Wyoming can now add another entry to its list of fun facts: According to my research, it had the highest rate of second-generation outmarriage of any state from 1880—1930.

As detailed in the previous posts in this series, my research measures historical assimilation by looking at how the children of immigrants in the United States chose to marry. After compiling millions of census records across six decades, I analyzed how many second-generation Americans chose to marry someone from a different immigrant community or whose family had been in the United States for at least two generations.

This entry focuses on how second-generation marriage patterns varied by geography, a topic I recently explored in a piece for City Journal. You can read that piece by clicking on the button below, or see a summary with graphics here.

Read the City Journal piece

Outmarriage was high in the South and the West, low in the Southwest, the Midwest, and the Northeast

Marriage patterns followed a surprisingly consistent regional divide but had some puzzles. After adjusting for the local concentration of second-generation Americans, the Coastal West had some of the highest rates of outmarriage while the Southwest had the lowest. The Mountain West significantly overperformed the neighboring Midwest, while the Deep South saw higher rates of outmarriage than the putatively cosmopolitan Northeast.

The states with the highest rates of outmarriage were Wyoming, Oregon, and Washington. The states with the lowest were Hawaii, Arizona, and New Mexico.

Some of this variation can be explained by small sample sizes and unique circumstances. In Hawaii many second-generation Americans were of Japanese heritage, who married within the Japanese American community at exceptionally high rates.

New Mexico, Wyoming, Idaho, and Arizona also had relatively small sample sizes. If we limit ourselves to states with large sample sizes, the top three states for outmarriage were Oregon, Washington, and Utah, while the bottom three were Texas, Wisconsin, and North Dakota.

Why was the West better for assimilation?

Three factors likely contributed to the West’s overperformance on outmarriage. First, there were plenty of young single men in these states during this time period but not as many young single women. This meant that the men couldn’t be too choosy about their partners, and as I covered in part 1 of this series, men were more likely to marry outside of their ethnic communities to begin with.

Second, the Coastal and Mountain West represented a more individualistic culture and new beginnings. In contrast, the ranching culture of the Southwest often led to tight kinship networks that were difficult to penetrate.

This had an effect even for notably insular groups. Mexican Americans had the lowest outmarriage rates of any major ethnic group in the data, but moving West had a significant effect. On average, only 17% of Mexican Americans in the Southwest married outside of their community. In California, 30% did.

The West also benefited from having a relative lack of immigrant and ethnic institutions. One of the only institutions in the Mountain West was the Mormon church, which seemed to break down ethnic barriers among its second-generation adherents. Continental Northern European groups like Swedish and German Americans tended to have higher outmarriage rates in Utah than they did in Midwestern states, where religious denominations often carried ethnic associations.

What about the South and the Northeast?

As in Utah, the Deep South’s high rates of second-generation outmarriage can partially be explained by a strong local culture that gave immigrants there something to assimilate into.

The South’s effect was best reflected in the outmarriage rates for Italian Americans, who were much more likely to marry outside of the Italian American community in places like Louisiana or the quasi-Southern state of Missouri than they were in places like Rhode Island or New York.

The Northeast had a distinct regional culture as well, but various factors created immigrant community enclaves that slowed outmarriage and other forms of assimilation.

These findings all reflect loose, general patterns and not hard or eternal truths. In my City Journal piece, I discuss the implications this research has for contemporary assimilation. Read it here, and subscribe to Points of Entry to get the final installment in this series in your inbox when it comes out.

Subscribe now

Proponents of the “decline” theory point out that Irish America has been shrinking in both absolute relative numbers; that powerful Irish American politicians like Ted Kennedy and Tip O’Neill are long gone; and that the 1998 Good Friday Agreement, which brought an end to the Troubles in Northern Ireland, ironically deprived Irish America of its primary animating cause.

Critics of this narrative counter that Irish American identity is changing and adapting but not vanishing, that a new wave of young Irish American politicians like Brendan Boyle and Brian Fitzpatrick are keeping Irish America politically relevant, and that the prospect of Irish reunification could reanimate diaspora interest in Ireland in the future.

Which side is right? I compiled the latest data on Irish Americans from the American Community Survey (ACS) and built an interactive guide to the current state of Irish America to find out. You can see the data for yourself here, but these are my five biggest takeaways.

Explore the Data

Irish American numbers have declined, but are holding steady.

In 1980, 15.3% of the U.S. identified as Irish in some way. As of the latest ACS data release from 2024, that number has gone down to 9.54%.

The total number of Irish Americans has held steady since 2015, at around 32 million. This identity is fragmented, however: Only 25.6% of Irish Americans identify solely as Irish. 33.8% identify as primarily Irish, while 40.6% say they are primarily something else in addition to Irish.

Irish Americans vastly outnumber recent Irish arrivals.

In 2024, about 143,000 U.S. residents were born in Ireland. Given that there were 32.4 million Americans who identified themselves as having Irish heritage, this means that there were roughly 227 Irish Americans for every Irish immigrant.

When I traveled Ireland in 2022 as part of my master’s program at Queen’s University Belfast, I consistently heard that young Irish people today, unlike previous generations, were more interested in destinations like Australia and Canada than the U.S. While the lack of fresh arrivals has probably helped solidify Irish American assimilation, in the long run it may also create a wider gulf between the Irish American diaspora and the Irish population.

Decline in Irish identity is due to changing norms, not immigration.

It would be easy to look at the data so far and assume that Irish American identity has declined due to increases in immigration from newer groups hailing from Latin America and Asia.

But the states with the greatest rates of decline in Irish identity have almost all been in Appalachia and the South, which have relatively low immigration rates. That indicates a shift among those states’ people of Irish descent, who now have a greater tendency to identify as “Scots-Irish” or just “American.”

In the future, we may see the opposite trend happen if more young people choose to actively identify with their Irish heritage regardless of how much descent they can actually trace back to Ireland.

There are two Irish Americas.

Among Americans who claim Irish ancestry, there are important demographic differences between those who are “solely” Irish and those who pair it with something else.

“Sole” Irish Americans are over a decade older on average and have a slightly greater concentration in the South. This population grew slightly from 2019 to 2024, from 8.1 million to 8.3 million.

Irish Americans who were primarily Irish and secondarily something else declined during the same period, from 11.5 million to 11 million.

The share of Americans who were primarily another ancestry and secondarily Irish grew from 12.8 million to 13.2 million from 2019 to 2024, and over time any significant growth in the Irish American population will happen in this sector.

Being Irish pays (unless you live in the Dakotas).

In nearly every state, Irish Americans earned more than the median state wage. The only exceptions were the Dakotas, where self-identified Irish Americans earned less.

(I confess I’m not sure why that is.)

If you have any ideas about what’s causing the Dakotas effect, feel free to comment below or email me at gguerra@niskanencenter.org. If you have no ideas about what’s causing the Dakotas effect but still have questions, compliments, or complaints to offer, feel free to do the same.

Wishing you all a merry St. Patrick’s Day, whether you are part of the 32.4 million or not!

Subscribe now

Despite this, there has been a dearth of easily accessible information about Iranian migration to the U.S. that would allow us to make informed decisions about security and humanitarian matters. To fix that, I built an interactive dashboard that tracks Iranian irregular migration across the Western Hemisphere in near-real time, pulling from government databases in Brazil, Colombia, Panama, Honduras, and the United States.

Explore the Dashboard

Here are the five things that matter most right now.

Encounters with Iranian nationals have virtually collapsed since Trump took office.

Since 2000, Customs and Border Protection (CBP) has recorded roughly 2,180 encounters with Iranian nationals. This is a small flow by any measure, but it grew from 50 annual encounters in Fiscal Year (FY) 2021 to a peak of 800 in FY 2024.

This pattern was mirrored throughout the Americas. In January 2025, transit countries recorded 452 Iranian migrants across the corridor — Colombia alone saw 180. Four months later, the combined total was 17. Colombia’s count fell to zero, Honduras dropped from 78 to 3, and Mexico went from 76 to 6.

Enhanced enforcement measures at every step of the journey have practically brought irregular Iranian migration through the Americas to a halt.

Recent strikes are unlikely to create a surge of Iranian migration to the U.S.

This collapse is unlikely to reverse itself even if a large number of Iranians choose to leave Iran in the coming months. The trip to the U.S. is estimated to cost tens of thousands of dollars, and as we’ll see, the likelihood of a successful asylum outcome is low.

Other countries, such as Turkey, Germany, and Canada represent far likelier options given their relatively cheaper costs, easier logistics, and clearer pathways. In the Americas, a small number may choose to apply for refugee status in Brazil or Argentina, but those seeking to do so with the intent of crossing by land to the U.S.-Mexico border will face steep odds.

Three scenarios are worth tracking. With the present enforcement measures in place, I expect the corridor to stay closed and Iranian flows to remain near zero. If one or two countries miss a surge of migrants, small upticks could follow but would be contained by remaining bottlenecks elsewhere. Even a full-scale displacement event from Iran would likely fall within historical parameters, with a ceiling of about 600 total encounters across all sources.

The security scenario policymakers should be worrying about isn’t a sudden rush of Iranian migrants arriving in the Americas.

The real security threat comes from established passport forgery practices.

The Iranian nationals who appear in this data carried Iranian passports, submitted to processing voluntarily, and identified themselves as Iranian. Screening these individuals is a manageable task for a competent border security apparatus, and they likely represent genuine humanitarian and economic migrants.

A more serious problem lies in the ongoing practice of passport forgery by Iranian smuggling rings. In April 2023, Brazil’s Federal Police disrupted a smuggling operation that moved Iranians through the hemisphere on forged Canadian passports — a full-journey operation with costs running into the tens of thousands of dollars per person. In an earlier case, seven Iranians were intercepted at the Brazil-Peru border carrying forged Israeli, Danish, and Canadian documents. Investigative reporting has identified the Tri-Border Area as a hub for Iranian document forgery networks, with fraudulent documents costing up to $60,000.

While asylum seekers who do appear in the Americas should be vetted, security resources would be better spent on document fraud detection, international law enforcement cooperation, and disruption of the smuggling networks that facilitate covert entry.

Most Iranian migrants are middle-aged men, some appear to be Afghan in origin.

The demographic profile of Iranians encountered across the corridor is narrow. Between 68% and 77% are male, depending on the transit country; two-thirds are over 30, and about 10% are minors.

Interestingly, the share of Dari speakers among Iranian-documented migrants in US immigration courts has been roughly 10% in the past two fiscal years. Dari is the primary language of Afghanistan, not Iran. A share of people presenting Iranian documents appear to be Afghan nationals who resided in Iran, a population of over three million.

Iranian asylum seekers face a tougher pathway.

The small number of Iranians who reached US soil before the corridor closed now face a sharply different adjudication environment.

In FY2024, immigration courts granted asylum to 85.6% of Iranian applicants. By the first quarter of FY2026, that rate had fallen to 27.0%, a drop of nearly 59 percentage points in under two years. Denial rates climbed from 14.4% to 73%.

The sample sizes are small (263 cases in FY2024, 122 in early FY2026), but the trajectory is clearly moving in the direction of increased denials.

The dashboard will continue to track these flows as conditions develop. The indicators to watch are Brazil’s airport entry data, Colombia’s transit numbers, and Honduran irregular encounter numbers. If displacement from Iran to the Western Hemisphere accelerates, those are where it will appear first, and the data from those sources will tell us whether the enforcement architecture holds or whether the corridor finds a new shape.

Subscribe now

Explore the Data

In total, 11 of our 33 medals involved an athlete with some sort of immigrant connection, whether that came in the form of being born abroad themselves or in the form of having at least one immigrant parent. Interestingly, immigrant connections were most common at the gold-medal level, where 50% of our medals were tied to athletes with immigrant connections, and least common at the bronze level, where only about 22% did.

Another interesting discrepancy was that our gold-medal-winning men’s hockey team was nearly a quarter second-generation, while I couldn’t find any athletes with an immigrant tie on the women’s hockey team.

Five of the six athletes with an immigrant connection on the men’s hockey team had heritage that traced back to Canada. Canadian Americans formed the largest group of medalists, thanks to their presence on the men’s hockey team, with no other group coming close.

Our second-generation representation even beat out heritage countries in two cases: Mexican and Argentine Americans won medals at the Olympics, but Mexican and Argentine athletes did not.

Explore the full dataset here.

Subscribe now

In the 1920 census, only 6.2% of Italian American men were married to Irish American women. In Illinois, where Capone and Coughlin settled so that Capone could pursue his career as a “used furniture dealer,” the proportion was an even lower 3.9%.

Subscribe now

Why did Irish and Italian Americans marry each other at far lower rates than we’d expect given their shared religion, cities, and socioeconomic status? As shown in Part 1, my research on census microdata from 1880 to 1930 covering 19.1 million people revealed that a little over a fifth of all second-generation Americans married someone from a different immigrant community. But, as we’ll see, there were clear patterns and preferences in who married whom.

The German-Irish Duopoly

Among second-generation Americans who married into another immigrant community, 35% married an Irish or German American. The most common pairing across all major groups featured a German American husband and an Irish American wife, and the top 10 most common overall pairings all featured either a German or Irish spouse.

Among the 20 largest second-generation ethnic groups, Mexican Americans were the only group that did not have either Irish or German Americans in their top two most common pairings — an indication of how geographically and socially separate the Mexican American marriage market was from the rest of the country.

Of course, raw frequency can be deceiving. German and Irish Americans had the largest populations, making up a combined share of 59% of the entire married second-generation American population. They were also widely dispersed, so some intermarriage was inevitable regardless of cultural affinity.

Once we adjust for factors like population size and geography, the makeup of which groups seemed to have a genuine affinity for each other shifts dramatically.

One Melting Pot or Three?

To find which groups genuinely had a preference for one another, I adjusted for geographic overlap. Since Norwegians and Germans both settled in Wisconsin in large numbers, for example, a baseline number of Norwegian-German American marriages would occur by proximity alone regardless of any cultural affinity.

Three distinct marriage clusters emerged, each internally cohesive and largely closed off from the others.

The British Isles cluster is the most intuitive. English, Scottish, and Welsh Americans married each other at two to three times the expected rate: English-Scottish pairings ran at 2.8x, English-Welsh at 2.6x, Scottish-Welsh at 2.3x. Canadians fit comfortably in this group. Irish Americans, meanwhile, were adjacent but not quite inside it — close enough to the door to knock, not quite invited to sit down.

The Scandinavian cluster is smaller and tidier. Norwegian and Swedish Americans married at 1.8x the expected rate, Swedes and Danes at 1.5x, Norwegians and Danes at 1.3x. Every pair was above chance, forming a mutual triangle of preference. Similar languages and shared Lutheran faith sustained a small but enduring Scandinavian American identity.

The Eastern European and Jewish cluster was the strongest of the three, with an average within-group affinity of 2.3x. It was dominated by the highest affinity scores in the entire dataset: Hungarian and Russian Americans married each other at 4.5 times the expected rate, Austrian and Hungarian Americans at 4.2 times, Austrian and Russian at 3.4 times.

These marriages, however, were probably not technically “cross-ethnic.” The likeliest explanation is that this cluster reveals the Ashkenazi marriage network, which operated across national-origin lines that the census drew. The son of Russian Jews marrying the daughter of Hungarian Jews was, from the community’s perspective, not exactly intermarriage. It’s a useful reminder that the census measures what the government could see, not necessarily what communities experienced.

The logic of attraction — shared language, shared faith, shared empire — maps neatly onto these three clusters. The logic of avoidance is harder to explain, and more revealing.

Left at the Altar

I ran the same adjusted analysis to find which groups married at the lowest rates once geography was accounted for. I then visualized the percentage of marriages that we would expect to happen between major groups based on population math but that failed to materialize.

The single most extreme avoidance in the dataset was between Irish and Russian Americans. For every 38 Irish-Russian American marriages the population numbers would predict, only one actually occurred. Despite sharing neighborhoods across the urban Northeast, Irish and Russian Americans almost never met at the altar. Irish-Polish (0.09x) and Irish-Austrian (0.12x) tell a similar story, suggesting an irreconcilable Catholic-Jewish religious divide.

Italian Americans had a different and in some ways more dramatic story. Italian-Russian avoidance (0.10x) was comparable to the Irish-Russian wall, but unlike the Irish, Italian Americans were comparatively insular across the board. Even marriage rates with French Americans, who were fellow Romance speakers, fellow Catholics, and European neighbors in the old country came in at just 0.47x, less than half the expected rate. Italy’s highest affinity with any group in the entire dataset was with Swiss Americans at 0.92x, still below chance.

The Irish-Italian American intermarriage rate deserves special attention because it’s the one that, on paper, shouldn’t be so low, and it challenges some previous theories about assimilation. Both groups were Catholic, concentrated in the same northeastern cities, and excluded from being part of the mainstream Protestant “establishment.” They had more in common with each other than most other pairs of groups in this dataset, and yet they married at just 13% of the expected rate — a level of avoidance nearly as severe as the Irish-Jewish divide.

The likeliest explanation is that the Irish and Italian communities occupied different rungs of the same ladder, and both knew it. The Irish had been arriving in large numbers since the famine in 1845 and had spent decades building institutional power within the American Catholic Church and urban political machines. Italian immigrants, arriving later and poorer, practiced a devotional Catholicism involving street festivals, patron saints, and public processions that many Irish Catholics regarded as harmful to their efforts to make Catholicism appear respectable. For Irish Americans who were steadily climbing toward the mainstream, marrying into a newer and poorer immigrant community risked taking a step backward. For Italian Americans, their tight-knit community was a protective asset, not something to be married out of.

What the Marriage Map Tells Us

One way to read this data is as an affirmation of cultural persistence across generations. In my view, that interpretation would be incomplete. While it is significant that British, Scandinavian, and Jewish Americans showed coherent cultural preferences in the marriage market, those three clusters accounted for only 28% of all cross-community marriages.

The majority, over 60%, involved groups that didn’t belong to any cluster at all. German Americans alone appeared in more than half of all cross-ethnic marriages, not because every group had a special affinity for Germans, but because German Americans were large in number, widely dispersed, and impossible to avoid in a marriage market shaped by proximity.

The avoidance patterns tell a similar story once you look past the affinity ratios to the raw numbers underneath. As we established, Irish and Italian Americans married at just 13% of the expected rate — but they still produced over 26,000 marriages. Irish and Russian Americans, the single most extreme avoidance at 0.03x, produced more than 5,000 marriages during this time period. Across the full dataset, 3.77 million second-generation Americans married someone from a different immigrant community, spanning 722 unique pairings. Nearly every combination you can imagine happened at least once.

If proximity did much of the heavy lifting as the data seems to indicate, then the same group should look very different depending on where it settled — insular in a state where it was concentrated, integrated in a state where it was dispersed. In Part 3, we’ll test that hypothesis state by state.

If you want to explore the data yourself before the next installment, the full interactive dashboard is here. If you have any questions you’d like me to take up in a future installment, email me at gguerra@niskanencenter.org or head to the comments section. And if you enjoyed this entry and would like to follow the series, I hope you’ll subscribe below.

Subscribe now

The Two Paths to Assimilation

My topline finding is deceptively simple. Among second-generation Americans with at least one foreign-born parent, 51.1 percent married someone from a different ethnic background.

But that majority conceals an important distinction. Of the 51.1 percent who married out, 28.7 percent married a third-generation-or-later American, someone with no traceable recent immigrant ancestry. This is the “mainstream” path, and the version of assimilation that Bill Niskanen’s parents took: an immigrant family arrives, children marry into the broader population, and ethnic identity fades within a generation or two.

The remaining 22.3 percent crossed ethnic lines but married into other immigrant communities rather than the mainstream, a different kind of boundary-crossing that we’ll return to in Part 2.

The Vast Range in Rates of Marriage into the American Mainstream

The aggregate numbers, however, obscure an enormous range in how different groups married into established American communities.

The dashboard ranks 25 origin groups by their rate of marriage into the mainstream. Before looking at the graphic below, take a guess: How likely were the children of Finnish immigrants to marry into a family with American roots, as Bill Niskanen’s father did?

If you guessed 15.8%, you should pursue a career in prediction markets.

Finnish Americans landed in the middle of the spectrum, which is where the most interesting findings are. Mexican Americans, at 14.2 percent, matched or exceeded the mainstream marriage rates of Italian, Hungarian, Lithuanian, Czechoslovakian, and Portuguese Americans. Danish Americans sat 15 points above Norwegian Americans despite their ostensibly similar cultural backgrounds. Swiss Americans, despite not having roots in an Anglophone country, ranked fourth, while Irish Americans, who did, landed six spots below them.

The extremes held fewer surprises. English and Scottish Americans had the highest rates of mainstream marriage, Russian and Polish Americans rounded out the end — but even the familiar endpoints gain new meaning once you see who falls in between.

What Opened the Door

What distinguished those who married into the mainstream from those who did not? Two factors stood out above everything else: having an American-born parent, and having parents from two different countries.

The American-parent effect was enormous — and strongest precisely where integration was hardest. Second-generation Russian Americans with two foreign-born parents married into the mainstream at 7.4 percent. When one parent was U.S.-born, that figure nearly quadrupled to 27.7 percent. The pattern is consistent and striking: having one foot in the door made the second step shorter.

Mixed-heritage parents told a similar story. When both parents came from the same country, mainstream marriage rates tracked the spectrum shown above. When parents came from two different countries — for example, a German mother and a Norwegian father — rates jumped across the board: 9 points for Germany, 13 for Norway, 22 for Italy.

The one exception was England, where mixed heritage actually lowered the mainstream rate by 6 points, probably because marrying into another immigrant community pulled English-origin children away from a mainstream they were already part of. For everyone else, parents who had already crossed one boundary raised children who crossed another.

What Held the Line

If having an American parent was the strongest accelerator of integration, gender was the strongest brake — and the pattern ran in one direction. In every major group except Norwegian Americans, women married within their heritage community at higher rates than men. For most groups, the gap was modest, running from 2 to 6 percentage points.

Italy was in a different category entirely.

Italian American men married within the Italian community at 58.7 percent. Italian American women married within the community at 85.1 percent. This 26-point gap was more than triple the next largest, and wider than the spread of Mexican, Swedish, German, and Russian Americans combined.

The raw gap alone doesn’t capture what was actually happening inside these families. Of Italian American women who married within the Italian community, 68.1 percent married a first-generation Italian immigrant. For Italian American men who married within the community, only 23.7 percent married an Italian immigrant.

The Italian case was extreme, but the direction was nearly universal. Across eight of nine major groups, women were more likely than men to marry within their community, and more likely to marry someone born abroad when they did. The melting pot did not operate equally for sons and daughters.

What the Melting Pot Actually Looked Like

The data paints the portrait of not one melting pot, but of dozens — heated unevenly and stirred at different speeds. While a slim majority of second-generation Americans married outside their heritage community overall, for some groups the share that did was overwhelming, and for others it was almost negligible.

Some factors mattered more than others in determining who entered the mainstream. Having an American-born parent could quadruple a group’s rate of integration. Being a woman could cut it in half. The same country, the same state, the same decade — entirely different outcomes emerged depending on which side of these lines someone fell on.

The aggregate numbers can take us only so far. They tell us that 51.1 percent married out, but they don’t tell us whom they married, or what invisible walls stood between communities that shared cities, workplaces, and pews. In Part 2, we’ll look at the patterns for the 22 percent who married between immigrant groups: who found each other, who avoided each other, and why Irish and Russian communities — even when they lived in the same neighborhoods — avoided intermarriage at a rate of more than 97 percent.

If you want to explore the data yourself before the next installment, the full interactive dashboard is here. If you have any questions you’d like me to take up in a future installment, email me at gguerra@niskanencenter.org or head to the comments section. And if you enjoyed this entry and would like to follow the series, I hope you’ll subscribe below.

Subscribe now