Straws in the Wind Suggest Immigrants Marry More than Natives

By David North on September 7, 2012

We have long known that immigrants have higher birth rates than natives of the United States, as my colleague Steve Camarota documented in a CIS Backgrounder a few years ago, but now there are some straws in the wind suggesting that they may also be more likely to marry than natives of the United States.

One reason for this is that marriage is a large — probably much too large — part of the immigration admissions formula in the United States. Of the million or so legal immigrants each year, fully 300,000 arrive as a direct result of marriage to a U.S. resident. I am not talking about the arrival of married couples as immigrants — that is an additional number. What is described here are the 300,000 new immigrants a year (on average) who are newly married to, or about to be married to, 300,000 U.S. residents. See this earlier blog of mine for more on the statistics on immigration through marriage.

Frankly, 300,000 looks like a very large number if you compare it with the total nation-wide number of marriages in a year, which was 2,077,000 in 2009, according to the Department of Health and Human Services.

There is an element of oranges and tangerines (if not oranges and apples) to the 300,000 and 2,077,000 numbers noted above; they are recorded by different agencies and some of the 300,000 take place overseas and do not overlap with the 2,077,000 while some of the 300,000 do.

What happens if you use a single data source, and totally unchallengeable numbers? I do not like to use Census data on such things because people can lie to the census taker on whether they are married or not; the issuance of a marriage license, on the other hand, is a hard fact, easily and routinely counted by reliable government agencies.

With all that in mind I have looked at the marriage license records in two locations, my home county, Arlington, Va., and Vermont, which has open marriage and divorce records. I decided to measure something that I doubt is tabulated anywhere: the relative incidence of three kinds of marriages:

  • Marriages of native-born to native-born (NB-NB);

  • Marriages of foreign-born to foreign-born (FB-FB); and most important,

  • Marriages of foreign-born to native-born (FB-NB).

By definition, all marriages fall into one of these three categories, with the bulk of marriage fraud taking place in the last category. There also can be some immigration fraud in the FB-FB marriages if one of the foreign-born partners has become a citizen, but because place of birth, not citizenship, is what you find in most marriage data sets you cannot study alien-citizen marriages from marriage license data, only FB-NB marriages.

With Arlington, thanks to the cooperative Clerk of the Court, Paul Ferguson, I examined paper records and drew a sample of 250 of the marriage licenses issued by the county in 2008. With Vermont, thanks to the help of Cindy Hooley, Vital Statistics Information Manager for the Vermont State Health Department, I worked from a printout of all the 5,320 marriage licenses issued in that state in 2007.

Arlington is an immigrant-rich environment, as Vermont is not, so I expected more aliens involved in marriage licenses there, and I was right. Further, matters are complicated in Arlington, which is just over the river from Washington, D.C., because many District residents find that the Clerk's Office in Arlington is more efficient than the comparable agency in D.C. As a result, Ferguson tells me, some D.C. residents come to Arlington for their licenses and some even get married there during lunch hour and go back to work in the afternoon — efficient if not very romantic.

For each jurisdiction, I first worked out the relative incidence of the three classes of marriage and this is what I found:

Class Arlington Vermont
NB-NB 45% 86.3%
FB-FB 28% 2.5%
FB-NB 27% 11.2%

What I find interesting in these numbers is the sharp difference between the FB-FB and FB-NB rates in the two jurisdictions. Straws, of course, but interesting ones. I guess with the relatively small FB population, young aliens in Vermont have fewer opportunities to marry other aliens, and must gravitate to citizen mates. This would suggest that integration through marriage is more likely to take place when the percentage of the foreign-born is small.

In my next step, I added up the native-born and the foreign-born participants in weddings. Then I compared those percentages to Census data on the extent of the foreign-born in the jurisdiction, with these results:

  Arlington Vermont
FB percent of all wedding participants 41.5% 8.1%
FB percent of population (Census) 27.8% 3.9%

In both jurisdictions, one with lots of immigrants and the other with few, the percentage of foreign-born wedding participants was much higher than the foreign-born share of the whole population. There are at least three factors at work here:

  1. A higher percentage of the immigrant population is probably in the marrying age range than is the native-born population, something I am still exploring;

  2. The cultural and social pressures toward marriage, as opposed to just living together, are probably stronger in the immigrant communities than in the population as a whole; and,

  3. Finally there is the factor that a NB-FB marriage is likely to produce a visa for one of the partners, which is not true in all NB-NB and in most FB-FB marriages; in many cases this is a valid visa because the relationship is genuine, but in the unknown number of fraudulent marriages, it is not.

The next step will be an interesting one: Will the divorce rates in Vermont be different for the three classes of couples? In other words will the NB-FB couples divorce at a faster rate than the couples in the other two groups?

That, for Vermont, is potentially knowable in the immediate future. It is potentially knowable for the Arlington class of 2008 — but only in the year 2033 — as Virginia is one of a number of states that favors secrecy (or privacy) regarding its divorce data. But even Virginia loses interest in privacy concerns after the passage of 25 years.