A quick and dirty estimate of Hurricane Harvey’s recovery cost

A quick and dirty estimate of Hurricane Harvey's recovery costI was born and raised in Houston. Besides 4 years as an undergraduate student in Austin, Texas, and 7 years as a graduate student in Pasadena, California, I’ve lived in Houston my whole life.

As a result, my personal focus has been on Hurricane Harvey nearly exclusively for a week now. On Thursday, I started to¬†see some estimates¬†trickle in of the storm’s recovery cost. Needless to say, they caught my attention.

Before I dug into those estimates, I thought it would be useful for me to make my own guess of Harvey’s economic damage. I’m no expert when it comes to disaster recovery. But I was curious what kind of guess I could make, using only some common sense assumptions.

To avoid suspense, here’s my estimate: $62 billion. It’s certainly wrong, like almost any prediction is. Still, I wanted to make an estimate now, explain what I was thinking, and then compare with the actual number at some point, likely years, in the future.

I broke the damages into four categories:

  1. Homes
  2. Infrastructure
  3. Automobiles
  4. Lost economic output

I’ll walk through my assumptions for each category, and conclude with some thoughts about what Harvey means for the greater Houston economy going forward.


Again, my damage estimate is quick and dirty. Reasonable people can absolutely disagree with these assumptions. I just went with my gut, and made what I think are reasonable choices.

We can start with homes. I’ll assume the average household has 2.5 people in it. I’m assuming there are more families of 3+ people than there are of single people, which gets me to 2.5 as an average.

I’m taking the greater Houston population to be 6.5 million. That’s a number Mayor Sylvester Turner used at several of his news conferences. With 2.5 people per household, greater Houston would have 2.6 million households.

I’ll assume 5% of homes were effectively destroyed by Hurricane Harvey. That’s just a wild guess. I can’t really explain it. It seems like a reasonable number, when I think of 1 out of every 20 homes suffering that kind of damage. A larger fraction of homes were impacted. Not all of them were destroyed. But if we try to squeeze all the home damage into an effective fraction of homes that were destroyed, I’ll assume 5% is reasonable.

Further, I’ll assume the average replacement cost of a home in greater Houston is $200,000. I chose that number first, and then learned via Google that the median home price in Houston is $289,000. So take out some costs for land, add in some costs for demolition, and $200,000 as an average home replacement cost seems reasonable to me.

If we start with 2.6 million households, multiply by 5% to isolate the homes effectively destroyed in the storm, and then multiply by $200,000 per home as a replacement cost, we get a total home repair/replacement cost of $26 billion.

Now let’s move onto infrastructure.


Of the four categories, I’m most lost here. How much does it cost to repair a broken storm sewer, or a malfunctioning water treatment plant, or a damaged school? I don’t know.

So I’ll make a really coarse assumption. I’ll back up and think of infrastructure on the whole. How much economic value should be assigned to each greater Houston resident for infrastructure? What is each person’s share of the value of roads, bridges, utilities, schools, etc.?

The only comparison I can quickly come up with is to a home. I have assumed the average replacement cost for a home is $200,000. If an average household has 2.5 people, then per capita home value is $80,000. Note that I’m using replacement cost as a measure of value, like companies do when reporting asset values on their balance sheets. I’ll simply assume infrastructure repair/replacement costs per person are also $80,000.

Why? Because I don’t have a great reason to deviate from that number. As a community, we share a lot of infrastructure here in Houston. This infrastructure, like our homes, is a fixed asset with an expected lifespan of decades. And it seems reasonable to believe that government invests the same order of magnitude in fixed assets that we invest ourselves, privately.

Could government invest twice as much as we do via our homes? Sure. Could government invest half as much? That seems reasonable too. But because I’m so out of my depth with infrastructure cost, I’ll simply assume the same dollar amount per person that I did with homes: $80,000.

If we similarly assume 5% of infrastructure was effectively destroyed, then we get the same damage estimate that we had for homes: $26 billion.


I’ve made some wild, high-level assumptions to this point, so why stop now? I’ll assume Houston has a uniform population, with people evenly distributed between 0 and 80 years old. In other words, 1/80th of the population is 0-1 years old, 1/80th of the population is 1-2 years old, and so on.

That means 80% of the population is greater than 16 years old, which is the legal driving age. Further, I’ll assume 80% of these driving-eligible adults have automobiles. With 6.5 million people, greater Houston would then have 4.2 million cars.

Like with homes, let’s assume 5% of cars were effectively destroyed. It’s probably true that fewer than 5% of cars were actually destroyed. But a larger fraction had less than total damage. So if we normalize it to a fraction that was effectively destroyed, I’ll go with 5%.

I’ll assume the average replacement cost for a destroyed car is $20,000. With 4.2 million cars, 5% of which were destroyed, we have automobile repair/replacement costs of $4 billion, rounded to the nearest billion.

Lost economic output

Finally, we have lost economic output. This is the economic activity that would have occurred, had the storm never hit. When people hunker down in their homes, they don’t eat at restaurants. They don’t watch movies in a theater. They don’t go to sporting events, or shop at the mall. Economic activity falls dramatically.

Here I used Google. The greater Houston GDP is $420 billion. The storm hit Houston hard over 5 days, from Saturday, August 26, through Wednesday, August 30. While the rains have now passed, there is still an enormous amount of recovery effort underway. Not all businesses have reopened. Those businesses that are open often don’t have a full inventory.

I’m going to have to make a guess here. I’ll say that Houston lost 5 full days of economic output because of the storm. No, not all businesses were closed the entire duration of the storm. But not all of them opened up immediately thereafter. And not all businesses that were open could offer a full complement of products and services. So I’ll split the difference and assume Houston lost all its economic output for the actual duration of the storm, or a full 5 days.

With a GDP of $420 million, 5 days of economic output is $6 billion, again rounding to the nearest billion. That’s my estimate of Houston’s lost economic output due to Hurricane Harvey.

Total damage estimate for Hurricane Harvey

Now I can add up the estimate for each category of damage:

  • Homes: $26 billion
  • Infrastructure: $26 billion
  • Automobiles: $4 billion
  • Lost economic output: $6 billion

That gives me a total damage estimate of $62 billion.

How reasonable is that number? According to Wikipedia, Hurricane Katrina cost $108 billion in 2005 US dollars. That would be $139 billion in today’s dollars.

So my quick and dirty estimate says that Hurricane Harvey would cost about 45% of what Hurricane Katrina did.

To me, that seems low. Houston is a much larger city than New Orleans. The greater metropolitan area of New Orleans has about 1.2 million people, or less than 20% of Houston’s greater metropolitan area. Hurricane Harvey broke countless rainfall records across a wide swath of land. Naively, I’d expect Harvey’s recovery cost to be at the same level of Katrina’s cost, not over 50% lower.

Still, no one knows at this point. We’ll find out in time. Knowing nothing about city planning or disaster recovery, I feel good about roughing out a $62 billion guess that at least seems to be in the right ballpark.

Won’t recovery spending be an economic boost for Houston?

Yes, but only in a limited away. Henry Hazlitt discusses this scenario in detail in his book Economics in One Lesson. He calls it the broken window fallacy.

When a window breaks, the window maker gets more business. True. The important point to remember, though, is that the money we spend repairing a broken window, cannot be spent in other ways. The added business for the window maker is visible. The lost business for the restauranteur, or the movie theater operator, or the electronics retailer, is invisible.

That’s the problem with this kind of economic spending. Yes, Houston residents will spend money on new floors, new furniture, new cars, etc. But this money didn’t materialize out of thin air. It’s coming from savings. It’s being pulled forward from the future.

Money that people would have spent hiring new employees, or going to Astros or Texans games, or supplementing their wardrobes, now gets spent to repair or replace stuff they already owned. Hurricane Harvey destroyed wealth. Even though we’ll see a rush of spending during the recovery, we’re collectively spending to build back the wealth that the storm took away from us.

And all else being equal, we should then expect to see less spending in the future. Houston has to sacrifice some of its future economic activity to recover from Hurricane Harvey today. Now, maybe we make investments that increase future productivity. Maybe we invest in our infrastructure in ways that create new jobs, or help existing municipal employees do more with the resources they have. In that case, we could make investments that will create new wealth. The storm may awaken us to new investment possibilities that otherwise would have gone unnoticed.

When it’s all said and done, though, destroying wealth is never good. If there are useful investments to make, we should have made those investments even in the absence of a storm. While we will absolutely recover, and grow far beyond where we were last week, we can’t pretend that Harvey was anything but an enormous economic setback.

Why did I write this post?

I wrote this post for two reasons.

First, I’m totally absorbed in all things Hurricane Harvey. I love Houston. I’ll forever be attached to it. This post was an opportunity for me to put to paper some of the thoughts that have floated through my mind for several days now.

Second, as scientists and engineers, we need to know how to dance with uncertainty. I know next to nothing about storm recovery. I’ve never thought seriously about these kinds of logistics. But with a little bit of effort, and some common sense assumptions, I can estimate damages within, probably, a factor of 2 or 3 of their actual totals. Not bad for having zero expertise in this area.

And it’s not just saying, Hurricane Katrina recovery cost $X, so I expect Hurricane Harvery recovery to cost $Y. I have a data-driven foundation for my estimate. My assumptions may be wildly off, but I have a sense of how to defend them. I think that kind of mental exercise is super helpful. I find myself in these kinds of situations at work all the time, where I have to quickly sketch out rough answers to unfamiliar questions. Hurricane Harvey recovery cost estimates are just one more example.

If you’re in Houston, or know anyone in Houston, I’m sending you my best wishes. If not, I hope you’re in a place that never sees the kind of damage Houston, and the greater southeast Texas region, just saw. It’s traumatic. Countless families have suffered unspeakable loss. My family and I were part of the lucky group that made it through mostly unscathed.

Here’s to a safe, healthy, and happy rest of 2017.

Leave a Reply

Your email address will not be published. Required fields are marked *