For the fourth year running, a key U.S. fertility rate has reached a record low, according to the most recent government figures. To some, this is cause for hand-wringing, as concerns arise that fewer births will spell problems for the nation’s economy; while others, concerned about limited natural resources, may look positively on the decline.
But aside from this debate, the question remains: Is this really a record low? The short answer is, it’s complicated, because there are different ways to measure fertility.
Three of the most commonly used indicators are the general fertility rate (GFR), completed fertility, and the total fertility rate (TFR). All three reflect fertility behavior in slightly different ways – respectively, in terms of the annual rate at which women are presently having kids; the number of kids they ultimately have; or the hypothetical number they would likely have based on present fertility patterns.
None of these indicators is “right” or “wrong,” but for years each measure told a different story about when fertility bottomed out. For the first time in decades, two of the three measures – the GFR and the TFR – now align, indicating that fertility hit a record low in 2018. Meanwhile, data for 2018 completed fertility is not yet available, but 2016 data indicates that it has been ticking up, not down, in recent years.
The latest report from the National Center for Health Statistics used the general fertility rate to show that for every 1,000 women of childbearing age – typically defined as ages 15 to 44 – there were 59.0 births in 2018. Like all the fertility measures discussed here, the GFR is not affected by the overall population size or the share of the population that consists of women of childbearing age. However, it is affected by changes in the age distribution among women of childbearing age; the higher the share of women in their peak childbearing years, the higher the general fertility rate will be, all else being equal (and vice versa).
The record-low GFR cited by the government is primarily a result of a decline in birthrates among women younger than 35. This may partly reflect long-lasting effects of the Great Recession of the late 2000s; when there is an economic downturn, people tend to postpone having children. It also reflects long-term increases in educational attainment and women’s labor force participation, as well as delays in marriage, which contribute to postponement of motherhood. However, there has also been a slight decline in the share of women who are in their peak childbearing years (ages 20 to 34), which may play a small role in the decline.
The second fertility measure is completed fertility, which counts the number of children a woman has in her lifetime. Typically, researchers collect fertility data for women ages 15 to 44. They then measure “completed fertility” as the number of children ever born to women ages 40 to 44, on the assumption that most women at this age are done having children. According to this measure, since 1976, the low point in U.S. fertility occurred around 2006, when women near the end of their childbearing years had had an average of 1.86 kids.
Because it is a retrospective measure, completed fertility summarizes childbearing patterns from recent decades but cannot provide direct insights into the fertility behaviors of younger women today. Also, because it doesn’t take into account the larger shares of women who are postponing childbearing, it may underestimate fertility for some, such as highly educated women, many of whom have children later in life. (To address this, the U.S. Census Bureau recently began collecting fertility data for women up to age 50.)
A third indicator, the total fertility rate, is an estimate of lifetime fertility, based on present fertility patterns. It reached a new low in 2018, when it was estimated that U.S. women would have, on average, 1.73 kids in their lifetimes. This beat out the prior low TFR of 1.74 that had stood since the inflationary period of the mid-1970s. The TFR is not an actual accounting of lifetime fertility, but instead is a hypothetical measure based on fertility information from one point in time, which is then projected into the future to estimate the number of babies a typical woman would ever have if her lifetime childbearing experience reflected current age-specific fertility rates.
For example, in 1980 the total fertility rate was 1.84. This means that girls age 15 in 1980 (and thus at the beginning of their childbearing years) were predicted to have, on average, 1.84 children in their lifetime, based on the assumption that their likelihood of having a baby at each subsequent age would reflect the age-specific fertility rates of 1980. So the assumption was that when they were 25 years old in 1990, this group of young women would have the same fertility rate as 25-year-old women did in 1980; when they were 35 in 2000, they would have fertility rates similar to the rate that 35-year-olds had in 1980; and so on.
This type of technique is called “age standardization,” and it’s wholly unaffected by the age makeup of women in their childbearing years. It will produce the same predicted lifetime fertility, regardless of whether there is an especially large share of women in the prime childbearing ages, or if the share in prime childbearing ages is very small.
The TFR is a powerful measure most commonly used to characterize “replacement fertility,” meaning the level of estimated fertility that is necessary for a population to reproduce itself, assuming no in-migration or out-migration. A total fertility rate of 2.08 is considered “replacement level” in the U.S. (While it might seem that “replacement fertility” for a couple should be 2, it is actually a bit higher to account for childhood mortality.)
All else being equal, the TFR can fluctuate dramatically if there is a shift in the timing of births among women in their childbearing years. On one hand, sometimes the TFR estimate proves accurate. As our chart shows, the 1980 TFR did a good job of predicting fertility – it almost exactly matches the completed fertility of women who had been 15 in 1980 and were approaching the end of their childbearing years in 2005.
But the total fertility rates during the baby boom dramatically overestimated U.S. fertility, because they reflected the fact that some women were having babies at ages younger than had been the norm, while others were perhaps “catching up” on births that were postponed due to World War II. For example, the TFR in 1960 was 3.65, but women who entered their childbearing years around that time actually gave birth to about 2.45 children in their lifetimes, according to 1985 completed fertility.
More recently, many women have been putting off having children. As a result, today’s TFRs may be underestimating what completed fertility in the U.S. and Europe will look like. Furthermore, research has revealed that because of its sensitivity to timing, the TFR typically overestimates fertility among Hispanic immigrants to the U.S., who experience their highest fertility rates within five years after migrating, regardless of their age at that time.
So, how do these three measures compare? Both the total fertility rate and the general fertility rate show similar trends. Moreover, both clearly reflect the heightened fertility of the baby boom, as well as the subsequent “baby bust.” And both show a rise in fertility just before the economic recession of 2007, followed by almost continuous fertility declines. While the GFR has hit a record low in seven of the last eight years, until 2018 the TFR remained above historic lows.
Meanwhile, available data on completed fertility reveals that from 1976 to 2016, the total number of children ever born peaked in the late 1970s as those women who came of age during the baby boom reached the end of their childbearing years. Since 2006, completed fertility has risen slightly, even as the other measures have continued to drop.
Fertility – however it is defined – is influenced by a variety of demographic, social and economic factors. Among the more important contributors to fertility levels are the age composition of childbearing women (most births occur between ages 20 and 34); the share of childbearing-age women who are married (married women are more likely to give birth than unmarried women); the economic climate; and the share of childbearing age among women who are born in another country (foreign-born women have higher birthrates than U.S.-born women).
Note: This is an update of a post originally published Feb. 24, 2015.