay claims even after borrowing.  I would not be surprised if it ended up with as little $700 million.  The probability of such losses occurring this summer would be about 7-9% if this were a normal hurricane season.  If, as climate experts agree, however, this proves to be a bad hurricane season the probability of TWIA going broke and unable to pay claims fully could rise to 10-14%.

Here’s the longer version.  I, by the way, am not alone in my alarm on this matter. TWIA itself raised the issue in its submission to the Texas legislature.  the Texas Public Finance Authority (TPFA) had trouble last year trying to help TWIA borrow. And several of the pieces of proposed legislation this session would have fixed this particular problem.  But all of these bills failed during the regular session. Governor Perry has thus far resisted calls that he add windstorm insurance reform to the agenda for a special legislative session.

if there is a storm that pierces the CRTF, TWIA will need to rely on post-event Class 1 bonds.  But, unless something has changed, per the Texas Public Finance Authority they won’t sell, at least not up to $1 billion authorized.  But if the Class 1’s don’t fully sell, then TWIA/TPFA is prohibited from selling the regular Class 2 bonds. (Section 2210.073). Instead, we go to the Class 2 Alternatives under section 2210.6136.  But if less than $500 million of Class 1 bonds have sold — which is likely to be the case —  the first $500 million of the  Class 2 bonds  are paid in the same problematic way as the Class 1 bonds (surcharges on TWIA policyholders).  (Section 2210.6136(b)(1)). And there is a serious question as to whether anyone will loan TWIA money on those terms. Why? Because as soon as substantial policy surcharges are issued on TWIA policies, some TWIA policyholders will either find other insurance, reduce the sizes of their policy, or simply choose to go bare.  This is particularly likely if a storm has impoverished many TWIA policyholders. And if enough TWIA policyholders reduce their premiums, the percent surcharge will need to go up to compensate in order to pay off the bonds.  But if the surcharge rate goes up, more TWIA policyholders will drop out.  And, we get into a death spiral.

But here’s the catch.  Under section 2210.6136(c), if TWIA/TPFA can’t sell every dollar of the $1 billion in Class 2 Alternatives, then TWIA/TPFA can not issue the class 3 bonds of $500 million.  The statute is crystal clear on this point.  And this means that TWIA has no Class 1 bonds, no Class 2 bonds, little or no Class 2 Alternative bonds and no Class 3 bonds.  The system has completely collapsed in a cascade of failures.  TWIA basically has no money beyond cash on hand, and the CRTF. That means policyholders will not be paid in full.  If the storm is bad enough, they won’t be paid even half of their legitimate claims.

Reinsurance — assuming that TWIA can get it — will not help a lot. The reinsurance will not kick in until losses exceed the “reinsurance attachment point.”  But the reinsurance attachment point is likely to be set on the false assumption that the post-event securities will succeed.  So, for losses less than the reinsurance attachment point, the reinsurance won’t pay at all.  TWIA will be just as bankrupt as if it did not have reinsurance at all.  Actually, it will be more bankrupt because  it will have paid $100 million in premiums.  And even if the storm is so bad that the reinsurance kicks in, there is still a gap between the top of the CRTF plus any post-event bonds and the reinsurance attachment point.  So, TWIA won’t have enough money to pay claims fully.

Why would Commissioner Kitzman do such a thing?

I’m not privy to her reasoning or all the facts, but there are concerns we have outlined before about pre-event borrowing such as a Bond Anticipation Note.  The problem with loans is that you have to pay them back — and at interest.  Thus, in the long run, particularly if interest rates rise or if TWIA is deemed high risk and thus charged high rates even now, borrowing perpetuates your insufficient capitalization.  Whatever the benefits in the short run — and there may have been many here that incoming Commissioner Julia Rathgeber will want to examine — it is not the ideal long run solution for insurance risk. It may well be that Commissioner Kitzman refused as her final act to be complicit in the bandaiding of TWIA in the hopes that a sufficiently obvious problem would spur the Governor to call a special session and the legislature to develop a sustainable fix.  If so, let us hope that gamble proves correct.


Just what we don’t need: an especially active hurricane season

Evidently the risk of blue tarps on people’s houses for long periods of time as they wait to collect from an insolvent Texas Windstorm Insurance Association following a moderate or severe tropical cyclone this summer or next has not been enough to motivate the Texas legislature to do anything serious about the problem during its regular session. Here’s one more piece of evidence, the new 2013 Atlantic Hurricane Forecast, suggesting that this failure to act leaves the Texas coast in grave danger and the rest of Texas at serious risk.  It strengthens calls for Governor Rick Perry to add windstorm reform to the agenda for a special session.

The National Oceanic and Atmospheric Administration’s Climate Prediction Center today joined other scientists (here and here) in predicting a worse than average hurricane season. They write:

This combination of climate factors historically produces above-normal Atlantic hurricane seasons. The 2013 hurricane season could see activity comparable to some of the very active seasons since 1995. Based on the current and expected conditions, combined with model forecasts, we estimate a 70% probability for each of the following ranges of activity during 2013:

  • 13-20 Named Storms
  • 7-11 Hurricanes
  • 3-6 Major Hurricanes
  • Accumulated Cyclone Energy (ACE) range of 120%-205%

The seasonal activity is expected to fall within these ranges in 70% of seasons with similar climate conditions and uncertainties to those expected this year. These ranges do not represent the total possible ranges of activity seen in past similar years.

Note that the expected ranges are centered well above the official NHC 1981-2010 seasonal averages of 12 named storms, 6 hurricanes, and 3 major hurricanes.

If, by the way, we assume a 50% higher number of tropical storms this year than average, the risk of TWIA becoming insolvent if it has $1 billion (after limited success selling post-event bonds) with which to pay claims goes from roughly 7% this year to roughly 10% this year. If we assume, as the most optimistic people do, that TWIA might have access to $3.5 billion (after successful post-event bond sales and some reinsurance), the risk of insolvency goes from roughly 2% to roughly 3%. All of these numbers are too high for comfort.

The only comfort one can obtain from this forecast of the 2013 Atlantic hurricane season is that it could, of course, be wrong.  Hurricane predictions, particular this far in advance, are not the most reliable forecasts out there.  NOAA has taken this into account, however, by saying that there is only a 70% chance that its forecast will be right.  Still, a 70% (or, technically greater) chance that we will have a greater than average number of Atlantic hurricanes this year, when TWIA, even with the most cheery assumptions, is still underfunded, should make one’s stomach churn as if in sympathy with the future waters of the Gulf.

My thanks to Houston Chronicle blogger extraordinaire, SciGuy Eric Berger, for bringing this news to my attention.



Colorado State publishes 2013 tropical storm risk for Texas

Colorado State University (CSU) has published one of its predictions for tropical cyclones along the Gulf Coast. Historically, the April predictions of this group aren’t very good. The statistical correlation between April prediction and summer reality is only 0.09. This result is  only a little better than chance.  Still, CSU’s work does provide some information on the risk facing our Texas coast. Its method is better, for example, than basing the current year on the number of named storms in the previous 2 years. By June 30, the predictions of this group for the remainder of hurricane season get a lot better. Unfortunately, the prediction for this hurricane season is that it will be worse than normal. That’s particularly troubling given that the leading insurer of windstorm risk in Texas has been found insolvent by its auditors and there is no legislative proposal currently gaining a lot of traction that addresses 2013 risk.

The interactive widget below shows the probability of named storms, hurricanes and intense hurricanes for each of the Texas counties for 2013 based on the April predictions of CSU. You can get the data underlying the widget here. For techies and statisticians, I’m assuming, along with CSU and many others, that the distribution of tropical cyclone landfall comes from the “Poisson family.”

[WolframCDF source=”” CDFwidth=”600″ CDFheight=”540″ altimage=””]

A hat top to the Houston Chronicle’s Eric Berger for bringing this forecast to my attention.

Eight Reflections on Sandy

1. Turns out a Category 1 hurricane can, under the right circumstances, do a heck of a lot of damage. So can tropical storm force winds.

2. Up until this week, New York City and New Jersey could (and, for all I know, did) make the sort of arguments that would have kept insurance reserves available to pay for Sandy far too low.  The argument would have been: no tropical cyclone has made landfall in New Jersey as a hurricane in 109 years, therefore the risk is low.  Acceptance of those arguments can result in greater insurer insolvency and policyholder shortfalls. Turns out the theoretical flaws with that kind of argumentation, which we hear from some Texas coastal politicians all the time, can come back to bite you in the rear.

Continue reading