A man dodges a wake in floodwaters during the aftermath of Hurricane Harvey Aug. 27 in Houston. (Brendan Smialowski/AFP/Getty Images)

There are three ways in which climate change will continue to make storms like Hurricane Harvey more disastrous, in the way that a basement full of gasoline cans will make a house fire worse.

The first is that scientists expect that climate change will increase the number of heavy precipitation events. Rainstorms will be more severe. More rain will fall, flooding places like Houston.

The second is that ocean temperatures are warmer, and warmer water helps build the strength of hurricanes.

The third is that warmer water also takes up more volume — meaning that oceans are rising even before you consider the increase in ocean levels that follows from the melting of onshore glaciers. Higher ocean levels make flooding more common and mean higher surges when storms push water inland.

This last point, about the increased likelihood of flooding, is not a secret and it is not a theory. Flooding on the coasts is more common now than it was in the past and sea levels in those places are higher than they used to be.

Flooding is one of the most common causes of billion-dollar events — natural disasters that do $1 billion or more in damage. Flooding has been an expensive problem long enough that, decades ago, private insurers generally stopped offering flood insurance for homes. So the government stepped up, creating the National Flood Insurance Program to provide insurance to homeowners in areas prone to flooding. Private insurers cover the homes, but the bills are ultimately paid by the federal government. People who couldn’t be insured for flood damage before can under the NFIP, a program managed by Federal Emergency Management Agency.

But there’s a problem. As more people have moved into flood-prone areas and climate change has increased the likelihood of flooding, the NFIP has been overwhelmed by severe weather events.

From 1978 to 2004, the program generally took in more in payments than it paid out in losses.

Then, in 2005, Katrina hit. Losses for which NFIP was responsible that year neared $18 billion — an amount equal to all of the premiums paid from 1991 through 2005.

Then, in 2012, Sandy hit. Losses neared $10 billion. The cost of losses in the years of Katrina and Sandy combined equals nearly half of all of the premiums paid from 1978 to 2016.

Then, in 2016, there was a slew of major floods, four of which were billion-dollar events. NFIP was overwhelmed.

Bolstering the NFIP became a big political fight in the wake of Hurricane Sandy. A small part of it was about the location of the storm, battering Democratic northeastern states. (Sixty-seven Republicans, including current House Speaker Paul D. Ryan (R-Wis.) voted against an NFIP funding bill in January 2013.) Most of it, though, was real concern about the failure of the program to pay its own bills. The NFIP is $24.6 billion in debt.

That problem, though, comes down to politics.

The Government Accountability Office describes the problem with the NFIP as follows: “Since the program offers rates that do not fully reflect the risk of flooding, NFIP’s overall rate-setting structure was not designed to be actuarially sound in the aggregate, nor was it intended to generate sufficient funds to fully cover all losses.” In other words, the NFIP charges less than it needs to.

There are two reasons for that.

The first is that FEMA dragged its feet on updating its maps to reflect the current threat posed from flooding. For years, the agency relied on decades-old maps created before the advent of new technological tools that would allow for better measurement. As such, people in areas at risk of significant flooding paid premiums based on an out-of-date estimate of how much risk of flooding existed. That is: They paid less than they should have. (The flood map for New Orleans had been last updated 20 years before Katrina.) Some people in areas that were likely to flood weren’t identified as living in flood-prone regions at all, so they didn’t have any insurance at all.

When FEMA started updating those maps several years ago, the reaction was predictable (and was why they weren’t eager to do the updates in the first place). Premiums for many people went up, and they didn’t like that. That includes a lot of people who live on expensive properties near the ocean, who had an outsized ability to influence the politics. The NFIP is meant to be a backstop for people at risk of flooding, but because it’s a government program it’s also subject to the vagaries of political influence. And there’s not much that is less politically popular than having the government tell you that you need to pay more money when nothing you see around you suggests anything has changed.

The NFIP is already baking in future problems along these lines. For the most part, the new FEMA flood maps don’t account for the effects of future changes to the climate. (New York City, near the Sandy epicenter, requested that FEMA’s new maps include future projections.)

“The floods of 20 years ago are not as bad as the floods are going to be 20 years from now,” Michael Gerrard, director of Columbia Law School’s Sabin Center for Climate Change Law told PBS’s “Frontline” in 2016. “But [the maps] only look at historic experience.”

The flood insurance program is due to expire this year. A bill to reauthorize the program was introduced in the Senate last month; Harvey’s arrival — and its landing in a Republican state with a lot of political clout — is expected by some to prompt approval.

Last month, President Trump rescinded an executive order signed by his predecessor that those receiving federal funds for building consider the future effects of flooding in their construction plans.

“Taxpayers have been made to shell out hundreds of billions of dollars in disaster-related spending over the past decade, including more than $136 billion for just the two years from 2011 to 2013,” an analyst from a conservative think tank said in a statement about the decision. “By contrast, evidence shows that every $1 spent on disaster mitigation can save $4 in post-disaster recovery and rebuilding costs.”

As James Surowiecki noted for the New Yorker in the wake of Sandy, though, politics makes that bad economic trade-off preferable. One study, he wrote, “found that voters reward politicians for spending money on post-disaster cleanup, but not for investing in disaster prevention, and it’s only natural that politicians respond to this incentive.”

They will probably continue to do so.