Hurricane Harvey Analysis

This story is going viral, having been shared and up-voted hundreds of times on Reddit and elsewhere: For years, engineers have warned that Houston was a flood disaster in the making. Why didn’t somebody do something?

Who is ‘somebody’? The flood fairy?

The most obvious answer is that insurance is cheaper than re-engineering everything. That also explains why the economic impact of the storm is so small despite all the headlines. The insurance companies eat the loss and respond by raising rates in accordance to their risk models being revised upwards.

It’s almost always easier to have a disaster destroy the property and have the insurance company pay up, than try to sell it on the market. One such scheme involves buying homes in areas at under the mean appraisal value and then pocketing the difference between the mean and paid value (buying a cheap home in a good neighbourhood). For example, if the mean value of homes in a neighborhood is $500,000, some will probably be worth $350,000 for the same square-footage; others $750,000. The person buys the $350,000 home, but gets it appraised for $500,000, which is the value of the insurance policy. If disaster strikes, he pockets the $150,000 difference. This can also be used for taking out loans against the inflated value of the home.

This creates perverse incentives. From The Weeds: the problem with the national flood insurance program:

SARAH: The Pew Foundation did a study on the flood program and they looked at these houses that have had multiple floods, and these are about 1 percent of houses that have flood insurance, but they make up 10 percent of claims.


One example they had, which was just astounding, was there’s this one $69,000 home in Mississippi that has flooded 34 times in 32 years and has received $663,000 in payments. If you have that kind of backstop, why not build a $69,000 house in a flood zone?

Despite all the doom and gloom and an estimated $50-$100 billion of damage, the S&P 500 posted its strongest week of 2017, gaining over 1%.

How come “100-year storms” happen every decade, it seems? Global warming? Fat tails? Both are wrong. The correct answer is that we’re simply better at measuring storms; we have more data to work with, and more accurate data. 100 years ago, record keeping was very poor, but also there was much less urban development. A storm that hits unimproved land is likely to be ignored. That also explains why it seems like there is more civil unrest now than in the past, when in reality we’re better at measuring and reporting unrest.