Get your dependency ratio off my lawn

Old people work more than they used to. This is important if you’re worried about what an aging population means for the economy.

When they taught me demography, I learned about the dependency ratio, which was the number of people presumed to be dependents (those ages 0-14 and 65+) relative to those presumed to be working (those ages 15-64). It’s a traditional measure, and a little archaic now that people spend much more time in school. But it’s nice because it sort of assumes that those “working age” adults are being productive whether they have jobs or not – it’s not just counting employed people – so it has an unstated recognition of (mostly women’s) unpaid labor.

In some economic work (see my paper here for an old review) people assume that non-employed women are being productive. But we don’t usually assume that about old people. That is, non-employed younger adults are assumed to be doing unpaid work, while non-employed old people are assumed to be really retired. I’m sure people are looking at the unpaid work of old people (I just haven’t yet). But their paid work profile has changed a lot, too – especially women’s.

This means the catastrophic view of productivity effects of again needs to be tempered by a better understanding of how much old people work. Here’s what I mean.

First, what the World Bank calls “Dependency Ratio, old,” which is the number of people age 65 and older as a percentage of the population ages 15-64. This is supposed to reflect the burden of age on the the young(er).* Here is it for the USA and the world (click to enlarge):

That’s the Baby Boom generation hitting older ages there at the end of the USA trend. As a result, the dependency ratio (old) has increased 30% in the USA since 1980, and the world is following.

But old people work more (or, we don’t label people “old” as early, you might say). Here’s the average annual hours of paid work for people in the USA ages 65 and older. Note this includes all those working no hours in the average, which is what you need to do if you’re interested in the total economic benefit/burden ratio (click to enlarge).

Since 1980, women ages 65-74 have increased their hourly employment hours by 138%, and men’s have gone up 44%. For the 75-plus community, the relative increases are even greater: 172% for women and 55% for men.

Now, if you add up those hours, you can calculate how much of a burden old people are relieving from the young by their employment hours. In this figure I calculate the total hours worked for each age-gender group and divide it by the total number of people ages 65 and older. Looking at the bottom blue area, for example, this shows that in 2015, the total population of men ages 65-74 did 166 hours of paid work for each person age 65 and older. Regardless of the size of the old population, then, there is that much less supporting of them to do (click to enlarge).

The per-person contribution of paid work hours from people 65 and older has increased 72% since 1980, from 206 to 354 hours per year. Most of the increase is from women’s employment, and it’s just starting. The oldest Baby Boom women, the women who led the increase in women’s employment over their careers, are still only 69 in 2015. Further, this measurement of paid hours may be an indicator of the unpaid productivity of these groups as well, as their health and activity levels improve.

It may be useful to track the population age composition over time (as in the World Bank data above), but it’s not reasonable to assume a constant level of dependency associated with people of different ages.

*Note: Of course, I use terms like “burden” in the classical demographic sense and tongue-in-cheek. I actually want more old people to live longer and work less, because that burden is what life is all about. But there is the issue of making sure everyone has their needs met.

On Asian-American earnings

In a previous post I showed that generalizations about Asian-American incomes often are misleading, as some groups have above-average incomes and some have below-average incomes (also, divorce rates) and that inequality within Asian-American groups was large as well. In this post I briefly expand that to show breakdowns in individual earnings by gender and national-origin group.

The point is basically the same: This category is usually not useful for economic statistics, and should usually be dropped for data on specific groups when possible.

Today’s news

What’s new is a Pew report by Eileen Patten showing trends in race and gender wage gaps. The report isn’t focused on Asian-American earnings, but they stand out in their charts. This led Charles Murray, who is fixated on what he believes is the genetic origin of Asian cognitive superiority, to tweet sarcastically, “Oppose Asian male privilege!” Here is one of Pew’s charts:

The figure, using the Current Population Survey (CPS), shows Asian men earning about 14.5% more per hour than White men, and Asian women earning 11% more than White women. This is not wrong, exactly, but it’s not good information either, as I’ll argue below.

First a note on data

The CPS data is better for some labor force questions (including wages) than the American Community Survey, which is much larger. However, it’s too small a sample to get into detail on Asian subgroups (notice the Pew report doesn’t mention American Indians, an even smaller group). To do that I will need to activate the ACS, which is better for race/ethnic detail.

As a reminder, this is the “race” question on the 2014 American Community Survey, which I use for this post:

There is no “Asian” or “Pacific Islander” box to check. So what do you do if you are thinking, “I’m Asian, what do I check?” The question is premised on that assumption that is not what you’re thinking. Instead, you choose from a list of national origins, which the Census Bureau then combines to make “Asian” (the first 7 boxes) and “Pacific Islander” (the last 3) categories. And you can check as many as you like, which is good because there’s a lot of intermarriage among Asians, and between Asians and other groups (mostly Whites). This is a lot like the Hispanic origin question, which also lists national origins — except that question is prefaced by the unifying phrase, “Is Person 1 of Hispanic, Latino, or Spanish origin?” before listing the options, each beginning with “Yes”, as in “Yes, Cuban.”

Although changes have not been announced, it is likely that future questions will combine the race and Hispanic-origin questions, and also preface the Asian categories with the umbrella term. This may mark the progress of getting Asian immigrants to internalize the American racial classification system, so that descendants from groups that in some cases have centuries-old cultural differentiation start to identify and label themselves as from the same racial group (who would have put Pakistanis and Japanese in the same “race” group 100 years ago?). It’s hard to make this progress, naturally, when so many people from these groups are immigrants — in my sample below, for example, 75% of the full-time, year-round workers are foreign-born.

Earnings

The problem with the earnings chart Pew posted, and which Charles Murray loved, is that it lumps all the different Asian-origin groups together. That is not crazy but it’s not really good. Of course every group has diversity within it, so any category masks differences, but in my opinion this Asian grouping is worse in that regard than most. If someone argued that all these groups see themselves as united under a common identity that would push me in the direction of dropping this complaint. In any event, the diversity is interesting even if you don’t object to the Pew/Census grouping.

Here are two breakouts. The first is immigration. As I noted, 75% of the full-time, year-round workers (excluding self-employed people, like Pew does) with an Asian/Pacific Islander (Asian for short) racial identification are foreign born. That ranges from less than 4% for Hawaiians, to around 20% for the White+Asian multiple-race people, to more than 90% for Asian Indian men. It turns out that the wage advantage is mostly concentrated among these immigrants. Here is a replication of the Pew chart using the ACS data (a little different because I had to use FTFY workers), using the same colors. On the left is their chart, on the right is the same data limited to US-born workers.

Among the US-born workers the Asian male advantage is reduced from 14.5% to 4.2% (the women’s advantage is not much changed; as in Pew’s chart, Hispanics are a mutually exclusive category.) There are some very high-earning Asian immigrants, especially Indians. Here are the breakdowns, by gender, comparing each of the larger Asian-American groups to Whites:

Seven groups of men and nine groups of women have hourly earnings higher than Whites’, while nine groups of men and seven groups have women have lower earnings. In fact, among Laotians, Hawaiians, and Hmong, even the men earn less than White women. (Note, in my old post, I showed that Asian household incomes are not as high as they look when they are compared instead with those of their local peers, because they are concentrated in expensive metropolitan markets.)

Sometimes when I have a situation like this I just drop the relatively small, complex group, which leads some people to accuse me of trying to skew results. (For example, I might show a chart that has Blacks in the worst position, even though American Indians have it even worse.)

But generalization has consequences, so we should use it judiciously. In most cases “Asian” doesn’t work well. It may make more sense to group people by regions, such as East-, South-, and Southeast Asia, and/or according to immigrant status.

The fathers behind teen births (or, statistical memes and motivated blind trust)

When makes people trust statistical memes? I don’t know of any research on this, but it looks like the recipe includes a combination of scientific-sounding specificity, good graphics, a source that looks credible, and – of course – a number that supports what people already believe (and want their Facebook friends to believe, too).

If that’s the problem, and assuming the market can’t figure out how to make journalism work, I have no solution except seizing the Internet and putting it under control of the Minister of Sociology, or, barring that, encouraging social scientists to get engaged, help reporters, and make all their good work available publicly, free, and fast.

Today’s cringe:

The blogger TeenMomNYC takes credit for creating this, and the Facebook version has been shared tens of thousands of times. Its popularity led to this story from Attn: “The Truth About Teenage ‘Baby Mamas’ is Quite Revealing.” (If anyone did want to study this issue, this is a neat case study, because she posted 8 “did you know” graphics on Facebook at the same time, and none of the others took off at all – why?)

I don’t know anything about TeenMomNYC, but I share her desire to stop stigmatizing and shaming young mothers. I wish her work were not necessary, but I applaud the effort. That said, I don’t necessarily think shaming young fathers (even if they’re not quite as young) is a solution to that, but that’s not the point. My point is, what is this statistic?

According to the footnote (thanks!), it comes from this 1995 National Academies report, and (except for changing “29” to “29.7”) it represents it accurately. From p. 205:

These data highlight an additional component of the sexual abuse picture— the evidence that an appreciable portion of the sexual relationships and resulting pregnancies of young adolescent girls are with older males, not peers. For example, using 1988 data from the NSFG and The Alan Guttmacher Institute, Glei (1994) has estimated that among girls who were mothers by the age of 15, 39 percent of the fathers were ages 20–29; for girls who had given birth to a child by age 17, the comparable figure was 53 percent. Although there are no data to measure what portion of such relationships include sexual coercion or violence, the significant age difference suggests an unequal power balance between the parties, which in turn could set the stage for less than voluntary sexual activity. As was recently said at a public meeting on teen pregnancy, “can you really call an unsupervised outing between a 13-year-old girl and a 24-year-old man a ‘date’?”

This is an important point, and was good information in 1995, when it cited a 1994 analysis of 1988 data, which asked women ages 15-44 a retrospective question. In other words, this refers to births that took place as early as 1958, or between 28 and 58 years ago. That is historical, and really shouldn’t be used like this today, given how much has changed regarding teen births.

The analysis is of the 1988 National Survey of Family Growth, a survey that was repeated as recently as 2011-2013. Someone who knows how to use NSFG should figure out the current state of the age gap between young mothers and fathers and let TeenMomNYC know.

Even if I didn’t know the true, current statistic, this would give me pause. Births to women before age 15 are extremely rare. The American Community Survey, which asks millions of women whether they have had a birth in the previous year, does not even ask the question of women younger than 15. The ACS reports there were 179,000 births in the previous year among women who were under 20 when interviewed, of which only 6,500 were to women age 15 at the interview. So that’s 3.7% of teen births, and 3 out of every thousand 15-year-old women. In 1958 this was much more common, and the social environment was much different.

Another issue is the age range of the fathers, 20-29, which is very wide when dealing with such young mothers. Look at the next phrase from the 1995 report: “girls who had given birth to a child by age 17, the comparable figure was 53 percent.” Realize that the great majority of girls who had a birth “by age 17” were 17 when they did, and the great majority of those men were probably close to 20. I’m not very positive about 20-year-old men having children with 17-year-old women, but it’s pretty different from 29-versus-13.

I can’t find the original source for this, but this report from the Resource Center for Adolescent Pregnancy Protection attributes this table to the California Center for Health Statistics in 2002, which shows that the father was age 20 or older  for 23% of women who had a birth before age 15. And of those, 93% were 20-24 (rather than 25+).

Anyway, this is a good case of a well-intentioned but under-resourced effort to sway people with true information, picked up by click-bait media and repeated because people think it will help them win arguments, not because they have any real reason to believe it’s true (or not true).

So I really hope someone with the resources, skills, and training to answer this question will produce the real numbers regarding father’s age for teen births, and post them, with accompanying non-technical language, along with their code, on the Open Science Framework (or other open-access repository).

Fixing the media and its economy is a tall order, but academics can do better if we put our energy into this work, reward it, and restructure our own system so that good information gets out better, faster and more reliably.

Related posts:

• No Black women are not the “most educated” group in the US
• International Women’s Day breathes new life into the 1% meme

Black women really do have high college enrollment rates (at age 25+)

The other day I reported on the completely incorrect meme that Black women are the “most educated group” in the U.S. That was a simple misreading of a percentage term on an old table of degree attainment, which was picked up by dozens of news-repeater websites. Too many writers/copiers and editors/selectors don’t know how to read or interpret social statistics, so this kind of thing happens when the story is just too good to pass up.

I ignored another part of those stories, which was the claim that Black women have the highest college enrollment rates, too. This is more complicated, and the repeated misrepresentation is more understandable.

Asha Parker in Salon wrote:

By both race and gender there is a higher percentage of black women (9.7 percent) enrolled in college than any other group including Asian women (8.7 percent), white women (7.1 percent) and white men (6.1 percent), according to the 2011 U.S. Census Bureau.

You know the rewrite journalists are playing telephone when they all cite the same out-of-date statistics. (That Census report comes out every year — here’s the 2014 version; pro-tip: with government reports, try changing the year in the URL as a shortcut to the latest version.)

But is that true? Sort of. Here I have to blame the Census Bureau a little, because on that table they do show those numbers, but what they don’t say is that 9.7% (in the case of Black women) is the percentage of all Black “women” age 3 or older who are attending college. On that same table you can see that about 2% of Black “women” are attending nursery school or kindergarten; more relevant, probably, is the attendance rate for those ages 3-4, which is 59%.

So it’s sort of true. Particularly odd on that table is the low overall college attendance rate of Asian women, who are far and away the most likely to go to college at the “traditional” college ages of 18-24. That’s because they are disproportionately over age 25 (partly because many have immigrated as adults). But, if you just limit the population to those ages 18-54, Black women still have the highest enrollment rates: 15.5%, compared with 14.6% for Asians, 12.6% for Hispanics, and 12.4% for Whites. Asians are just the most likely to be over 25 and not attending college, most of them having graduated college already.

This does not diminish the importance of high enrollment rates for Black women, which are real — after age 25; the pattern is interesting and important. Here it is:

Under age 25, Black women are the least likely to be in college, over 25 they’re the most likely. This really may say something about Black women’s resilience and determination, but it is not a feel-good story of barriers overcome and opportunity achieved. And, despite her presence in the videos and stories illustrating this meme, it is not the story of Michelle Obama, who had a law degree from Harvard at age 24.

This is part of a pattern in which family events are arrayed differently across the life course for different race/ethnic groups, and the White standard is often mistaken as universal. I have noted this before with regard to marriage (with more Black women marrying at later ages) and infant mortality (which Black women facing the lowest risk of infant death when they have children young). It’s worth looking at more systematically.

ADDENDUM 6/29/2016: Cumulative projected years of higher education

If you take the proportion of women enrolled in each age group, multiply it by the years if the age group (so, for example, 18-19 is two years), and sum up those products, you can get a projected total years in college (including graduate school) for each group of women. It looks like this:

Note this makes the unreasonable assumption that everyone who says they are enrolled in college in an October survey attends college for a full year. So, for example, Asian women are projected to spend 6.2 years in college on average between ages 18 and 54. What’s interesting here is that Black women are projected to spend more years in higher education than White women (5.5 versus 4.9). But we know they are much less likely than White women to end up with a bachelor’s degree (currently 23% versus 33%). This has to be some combination of Black women not spending full years in college, not going to school full time, or not completing bachelor’s degrees after however many years in school. Attendance may be an indicator of resilience or determination, but it’s not as good an indicator of success.

Life table says divorce rate is 52.7%

After the eternal bliss, there are two ways out of marriage: divorce or death.

I have posted my code and calculations for divorce rates using the 2010-2012 American Community Survey as an Open Science Framework project. The files there should be enough to get you started if you want to make multiple-decrement life tables for divorce or other things.

Because the American Community survey records year of marriage, and divorce and widowhood, it’s perfectly set up for a multiple-decrement life table approach. A multiple-decrement life table uses the rate of each of two exits for each year of the original state (in this case marriage), to project the probability of either exit happening at or after a given year of marriage. It’s a projection of current rates, not a prediction of what will happen. So, if you write a headline that says, “your chance of divorce if you marry today is 52.7%,” that would be too strong, because it doesn’t take into account that the world might change. Also, people are different.

The divorce rate of 52.7% can accurately be described like this: “If current divorce and widowhood rates remain unchanged, 52.7% of today’s marriages would end in divorce before widowhood.” Here is a figure showing the probability of divorce at or after each year of the model:

So there’s 52.7% up at year 0. Marriages that make it to year 15 have a 30% chance of eventually divorcing, and so on.

Because the ACS doesn’t record anything about the spouses of divorce or widowed people, I don’t know who was married to whom, such as age, education, race-ethnicity, or even the sex of the spouse. So the estimates differ by sex as well as other characteristics. I estimated a bunch of them in the spreadsheet file on the OSF site, but here are the bottom lines, showing, for example, that second or higher-order marriages have a 58.5% projected divorce rate and Blacks have a 64.2% divorce rate, compared with 52.9% for Whites.

(The education ones should be taken with a grain of salt because education levels can change but this assumes they’re static.)

Check the divorce tag for other posts and papers on divorce.

The ASA-style citation to the OSF project would be like this:  Cohen, Philip N. 2016. “Multiple-Decrement Life Table Estimates of Divorce Rates.” Retrieved (osf.io/zber3).

Delayed parenting and anti-poverty policy

Here’s a preview of talk today at Brown University’s population center.

My basic argument is that policies intended to prevent poverty by delaying parenthood are mostly misplaced, especially with regard to Black women. Not that delaying parenthood is bad per se, but delaying parenthood in the absence of other improvements in people’s conditions is ineffectual in the aggregate, and actually harmful for some populations.

The delayed childbearing argument features prominently in the recent “consensus” on anti-poverty strategy reached by the American Enterprise Institute / Brookings working group I wrote about here. They say:

It would be better for couples, for children, and for society if prospective parents plan their births and have children only when they are financially stable, are in a committed relationship (preferably marriage), and can provide a stable environment for their child.

Isabel Sawhill, a leading proponent of delayed childbearing as anti-poverty strategy, says in her book Generation Unbound, that she is not telling poor people not to have children, but she sort of is. She writes:

It is only fair to expect parents to limit the number of children they have to something they can afford.

The evidence I offer to help argue that this approach is unhelpful includes this paper (the actual new research for the talk), which shows the risk of infant mortality rising with parent age for Black mothers, a pattern strikingly different from White and Hispanic mothers’ (see a discussion here). Here’s that result:

Adjusted Probability of Infant Death, by Maternal Age: White, Black, and Mexican Mothers, U.S., 2013. Predicted probabilities of infant death generated by Stata margins command, adjusted for plurality, birth order, maternal education, prenatal care, payment source, and cigarette smoking during pregnancy. Data source: 2013 Period Linked Birth/Infant Death Public Use File, Centers for Disease Control.

Of course, infant mortality is thankfully very rare, but it’s the extreme measure for the underlying pattern of women’s health. When infant mortality in a group is higher, their average health is usually worse.

I’m adding to that the following descriptive figures on children’s poverty rates according to how old their mothers were when they were born. This is by necessity limited to children who are still living with their mothers, because I used the Current Population Survey. I show this for all children (black lines), and then for those whose mothers have never married (red lines). The solid lines are official poverty-line rates, and the dotted lines use the Supplemental Poverty Measure. The latter shows lower poverty rates for children whose mothers were younger, because it reflects transfer income and welfare support as well as income from unmarried cohabiting partners.

For children overall (black lines), being born to an older mother appears beneficial in terms of poverty rates. This fits the standard story, in which delaying births allows women to go further in school and their careers, and get married, as well as being more mature and so on. However, for those whose mothers remain unmarried the relationship is much weaker, and there is no relationship to the SPM. To me this undermines the policy of delay with regard to women who have low probability of marriage during their child-bearing years. Which brings me back to Black women.

I estimated the same pattern by race/ethnicity, this time just using the SPM, in a model that controls for child age, sex, nativity, geography, and mother’s marital status (ever- versus never-married). I didn’t control for education, because schooling is also an outcome of birth timing (so if young mothers don’t go to college for that reason, this would show them more likely to be poor as a result). Here’s the result:

For White women there is a strong relationship, with lowest poverty rates for children whose mothers were in their 30s when they were born. For Black and Hispanic women the relationship is much weaker (it actually looks very similar when you control for education as well, and if you use the continuous income-to-needs ration instead of the poverty-line cutoff).

My conclusion is that I’m all for policies that make family planning available, and U.S. women should have better access to IUDs in particular (which are much more common in other rich countries) — these need to be part of better medical care for poor people in general. But I don’t favor this as a poverty-reduction strategy, and I reject the “responsibility” frame for anti-poverty policy evident in the quotes above. I prefer education, jobs, and income support (which Sawhill also supports, to her credit). See Matt Bruenig on the Brookings “Success Sequence” and my op-ed on income support.

Ideals and intentions

Consider this from Sawhill. In her book Generation Unbound, she writes:

‘poor and minority women … themselves do not want to have as many children as they are currently having. Unintended pregnancy rates are much higher among the poor, minority groups, and the less-educated … [free, better contraception] can help poorer and less-educated women align their behavior with their intentions.’ (p. 138)

I think we need to take a little more complicated view of intentions here. She is referring to what demographers call “unintended” births, which means the woman recalls that she was not intending to get pregnant at the time — she either wanted to get pregnant some time in the future, or never. As you can see, such unintended pregnancies are very common:

However, most poor women think the ideal family size is large. Among young women, 65% of women who didn’t finish high school, and 48% of those with high school degrees but no BA, believe 3 or more children is the ideal for a family:

For lots of their births, poor women were not ready, or not planning to get pregnant. But it’s also common for poor people to never achieve their ideal conditions for having children — good job, marriage, housing, education, and so on. In that case, with the clock running on their (and their mothers’) health, unintended childbearing is more complicated than just a behavior problem to be solved. It may reflect a compromise between unachievable goals.

In addition to making sure everyone has the reproductive healthcare they need (including more effective contraception), I think we should also help people achieve their long-term ideals — including having the children they want to have — rather than (just) help them realize their short-term intentions.

Old people are getting older and younger

The Pew Research Center recently put out a report on the share of U.S. older women living alone. The main finding they reported was a reversal in the long trend toward old women living alone after 1990. After rising to a peak of 38% in 1990, the share of women age 65+ living alone fell to 32% by 2014. It’s a big turnaround. The report attributes it in part to the rising life expectancy of men, so fewer old women are widowed.

The tricky thing about this is the changing age distribution of the old population (the Pew report breaks the group down into 65-84 versus 85+, but doesn’t dwell on the changing relative size of those two groups). Here’s an additional breakdown, from the same Census data Pew used (from IPUMS.org), showing percent living alone by age for women:

Two things in this figure: the percent living alone is much lower for the 65-69s, and the decline in living alone is much sharper in the older women.

The age distribution in the 65+ population has changed in two ways: in the long run it’s getting older as life expectancy at old age increases. However, the Baby Boom (born 1946-1964) started hitting age 65 in 2010, resulting in a big wave of 65-69s pouring into the 65+ population. You can see both trends in the following figure, which shows the age distribution of the 65+ women (the lines sum to 100%). The representation of 80+ women has doubled since 1960, showing longer life expectancy, but look at that spike in the 65-69s!

Given this change in the trends, you can see that the decrease in living alone in the 65+ population partly reflects greater representation of young-old women in the population. These women are less likely to live alone because they’re more likely to still be married.

On the other hand, why is there such a steep drop in living alone among 80+ women? Some of this is the decline in widowhood as men live longer. But it’s an uphill climb, because among this group there is no Baby Boom spike of young-olds (yet) — the 80+ population is still just getting older and older. Here’s the age distribution among 80+ women (these sum to 100 again):

You can see the falling share of 80-84s as the population ages. If this is the group that is less likely to live alone the most because their husbands are living longer, that’s pretty impressive, because the group is aging fast. One boost the not-alones get is that they are increasingly likely to live in extended households — since 1990 there’s been a 5% increase in them living in households of at least 3 people, from 13% to 18%. Finally, at this age you also have to look at the share living in nursing homes (some of whom seem to be counted as living alone and some not).

In addition to the interesting gerentological questions this all raises, it’s a good reminder that the Baby Boom can have sudden effects on within-group age distributions (as I discussed previously in this post on changing White mortality patterns). Everyone should check their within-group distributions when assessing trends over time.