The Border Wall Still Cannot Pay for Itself

Steven Camarota of the Center for Immigration Studies (CIS) responded to our criticism of his claim that the border wall will pay for itself. Most of Camarota’s comments confuse the multiple and different simulations that I published with David Bier. He only responds to a handful of our points and then spends most of his space attacking a section called "A Better Cost Estimate Should Include These Variables.” We did not incorporate any of the suggestions from that section into our corrected version of his fiscal analysis.

The only changes we made in our headline findings, relative to Camarota, were that we adjusted for the border crosser age of arrival in 2015, adjusted for the education level for 2015 border crossers, and used an actual cost estimate for the border wall. We also copied Camarota’s methods for our additional simulations but clearly stated the changes we made and why.

Camarota’s comments are in the block quotes, my responses are below.

“[D]espite the Cato blog post being titled 'The Border Wall Cannot Pay for Itself', their own cost estimates would simply mean that a border wall would have to stop 16 to 20 percent of those expected in the next decade to pay for itself (as opposed to 9 to 12 percent in my estimate).”

Camarota misread our response. The point of generating a new estimate from his assumptions was to demonstrate how flawed his report was by showing that small changes drastically change his results.  These are not our “own estimates,” but rather, they would have been his estimates if he had bothered to use more up-to-date and precise numbers.  Instead, Camarota pretends that our updates are a comprehensive fiscal cost estimate despite the fact that we have an entire section dedicated to explaining what sorts of other factors a good estimate would need to include.

“Cato argues for excluding state and local costs. Cato makes the argument that costs at the state and local level should not be counted, even though this information is available from the NAS study and I included it in my analysis. The only reason they give for not including these costs is that 'the federal government will actually be paying for the wall.' This is a very odd argument. The federal government often considers the costs of its policies at the state and local level, so why should building a wall be any different? These costs are real and have to be paid for by the same taxpayers who pay for the federal government.”

Camarota’s comment is perplexing. In the “Calculating the Fiscal Cost Section” of our blog, we used the average net present value flows for consolidated federal, state, and local governments in Table 8-12 of the NAS report. Camarota used that same table in his paper. We even averaged the net fiscal costs for all eight tables like Camarota did. The only exception is that we controlled for the age of the border crossers.  Camarota’s passage is actually criticizing one of the three additional simulations we ran later in the blog with different assumptions.  A person reading his criticism would inaccurately assume that we used a different table from the NAS than we really did.

“[T]he Cato authors argue that my analysis assumes that legal and illegal immigrants cost the same. For example, they say my analysis assumes that illegal immigrants will retire in the United States at the same rate as legal immigrants. In fact, my analysis very much takes this into account. Nowrasteh and Bier do mention the reduction in fiscal costs associated with illegal vs. legal immigrants that I included in my analysis, but they do not seem to understand the implications.”

Camarota’s statement is false. We never argue that legal and illegal immigrants impose the same fiscal costs.  Camarota does attempt to adjust downward the cost of border crossers, but he drew his estimate from a 2013 Heritage report that provides an estimate for a single year, not a lifetime. Thus, it does not take into account the emigration rate of each group.  The NAS report takes into account only the average emigration rate for all immigrants and not the emigration rate for illegal immigrants, meaning that this is Camarota’s assumption as well.

Furthermore, Camarota does not respond to our point that the Heritage report is methodologically incompatible with the NAS report.  Heritage’s report focuses on households headed by illegal immigrants while the NAS estimate measures individuals.  NAS also discounts a 75-year projection to the present value while the Heritage report does not discount a 50-year projection and, thus, reports a meaningless figure.  There is no sound justification for combining the figures from these two incompatible reports.

“Cato inflates cost of the wall. Cato argues the cost of the wall will be much higher than the $12 to $15 billion Senate Majority Leader Mitch McConnell (R-Ky.) has said Congress will spend, and the senator is certainly in a good position to know what Congress is likely to spend. A wall is not an entitlement program that grows on its own without Congress specifically allocating money. Further, Congress and the president will determine the structure, design, and length of the wall, as well as spending levels. In some sense "the wall" is whatever Congress and the president decide.”

David Bier and I decided to rely on actual DHS cost estimates that included maintenance and eminent domain.  Camarota relied on a quote by Senator Mitch McConnell.  Camarota confused what Senator Mitch McConnell said the Senate would spend on a border wall with what a complete border wall would actually cost. The two are not the same. Camarota then assumed that whatever Congress decides to spend would complete whatever project Congress sets out to complete.  Following Camarota’s line of thinking, if Congress wanted to build a complete border wall out of sunshine and puppy dogs then it will be so because they decreed it.

Camarota twists the words of a Senator to fit his own meaning while we take the average per-mile costs of construction and maintenance. The reader can decide which method produces a fairer cost estimate.   

"Cato recalculated the education level of illegal immigrants in order to reduce their costs, but they do not explain how they did this."

Camarota correctly guessed how we estimated the education of illegal border crossers.  This is Camarota's strongest point but it only accounts for less than half of the difference in our estimates.  Adjusting for the age of arrival accounts for most of the difference between our Camarota-inspired estimate of -$43,444 and Camarota's actual estimate of -$74,722 (more on this below).  Adjusting for age of arrival is important.       

Camarota did not respond to some important points:

• Cato’s adjustment for illegal immigrant age of entry. This minor adjustment accounts for slightly more than half of the difference between CIS and Cato and means that each border crosser produces a -$59.210 net fiscal impact. That means the border wall would have to deter about 739,092 border crossers without incurring additional costs to pay for the wall. That is approximately 44 percent of all estimated future border crossers over the next decade – more than twice as high as Camarota’s worst-case-scenario estimate.  
• Illegal immigrant border crossers are younger than what Camarota estimates if Border Patrol apprehensions data is meaningfully related. The surge in UAC since 2010 has lowered those ages even further, making the net fiscal impact more positive. Age is an important adjustment that Camarota should take account of.
• Border crossers are down in the first few months of the Trump administration. This might or might not continue depending on whether President Trump’s words turn into action as well as myriad economic factors. Recent research by Warren and Kerwin found that approximately 140,000 border-crossers entered annually from 2011-2013. If those lower numbers hold then the 1.7 million estimated border crossers over the next decade that Camarota relies upon may already be too high without factoring in President Trump’s other non-wall immigration enforcement actions.  In that case, the border wall will have to deter a much larger percentage of border crossers than he claims even without any changes to his model.

Camarota’s response to our blog is disappointing. He is correct that we are arguing that illegal immigrants have a smaller negative fiscal impact relative to legal immigrants when controlling for age and education. His estimates of new border crosser education levels could also be better than ours is. However, Camarota misses our biggest criticisms when he ignores actual border wall cost estimates and refuses to acknowledge that a border crosser’s age of arrival is important to determining his net fiscal impact.  There is not a good reason for relying on Senator McConnell’s quote for a border wall cost estimate while ignoring real cost projections and failing to adjust for the age of arrival of the border crossers.