A Semi-Secret Immigration Agency: USCIS's Administrative Appeals Office

All decisions made by U.S. Citizen and Immigration Services (USCIS) can be divided into three parts:

1. Approvals, which happen about 85 percent to 90 percent of the time;
2. Denials of cases so feeble that there is no appeal; and
3. Appeals of some of the denied cases, the vast majority of which are denied again on review.

This report is on the third group of cases (for which there is ample documentation, as there is not on the others) and on the all-but-unknown agency within USCIS that handles most appeals, the Administrative Appeals Office (AAO).

But before we get into the AAO operation, it would be useful to describe how the agency fits in with the larger question of how different kinds of migration-related appeals are managed within our immigration system. Needless to say, since this relates to immigration, it is complex.

If an alien wants a benefit of some kind, or a would-be employer wants alien workers, the applicant must file the right piece of paper with the right governmental unit. If the application (or petition) is denied at the staff level, then there is usually — but not always — an appeals route, and we have identified at least seven such semi-judicial arrangements. (There may well be more.)

These include:

• Within the Department of Justice: the Executive Office for Immigration Review (this includes the immigration judges) and the Office of the Chief Administrative Hearing Officer (a tiny unit);
• Within the Department of Homeland Security: the AAO, along with separate in-house reviews for disputed naturalization cases and for the (few) cases decided against applicants to ICE’s Student and Exchange Visitor Program (SEVP) office;
• Within the Department of Labor: the Board of Alien Labor Certification Appeals (BALCA); and
• Within the State Department: a system of in-embassy reviews of some denied visas, more of a supervisory operation than an appeals system.

Of these review systems, the immigration judges are the most formal and the most expensive. A judge (always wearing a robe), a DHS attorney, the alien, and a government-supplied interpreter, if needed, and sometimes the alien’s attorney, spend an hour to an hour and a half at each substantive hearing. If the petitioner is unhappy with the outcome, he or she can appeal to the Board of Immigration Appeals (BIA), which then uses a paper review; sometimes this is done by a single judge, and sometimes by a panel of three.

In contrast, the other systems, including that of the AAO, use one-level paper reviews only, and usually no hearings, to conduct their reviews. Further, AAO is staffed by experienced DHS adjudicators who are not necessarily lawyers.

Sometimes these cases are carried into the federal court system, but most are settled at the semi-judicial levels just described.

The Caseloads of These Agencies. The immigration judges completed 231,659 cases in 2020, according to a tally of new cases and total completions within a Justice Department data set. This indicates that these judges are handling more than 20 times as many cases as AAO, which published a total of 152,453 decisions for 2005 through June 21, 2021 (see Tables 2 and 3). The AAO decisions thus come to about 9,500 a year. Most of the AAO decisions are regarded by the system as non-precedent ones, meaning that other AAO officers need not make their decisions based on them. A handful of precedent decisions are designated by the leadership of USCIS each year. (This report deals with the far more numerous non-precedent decisions.)

The other five systems for handling review of initial civil servant decisions have totals that would appear to be lower than those of AAO.

Another distinction should be made between the work of immigration judges and that of the other six systems. Immigration judges can and do make deportation decisions, which are more likely to be enforced if the alien is in detention at the time of the decision. The other six systems make decisions about granting or denying a benefit, or in some cases setting the size of financial penalties, but deportation, or relief therefrom, is not on their agendas. (The denial of an immigration benefit may, later, lead to a deportation, but these are separate systems.)

AAO’s Decisions. As noted earlier, AAO decisions on appeals to it are usually negative – the staff position is confirmed and the applicant loses. But it is more complicated than that. Sometimes an AAO officer decides that more work needs to be done and in those cases the officer remands the decision back to the agency staff to work through it again, usually with some advice as to how to proceed.

In order to get a system-wide view of AAO’s activities, we plowed through 80 different categories of cases, selecting 33 categories for further attention. Of the 47 classifications excluded, eight dealt with the 1986 IRCA legalization program; these cases are long in tooth, and relate to by-gone issues. We dropped 21 other classes on the grounds that they had 19 or fewer entries over a period of some 16 years; in three categories (see Table 3), there were no cases at all. We eliminated others on the grounds that they involved relatively minor issues, such as the award of a refugee travel document or a special certificate of naturalization. Our study dealt with categories with 102,426 decisions; the excluded categories had 50,027 decisions.

The three largest groups of appeals examined, and the only ones with more than 10,000 decisions published, were these (including the AAO code):

• Temporary Protected Status (M-1): 22,349;
• H-1B petitions for nonimmigrant workers (D-2): 15,633; and
• Immigrant Petition for Alien Workers (Professional and Other Workers) (B-6): 15,107.

Most of those in the third grouping are H-1B workers seeking green card status.

So what did AAO say in the significant categories? We decided to read 20 decisions in each of the 33 categories included and found that AAO dismissed the appeals in fully 78.5 percent of the cases, a remand was ordered in 13.5 percent of them, and in only 8.0 percent did the applicant have his or her appeal sustained. Bear in mind that this was not a random sample of all cases presented to USCIS — this was a study of what happened to cases that already had been rejected by the staff. Further, it should be noted that many of those with really poor cases did not appeal their denials.

These percentages are, in a sense, an appropriate mirror image of USCIS decision-making generally; if a high percentage of cases are approved, as they are, then most of those that are not approved (and subsequently appealed) are presumably correctly decided.

Looking at these numbers in another way, we found that in 17 of the 33 groupings, there were 17 or more dismissals, usually out of 20; the staff below and the AAO officers were in almost complete agreement (see Table 1).

In only two instances were there fewer than 10 dismissals. In one of these two (AAO code A-7, Adjustment of Alien in T nonimmigrant status), all of the aliens said that they had been victims of human trafficking, and in the other, another victims’ category (AAO code C-6, Petition for Special Immigrant Juvenile), the aliens contended that they had been abused or abandoned by their own, usually illegal, alien parents.

We can see a social utility in rewarding trafficking victims who come forward and help law enforcement to end such practices. Why being kicked around by your own parents, usually out of status themselves, should be rewarded with a green card has always been a mystery.

An examination of the special juvenile decisions showed that in several of them the AAO regarded what would happen to the youngster were he or she to be sent to their native country as too dangerous for the people involved, where, in some cases, there would be no guardian or parent to help them.

As to the trafficking cases, in several the applicant, a relative of another applicant who had been denied, appealed following a reversal of that denial, and the relatives’ petitions were then accepted. In others, the AAO had a more tolerant definition than the staff had of losing eligibility because of an absence abroad of more than 90 days. (In many immigration cases, one must show a long-term stay in the States; being abroad for more than 90 days without a good reason can lead to the denial of a benefit.)

Applicants’ Criminal Records. We were interested in how many of the applications had been turned down because of applicants’ criminal records. In the vast majority of the study groups, we found one or no such references, but the minority of categories, where there were more, are of interest.

These were the alien groups in which a criminal record was most likely to be mentioned, using both the AAO codes and the descriptions of the categories, out of 20 decisions:

• Applications for waivers (criminal and related) (H-2): 18;
• Petition for U nonimmigrant status (crime victims) (D-14): 10;
• Application for Temporary Protected Status (M-1 ): 8
• Adjustment of status, Cuban nationals (A-2): 7
• Petitions for dependents of U-1 visa holders (D-15): 7; and
• Petitions filed for CNMI-only temporary workers (D-16): 7.

The category with the most references to criminal records (H-2), was the logical one; this is also one of the most-populated categories covered in our work, with 7,089 entries over the years, and the fourth biggest in Table 2. Of our 20 H-2 cases studied, 18 had criminal records. Our sample also had eight remands, which is well above the average.

Most of these eight remands were tough judgement calls. What do you do, for example, if the man seeking the waiver had a serious, but not violent, criminal record, and his wife, for whom he was the sole support, had been through six surgeries and was pretty sure that she would die were she to be deported to Sri Lanka along with her husband? This was one of the remands, with the text supporting the waiver, but the AAO person not making the call, as he or she could have done with a sustaining decision.

Interestingly, most of the categories with the highest mention of criminal records were either at the top of the income scale (one), or at the bottom (three).

At the high end of the scale we found Petition for CNMI-Only Nonimmigrant Transitional Worker (D-16), with the initials standing for the Commonwealth of the Northern Mariana Islands, a U.S. territory just north of Guam in the Western Pacific. All of the criminal records were attached to employers, not workers. This low-resource, high-corruption territory only recently came under Mainland immigration law and has, by far, the most foreign workers, proportionately, and the worst working conditions, of any place under the U.S. flag. The seven cases that we encountered in our sampling of 20 all related to a single once-prosperous casino, which has been ruled a criminal enterprise. It is controlled by Mainland Chinese, has been the subject of numerous FBI investigations, and has since been closed.

In another Chinese-dominated, high-income category we found a different kind of problem, which is not reflected directly in the tables. Of the 20 cases within the AAO code B-7 Immigration Petitions by Alien Entrepreneur (i.e., investor) category, while there was only a single reference to a criminal record, in many of the cases an alien’s application was turned down because the alien was not able to adequately explain where he or she got the $500,000 needed (at the time) to qualify as an alien investor. One must prove in the EB-5 (immigrant investor) program that one secured the money involved legitimately. Given the controversial nature of the program, we took a look at a larger group of these cases, 41 in all. We found that the issue of the legitimacy of the funds was raised in 31 of the cases, or more than 75 percent of them. Of the 31, 28 resulted in denials, two in sustained appeals, and one in a remand. China provides the overwhelming majority of the would-be investors in the EB-5 program.

At the other end of the income spectrum, we have relatively high numbers of criminal records in the decisions about two phases of the U (crime victim) program and in the TPS program. In all three, the applicants are largely illegal aliens from low-income countries who claim a particular kind of victimhood. These include those seeking U-1 nonimmigrant visas as crime victims, and the follow-on grants of green cards to those with the U nonimmigrant visas. The third set of victims happened to be in the U.S., usually, but not always, illegally, when a revolution or earthquake or big storm damaged their home country, such as Haiti or El Salvador. They were in the TPS program.

The Biden administration is seeking to expand both the U visa and TPS programs.

Reasons for the Underlying Denials in Specific Categories. Some of the AAO decisions were easier to make than others.

At the easy end of the scale, some appeals, often filed without the help of a lawyer, simply said that the staff decision was wrong, but presented no evidence to support that statement. These got summary dismissals. The wording of these denials is pure boiler-plate and requires little of the officer signing the document.

In other relatively easy cases, the aliens’ appeals and legal timing were out of sync; many simply were filed too late. In others, different kinds of timing issues presented themselves. Some of the special juveniles aged out, by filing after they turned 21. In a mirror image of that problem, some of the would-be battered parents (AAO code B-9) reported that they had been attacked by their own children, but those attacks came before the assailants were 21, and hence did not count. (The abuse has to be at the hands of U.S. citizens who are 21 or older.)

At the other end of the easy-to-difficult spectrum were ones dealing with complex economic issues among would-be immigrant investors; the worst ones for the AAO officers were the ones in which criminal records, on one hand, had to be balanced against real hardship issues on the other. Generally, the more difficult the decisions, the longer the text and the more work to be done by AAO.

Conclusions. Our sense is that AAO’s work overall is adequate, that justice is handed out more or less as it should be, and that the high rates of denials are appropriate at the appeals level, given the generosity of the underlying staff decision-making process.

However, many reporting problems are present, as we next note, and the AAO system is needlessly self-sabotaged by its obsession with privacy.

Methodology. The data on which this study depends was difficult to work with. It resembled, if there is such a thing, a multi-compartmented dumpster with a near-paranoia for privacy.

It is not AAO’s fault that immigration law, particularly in the U.S., is a many-splendored thing, having been written (largely by migration advocates) over the years in a piecemeal way, with many a special provision for a special class of migrants or their employers. So having 80 or so different categories of appeals comes with the territory.

But this dumpster could be better organized. In some categories, if not all, the cases show up in more or less reverse chronological order, but with many older cases mixed in with new ones. One cannot count on any predictable sequencing of cases, which makes year-to-year comparisons difficult if not impossible. There are also some cases that show up more than once. Then there are often sprinklings of other kinds of cases than those supposed to be in a given set.

For example, we wanted to include the AAO set of decisions on NAFTA workers, which is AAO category D-11. These are people from Canada and Mexico who could qualify for H-1B, were they to apply that way. But we ran into a problem: The first 10 cases we summoned up, in our usual 4-8-12-16 etc. sampling included not a single NAFTA case but eight H-1Bs, one L-1A, and one H-2B. We also looked at the first case in the set; it was regarding an H-2B petition. We felt we had to eliminate the category and did so. Here’s a link to the start of that set of decisions.

Although all the examined D-11 cases were misfiled, they also showed up in an odd chronological order; the first case was 2020, and the next several dozen all dated back to 2005, but within 2005, they were not in consistent reverse order.

And no other semi-judicial system in the U.S. worries so much about privacy. I am not sure that I object to the aliens being anonymous, but with AAO, so are the judges, the lawyers, and any corporations involved. In recent years, the location of cases, the nationality of the aliens, and sometimes the birth dates of the aliens are scrubbed from the record. So are specific financial transactions. In earlier years, the privacy virus was always evident, but some of the older cases published locations, birth dates, and sometimes the nationality of the alien.

In contrast, the approval/denial averages on asylum cases for individual immigration judges has been published by an outfit associated with Syracuse University, the Transactional Records Access Clearinghouse (TRAC) for all to see.

Once we established the 33 classes of decisions to study, moving backwards from the most recent cases, we examined every fourth case until we reached 20 valid ones; thus it was case four, then case eight, then case 12, etc. If, when at the level of the 12th case in a sequence, for instance, we found a duplicate or a mismatch, we took the next case and then resumed with the 16th one. If there were fewer than 80 cases in a category, we adjusted our sampling to reflect the lower number, but we had to do that in only a few categories.

In the tables that follow, we list the total number of cases on file as of June 21, 2021, with Table 2 covering the categories that were studied and Table 3 listing those that were excluded.