The American Enterprise Institute on immigration, 2015 and 2025

The American Enterprise Institute in 2015:

“[Nicolas] Eberstadt [of the American Enterprise Institute] sees US demographic trends as mostly positive. The US, the world’s third-most-populous country (321.4 million people) and largest economy (GDP of $18.1 trillion), is projected to have modest population and working-age population growth over the next 20 years. And its population will age more slowly than in other OECD countries. The US still has a positive replacement-level fertility rate, augmented by continued immigration, including an influx of highly educated immigrants at a rate above the OECD average, he said. Eberstadt sees immigration in the US and Canada as a “fantastically positive experience” for those countries.” (Quoted here.)

The American Enterprise Institute in July 2025:

Report: “Immigration Policy and Its Macroeconomic Effects in the Second Trump Administration”

This AEI report projects that U.S. net migration in 2025 will fall to between −525,000 and 115,000 and reduce GDP growth by 0.3–0.4 percentage points. Its model estimated that if low immigration persists, GDP in 2034 would be roughly 1–2 percent smaller than under normal migration levels. That is to say, the GDP in 2034 would be about 41 trillion with pre Trump prevailing immigration and about 40 trillion without.  Thus its model expects a modest adverse impact.

 

 

 

How many unauthorized persons in the US now?

The number of unauthorized persons in the U.S. on December 2000 was 11-12 million. The number as of end of October 2025 is about 14 million. Below is how I arrived at this estimate.

Governmental and non-governmental organizations have differed by a lot on how many unauthorized persons are in the U.S. The disparity is due to several factors, including vague counts of persons permitted to enter temporarily under Biden and pure guesses (using modelling) as to those truly undocumented. 

Prior to the Biden administration there was a consensus that about 11-12 million persons were in the United States.  That included DACA recipients (about 600,000) and asylum applicants.  Both categories are persons here without legal status; their deportation was suspended.   There were small numbers of Temporary Protected Status (TPS) and humanitarian parole.

During the Biden administration the following happened to swell the total number. The accounting expressly included persons whose legal status is vulnerable, asylum applicants, TPS and humanitarian parole (DACA numbers did not materially change). It is important to keep in mind that estimates of total “unauthorized” include these categories. They are in fact documented. The core number of truly undocumented – that is without any temporary protection – may be around 10-11 million.

The Biden years saw:

*Those crossing illegally or overstaying temporary visas – I see no count for these. A guess: upwards of 250,000 a year, or 1,000,000 in total

*Increase in asylum cases: about 2,500,000 million

*Temporary protected status: about 500,000 + additional

*Humanitarian parole: about 600,000

LESS those deported /removed or self-deported under Biden: roughly 1,500,000. This assumes a number who left voluntarily, as has been the pattern for decades.

Thus, total in December 2000: 11-12 million

Change under Biden: 4.6 million less 1.5 million = 3.1 million

Total in December 2024: about 14.5 million

In the first nine months of the Trump administration – DHS’s released estimates have been deemed unreliable. Assume 400,000 have been deported and a large number self-deported. The removal of TPS status does not by itself cause the total number to go down. Assume the arrival of undocumented persons is relatively very low, say 100,000.

Thus, as of October 31, 2025: about 14 million.

Here is an informative estimate of the number of unauthorized as of mid 2023. Here is Pew’s estimate for 2023.

The rising trend of the Republican Hispanic voter has reversed

Republicans began to receive more Hispanic votes – until November 4, 2025. The number of eligible Hispanic voters is growing solidely. But be aware of the Hispanic voting deficit!

Background to 2024/2025: In the 2010s, Democrats had been relying on the support of roughly 90% of Black voters and 70% of Hispanic voters. This implies that per the Dems, the Republican should not expect to receive more than 30% of the vote.

In 2016, according to Edison Research, Trump received 28% of the Hispanic vote

In the 2018 mid-term Congressional elections, an estimated 69% of Latinos voted for the Democratic candidate and 29% backed the Republican candidate. 27% of Latino voters said they were voting for the first time, compared with 18% of black voters and 12% of white voters. This has been a trend for some time, as Hispanics are coming into adulthood at a relatively faster rate than are others.

In 2020, Trump won 32% of the Hispanic vote. Thus, the Democrats in the late 2010s achieved its target of 70% of the Hispanic vote, but the percentage was marginally declining. (Go here.)

In the 2022 mid-term Congressional elections, Pew Research estimated that 39% of Hispanic voters cast their ballots for Republican candidates, while 60% supported Democrats. This should have caused flashing red lights.

The Hispanic vote in November 2024

Trump won 46% of the Hispanic vote in the 2024 election. This percentage is 7% points higher than the 2022 mid-terms and 14% higher than the 2020 election. The swing was heavily among Hispanic men: per Edison Research, by 55% up from 36% for Trump in 2020.

American Electorate Voter Poll put Trump’s support among Latino men at 43%. The Navigator Research post-election survey said received 50% of Hispanic men. According to Edison Research, 38% of Hispanic women voted for Trump in 2024, up from 30% in 2020.

The Hispanic vote in November 2025

In the governor’s race in Virginia and New Jersey about 33% of the Hispanic vote went to the defeated Republican candidate. That below the percentage in the last non-presidential election (2022). In Hudson County, NJ, which has a 40% Hispanic population, the Republican candidate received only 25% of the vote, compared to 34% in 2000.

The Washington Post’s graph shows that with a 5% Hispanic share of the population, the total vote went Dem by about 2%, and with a 35-40% share the total vote went Dem by about 15%. there are some extreme outliers.

In sum, the surge of Hispanic approval of Republicans, dramatically evident in 2022 and 2024 voting, appears to have reversed. Is this is due to pocketbook issues, which past polling in early 2025 and before would suggest, or is it in part due higher concern about ICE enforcement?

The Hispanic voting deficit

One important aspect of the Hispanic vote for which we do not have the best information: the percentage of voting eligible Hispanics who actually vote is much lower than for whites (something like 50% vs 70%). And white people is the only race/ethnic group which today supports current ICE practices.

Nationwide, the number of voting eligible Hispanics is growing rapidly. Between 2020 and 2024, their numbers rose from 32 million to 36 million – about 4% a year, vs. an absolute decline in eligible white voters.  Hispanics make up about 22% of the white eligible vote, but only about 15% of these Hispanics vote. How this voting deficit effects Congressional races is a complicated question. But it should be worrisome to national Republican strategists.

For many facts on voting activity by race/ethnicity, go here.

What is driving the Hispanic vote?

Is ICE driving Hispanics to vote for Democrats? If so, how serious is the loss of Hispanic support for Republicans for the November 2026 elections?  This is the first of several posts about the Hispanic vote.

Trump’s immigration policy has split the electorate between poles. Many Republicans think that Trump has not gone far enough. Democrats – too far. (This is an April 2025 poll.)  I’m looking here at sentiment among Hispanics, interested in how this issue may be contributing to a wave of Hispanics who voted for Trump in November 2025 and are shown to switch to Democratic candidates in this week’s November 4 elections.

Polls taken through mid 2025 consistently show among Hispanics relatively high levels of concern about deportation and criticism of ICE action – but I have not seen any report on if concern about the economy and concern about immigration law enforcement stack up relative to each other in impact on voters. Polls such as this one fairly consistently report that Hispanics are more concerned aout their economic well being than about immigration – especially as immigration is typically characterized as legalization of unauthorized persons. What we don’t know is if a combination of fear and anger may touch visceral emotions and spur Hispanics to vote (their rate of voting is low, that is for a follow up posting) and to vote against Republicans.

Opposition to ICE enforcement

June – July polls by various sources show that 60-70% of Hispanics disapprove of ICE raids and other enforcement measures.  Among Asians and Blacks, adverse reaction is slightly less. But notably among Whites, as many support ICE raids as oppose them.

Fear of deportation

Through the 2000’s and 2010, roughly half of Hispanics said to Pew Research pollsters that they were worried that a family member or some one close to them might be deported.  In March 2025 Pew found that 51% of foreign-born Hispanics and 36% of U.S. born Hispanics were worried. 50% of the 18- 49 year old age group were worried.

In a Pew poll in June 2025, 47% of Hispanic adults, 15% of White adults (non-Hispanics, 26% of Black adults, and 29% of Asian adults were worried.

We take from this that a lot of Hispanics are worried about deportations, including many long time residents. For example, 29% of Hispanics both of whose parents were born here – that is, third generation Hispanic immigrants, are worried.

 

The case for skilled worker immigration

The report “Exceptional by Design” is not a very original statement in support of skilled immigration, but it synthesizes many studies into one document. It covers the potential use of more of those who come for higher education study and the use of early and mid-career immigrants.  Its argument:

  • High-skilled immigrants accelerate innovation—their authorship of 36% of U.S. innovative output since 1990, with each 1% increase in their population share raising patents per capita by 9-18%
  • They comprise 19% of National Academy of Engineering members, 24% of National Academy of Sciences members, and 34% of Nobel Prize winners.
  • The semiconductor sector where immigrants founded Fairchild Semiconductor, which spawned Intel and Apple, and now account for 34% of patents in strategic fields, enhancing U.S. exports and competitiveness.
  • 80% more likely than native-born adults to start businesses, founding not just boutique firms both also 55% of unicorn startups (valued over $1 billion),
  • They expand job opportunities more than they compete for them, with H-1B workers earning a median $118,000 (surpassing 90% of U.S. workers).
  • They make high-skilled labor less scarce, reduce income inequality, accelerating wage growth faster for low-wage workers while still elevating high-skilled pay.

Here are past postings on the penetration of skilled immigrants (mainly STEM and medicine) in the Boston area, the rise of foreign-born STEM workers over time, and polls showing strong support for skilled worker immigration.

 

Foreign born workers and the supply of software engineers: Japan vs U.S.

The U.S. information technology sector (computer design, software, AI, etc.) is a major force in the world. When one compares its IT workforce with those of other countries, the role of foreign-born workers stick out.

The United States has 30% – 50% more software engineers per total workforce than does Japan. This is despite the relatively high penetration of numeracy and adaptive skills in Japan compared to other countries (go here).

One major factor behind the U.S. lead is the very high level of foreign-born information technology workers in the United States. The U.S. greatly depends on foreign-born IT workers. 39% of its software engineers are foreign born, compared to 4% in Japan. Japan depends are less on foreign workers throughout its economy than does the U.S (about 3% vs 19%).

For context, the U.S.’ use of foreign-born for software engineering (39%) is much higher than in the U.K. (18%).  There are roughly one million foreign-born software engineers in the U.S. How do they get here?  Among the roughly 700,000 H1-B workers. About half, or 350,000 pf them are software engineers. As I’ve posted before, there is a transnational IT workforce in the U.S. and India. This means that besides having a relatively lot more software engineers compared to Japan, the U.S. has a distinct advantage of being more able to draw easily upon a major supplier of software engineers in the world – India. Language barriers alone create a big barrier for Japan to beef up its software engineering workforce.

 

 

 

Palantir’s immigration work  

The New York Times published on October 30 an interview by Ross Douthat with Shyam Sankar, Palantir’s chief technology officer. The immigration segment centers on Palantir’s work with ICE. Sankar says that ICE enforcement, paired with oversight, strengthens democracy rather than undermines it. Douthat: “From the point of view of someone who is a skeptic of Palantir, you are putting a lot of power into the hands of people that some don’t trust.”

Palantir began working with ICE in 2011 to integrate data across DHS, not to collect new information. The software allows agents to see connections among existing records on border encounters and benefits applications. He defends this as improving efficiency, not expanding surveillance. Sankar: “Why are we using the human to search 20 different systems? That’s clearly a problem that technology can do.” Douthat responds: “Does that include a person’s past address?”

Sankar says that Palantir partners only with democratic allies and avoids projects without privacy safeguards, such as the U.K.’s proposed digital ID. He argues that in the U.S., enforcement policies reflect democratic will and Palantir’s job is to ensure those policies are carried out lawfully and transparently. Sankar: “What ICE is doing was voted on at the ballot box. That seems like a functioning democracy….the lack of enforcement breeds nihilism, and that’s not tenable.” Douthat: “So Palantir is more comfortable with mass deportations in a world where Donald Trump campaigned on them and won?”

Sankar’s family fled violence in Nigeria. He says that immigration strengthens America when it is legal and includes assimilation.  Sankar: “We came here legally. We believe in America… but legal immigration. There are rules to follow.” Douthat: “Do you see any tension between your own personal narrative as an immigrant and the fact that Palantir is working to effectuate mass deportations?”

The drug cartel bust that wasn’t

There is little system-wide analysis of who ICE is detaining except for a flock of articles mid summer (one of them noted at the end.) Here is the outcome of an investigation into a big deal cartel bust that happened in late August. The reality of the bust came out when formal charges were made against those arrested.

As of today, late October, a slim majority of Americans appear to still believe that  illegal immigration = a major crime problem.

A Boston Globe Spotlight investigation revealed that the DEA’s August raids in Franklin, New Hampshire, promoted by the Trump administration as a major strike against Mexico’s Sinaloa Cartel, mostly swept up addicts, small-time dealers, and homeless people.  “These are high-level arrests, not low-level retail distribution. They are members of the Sinaloa cartel,” said Jarod Forget, special agent in charge of the New England Division. Fox News reported the entire 23 state operation against the Siniloa cartel led to 613 arrests nationwide.

Federal agents claimed to have arrested 171 cartel members in New England. But many were charged with minor offenses like shoplifting or drug possession. In 8,800 population Franklin, what the administration billed as a cartel bust was actually an enforcement operation resulting in 27 arrests, targeting fentanyl and meth distribution, and the clearing of a homeless encampment in Franklin. Franklin’s police chief said he recognized 85% of the people from his patrol days.

From The Marshall Project  (8/15/25)

“People with no criminal convictions at all make up two-thirds of the more than 120,000 people deported between January and May. For another 8%, the only offense on their record was illegal entry to the U.S. Only about 12% were convicted of a crime that was either violent or potentially violent. The numbers contradict officials’ continued claims that immigration enforcement is focusing on the “worst of the worst” criminal offenders. The numbers are estimates from a Marshall Project analysis of Immigration and Customs Enforcement data, provided to the Deportation Data Project in response to a FOIA request.”

 

 

How deportations happen — the legal issues

Most deportations from the U.S. happen in one of the following three ways.

Formal removal proceedings (immigration court)

In FY 2024, about 325,000 persons received removal orders by an immigration court.

Persons who are brought into this process typically entered the U.S. illegally and have been in the U.S. for over two year, overstayed their visa, are subject to removal due to a criminal conviction, or are denied asylum. Typically the person has resided in the U.S. for years.

DHA services the person a Notice to Appear before an immigration court, listing immigration law violations. An immigration judge within the Department of Justice’s Executive Office for Immigration hears the case.  The person can contest removability and can bring a lawyer. Between the date of issuance of the Notice to Appear and a decision by the court can be over a year.

Around 35–45% of these cases brought to a court result in no removal order. The most common reasons for non-removal are asylum status or a related protection; the judge is persuaded by long residence or other personal factors; a family petitions; or DHS decides in the end not to deport. A person can appeal a court ordered deportation to the Board of Immigration Appeals.  A review and decision by this court can take over a year.

A person given a court removal ordered is barred from legally entering the U.S. for ten years.

Expedited removal

In FY 2024, roughly 300,000 persons were removed by expedited removal.

DHS deports a person without going to an immigration court generally if the person illegally entered the U.S. (entered “without inspection”) and has been in the U.S. for less than two years. The two-year threshold is firm, written into law. A person detained for expedited removal as a right to a “credible fear” hearing by a U.S. Citizenship and Immigration Services asylum officer. Removal can take days or weeks. A person subjected to expedited removal may not legally enter the U.S. for five years. As visa overstays legally entered the country, they are not subject to expedited removal.

Legal entry, overstay and self-deportation

There are no government estimates of persons who self-deport. However, let’s look at persons who enter legally for an expressly limited time, overstay, and then leave on their own volution without a DHS encounter.

If a person subject to a visa overstays their visa by under six months, there is no formal bar to gain another visa to enter. But an overstay of over six months imposes a bar of at least three years. If the overstay is for, say two years, the bar raised to ten years.

If a person entered without being legally required to have visa, and beyond the maximum stay of 90 days, say for under 180 days, they are not barred from re-entry. But if they stay for more than 180 days beyond the 90 days they are barred for three years.

DACA status: A Russian doll of legal reasoning

The legal status of DACA can be compared to a Russian matryoshka. A federal judge declared it illegal 2021; it still operates; and when the Supreme Court will receive it for review is uncertaion.

Deferred Action for Children Arrival was created on June 15, 2012 not by Executive Order of President Obama bur rather by a Department of Homeland Security  memorandum, “Exercising Prosecutorial Discretion with Respect to Individuals Who Came to the United States as Children,” issued by: Janet Napolitano, Secretary of Homeland Security.

On July 16, 2021, in State of Texas, et al. v. United States of America, et al, The United States District Court for the Southern District of Texas, Brownsville Division (Judge Andrew S. Hanen) ruled that the DACA program was illegal because admininistrative memoranda from DHS were not intended for such a broad scope, and because the program violated the Administrative Procedures Act.

In response, the Biden administration undertook a cure. It e-established DACA in federal regulation (8 C.F.R. §§ 236.21 – 236.25); kept the same eligibility criteria as the 2012 memo; clarified that DACA does not confer lawful status, but provides deferred action and work authorization, and created a clearer administrative record to defend the policy in court.

But in September 2023 Judge Hanen again ruled that the program was defective. Among his objections was that DACA exceeds the powers Congress granted to DHS under the Immigration and Nationality Act. Congress never authorized DHS to grant “deferred action” and work authorization to such a broad class of undocumented immigrants. The government had “creates a program that Congress has repeatedly declined to enact.”

But Hanen allowed the program to continue for persons already enrolled – about 540,000. Had the Administration been permitted to continue enrollment and to broaden eligibility, over one million would be enrolled as of the fall of 2025.

On appeal by the plaintiffs, the Fifth Circuit on January 17, 2025 found that the DACA regulation is unlawful but limited the scope of its ruling to Texas and maintained a stay to allow current DACA recipients to continue renewing their benefits while the litigation continues. The Appeals court expressed caution noting that the welfare of many persions was at stake,

The case is not yet ripe for an appeal to the Supreme Court. The program has been operating for over a dozen years, efficiently, without scandal, and with broad public support. I suspect The SC doesn’t want to touch this box of explosives.