You often hear about baby names being cyclical, and lately how “vintage names” are all the rage among new Moms[1]. The Social Security Administration puts out annual data on baby names, and this data goes back to the 1880s. As part of another post I’m working on, I recently took a look at this data (for now, I just analyzed boys) and found some interesting conclusions. You can find the data here: https://www.ssa.gov/oact/babynames/.[2]
First, I found that 570 different names for boys have been in the top 200 most popular names in one decade or another since the 1880s.
Of those names, only 43, or 7.5%, have been in and out of the top 200 more than once. I define this slim contingent as “come back names”. That means most names had their popular moment and then faded away.
On that theme, of these 570 names, 549 of them, or 96%, have at one point or another fallen out of the top 200 after at some point being in it
This leaves only 21 names (3.6%) that have been in the top 200 in every decade since the 1880s and including the year 2022. These include: Andrew, Anthony, Benjamin, Charles, Daniel, David, George, Henry, Jack, James, Jesse, John, Jose, Joseph, Michael, Nicholas, Robert, Samuel, Thomas, Wesley (maybe the most surprising?), and William.
The most popular name in American history? Judging by the highest average ranking, the answer is James, closely followed by William.
While we’re on the subject of most popular names, perhaps the most interesting conclusion of this analysis is that from the 1880s through the 1910s (so call it 40 years), boys in the US had essentially the exact same top ten names, though the ordering shifted a bit by year. Considering the decline of name concentration that would happen in the 2nd half of the 20th century, this is remarkable, especially considering the country’s population nearly doubled over this period (from an average of about 57M in the 1880s to an average of ~99M in the 1910s) and dealt with a huge wave of immigration. The table below shows that through the 1910s, at least 8 of the top 10 names from the 1880s were still in the Top 10 up until World War 1. Nothing like this has ever happened since.
Another conclusion, as noted above, is that baby names are fashionable, not cyclical. I like this word better because while it’s true some baby names are cyclical, most have their moment of popularity (which to be fair could be multiple decades), and then fade away. Recall the “come back name” group I highlighted above (which we defined as names that were in the Top 200 multiple times but with at least one decade of absence in between) was only 43 names (or 7.5% of the 570). It’s therefore hard to say in general that baby names are cyclical. The ones that come back are cyclical, sure. But I like “fashionable” better because while fashion too is cyclical, at least some fashions never (or at least not yet) come back. Thus, for every chart that looks like this:
Or this:
You get about 10 that look like this:
You sometimes hear about a 100 year “rule” for baby names (https://nameberry.com/blog/1910-baby-names-testing-the-100-year-rule). First, given less than 10% of baby names “come back,” I think a better word for “rule” is “tendency.” Additionally, of the 43 names that I found “came back,” I actually found that the average “come back” period is more like 60 years (so figure two generations), rather than 100, as the table below shows.
Another interesting conclusion I found is that over time, name popularity “fashions” seem to be happening faster. The below chart shows the number of names falling outside of the top 200 each decade after initially being in it. So for example, a number of 40 for the series of the 1880s in the 1st decade would mean that 40 names that were in the top 200 in the 1880s would no longer have made the list in the 1890s. Similarly, a number of 100 in the 8th decade after indicates that 80 years later, the “class” of 1880s most popular names has only half of its original list remaining in the Top 200.
Note that each of the lines in the chart below appears to be shifting up and to the left, indicating that names are falling out of the Top 200 faster than they used to. This is happening quite consistently over time.
This next chart shows the average popularity ranking of those “survivor” names that are still in the top 200 each decade after initially being there, again, by class. Notice the lines on this chart are now moving lower more quickly, indicating that those names that do survive are getting popular faster. It seems that names that do survive now either get more popular more quickly, or they fade away.
You might think that this second chart would be subject to “survivorship bias,” which would argue that of course those names that have remained in the top 200 are getting more popular. After all, they’re the ones still on the list! But there’s no reason that the names still on the list would necessarily get more or less popular just by remaining on the list. All initial Top 200 lists will by definition have an average ranking of 100, and as the chart above shows, in the second decade, those names still on the list have historically had an average ranking of between 90-95 on the Top 200. The fact that the survivors are not only staying on the list, but are actually getting more popular, however, is notable. It is possible, in fact you’d think probably likely, that the average ranking of survivor names would trend in the same direction (on average) as those who have fallen off the list, and therefore see their popularity ranking go lower. But this doesn’t seem to be happening. The fact that the lines are going lower (indicating an improving average ranking and higher popularity), and doing so more quickly, suggests that the names that are still on the list are actually getting more popular over time, and are getting more popular more quickly than in prior “classes” of names. Each class of name’s survivors still generally peaks out at the same level of popularity, but the more recent classes of “survivors” reach that peak more quickly, suggesting people are cycling in and out of baby names more quickly than they did in prior decades.
We touched on name concentration earlier, so let’s return to that topic. Below are charts of the top Male and Female Names as a percentage of all births, and then of the top 10 and 200 names as well.
Notice how the top name, the top 10 names, and the top 200 names as a percentage of total births have all seen significant declines since the 1940s. This is interesting for a couple reasons:
First, because in the decades prior to this (say from the 1900s to the 1940s), there was a fairly lengthy period of consistently high “name concentration.”
Second, because the 1950s is often viewed as the height of American conformism.
Notably though, the concentration in the Top 200 names, while also eventually declining in a similar way as the Top 10, held up at the plateau longer than the Top 10 did, suggesting greater name selections outside of the Top 1, 5 and 10 in the 1950s but not necessarily outside the Top 200. People still seemed to want popular names in recent decades, just not the most popular names. Even this, however, is changing, as clearly indicated by the chart below, where even the top 200 has lost a lot of “market share”:
Lastly, notice the significant differences in name concentration between boys and girls. This one stumped me and I will need to maybe tackle this one in a future analysis.
Importantly, it does not seem like population growth alone is driving greater name variability. The country roughly doubled in size over the 1900-1940 period (from 76.3M to 132.2M), and it did so again from the 1950s to 2010 (from 158.8M to 309M). This therefore presents two periods where the country’s population doubled, yet in one period (1900s – 1940s), name concentration was actually quite stable, with the top name accounting for between 5.3% and 5.8% of total births and the top 5 names consistently comprising in the low 20s percent of all boy names. In the other period (1950s to 2010s), the top name declined from 5.3% to 1% (which is also what it was in 2022), and the top 5 names fell from 20% share in the 1950s to 4% in 2022. That is just a stark contrast.
Several things seem like possible culprits for this.
First, immigration. Despite significant immigration in the second half of the 19th century, most immigration during this period tended to be more from Western European countries[3] (predominantly Germany, England and Ireland), which are generally much more Christian, and much closer to our roots as an English colony (and thus still much more likely to name their baby John than someone from Asia, for example). This may have started to shift during the Progressive Era (in this case, 1900-1915), where “unlike earlier immigrants, the majority of the newcomers after 1900 came from non-English speaking European countries.” “The principal source of immigrants,” the Library of Congress adds of this period, “was now southern and eastern Europe, especially Italy, Poland, and Russia, countries quite different in culture and language from the United States.”[4] As the below chart from the Population Reference Bureau (PRB) shows well[5], the sheer amounts of immigrants entering the country just after 1900 (notice the massive spikes around then), in tandem with where they were coming from, almost had to disrupt baby name concentration.
The change in where immigrants were coming from also ties nicely with the significant decline in name concentration that began around the 1950s. According to PRB, in 1965, immigration law changed, with national origin preferences favoring Europeans being eliminated. Additionally, “during the 1970s, the origins of most immigrants changed from Europe to Latin America and Asia,” and more significantly, “between 2000 and 2009 over three-fourths of the 10 million immigrants admitted were from Latin America and Asia.”[6] As immigrants from Asia and Latin America became more prominent origin regions then, given those cultures are more distinct from even Eastern Europe (and certainly from Western Europe), we shouldn’t be surprised to see greater name variability as a result.
Importantly though, it does not appear as though immigrants’ religiosity was a driving factor in the declining concentration of Kingly / Biblical names early in the 20th century. Consider that the World Christian Encyclopedia states that in 1900, North America was 97.1% Christian[7], and according to Gallup, in 1956, the number was still at least in the mid 90s[8]. Experience, and both anecdotal and empirical evidence suggests immigrants today still tend to be more religious than the average non-immigrant American[9], though increases in immigration from those belonging or subscribing to non-Christian sects may also be contributing to greater baby name diversity today. Latin Americans, for example, are actually more Christian (and definitely more Catholic) than the rest of America, but non-Christians now make up about 1 in 4 legal immigrants, with other notable non-Christian religions being Islam, Buddhism and Hinduism, all of which derive from markedly different cultures than the West.[10] Thus, levels of religiosity are likely not the driving factor in name diversity from immigration, but rather the cultures stemming from these non-Christian religions that are probably resulting in immigrants in the 20th century contributing to a wider array of baby names.
What about a decline in overall religiosity? Nietzsche had already declared that “God is dead” in 1882[11], after all, but as noted above, whatever decline was happening at that point likely wasn’t being driven by immigrants. But while this trend away from religion may have been emerging at the turn of the 20th century, it seems too nascent to have been much of a factor until the 2nd half. In addition to the fact that America remained a Christian dominant nation through at least the middle of the century, a late 1950s survey of religious trends in the United States found that the “membership of the religious bodies” (i.e. churches) had actually increased significantly (as a percentage of the population) during that period, rising from 39% of the population in 1906 to almost 60% in 1956.[12]
Things have definitely changed since then, however. Today, only 80% of Americans have a religious affiliation, including 7 in 10 Americans who still affiliate with Christianity[13]. Additionally, Gallup’s surveys of Church Membership also show marked change, with only 47% of Americans belonging to a church in 2020, down from 70% as recently as 1999, where it generally was for the prior six decades[14]. I have personally witnessed this exact phenomena in my own family. Church-going has also declined as a result, with only 31% of people attending a religious institution once a week or nearly weekly today compared to more than double that percentage for their parents[15].
Since 1944, Gallup has also asked people if they believe in God, and in that year, 1947, 1953, 1954, 1965, and 1967, at least 94% of Americans said yes, and as recently as 1965 and 1967, that number was 98%. The next time this question was asked in 2011, the number fell to 92%, and when Gallup most recently asked this question in July of 2022, the number fell to 81%, down from 87% in 2017 and a new low in the history of the poll[16]. Thus, it's possible that declines in religiosity help explain reduced concentration of names in the 2nd half of the 20th century, but its not clear that was a main culprit that broke the Kingly / Biblical name concentration in the late 1880s through World War I.
A third hypothetical factor has to also be the migration from farms to cities. First, some perspective. By 1880, significant migration to cities (or in other cases, development of cities) had already occurred, with over ¼ of the country’s population living in “urban” areas[17]. By 1900, that number was up to 40%, and by the end of World War I, that number was over half. After the next World War it reached 60%, and by 1970, it reached 74%. It is notable, though not the subject of this post, that from 1870 to 1970, the share of the population increased by almost 50 percentage points (26% to 74%). Since 1970, however, a period of 50 years, that share has only increased by roughly 6 percentage points (from 74% to 80%), indicating America’s migration to cities may be over. The 2020 census appears to be the first census on record where the share of the urban population actually declined compared with the prior census.
Why would migration to cities potentially be a factor in affecting baby names? For several reasons. First, sociability is greater in cities, and information travels faster. If people are talking and socializing more, they learn what other people are doing more quickly, including what they’re naming their babies. It perhaps shouldn’t be a surprise then that baby names are falling in and out of fashion more quickly, as we showed above.
There is another, and likely related factor here, which is a potential rise in individualism. This was likely accelerated by the rise of migration to cities, where being surrounded by so many other people makes it harder for us to stand out, and therefore harder to dignify ourselves as individuals. Given this, it probably shouldn’t be a surprise that you tend to see more “culture” in the city, or people otherwise doing things to stand out or not conform. Cities may have more brick walls in them, but they have fewer people who see themselves as “just another brick in the wall.”
Cities also breed competition in this regard, which also lights a fire under us to thrive (spend a couple days in New York City and you’ll feel what I’m talking about). This part is no doubt much harder to prove, but it seems likely to have been a factor in greater baby name variability too. The harder part though is explaining why individualism reached such a tipping point in the 1950s, given so much migration to cities had already happened by then. Perhaps the ending of World War II and the Depression allowed individuals to breathe as individuals again after such massive collective political efforts the prior two decades. Or, perhaps this is less of a factor, and immigration from non-Western nations a larger one.
In summary, this analysis of baby names provides the following conclusions:
First, baby names are more fashionable than they are cyclical. Few baby names actually “come back.” Far, far more see a random surge in popularity, followed by a subsequent decline. It’s certainly true that more may come back in the future, but this has not happened historically. Only time will tell if a given name re-emerges in the future.
These phases of baby name popularity are happening faster. This phenomena suggests that people are still seeking out the “cool” or “trendy” names, but they become popular faster and subsequently de-popularize faster than in prior decades.
Greater name diversity suggests enhanced diversity within our culture, which likely stems from greater immigration of non-western cultures in the 20th century, particularly in the second half.
[2] Note that this data comes from those filing for a Social Security card. Given the Social Security Administration was only created in the 1930s, the SSA states “All names are from Social Security card applications for births that occurred in the United States after 1879. Note that many people born before 1937 never applied for a Social Security card, so their names are not included in our data.” Statistically though, there’s no reason why Johns or Andrews born in the 1880s would be any more likely to file for a Social Security card in the 1930s than people with names that are less popular. Additionally, even in the 1880s the 200th most popular name had 725 people who had filed for a SS card with that name (Jasper), which is not exactly a small amount of people. It’s therefore not clear to me why there would be kind of adverse or survivor selection bias or any other problems with the older portions of the data.
[3] See Immigration to the United States, 1851-1900 from Library of Congress. https://www.loc.gov/classroom-materials/united-states-history-primary-source-timeline/rise-of-industrial-america-1876-1900/immigration-to-united-states-1851-1900/
[4] See Immigrants in the Progressive Era; Library of Congress; https://www.loc.gov/classroom-materials/united-states-history-primary-source-timeline/progressive-era-to-new-era-1900-1929/immigrants-in-progressive-era/
[6] See PRB report cited above
[7] See here for reference: https://www.gordonconwell.edu/blog/christianity-in-north-america/
[9] See church going source data below
[11] See Thus Spoke Zarathustra by Friedrich Nietzsche, published in 1882.
[12] See 1900-1950 Survey: Religious Trends in the United States; Christianity Today, Richard C. Wolf, April 27th, 1959; https://www.christianitytoday.com/ct/1959/april-27/19001950-survey-religious-trends-in-united-states.html
[13] https://news.gallup.com/poll/467354/childhood-churchgoing-habits-fade-adulthood.aspx#:~:text=Across%20all%20of%20Gallup's%20polling,or%20to%20regularly%20attend%20services.
[15] https://news.gallup.com/poll/467354/childhood-churchgoing-habits-fade-adulthood.aspx#:~:text=Across%20all%20of%20Gallup's%20polling,or%20to%20regularly%20attend%20services.
[17] The definition of Urban has changed over time. Early on when the Census Bureau first started releasing this data, the threshold was 8,000 people or more, though in some decades it was lowered to 4,000. See The Development of the Urban-Rural Classification in the United States: 1874-1949, by Leon Truesdell, Chief Demographer of the Census Bureau; https://www2.census.gov/library/publications/1949/demographics/p23-001.pdf
Comments