President Donald Trump is playing with the idea of declaring a national emergency to build a border fence despite congressional opposition.

This would make Trump the second president in a row willing to cut Congress out of the legislative process if it doesn’t agree to his priorities on immigration, and is a very bad idea.

It would functionally be an end run around Congress’ power of the purse; create yet another precedent for “pen and phone” governance, which is not how our system is meant to work; and probably not achieve his substantive and political goals.

For all that, a move to build the wall unilaterally wouldn’t be nearly as brazen as the Obama-imposed amnesty for so-called Dreamers. The Obama administration simply wrote legislation on its own authority. Trump would at least be relying on congressional statute.

Although declaring a national emergency sounds frightening, there is ample — indeed absurdly extensive — precedent for it. The nation is awash in more than two dozen, little-noticed declared national emergencies. So we have room for a national emergency at the border.

The next step, which would presumably be reallocating military funds to building the border fence, is trickier. The administration can perhaps rely on statutes enabling the president to spend on military-related construction under his emergency powers.

Yet no one to this point has thought of the border fence as a military project. It has been built up over the years with civilian funds by civilian workers. Yes, the National Guard and, at the moment, the active-duty military have been deployed to the border, but in a logistical or backup role, and largely as symbolism.

The border crisis is not amenable to a military solution. Because the rules around how we handle Central American families are so latitudinarian, we could deploy the 82nd Airborne to the border, and migrants would simply surrender to the troops the way they currently do to Border Patrol agents and (correctly) expect to be waved into the country.

Even if technically legal, an emergency declaration and a move to unilateral spending would obviously be pretextual.

First, there’s the timing. Trump didn’t declare an emergency over the past two years, even though his administration was rightly already fixated on the influx of Central American families and minors.

Then there’s the fact that the natural response to a national emergency isn’t to say, “Let’s undertake a yearslong building project.”

Legalities aside, a declaration of national emergency won’t achieve what Trump wants, unless his goal is simply getting out from under the shutdown. That’s easy. He can say he’s going it alone under his emergency powers and agree to open the shuttered parts of the government, then fight it out in the courts.

But in terms of his substantive goal of building a fence, and his political goal of building enough of it to showcase in his 2020 re-election campaign, it would get him very little. A district court somewhere would immediately issue an injunction blocking the action.

Once the administration gets to the Supreme Court, it might have a chance to prevail if the court concludes that it shouldn’t second-guess the chief executive on questions related to national security. But when would such a decision get handed down? Sometime in 2020?

Assuming Trump wins at the high court, it would take time to get anything going on the ground. Progress on the fence probably wouldn’t be any more advanced by November 2020 than it would be otherwise, and perhaps less advanced than if the president simply got some inadequate compromise out of Congress.

In a different world, Congress as a body would be more exercised about a potential emergency declaration, but Congress is not a self-respecting institution.

It is still a mistake to try to take advantage of its laxity. Such a move would strain our system, and probably not even work.

Rich Lowry is the editor of the National Review.