Apples and Oranges

Center for Migration Studies uses incomparable data sources to challenge CIS findings

By Steven A. Camarota on September 9, 2016

Download a PDF of this Backgrounder.

Steven A. Camarota is the director of research at the Center for Immigration Studies.

On August 31, the Center for Migration Studies (CMS) published a report by noted researcher Robert Warren arguing that a CIS Backgrounder from June of this year incorrectly found that both legal and illegal immigration increased in 2014 and 2015.1 CMS argued that the recent surge was “illusory” because it was exclusively due to an increase in legal immigration — mainly temporary, long-term “non-immigrants” and legal immigrants returning to the United States. In our view the CMS analysis is not correct. CMS mistakenly compared administrative data that does not show new arrivals to Census Bureau data that does. Further, CMS’s estimates of returning legal immigrants are for 2011 to 2013, not 2014 and 2015; and the very evidence it cites shows that returning immigrants do not show up as new arrivals in Census data and so cannot account for the surge. Although it was not the focus of our original analysis, we made clear that our estimates of illegal immigration were and are preliminary; as new information becomes available we may need to revise them. At present we stand by our original conclusion that both legal and illegal immigration increased significantly in 2014 and 2015.

Did Immigration Increase?

The Center for Migration Studies report states that the increase in both legal and illegal immigration in 2014 and 2015 shown in the Census Bureau data is “well within the bounds of normal annual fluctuations observed in Census Bureau survey data over the past 15 years.” It is not clear what this statement means because the CMS report also acknowledges what all other researchers have found — immigration has fluctuated a great deal over the last 15 years. The entire point of our June report was to argue that new arrivals (legal and illegal) are now much higher than in the recent past, and maybe even match what they were 15 years ago. So, yes, the number of arrivals is not unprecedented in the last 15 years. But those 15 years have seen truly dramatic fluctuations.

As we show in Figures 2 and 3 and Tables A1 and A3 of our earlier report, the higher level of immigration in 2014-2015 found in the Census Bureau’s Current Population Survey is statistically significant compared to any two- year period since 2000-2001. (The Current Population Survey (CPS) groups arrival data in two-year cohorts.) The Bureau’s much larger American Community Survey (ACS), which reports data only for the first part of 2014, also shows a statistically significant increase in new arrivals for the first half of that year. As we will see, CMS implies that the large increases in 2014 are not that meaningful, but it never disputes the increase itself.

Was the Increase All Temporary?

We certainly agree that legal immigration is up, particularly long-term temporary visas. We base that conclusion in our report on our examination of permanent and temporary visas issued overseas by the State Department. In our view, data from State provides some insight into whether new arrivals have increased because those who receive new visas typically then proceed to come to the United States, though they are not required to do so.2 And while legal permanent immigration has increased, we state in our report that “Much more important in terms of numbers, however, is the increase in long-term temporary non-immigrant visas in the last two years that allow for stays in the United States for a period long enough to show up as new arrivals in Census Bureau data.” Long-term “non-immigrants” includes such visa categories as guest workers, foreign students, and cultural exchange.

CMS claims that almost all of the increase shown in Census Bureau data is likely temporary non-immigrants, with returning legal immigrants playing some unspecified role. It says that “essentially all” of the increase in Census Bureau data could have been due to an increase in temporary admissions. Note that it only deals with the 2014 increase, it never really addresses the evidence that immigration was also higher in 2015.

CMS’s primary evidence is not State Department data on visas issued overseas to those who then come to the United States.3 Instead, CMS uses the DHS Yearbook of Immigration Statistics, which shows that non-immigrant admissions increased in 2014 compared to 2013. But the footnote in Table 25 that CMS cites from the Yearbook very clearly states that “Admissions represent counts of events, i .e ., arrivals, not unique individuals.”4 (Emphasis added.) In other words, the arrival data in the Yearbook includes repeat entries of the same people, including those already in the country for many years.

Non-immigrant admissions reported in the Yearbook are most definitely not a measure of new arrivals, as is the Census Bureau data used in our analysis. Non-immigrant visas allow for multiple entries and these reentries of the same people are what is reported in the Yearbook. We can see this clearly by looking at foreign students. In 2012, DHS estimated that the entire foreign student population in the country, not just new entrants, was 720,000.5 But the Yearbook shows that 1.65 million students were admitted in just 2012 — more than double the entire foreign student population in the country.6 Admissions data is not a measure of new arrivals.7

Despite all this, CMS assumes there is a clear relationship between admissions of non-immigrants in the Yearbook and the number of new non-immigrants coming for the first time who then show up as new arrivals in Census data. It even estimates the total size of the non-immigrant population based on these admissions figures. We strongly disagree with CMS’s assumption that Yearbook non-immigrants and Census arrival data are comparable data sources. Many factors can increase or decrease the number of times non-immigrants choose to come and go during a year. Conditions in the United States, conditions in the home country, or simply an increase or decrease in the cost of international air fares can affect an individual’s decision to leave and return to the United States and the number of times they do so in a year. But even if it were possible to estimate new non-immigrant entrants form Yearbook admissions, there is another problem. For CMS’s analysis to work, there has to specifically be a clear relationship between Yearbook non-immigrant admission data and the arrivals shown in Census Bureau data, but there is none.

Incomparable Data

If the CMS argument that non-immigrant admissions as shown in the Yearbook can entirely or nearly entirely explain an increase in arrivals shown in Census Bureau data, then there should be some relationship between trends in Yearbook admissions data and what the Census data shows. This is doubly true because CMS argues that permanent immigration and new illegal immigration have been stable. So increases in temporary immigration as shown by Yearbook admissions data should be apparent in Census Bureau data. But Figure 1 shows this is not the case. (The sharp upturn in 2010 was due to an accounting change by DHS.8) Putting aside the change from 2009 to 2010, it is still the case that admissions and Census arrivals do not move in tandem.


Graph: Number of New Arrivals in the ACS and Non-immigrant Admissions, 2006-2014


It is true that the ACS shows an increase in arrivals from 2013 to 2014 and so does Yearbook admissions data. And that is what the CMS report focused on. But in other years there is no relationship. Figure 2 reports changes from the prior year in Yearbook admissions data and arrivals from the ACS. The figure shows that the direction of the changes year-to-year from the two data sources are not even the same in most years, with Yearbook admission data showing an increase when ACS data shows a decrease or vice versa. In fact, the statistical correlation between arrivals and admissions present in Figures 1 and 2 is roughly zero.


Graph: Change from the Prior Year in ACS arrivals and non-immigrant admissions

None of this is surprising given that Yearbook admissions are not new arrivals. This is a critically important point because CMS’s argument centers on the idea that non-immigrant admissions shown in the Yearbook can explain the increases in arrivals shown in Census Bureau data. All the evidence shows that Yearbook admissions of non-immigrants and Census data are not comparable data sources. And it is a mistake to use admissions to estimate new arrivals that are comparable to Census Bureau data.

No Evidence Returning Immigrants Caused Surge

In addition to non-immigrants, CMS also argues that some share of new arrivals in Census data are legal immigrants who had left in prior years and are now returning. (CMS also estimates a tiny number of returning illegal immigrants.) There is no question this happens; we even mention it in our original paper. But measuring it is not easy. As the CMS report itself states: “Unfortunately, no statistical information about the return to the United States of former immigrants is collected.” There are two important problems with CMS’s argument in this regard. First, based on a methodology it developed, CMS provides estimates of returning legal immigrants and illegal immigrants for only the years 2011, 2012, and 2013. CMS presents nothing for 2014 and 2015, the two years of the surge. In our view, it is unclear if this type of “re-immigration” (my word) to the United States played a role in the 2014-2015 surge. But at present we do not see any evidence that it contributed significantly to the surge.9 More important, at present CMS has no data for these years either.

A second problem with arguing that re-immigration helps explain the surge is that it is not clear how many returning immigrants might show up in Census data as new arrivals. The CPS and ACS ask respondents what year they came to “live” in the United States.10 If someone lived in the United States from 2000 to 2009 on a green card and then left and came back in 2014, they would have to tell the Census Bureau that they came to live in the country in 2014. When discussing its estimate for re-immigration, CMS makes clear that returning immigrants do not give the year they came back as the year they came to live in the United States. But CMS still argues that “some” of these returning immigrants “probably” give the year of return as their year of arrival. That is the extent of its argument.

Although CMS devoted a significant share of its report to measuring re-immigration, which itself is certainly interesting, it is unclear why it did so. CMS presents no data for 2014 and 2015 and does not show that returning immigrants account for new arrivals in the CPS or ACS.

A Different Perspective on Immigration

It is worth noting that a secondary argument in the CMS report is that the significant increase shown in Census data is not very meaningful because so many of these arrivals are long-term temporary “non-immigrants” (e.g. guest workers and foreign students) and returning legal permanent immigrants. CMS go on to write: “Increases in these two categories do not support the CIS statement that the scale of new immigration is clearly enormous.” We could not disagree more. Even if the increase was caused entirely by legal immigration, which is not the case, we still think the findings are very important.

A significant increase in returning legal immigrants, foreign students, and guest workers has profound implications for the nation’s labor market, public coffers, universities, physical infrastructure, and a host of other areas. There is also the question of whether we can properly vet so many new entrants. Furthermore, nearly half of all new green cards (permanent residency) go to people who originally came as a non-immigrants — temporary immigration often becomes permanent. CMS is surely mistaken when it argues that for non-immigrants, “whatever their role in helping to create a surge, that role is temporary.”


We disagree with the CMS analysis arguing that the surge in new arrivals in Census data is entirely explained by increases in legal immigration, particularly temporary non-immigrant visas. Our analysis indicates that the surge shown in Census Bureau data was driven by an increase in temporary and permanent legal immigration, as well as by an increase in illegal immigration. The primary problem with the analysis in the CMS report is that is uses administrative data that shows multiple admissions of the same person, not new arrivals. It then compares these admissions to Census Bureau data that does measure new entrants.

The Current Population Survey shows that new arrivals surged 39 percent from the 2012-2013 period to the 2014-2015 period. It also shows that immigration from Latin America increased by 58 percent, compared to a 29 percent increase from everywhere else. Latin America has traditionally been the primary sending region for illegal immigrants to the United States. The American Community Survey shows that new arrivals surged 24 percent from the first half of 2013 to the first half of 2014. Yet the CMS report argues that none of this increase was due to an uptick in illegal immigration. Our analysis of visas issued overseas at American consulates and other data indicates that non-immigrant arrivals have increased, but it is still the case that roughly one-third of all new arrivals are illegal immigrants. Of course this conclusion remains preliminary. As more data becomes available we may revise our estimates. But at present we stand by our original finding that both legal and illegal immigration surged in 2014 and 2015.

End Notes

1 Steven A. Camarota, “New Data: Immigration Surged in 2014 and 2015: More than three million legal and illegal immigrants settled in the United States in the last two years”, Center for Immigration Studies, June 2016; Robert Warren, “Surge in Immigration in 2014 and 2015? The Evidence Remains Illusory”, Center for Migration Studies, August 31, 2016.

2 Because visa holders do not have to come immediately, the number of visas issued by State is an imperfect measure of likely new arrivals. But it does correlate with Census arrival data.

3 If we focus on the same temporary visa categories shown in Table 1 of the CMS report, there was a 121,000 or 11 percent increase in visas issued overseas by the State Department from 2013 to 2014. Of course it must be remembered that some of these individuals did not come to the United States, as their visas are good for a number of years. But this increase, while not trivial, cannot account for the huge increase in new arrivals shown in the CPS, especially for Latin America. State Department data can be found here.

4 Table 25 from the 2014 Yearbook of Immigration Statistics can be found here.

5 The estimate can be found in Table A2-1 in Bryan Baker, “Estimates of the Size and Characteristics of the Resident Nonimmigrant Population in the United States: January 2012”, Department of Homeland Security, Office of Immigration Statistics, February 2014.

6 CMS reports these same numbers in Table 1 of its report.

7 It is unclear why CMS leaves the reader of its report with the impression that the admissions figures show new entrants, which is not the case. In the discussion that follows we confine our analysis of non-immigrant admissions to the same visa categories CMS used in Table 1 of its report.

8 The Yearbook states that, “beginning in 2010 the number of I-94 nonimmigrant admissions greatly exceeds totals reported in previous years due to a more complete count of land admissions.”

9 Table 2 in the CMS report shows an 87,000 increase in re-immigration of legal permanent immigrants from 2011 to 2012 and a larger increase of 117,000 from 2012 to 2013. But ACS arrivals show the increase 2011 to 2012 is twice the size of the increase 2012 to 2013. This is not the pattern we would expect to see if re-immigration was reflected in ACS arrival data.

10 The ACS questionnaire can be found here. The CPS questionnaire can be found here.