Marrying Older, But Sooner?

Print This Post Print This Post
Bookmark and Share

Written By: Jonathan Vespa

The age when Americans marry for the first time has risen to its highest point since the 1950s. Although true, there is more to the story. With life expectancy increasing, Americans today are actually marrying sooner in their lifetime, despite marrying at older ages.

Taking the long view

Although it is common to use the 1950s as a comparison period, doing so exaggerates how much the age at marriage has really risen (Figure 1). Looking at trends since 1890 reveals a U-shaped curve in which the 1950s and 1960s stand out as the exception for marriage, not the norm (estimates for these two decades are not significantly different from one another).

By the end of the 19th century, men married for the first time at 26 years old, three years later than they did in the decade following the Second World War. By 1900 their age at marriage began falling, and it took a full century before returning to its 1890 level. For women, it took 90 years (Figure 1). The idea then that our great-great-grandparents married when they were little older than teenagers is little more than myth.The rising age at marriageThe rising age at marriage is not nearly as large if 1890 were the benchmark. Today the typical man marries when he is 29 years old. That is six years older than the historic low in 1956, but only three years older compared with 1890. The pattern for women is similar.

Age at marriage has risen in tandem with life expectancy

By itself, the rising age at marriage toward the end of the 20th century looks startling. Since 1975 it has risen 6 years for men and women. Nonetheless this increase should not be surprising. Because we are living longer, we should expect to be marrying at older ages.

In 1890 the average lifespan for men was just 43 years, but they did not marry until 26 (Figure 2). In other words, they did not marry for the first time until well over half way through their lifetime (Figure 2). Today, men can expect to live to 76, yet they marry at 29, about one third of the way through their lifetime (Figure 2).

Marrying older, but sooner? Graph showing men's first age at marriage from 1890-2010.If the proportion had remained at the same level as in 1890, men would not marry until they were 46 years old — 17 years later than they do today. Women would not marry until they were 40 (Figure 3).

Marrying older, but sooner? Graph showing women's age at first marriage from 1890-2010.Viewed in this light, the gains to life expectancy during the 20th century dwarf increases in the age at marriage. What is more, calling the rising age at marriage a delay ignores life expectancy, which is not accounted for when we compare ages in absolute terms. Thus despite marrying at older ages, Americans are actually marrying sooner in their lifetime than they did a century ago.

For more information, visit estimated median age at marriage and the Census working paper on historical marriage trends.

This entry was posted in Families, Marriage and tagged , , . Bookmark the permalink.

7 Responses to Marrying Older, But Sooner?

  1. Deborah Stein says:

    It would be really interesting to also look at age at the birth of the first child over time, especially since the period of fertility is not expanding as much as life expectancy.

  2. Mark P says:

    It’s misleading to measure lifespan using life expectancy at birth. This adds quite a bit of infant/childhood mortality into the mix, which is the cause of almost all of the increase in life expectancy over the last century. Life expectancy at 21 would be a better indicator of lifespan, but I suspect that your results will not be so striking if you were to use this metric.

  3. Jonathan Vespa says:

    Right, life expectancy at birth does reflect infant mortality. But it’s not as skewed as you might think.

    Using life expectancy during childhood — instead of at birth — helps cut out bias from infant mortality, but the differences aren’t dramatic. Using expectancy during childhood, the estimate of time spent before 1st marriage is still above 50% for men in 1890. After 1890 there’s only a few percentage point difference between using expectancy at birth versus childhood, and that difference trails off by 1930.

  4. Susan R says:

    I think it is interesting to note that the difference in the age of men and women at first marriage has reduced by half. This article downplays that the age of women at their first marriage has changed much more dramatically than it has for men. The gap closed gradually from 4 years to 2 years from 1890 to the 1950 and has remained fairly constant since that time.

    I am glad I wasn’t imagining this trend while I was doing marriage license indexing earlier this year. I think our forbears were much more conscientious of being financially stable before marriage.

  5. Steve says:

    Do the average ages being used account for declining rates of infant mortality? For this type of comparison it’s much more useful to look at average age of those who live more than one year, otherwise the average age doesn’t represent the life expectancy for someone of age to be making a decision about marriage.

    • Jonathan Vespa says:

      Steve, thanks for the post. We’re happy to see that people are interested in and thinking about the data.

      We used life expectancy at birth to show a consistent, straightforward comparison across time. There are other good ways to present the data, too.

      As another poster commented, infant mortality lowers the average life expectancy. It’s most noticeable around the turn of the 20th century, but negligible for the population as a whole in later decades. Using life expectancy during childhood doesn’t change the overall trend, however. Both men and women married quite a bit later in their lifetimes a century ago than they do today.

  6. Kristi Williams (@kristexanite) says:

    Great work, Jonathan. You continue to make OSU Sociology and OSU IPR proud!

Leave a Reply

Your email address will not be published. Required fields are marked *

*