Forty years ago the economists Finn Kydland and Ed Prescott wrote a paper (for which they later won the Nobel Prize) observing that there are situations when the government makes a promise it can’t be expected to keep, and that policy inevitably reflects that reality.

For instance, they observed, the best policy regarding floods would be to simply disallow all construction in a floodplain, or at least forswear any public financial assistance to people who do build there and subsequently have property damaged by a flood.

However, such a promise cannot be kept: the political pressures to help victims will be too great to resist. What’s more, people recognize that reality, so they fail to heed the government warnings and build in the floodplains anyway, which in turn forces the government’s hand by putting it in a situation where it finds itself obligated to provide assistance after a flood.

And given that we have development along floodplains across this country, from Miami Beach to the Illinois river in Mossville, it has become inevitable that we will repair and rebuild those communities when they’re damaged by a flood, as well as spend money to prevent flooding from occurring in the first place.

That reality is a main reason that we have a National Flood Insurance Program, and why it is $24 billion in the hole — a number that’s undoubtedly about to rise precipitously. The government did not intervene in the flood insurance market because the markets failed: the property and casualty insurers priced its insurance above what people wanted to pay for it, and the government stepped in to sell it at a price that homeowners would pay, which subsequently entailed it taking billions of dollars of losses. These same insurers are anxious to sell flood insurance policies today, but homeowners would — naturally — prefer to get a subsidized price from the government.

The time-consistency problem that makes flood insurance so difficult also applies to the soon-to-be expired DACA, the Executive Order issued by President Obama that allows the adult children of illegal aliens who were brought into this country by their parents to legally remain in the U.S. to attend college and, later, obtain a work permit to seek employment. In this case, however, the government is employing a different rational, and one that happens to be economically costly.

Our current immigration situation, whereby we let about 1 million immigrants enter the country legally each year and a similar number come illegally, is not what anyone would say is the right immigration strategy. Illegal immigrants often do obtain employment but typically in jobs that do not pay well, and they often do not pay all taxes. The tenuous nature of the jobs made available to illegal immigrants means they don’t typically invest in new skills or training.

The optimal immigration strategy would be to have no illegal immigrants. Jeff Sessions would achieve that by making massive new investments in border enforcement until it is nearly impossible to enter illegally. My way would be to let every immigrant who arrives in the U.S. receive permission to work.

Since we eschew either extreme we are left with an estimated 13 million illegal immigrants, including the 800,000 DACA recipients. Their presence begs a question: what is the right immigration strategy given that we have these educated, foreign-born illegal immigrants already in our country?

If we follow the rationale we apply to flood insurance, we should recognize that we’ll likely never have an optimal immigration strategy, and that these DACA recipients will not leave the country when they become illegal. Hence, we should choose a policy that is consistent with our goals, given the current reality. For the so-called Dreamers, who have either completed some form of post-high-school education or are currently receiving such education, that would be to allow them to keep their work permit and remain in the labor force.

What the federal government is doing with DACA is akin to it telling Miami Beach that it had warned those people that it was susceptible to floods and that they’re on their own.

Miami Beach isn’t going anywhere, and the federal government would rebuild it if the next hurricane wiped it out because at this point, that’s the best outcome for the economy. The DACA recipients aren’t going anywhere either, and the best outcome for the economy would be for the government to recognize this reality and allow them to legally continue to study and obtain employment.

Earlier this year Logan Albright and I estimated that the repeal of DACA would cost the economy over $200 billion in the next decade and the U.S. government roughly $60 billion via reduced incomes and tax revenues from DACA recipients. That the president repealed it just as three hurricanes hit the U.S. and inflicted roughly the same amount of damage to the U.S. economy is an unfortunate coincidence that only further illustrates the folly of making rules for a world that does not exist.