Mario Aguilar shows photos of his four sons as he waits for work outside the Casa del Migrante shelter for migrants in Tijuana, Mexico. Aguilar, a Mexican citizen, had been living in the United States for 28 years before he was deported.

On November 20, 2014, US President Barack Obama used his executive authority to make sweeping, though non-permanent, changes to the US immigration system without legislative action by Congress. The resulting Immigration Accountability Executive Actions delay the deportation of over four million unauthorized migrants living in the United States. To qualify for the new category of deferred deportation, migrants must have been living in the United States for more than five years without certain criminal convictions and have US citizen or legal permanent resident children. The president also de-prioritized the deportation of some categories of unauthorized migrants and streamlined the process for people with US family members to obtain visas to stay in the country legally.

At the same time the executive actions did not address existing border removal policies that punish and provide no relief to families that have already been separated by harsh and overly inclusive deportation policies. Since 2008, the Obama administration has deported more than two million non-citizens. The government classifies a growing number of these deportations as “border removals,” a category that appears to comprise anyone removed after apprehension by US Customs and Border Protection (CBP), including the Border Patrol, despite a person’s previous ties to the United States. The executive actions give priority to the often summary deportation of noncitizens apprehended at the border unless they qualify for asylum or other relief or unless there are “compelling and exceptional” factors indicating the person is not a threat to national security, border security, or public safety. This meager loophole appears to do little to end summary deportations of people apprehended at the border with deep ties to the United States.

Human Rights Watch has previously documented the harms caused by the Obama administration’s aggressive immigration enforcement at the border on people seeking to join their US citizen children or other loved ones. The present analysis is based on CBP data Human Rights Watch recently obtained through the Freedom of Information Act on the nearly 700,000 adult and child migrants that the CBP apprehended along the US border in fiscal years 2011 and 2012, the most recent data Human Rights Watch obtained.

Human Rights Watch estimates that more than 101,900 of these migrants – about 15 percent – were a parent of a US citizen child. Collectively, these migrants say they are the parents of more than 202,000 American children. Though having a US citizen child does not mean in all cases that a non-citizen has substantial ties to the US, it is an indicator of the strong possibility of such ties.

This rate applies only to apprehensions at the border during these two years. CBP apprehensions accounted for approximately 52 percent of non-citizens apprehended by immigration authorities in fiscal years 2011 and 2012, with the others apprehended in the interior of the country by Immigration and Customs Enforcement’s offices of Enforcement and Removal Operations or Homeland Security Operations (ICE ERO). The estimated numbers and rates of parenthood status for those apprehended by CBP cannot be generalized to represent all immigrants apprehended by the Department of Homeland Security.

Parents of US Citizen Children Often Subject to Summary Deportations at the Border
Few parents of US citizen children apprehended at the border are processed for deportation by CBP in a way that allows them the chance to explain their case to an immigration judge. Sixty percent of parents with US citizen children apprehended at the border are processed under two summary procedures – expedited removal and reinstatement of removal – that allow them no opportunity to make claims about their ties to the United States.

Expedited removal allows for the summary expulsion of non-citizens who have not been admitted or paroled into the US, who have been in the US for less than two years, and who present fraudulent documents or have no documents. Twenty percent of apprehended parents – the parents of 36,000 US citizen children – were processed for deportation using expedited removal in fiscal years 2011 and 2012. Migrants, including parents, processed under expedited removal may be deported immediately, have no right to see an immigration judge unless they present a credible fear of returning to their country of origin, and upon deportation will be barred from returning legally to the US for at least five years, and often much longer.

Migrants – including parents of US citizen children – who enter the US after a previous deportation may face the reinstatement of the previous removal order, a similarly accelerated procedure in which they may be deported immediately and are not able to receive a new hearing in front of an immigration judge unless they present a reasonable fear of returning to their country of origin.

Consistent with previous Human Rights Watch research, the new data suggest that parents of US citizen children try again and again to rejoin loved ones. In the time period analyzed, parents of US citizen children were 1.77 times more likely than non-parents to be processed under a reinstatement of a previous removal order. Human Rights Watch estimates that more than 40,000 parents of US children apprehended by CBP – 40 percent – in 2011 and 2012 were processed with a reinstatement of a previous removal order. Of the total number of migrants placed in reinstatement of removal during the two years, more than 20 percent were parents of US citizen children.

Nearly a third of apprehended migrants, almost all from Mexico, are removed through a process called “voluntary return,” in which Border Patrol agents allow a non-citizen to depart voluntarily without going through a formal deportation. Processing time is shortest for voluntary return, and the non-citizen will not face immigration penalties, such as being barred from entering the US legally at a later date. Parents of US citizen children were deported through voluntary return at similar rates to people who were not parents of US citizen children.

Fewer than 10 percent of the parents of US citizen children apprehended by CBP in the years under analysis were allowed a hearing before an immigration judge in which they could request relief from deportation based on their close family ties.

Parents of US Citizen Children Prosecuted at Higher Rates for Illegal Entry, Re-entry
Illegal entry and illegal re-entry, taken together, are now the most prosecuted federal crimes in the United States. In “Turning Migrants into Criminals,” Human Rights Watch detailed the human impact of these prosecutions, including the impact on American families. Nearly one in four apprehended migrants in the dataset Human Rights Watch used for that report – which includes those with and without US citizen children – faced criminal prosecution for illegal entry, re-entry, or both, in addition to deportation. Reports that the Obama administration’s executive action plan is expected to increase security resources at the US border with Mexico do not specify any changes to the administration’s policy promoting these prosecutions. The new 2011 and 2012 data analyzed by Human Rights Watch show that these prosecutions disproportionately affect people who are parents of US citizen children.

One in five of migrants charged with the crime of illegal entry or re-entry say they are the parent of a US citizen child. Over the course of the two fiscal years under analysis, this adds up to more than 32,000 parents of US children referred for prosecution for entering the country where their children live. Parents of US citizen children were 1.35 times more likely than non-parents to be charged with an illegal entry or re-entry crime.

In the two years under analysis, more than 17,000 parents of American children were charged with illegal entry and another 16,000 with illegal re-entry. Nearly 1,400 parents were charged with both.

A migrant convicted of illegal entry or re-entry may be targeted for deportation as a “convicted criminal.” Indeed, in 2012, 24 percent of so-called “criminal aliens” deported had immigration crimes like illegal entry or re-entry as their most serious conviction. In fiscal years 2011 and 2012, only 0.2 percent of parents of US citizen children were charged by the CBP as deportable because they had been convicted of a more serious “aggravated felony.” Only 0.2 percent were charged with a drug trafficking offense. Parents of US citizen children account for 6 percent of the migrants charged with drug trafficking.

In response to a Freedom of Information Act request, CBP provided Human Rights Watch with a dataset containing information on more than 6.7 million border apprehensions between October 1, 2003 and September 30, 2012. For this analysis, Human Rights Watch only used data from October 1, 2010 through September 30, 2012, because of the level of incomplete data as described below. The unit of analysis is each “apprehension” of an individual. Therefore, individuals apprehended multiple times would appear in the dataset as multiple entries. The data was provided as a 1.28 GB Microsoft Access database of 11 variables: date of entry, date of apprehension, location of apprehension, country of citizenship, gender, marriage status, number and nationality of children, DHS status, entry status, disposition, and statute under which the migrant was charged.

Data Cleaning
The data provided by CBP required a great deal of data cleaning and processing before it could be analyzed. Several variables in the database were clearly open-text variables, which allowed officers to type in their entries. These resulted in extremely “dirty” data. For example, the “city/state apprehension location” variable had 34,279 unique values, including many different entries for the same place (“El Paso, TX,” “elpaso, texas,” etc.). Following data cleaning, 691 city/state apprehension locations were used in the final dataset, though some locations such as “MILE MARKER 5 AT FR 35” did not correspond to a city/state and were coded as “N/A.”

The statute variable, while originating from what is likely a drop-down menu, and not open-text, was provided in a complex manner. Any individual apprehension may have been related to multiple statutes, and all relevant statutes were combined into a single variable for each apprehension. The dataset had 7,037 unique combinations of statutes. The Human Rights Watch analysis examined eight statutes of interest. For each, a binary “true/false” variable was created identifying whether the apprehended individual was charged under each statute.

Number and Nationality of Children
CBP included a variable for the self-reported number and nationality of children of the person apprehended. It is unknown how this variable may be affected by self-reporting bias. Some migrants may have feared telling a CBP officer about their children while others may have lied to influence their case. This analysis does not adjust for any bias and treats border patrol agents’ entries as true.

For the 1.9 million apprehensions for which this variable was complete, the child variable was entered into CBP’s database 100,038 different ways. Data was cleaned thoroughly, and though it was not feasible to recode every entry, more than 99 percent were re-coded. The child variable was eventually re-coded into two variables: number of US citizen children and number of non-US citizen children. The number of non-US citizen children was not used in the missing data imputation or any other analysis.

Missing Data and Imputation
Missing data in CBP’s database was of greater concern than “dirty data.” On this point, the CBP has vastly improved in recent years. Several variables – DHS status, marriage status, and number and nationality of children – were about 80 percent incomplete in 2004, but the level of completeness has improved nearly every year since. By 2012, DHS status and marriage status were nearly complete. The statute variable is also missing in a number of cases, but nearly all relate to people who were deported as “voluntary returns” and therefore not charged under a statute.

The level of completeness for the “number and nationality of children” variable is still of serious concern. In 2012, 25 percent of people apprehended were still missing data for this variable. This data is not missing simply because the apprehended migrant did not have any children as the vast majority of the completed data was entered in some version of “no children.” Human Rights Watch believes the data to be missing at random, due to the officers simply not being thorough, as there is no clear evidence that the data is missing at higher rates in certain CBP offices – as determined by apprehension locations, or on certain dates, or due to the demographics of those apprehended. Rather than remove the records with missing data from the analysis, which would distort the findings, Human Rights Watch used an analysis called missing data imputation to estimate the number of US citizen children.

Imputation requires using the data that is present to estimate the values for the missing data. It makes the assumption that apprehended migrants with missing data for the child variable will have children at similar rates to the other migrants in the database. Human Rights Watch used the Multivariate Imputation by Chained Equations (MICE) package in R to conduct the imputation. The number of US children variable was imputed using predictive mean matching (pmm). The variables used to impute the missing data were location of apprehension, date of apprehension, gender, nationality, marital status, statute, and disposition. Five imputations were run, resulting in a range of values for each individual. Among the five imputations, the total aggregate number of apprehended migrants with US citizen children ranged from 101,425 to 102,752. The imputation with the median value was used in the analysis for any cross-tabulations between the child variable and other variables.