On Friday morning, President Trump was expected to sign an executive order declaring a state of national emergency on the border with Mexico, potentially allowing him to allocate money to build a barrier that Congress was loath to fund. It’s Trump’s last-ditch effort to deliver on a key campaign promise, but instead of an extensive 30-foot concrete wall paid for by Mexico, under this plan the United States will get 30-foot steel-slat barriers in certain areas paid for by dipping into existing funding buckets within the government.

By now, the political and legal challenges of this action are well-established. A president who tries to declare an emergency to get around a rejection from Congress sets a precedent by which any president might declare other emergencies to move money around for any campaign pledge. This, it’s probably unnecessary to point out, is not really what the Founding Fathers had in mind.

Often overlooked in that debate, though, is how odd this solution is.

On the campaign trail, Trump could generally wave at the border and declare that a wall was needed. Now, though, he’s arguing that there’s an urgent crisis that demands action. That argument is tough to defend in light of the data on what’s happening on the border — and that this urgent crisis can be fixed only by a project that will span months to years.

The crisis

Trump has pointed to the surge in apprehensions on the border with Mexico, particularly of families seeking entry into the United States, as evidence of a looming crisis. It’s true that there was a surge in apprehensions — but to a little over the levels we saw in 2016. The number of apprehensions plunged early in Trump’s administration, so the recent surge looks significant mostly because it’s being compared with the aberration of that brief decrease.

Trump first floated the idea of taking executive action in November, when the number of apprehensions was at a peak. Since then, it has declined two months in a row. (This decline isn’t uncommon for the season.)

The scale of that graph is also deceptive. If we compare it with the peak in apprehensions — March 2000 — we get a better sense for how this recent surge compares with apprehensions historically.

The number of apprehensions (and, therefore, illegal border crossings) declined sharply after the early 2000s. Why? In part, certainly, because of an effort to build walls and barriers that now exist. In 2006, there was a new law enacted by President George W. Bush that added hundreds of miles of new barriers.

All the barriers are in red below.

Monthly apprehensions, the vast majority of which are on the border with Mexico, were already dropping at that point but fell further after the enactment of the new barrier construction.

This is part of the question: Are more barriers needed beyond what exists? The debate isn’t over whether, say, your house needs any walls. It’s whether more walls are necessary.

We’ll come back to that 2007 construction in a moment.

The political fight — and the courts

So let’s assume that there is an urgent crisis on the border that needs to be addressed with a wall, right this minute.

The first delay to that plan is that Trump’s declaration of a national emergency is not the final word on the subject. As The Washington Post’s Greg Sargent has reported, Congress can (and probably will) immediately move to block the action. According to the Brennan Center for Justice’s Elizabeth Goitein, the Democratic House can quickly pass a bill challenging the declaration that will then need to be acted upon by the Senate within 36 days. Unless the Senate acts quickly and in support of Trump’s move, that could immediately add a month to the time frame.

Separately, it’s expected that legal challenges to Trump’s action will be filed which could (and likely would) eventually wind up at the Supreme Court. There’s precedent for this: In 1952, President Harry S. Truman signed an executive order nationalizing the country’s steel mills, a move that wound up before the highest court.

How long could this fight take? Well, in 1952, it took about two-and-a-half months, from Truman’s signature in late March to the court’s decision in June.

The fight over eminent domain

That’s probably not the more problematic fight for the government. Much of the land on the border, especially in Texas, is privately held, as this map from the Congressional Research Service demonstrates. All of that light-yellow area is nonfederally held.

Some portion of that land will need to be obtained by the government, using the process of eminent domain. Under eminent domain, the government can seize private land, compensating the original landowners.

But this is not a simple process. After Bush signed that law in 2006 authorizing new barrier construction, a number of landowners sued to stop the seizure.

In 2017, reporter Noreen O’Donnell found that, of 300 eminent domain court fights that resulted from the enactment of that law, 90 were still pending a decade later. NPR profiled one such landowner, Eliosa Tamez, who fought the construction of a barrier on her property. That case, United States of America v. .26 Acres of Land, took seven years to resolve — in the government’s favor.

Since much of the new construction will be in places where there is no existing fence, expect many more such fights.

The actual construction

All this sets aside the other question: how long it takes to build a wall.

In a call with supporters Friday morning, White House adviser Stephen Miller reportedly said that the construction would happen at a speed that would “shock” people. That seems optimistic.

In California last year, workers replaced an existing barrier with new bollard fencing (the vertical steel slat barrier that Trump has recently embraced). That project took from February to October, replacing two miles of existing barrier.

But that was probably unlike many of the projects that will be necessary. It was replacement of an existing barrier, necessitating removal of what was there — but it also meant that the terrain was already prepared for barrier construction. In places where no barrier has been built, the terrain will often need to be prepared for construction. There’s already an ongoing legal fight over an effort to build a wall through a butterfly preserve on the border where the government began clearing trees and clearing brush to start construction.

And that’s probably a fairly simple preparation effort. Compare it to some alternatives.

For example, in 2009, the Associated Press reported last year, “the federal government spent about $16 million a mile on a 3.5-mile (5.5-kilometer) stretch in San Diego, using about 2 million tons of dirt to fill in a canyon known as Smuggler’s Gulch. The earthen dam was then topped with layers of fencing.”

The AP described another project: building a floating 16-foot-high fence through sandy terrain that was used to film scenes from “Return of the Jedi.”

Many of the places where barriers would be built under Trump’s plan are presumably not low-hanging fruit. In the places where barriers are needed and could be built rapidly, they likely have already been built.

“[C]ontracting, designing and building new wall systems complete with updated technology could take years,” the AP determined, “and past experience has shown such work can be complicated and costly.”

The total time frame

We began this article with a look at how apprehensions have tracked on the border to show the variability in those numbers. If Trump gets through the political and Supreme Court fights that would likely follow an emergency action (and take several months), eminent domain fights and tricky construction solutions would almost certainly mean years before Trump’s desired barriers are in place.

By which time, the number of people seeking entry into the United States might already have fallen dramatically. Or, perhaps by then, there will be a real, significant surge mirroring that seen in 2000.

Enough time will pass that, who knows, Trump could get an unquestioned crisis on the border.