From the article:
Officials said the raids targeted known criminals, but they also netted some immigrants without criminal records, an apparent departure from similar enforcement waves during the Obama administration. Last month, Trump substantially broadened the scope of who the Department of Homeland Security can target to include those with minor offenses or no convictions at all.
Trump has pledged to deport as many as 3 million undocumented immigrants with criminal records. …
If the Administration does undertake removal of 3 million undocumented immigrants (whether there are 3 million undocumented with criminal records is up for debate), there will be substantial fiscal costs. The American Action Forum, headed by CBO’s former head Doug Holtz-Eakin, estimated the 20 year cost of deporting all 11 million at between $419.6 to $619.4 billion. One time cost (so excluding the subsequent 20 year’s of expenditures) would range from $103.9 billion to $303.7 billion.
Assuming linearity, this means the 20 year recurrent costs will be only $114-$169, and one removal time cost of $28-$83 billion. That can be added to the $21.6 billion for the completion of the wall .
In the short-run, implementation of mass deportations will inject spending into the economy, stimulating the economy. It’s not quite “military Keynesianism”, but is in economic terms equivalent. The net impact on GDP will depend on the extent of monetary accommodation, the demand for US Treasurys from the private sector and the foreign official sector, and (over the medium term) the economy’s proximity to full employment.
In the long run, assuming linearity, potential GDP will be 1.6% lower than otherwise in twenty years.
Digression: The President has asserted he will negotiate down the price of the wall ; my question — why does he care about the price tag if the Mexicans are ultimately going to pay? Is it altruism?