DHS Claims 622,000 Deportations, Here's Why That Number Might Be Meaningless
The Trump administration claims record deportations but something doesn't add up. I analyze ICE and DHS data to test whether the numbers are plausible—and find reasons to be deeply skeptical.
On December 19, the Department of Homeland Security announced that it had deported 622,000 people since the start of the Trump administration. If true, it would represent the largest forced relocation of immigrants since the institutionalization of U.S. immigration controls in the late 1800s and provide evidence that Trump is delivering on his “mass deportation” agenda.
But something is off. Way off.
The administration has not provided a shred of evidence for how it is producing this number, whether this number is comparable to previous years, or why this number differs so remarkably from ICE’s reported 344,000 deportations over the same time period. In fact, based on ICE data alone, ICE ERO only deported 21% more people in FY 2025 (329,018) compared to FY 2024 (271,480) under the Biden administration. This is well within the level of annual variation in deportations during previous Democratic and Republican administrations, well below Obama’s highest deportation numbers, and thus hardly unprecedented.
What’s really going on?
To untangle the Trump administration’s puzzling claims about deportation requires careful attention to the technicalities of immigration law, how statistical categories are defined, and how data, rather than being simply an objective description of “what is,” is always politically and rhetorically produced in ways that may be more or less defensible. We also have to grapple with how government secrecy foments confusion and undermines public trust, even if we later find that its internal methods for coming up with numbers are at least minimally well-reasoned.
Make no mistake: understanding how many people have been deported under the Trump administration is essential. No single number is more important. No metric is more emblematic of our ability to assess the Trump administration on immigration. This is true whether you voted for Donald Trump because you support mass deportation and you want to see if he’s living up to his promises, or whether you oppose Trump’s immigration policies and want to quantify the scope of the damage he’s done.
For DHS’s deportation number to be meaningful and defensible, rather than pure propaganda, requires a minimal level of transparency. The administration has failed on this account. In response to need for clarity—and motivated by the many questions I’ve received in recent days on this topic—this essay will unpack what we know about deportations numbers, how the counting of deportations is more complex than it seems at first glance, and assess the plausibility of various approaches to reverse engineering the administration’s claims. In addition to serving a “fact-check” function, my hope is that this exploration offers pedagogical value for anyone who is interested in leveling-up their immigration research skills.
My approach in this article is fairly straightforward. I begin by dissecting what various sources of deportation statistics say, identifying incongruities and gaps in what we know, then applying this understanding to DHS’s claims to see if we can make them make sense.
Counting Deportations is Harder Than It Looks
Before we can look the numbers, we need to understand what we’re counting. Let’s start with the term deportation. Most people understand that “deportation” means that the government kicked someone out of the country who didn’t have legal authorization to be here. However, the term deportation ceased to carry a precise legal meaning in 1996 when the law changed the term from deportation to removal. This was not only a rhetorical change. Prior to 1996, immigration law distinguished between deportation, which took place inside the country, and exclusion, which took place at the borders and ports of entry. Colloquially, deportation and removal are interchangeable, but legally and statistically, we must be careful. From this point forward in this essay, I will stop using removal and deportation interchangeably and only use them in their most precise senses.
What is a removal? Since we are focusing on removal data, I will use the definition that DHS includes in various reports and data publications. I will be referencing two main sources of data throughout this article: data and reports from Immigration and Customs Enforcement (ICE), and data and reports from the Office of Homeland Security Statistics (OHSS); the latter is a newer agency outside of ICE (but within DHS) that does more sophisticated statistical analysis of immigration data across the entire agency.
The most succinct definition of a removal comes from OHSS’s reports, which defines it as follows: “Removals are the compulsory and confirmed movement of an inadmissible or removable noncitizen out of the United States based on an order of removal. A noncitizen who is removed pursuant to a removal order has administrative or criminal consequences placed on subsequent reentry owing to the fact of the removal.” Removals are distinct from a separate category called “returns.” According to OHSS, “Returns are the compulsory and confirmed movement of an inadmissible or deportable noncitizen out of the United States not based on an order of removal.” Removals and returns are similar in that they represent a moment of immigration control, but they are different in that removals are based on removal orders and returns are not.1
One key source of removal data comes from ICE’s biweekly detention spreadsheet, which provides a running fiscal-year-to-date total of the number of removals effectuated2 by ICE Enforcement and Removal Operations (ERO). In the footnotes tab, ICE describes its reported removal data in the following way:
“ICE Removal Data Include Returns and Expulsions. Returns include Voluntary Returns, Voluntary Departures and Withdrawals Under Docket Control.”
We do not need to define every single term here. Just note that ICE’s reported removals number includes removals and returns. This is our first example of slippage between distinct legal categories and precise statistical categories. There is nothing nefarious here; it’s justifiable to lump categories together for reporting convenience (I do it all the time).
The real question we need to answer is: how much do these other non-removal elements within ICE’s removals data drive the overall numbers? Using TRAC’s detailed data on ICE ERO removals, the answer is: not much. Since 2015, the number of removals labelled as some form of return fluctuates between 1% and 3% of total records in its removals data each fiscal year. Based on this data, we can say that although ICE’s biweekly removals data includes both removals and returns, it is defensible to refer to ICE’s reported removals data as just that: removals data.
Now that we answered that question, we can move on to the next logical question: does ICE ERO removals constitute the entire universe of removals from the United States? It is hardly unfounded to assume that ICE’s Enforcement and Removal Operations (ERO) is responsible for all removals—it’s right there in the name. ICE ERO’s website says clearly:
“As part of its critical mission, ERO manages all aspects of the immigration enforcement process, including the identification, arrest, detention and removal of aliens who are subject to removal or are unlawfully present in the U.S.”
And over the years, the number of removals reported by ICE ERO have been treated as definitive, whether that data came from ICE’s biweekly spreadsheets or from ICE’s annual enforcement reports. But in fact, the answer is: “no.” ICE is not the only agency within DHS that does removals. ICE’s 2023 annual report includes the following footnotes:
“This chart represents removals performed by ICE and does not capture total removals or returns by DHS.”
“Other DHS components, primarily CBP, also return and remove noncitizens. Those returns and removals are not included in the ICE Annual Report.”
ICE’s 2018 report includes another relevant footnote:
“ICE removals include removals and returns where aliens were turned over to ICE for removal efforts. This includes aliens processed for Expedited Removal (ER) or Voluntary Return (VR) that are turned over to ICE for detention. Aliens processed for ER and not detained by ERO or VRs after June 1st, 2013 and not detained by ICE are primarily processed by the U.S. Border Patrol. CBP should be contacted for those statistics.”
From these reports we learn that while ICE ERO is tasked with removals as a key part of its mission, the total number of removals effectuated by the Department of Homeland Security (DHS) in a given year equals ICE ERO removals plus some other number of removals.
Total DHS Removals = ICE ERO Removals + Some Other Number of Removals
As the ICE footnote above says, many of these “other removals” come from Customs and Border Protection (CBP)—more specifically from Border Patrol and the Office of Field Operations (OFO). Most people have heard of Border Patrol, but OFO is less well-known; OFO is the agency that manages ports of entry and they, too, play a significant border enforcement role even though they are not part of Border Patrol.
How many removals does CBP effectuate each year? CBP has several tables and dashboards that allow the public to see monthly metrics on encounters, apprehensions, and processing dispositions. But you’ll notice that CBP does not report the total numbers of removals. That’s because—remarkably—the agency does not have a field in its data for removals. Instead, CBP removals data has to be inferred from a combination of other data points using research and analysis. It’s not just that CBP doesn’t publish this data, the data itself simply does not exist—or at least does not exist in any format equivalent to ICE ERO’s data.
Trump Cuts Off Key Deportation Data Source
Only OHSS has the combination of expertise and systems access to produce estimates of CBP removals and add those numbers into total DHS removals. For this reason, OHSS is a lynchpin in this analysis. Until the Trump administration prevented OHSS from publishing public data, OHSS’s monthly enforcement reports provided a crucial, non-partisan source of data on removals that filled in this crucial gap in public understanding.
OHSS’s data tables also complicated how we think about immigration data in ways that will become immediately relevant. We learned above that ICE ERO removals include removals and returns, and that ICE ERO removals are only a fraction—a currently unknown fraction—of all DHS removals. OHSS adds another layer of complexity. OHSS uses a meta-category of “repatriations” to capture the full universe of immigration enforcement actions across the entire DHS. The total number of repatriations includes distinctions for removals, returns, and Title 42 expulsions aggregated to the level of DHS—not the level of CBP or ICE. OHSS does also provide data on ICE specifically—but like ICE’s biweekly spreadsheet, no distinction is made between removals and returns; the two are still lumped together. OHSS does, however, break out total returns into what it calls “enforcement returns” and “administrative returns”—a distinction that may help us in our inferential study later.
To summarize, OHSS provides total removals and returns, and returns by type, at the DHS level, and reproduces ICE’s approach to reporting removals data—but OHSS does not break out removals and returns separately by CBP and ICE.
To see the data for yourself, go to OHSS’s online dashboard or download the last of OHSS’s spreadsheets from January 2025 (current through November 2024) below.
Comparing Deportation Data Sources
At this point, we’ve covered all of the available data sources. Now let’s compare the counting for removals across ICE’s biweekly spreadsheets, ICE’s annual enforcement reports, and OHSS’s monthly enforcement tables.
Looking at FY 2014 through FY 2024, these sources tell a consistent story about ICE removals. In FY 2024, ICE removed about 271,500 people, as reported in ICE’s annual report and OHSS’s data tables. Measured against these sources, ICE’s biweekly removal numbers are reliable real-time data. If we take the data closest to the end of each fiscal year, then use a simple daily average to project through September 30, the results run close to—but just slightly lower than—ICE’s annual reports and OHSS’s data.
The more interesting pattern is the gap between ICE removals and total DHS removals. Using OHSS data, in FY 2024, ICE reported 271,480 deportations. However, OHSS found another 58,510 removals somewhere (mostly likely in CBP’s data) that added up to 329,990 removals for DHS as a whole. That means ICE accounted for 82% of all DHS removals that year.
Since OHSS reports on returns, and since these numbers are huge, we should discuss those, too. Continuing with our discussion of FY 2024, OHSS says that DHS effectuated a further 355,490 “enforcement returns” and 92,310 “administrative returns.” The total number of repatriations under the final full fiscal year of the Biden administration came to 777,580. No Title 42 expulsions were conducted that year.
How do we make sense of OHSS’s data and why does it matter for this study? Simply put, OHSS’s far more comprehensive data on enforcement actions taken across DHS helps us to understand just how complicated the data landscape is here. We can see that ICE ERO removals constitute most—but not all—DHS removals. And we can see that not all enforcement actions are considered removals. Depending on how you slice it, different combinations of categories gives you different numbers, each of which could be used to illustrate the quantifiable outcomes of an administration’s immigration policies. OHSS says that Biden deported 330,000 people in FY 2024, not 271,500. But all told, DHS took 685,280 enforcement actions (total removals plus total returns), and repatriated three-quarters of a million people.
This complexity has led some researchers to reintroduce the term “deportation” to represent total enforcement actions beyond simply removals. An MPI article from 2024 defines the term deportation as “A non-legal term to describe immigration authorities’ removal or the enforcement return of a noncitizen from the United States.” Clearly these terms are taken straight from OHSS.
Time to apply all of this to the Trump administration’s claims.
Assessing DHS’s Claims About High Deportation Numbers
At the beginning, I cited DHS’s claims that it deported 622,000. But actually the phrasing in the similar announcement of 605,000 deportation on December 10 is clearer. In that press release, DHS said,
“Since January 20, 2025, DHS enforcement operations have resulted in more than 605,000 deportations.” (DHS’s emphasis.)
You now know enough about the technicalities at stake to catch two things. First, the statement attributes the number to DHS, not ICE ERO. So if this statement is actually about removals, we know that DHS is arriving at its 600,000+ number by adding some unknown “other number” of removals to ICE ERO removals. Of course, because the administration cut off OHSS reports, we have no way of knowing what the “other” number is, we only have the ICE ERO numbers.
Second, the press release says deportations, not removals. When I first read the announcement, I assumed that DHS was using the term deportations as an equivalent substitute for removals in order to be more legible to the public. It is Trump’s style, after all, to choose inaccurate simplicity over accurate complexity. But on deeper analysis, it’s clear that they are using deportations strategically to obscure the composition of what they are counting in the data—which means we have to try to reverse engineer what they are counting based on what we learned from OHSS. So let’s try that.
As I said in the introduction, based on ICE’s biweekly detention spreadsheet, for FY 2025 (October 1, 2024 to September 30, 2025), ICE ERO effectuated an estimated 329,018 removals. But the Trump administration’s claim doesn’t cover the full fiscal year—it excludes the first three and a half months (October 2024 through January 19, 2025) when Biden was still president and includes data through the time of the press release. I will work with the 605,000 number, because that statement was made on December 10 and the ICE biweekly spreadsheet from December goes neatly to December 13—so the most equivalent timeframe for analysis.
Using ICE’s biweekly spreadsheets, we can subtract the 76,939 removals that occurred before Trump took office from the total (329,018) to get 252,079. Then we add in 92,181 removals reported through December 13 for a total of 335,222 ICE removals attributable to the Trump administration from January through December 2025.
This method accounts for barely more than half—54%—of DHS’s claimed deportation number. We are still a long way from figuring out how DHS is getting to 605,000 or 622,000.
We said before that ICE ERO removals do not constitute the total universe of DHS removals or DHS enforcement actions. So let’s try to get there.
If DHS is using the term “deportations” as a straight substitute for “removals,” we could try to estimate this number using known ICE ERO removals. Using the table above, we observe that ICE ERO removals made up between 69% and 82% of the total DHS removals. Take 335,222 and divide it by 0.69 and 0.823 to get a range of 408,807 to 485,829 total DHS removals over this time period.
This is still far short of DHS’s claimed 605,000 “deportations” at the scale of 119,000 to 196,000 removals.
Explaining DHS’s Massive Discrepancy
What accounts for the differences? Here are two plausible explanations and implications for each.
The number of DHS removals counted under the “other” category has ballooned so significantly that our estimates based on historic data do not apply. Take at face value, this would mean that CBP is doing so many removals that it has significantly driven down ICE removals to a much smaller fraction of the whole. This would be odd, if true, since it contradicts the widely reported findings that border enforcement is as recent historic lows. It is therefore also possible that some other non-ERO and non-CBP entity is doing huge numbers of removals—but I don’t know who that would be.
DHS is using “deportations” in an expansive sense that includes much more than just removals—including returns, etc. If this is true, the public needs to understand that when the Trump administration says “deportations,” they don’t really mean deportations in the colloquial sense that we still use it. Rather, they mean various enforcement actions including but beyond removals. This could include an increase in so-called “self-deportations”—but again, those are not necessarily removals, strictly speaking.
Here’s why I find neither of these explanations satisfactory. For #1, I don’t have any evidence for the scenario that there is some other non-ERO agency doing a bunch of removals. It’s possible, I suppose—but if so, it seems like someone would have broken that story. For #2, if DHS really was using an expansive definition of deportations as “removals plus enforcement returns,” then DHS is severely underselling its enforcement activity—which I can’t imagine they would do. Just to illustrate this point, if 605,000 “deportations” includes removals and enforcement returns, then Biden already exceeded this in FY 2024 with 685,280. Using this count, Trump is coming in far under Biden. That doesn’t make sense.
DHS’s Deportation Data Assertion Is Meaningless
There is a third scenario that I’m starting to think is the most likely. It’s plausible that DHS has thrown out everything we know about removals and returns data and has come up with their own secret math to make deportations look really high. If this is true, it would completely undermine the administration’s integrity on immigration data.
DHS isn’t Coca-Cola—it doesn't get to have a secret formula and expect the public to still buy its product. Without evidence of what’s being counted, the claim of “622,000 deportations” is utterly meaningless and totally disconnected from any plausible interpretations of enforcement data. It certainly cannot—and should not—be repeated as if this number reflects actual deportations (i.e., removals). We have no evidence to support that and plenty of counter evidence against it.
Finally, whatever is included in DHS’s number, it cannot be compared over time to past years and past administrations. Any claims that Trump deported 622,000 people compared to Biden’s 271,500 or 330,000 (depending on which number you use) is an apples to oranges comparison—not apples to apples.
A Warning Label on DHS’s Deportation Claims
The simple fact is that the administration’s claim doesn’t square with available data, even with the most generous counting methods. I’d genuinely like to understand what’s included in that 603,000 number. From a research perspective, this is crucial information. From a public understanding perspective, it’s critical that the public has access to trustworthy data points. For policy analysts, reporters, and policymakers who need to assess what’s actually happening, this is a straightforward question that should have a straightforward answer.
There’s no good reason to hide the methodology if the administration is using defensible counting methods and isn’t engaged in outright dishonesty. If the administration wants to politicize deportations and compare their record to previous administrations, they need to be honest about what they’re doing. However they are counting deportations now, they need to reconstruct those numbers going back 10 years to provide an appropriate apples-to-apples comparison with previous administrations, otherwise they aren’t justified using this number in a comparative frame.
If the administration wants to provide an explanation for this number, I’d be glad to hear it. And if anyone at the relevant agencies wants to explain what’s going on, either publicly or on background with me, I’m genuinely curious about the factual basis for this claim. But until we have clarity, DHS’s purported overall deportation number should be treated as profoundly suspect, if not mere propaganda dressed up as objective data.
Consider this essay your warning label on DHS’s deportation data claims that will remain on DHS’s claims until they provide an explanation and the underlying data. No reporter, researcher, or policymaker is justified in accepting DHS’s claims at face value until the agency explains its methodology.
And if I’m the one who missed something, let me know and I’ll update the post with corrections or further explanation.
For Further Reading
Support Public Scholarship
This newsletter is only possible because of your support. If you believe in keeping this work free and open to the public, join the 100% of surveyed paid subscribers who said this newsletter is a good steward of their investment. You can read more about the mission and focus of this newsletter and learn why, after three years, I finally decided to offer a paid option.
For our current analysis, we do not need to consider expulsions in detail. Expulsions at the border during the COVID-19 pandemic were justified using public health laws under Title 42 rather than immigration laws under Title 8. If you dive into the data further, you’ll see a distinction between Title 8 removals and Title 42 expulsions.
In an effort to be rigorously precise in this article, I will use the burdensome yet most accurate governmentalese term “effectuated” in this article to refer to the number of deportations completed rather than describing it as “people” deported. It is possible for a single person to be represented more than once in the data if they have been deported more than once.
I am not using the higher 85% for FY 2025 because it only includes two months of data.
Hi Austin,
Thanks for this your consistently rigorous and clear analysis. One question I have regarding your ICE removals calculation for calendar year 2025:
How were you able to derive ICE removals on a calendar year basis for 2025 using the ICE biweekly spreadsheet? The spreadsheet gets updated without a historical trail, so I am unable to locate the ICE removal data that was available as of December 2025, which would be required to derive CY 2025.
Am I missing something?
You mentioned that a ballooning of CBP removal actions is unlikely to be an explanation for the discrepancy. To be clear, this couldn’t be explained by historically increased use of CBP forces in the interior of the country in addition to ICE, correct? My assumption is that their arrests are still transferred to ICE detention where a “deportation” would count as an ICE ERO removal, so that the “higher share of CBP removals” explanation still wouldn’t hold water.
Additionally, what of the expansion of expedited removal? CBP could perform those at the border before; if the border zone is the entire country now, could many interior expedited removal actions be counted as CBP removals?