How Many Separated Children are in Custody?

The Department of Homeland Security’s own reports indicate that over 4,200 children have been separated at the border, much higher than the current official figure of fewer than 3,000.  Also there must be many more than 102 children under five years of age.

Since the introduction of the policy of “zero-tolerance” on our southern border, the Department of Homeland Security (DHS) has released a set of inconsistent reports about the number of children taken into custody and the number reunited with family members.  Now that a Federal judge has ordered that all the separated children be reunited expeditiously, the Department of Health and Human Services, in whose custody parent-less child immigrants are placed, now states that the number in their care is “fewer than 3,000.”  That figure is not consistent with the existing reports from DHS.

In a telephone call with reporters DHHS Secretary Alex Azar said on July 5th that an extensive review led him to believe his agency had custody of “fewer than 3,000” separated children.  Yet a week earlier DHS had reported just over 2,000.  Azar attributed this discrepancy to the court order requiring “officials to go back further in time and comb through thousands of cases to find any separated children.”

We have one good estimate of the number of children taken prior to zero-tolerance.  On June 29th DHS told NBC News that a “pilot program” had already separated 1,768 children through February, 2018,  a figure which had probably grown to 2,000 by the time zero-tolerance took effect on May 7th.

Also the last DHS report on the number of separated children ends on June 9th, eleven days before the President’s ordered an official end to separating families.  Extending the DHS figures through June 20th adds nearly 700 more children to the total.  In the sections below I will review the DHS reports then use them to estimate the total number of separated children.

The Evidence

Here is a compilation of the varying reports on the number of separated children from the Department of Homeland Security:


According to the NBC News report, nearly 1,800 children were already in custody by the end of February, 2018, weeks before the Attorney General ordered “zero-tolerance” for all migrants.  Separations were fairly rare in this period though, averaging fewer than five per day between last October and February of this year.

DHS issued two overlapping reports after the imposition of zero-tolerance.  One covered the period from April 19th until May 31st; the other covered May 5th through June 9th.  Even this second report did not include the entire period of zero-tolerance which lasted until President Trump’s order on June 20th.

Reports since then have been inconsistent.  On June 20th DHS reported 2,053 children in Health and Human Services custody; a week later that figure fell by just six to 2,047.  Yet in between those two reports, on June 23rd, DHS reported that over 500 children had been reunited with their parents, which should have brought those totals down considerably.  Finally we have Secretary Azar’s announcement of fewer than 3,000 children in custody as of July 5th.


The Estimates

I have tried to reconcile these figures as best I can.  I use the daily separation rate for periods when it is can be calculated to estimate the number of separated children for periods when data are not available.   Including the “pilot” program, and unreported periods during zero-tolerance, I estimate that nearly 4,300 separated children remain in government custody, more than forty percent higher than Secretary Azar’s figure:


As an example of the calculations involved, take the number of children before zero tolerance began.  Official DHS figures cover the period from October, 2016, through February, 2018.  That leaves some 66 days between the end of that report and the onset of zero tolerance on May 7th.   I use the estimate of 4.7 separations per day from the second DHS report to estimate that another 309 children were separated from their parents during those 66 days.  The same method is used to generate the other figures marked “Estimate” in the table above.  Note that using the 4.7 per-day rate is probably a conservative estimate of the rate of separations between March 1st and May 5th.

On June 23, DHS reported that 538 children had been reunited with their parents “as part of the Zero Tolerance initiative.”  I have attributed all those children to the zero-tolerance period.  (There is no data on the number of reunited children taken before zero-tolerance began. I return to that issue below.)  If we assume that reunifications continued at the same pace through July 5th, another 126 children would have been returned to their parents between June 23rd and July 5th, raising the estimated total to 664.

Adding together the estimates for separations before and during zero tolerance gives us the total number of children reportedly detained at the border since late 2016. Subtracting the estimated number of family reunions provides the estimate for how many separated children likely were still in government custody as of July 5th.  That figure of 4,290 is some forty-three percent higher than Secretary Azar’s estimate of  “fewer than 3,000.”

I have not found figures for reunifications prior to zero-tolerance.  Some of the 2,077 children taken before May 7th may have been reunited with their parents.  But even if we assume as many as a quarter of them were placed back in their parents’ arms, that would still leave the total number of children separated from their parents closer to 4,000 than to Azar’s 3,000.


How many children are under five years of age?

The court order mandated that children under the age of five had to be reunited with their parents fourteen days after the order went into effect, or July 10, 2018; for older children the deadline was thirty days (July 26).  Azar told reporters in the telephone call-in that only about 100 of the children in DHHS custody, or about three percent, were under five.  This past weekend, under order from Judge Sabraw, the government turned over to the ACLU a list of 102 children that purportedly included all those under five in the government’s custody.  This figure is implausibly low.

First, even if the proportion of young children was only three percent, my estimate of 4,290 total separations implies that the total number of young children should be closer to 143.  However the three percent figure itself is highly suspect given the age distribution of children living in the three countries from which most asylum-seekers emigrate — El Salvador, Guatemala, and Honduras.

The CIA’s World Factbook reports the “population pyramids” for every country.  Here are the charts for those nations:


The bottom four bars of each graph depict the numbers of boys and girls in four age groups, 0-4, 5-9, 10-14, and 15-19.  In both Guatemala and Honduras about a quarter of all children and teens are under the age of five.  For El Salvador, whose birth rate has slowed in comparison to the other two countries, young children constitute a somewhat smaller share.

Now we can certainly imagine that worried parents might be reluctant to carry a two-year-old child across hundreds of dangerous miles to seek refuge in the United States.  That would imply that the age distribution of the separated children skews somewhat older than the source populations from which they are drawn.  Even with such an older skew though, Secretary Azar’s figure of just three percent remains highly dubious.  If the proportion of separated young children were just ten percent, or half or less of the actual proportion of very young children living in these countries, there should be over four hundred children subject to the July 10th deadline, not 102.

One other feature of the DHS reports also bears mentioning.  Notice in the first chart that the numbers of separated children and parents are nearly equal. That ratio is consistent with either single parents carrying a single child, or couples with two children. It is unlikely that most of these are single-parent families given the high proportion of Catholics in the source countries and the consequent low divorce rates like those reported for Guatemala. Some of these parents might have left their spouses behind to look after other children or have come to be reunited with a spouse already in the United States.