A person walks through the flooded streets of a trailer park in Key Largo, Fla., in the aftermath of Hurricane Irma on Sept. 11. (Wilfredo Lee/AP)

With Hurricane Irma pounding Florida with historic intensity and Hurricane Harvey leaving unprecedented levels of flooding in parts of Texas, Congress is back in session — and, coincidentally, looking at reauthorizing the National Flood Insurance Program (NFIP). NFIP runs on five-year authorizations from Congress; the current one expires September 30.

Here are the five things you need to know about this perennially problematic policy.

1. What is the National Flood Insurance Program (NFIP)?

NFIP is a program administered by the federal government, now housed in the Federal Emergency Management Agency (FEMA). According to FEMA, the NFIP:

… aims to reduce the impact of flooding on private and public structures. It does so by providing affordable insurance to property owners and by encouraging communities to adopt and enforce flood plain management regulations.

2. Most insurance is provided by private firms. Why is flood insurance different?

In 1968, Hurricane Betsy struck the Gulf Coast, killing 76 people and leaving behind more than $1.5 billion in property damage, which would be more than $10 billion in 2016 dollars. Virtually none of that was covered by flood insurance — leaving the historic storm to be paid for by relief legislation.

Why? Most private insurers had left the flood insurance market after the historic Mississippi River Flood of 1927. Insurers generally see catastrophe insurance as a bad bet: Premiums that reflect actuarial risk are prohibitively expensive, because the risk pool is concentrated and not large enough to effectively diffuse costs. In an important sense, this is precisely the point: prohibitively expensive insurance (or the inability to get insurance at all) should have encouraged people not to move to, build or invest in flood-prone areas.

But at the time, a number of policymakers and observers argued that the lack of private-sector flood insurance was a market failure — and should be corrected by the government. As a result, Congress enacted the National Flood Insurance Act of 1968, which was intended to make affordable flood insurance available to those who need it. Because of the actuarial problems just mentioned, however, the only way to make insurance affordable was to subsidize premiums.

3. Has NFIP been successful at achieving its goals?

It has not. NFIP has a unique policy history of failing to deliver on its stated goals. It routinely runs large deficits because few of those who need it buy it; those who do pay steeply subsidized premiums.

Worse, NFIP actually creates incentives for people to move to and invest in flood-prone areas, increasing the number of properties likely to be flooded. A huge fraction of the program’s payouts have gone to the same properties over and over again; NFIP calls these “Repetitive” and “Severe Repetitive Loss Properties.” That’s precisely because the policy actually subsidizes rebuilding — and offers no incentive to relocate out of flood zones.

So why does NFIP subsidize premiums, if that causes deficits? The idea is that offering discounted rates and making up the difference with tax dollars will encourage the customers who need it to actually buy it. Unfortunately, NFIP fails at even this goal.

That’s in part because most people do not have a thorough understanding of their property’s flood risk. For instance, many people think that if they live behind levees, they are safe from flooding and thus don’t need insurance. The flood risk maps that NFIP uses to calculate flood risk are badly out of date.

What’s more, flood policy experts have begun to realize many people forego insurance in part because of Congress’s implicit promise that it will deliver humanitarian relief after any major disaster.

As a result, a very small percentage of those at risk actually have flood insurance. Take Harris County, Tex., where Houston is situated — only 80 feet above sea level, in a hurricane zone — where only about 15 percent of its structures are covered by flood insurance. Nationwide, market penetration hovers around 50 percent of properties in 100-year flood plains. Even in very high-risk areas, such as New Orleans and South Florida, penetration rates typically range from just over 50 percent down into single digits. Low participation rates means the risk pool remains too small to absorb losses from major flood events.

4. Why doesn’t Congress fix NFIP?

Congress has tried to improve NFIP from time to time over the last 50 years. Major floods almost always lead to small adjustments, usually aimed at getting more at-risk property owners to buy it.

In a recent article, I argue that Hurricane Katrina made NFIP’s biggest problems too big to ignore, leading Congress to pass a far-reaching reform called the Biggert-Waters Act of 2012. Biggert-Waters phased out subsidies for all properties and eliminated them for repetitive loss properties, took steps to educate the public about flood risk, all new policies would reflect actuarial risk, provided funds to update flood risk maps and more.

But only 14 months later, Congress passed the Homeowner Flood Insurance Affordability Act of 2014, which undid most of the Biggert-Waters reforms by reviving subsidies, increasing deductibles and the like.

Why? The passage of the Biggert-Waters Act made flood policy highly salient: The news media gave a fair amount of attention to the fact that the end of subsidies would make insurance much more expensive for individual property owners. At-risk owners — many from relatively affluent coastal areas — lobbied Congress to keep flood insurance “affordable.”

In Congress, electoral incentives matter a lot. As political scientist R. Douglas Arnold famously put it, the “first law of congressional behavior” is “never impose costs on one’s constituents that might be traced to one’s own individual actions” — and which might jeopardize chances for reelection. When few people were paying attention to flood insurance, Congress could alter the policy relatively freely. But once the policy changed and voters saw what would happen to their costs, members changed their tune. Those voters are dispersed more widely through the country than you might imagine, along every river and big stream in the country, in addition to the coasts.

In short, Congress reversed course because enough of them were unwilling to make voters unhappy by forcing them to pay for their real flood risks.

5. So what now?

Unfortunately, there’s no easy answer. Subsidized premiums and post hoc disaster aid shift flood risk from individual owners to taxpayers at large and encourage development in flood-prone areas. After disasters, people rebuild right back in these areas, making the next major loss inevitable.

NFIP could be designed to discourage people and businesses from living and building in flood zones — and to help with the costs for those who are flooded nevertheless.

Many of the reforms in Biggert-Waters were aimed at precisely that goal. But many people want the NFIP to make flood insurance “affordable.” And so Congress will almost certainly continue kicking the NFIP can down the road.

Logan Strother is CKF postdoctoral fellow and a visiting scholar in the Program in Law and Public Affairs at Princeton University.