06 April 2008

Wealth and Children

Today's USA Weekend Sunday newspaper supplement story "More money, more kids" by Rochelle Sharpe, notes that "Since the 19th century, the nation's poorest families have traditionally had more children than those with large incomes." But also notes that the trend seems to have reversed.

Fewer families whose incomes fall in the bottom 10% of the population are having three or more children: 33.7% of such families had at least three kids in 2006, down from 39.8% in 1990, according to Census surveys.

By contrast, the wealthiest families are having more children. In 2004, an estimated 41.3% of the wealthiest families had at least three children, much higher than in 1995, when an estimated 29.3% of families earning in the highest income bracket (that year, $300,000-plus) had three or more kids.

A source quoted for the story suggests causes like the move of women into the workforce, child care costs, a more secular society, and the inability of children to bring into to a family as they did in the 19th century.

This might be true, but I'm skeptical of the reasons cited because the timing isn't right. The earning capacity of children changed sometime between sixty and a hundred and twenty years ago, and has differed little in the last decade or two.

The move of women into the work force was a decisive change from about 1965 to 1985, but has leveled off a great deal since then, with the more recent shift being towards a greater balance between work and family.

A move towards a more secular society could explain a decline in fertility (although some recent studies suggest that decreased fertility makes a society more secular rather than the other way around), but recent Pew Foundation figures suggest that the shift towards a non-religious take on life has been quite equitably spread across almost all demographics.

My suspicion is that the shifts are more medical than directly economic.

On one hand, improved sex education and greater availability of contraception, among other things, has reduced teen pregnancy to an all time low over a a time frame of the last couple of decades. Pregnancy teens are disproportionately poor and disproportionately likely to become poor, so a meaningful reduction of teen pregnancy is likely to produce a significant reduction in the number of poor families with three or more children. While probably a smaller effect, the combination of declining teen fertility and the availability of legal abortion has already reduced the number of domestic adoptions, an option that made it possible for working and middle class families to increase their family size in earlier years.

On the other hand, the dramatic increase in family size among very high income families probably reflects the increased availability of two important ways for older parents, who are disproportionately high income, to add children to the family. One is fertility treatments which not only make it easier for women who have postponed child bearing to have careers to have children, but also is a major factor behind a rise in the number of twins and higher order multiple births. The other is the increasing availability of foreign adoption as an option, but only for couples that can afford the expensive process. Both of these developments are quite recent.

By age, fertility is up dramatically for women aged 30 and up, and down significantly for women under 30, since 1990. The teen fertility rate has dropped by about 33$, it has dropped by about 12% for women in their early 20s, it has dropped by about 4% for women in the late 20s, it is up by more than 15% for women in their early 30s, it is up by about 50% for women in their late 30s, it is up by about 60% for women in their early 40s and it has roughly tripled for women in their late 40s.

Fertility rates have dropped only slightly for Cuban-American, non-Hispanic white, and Asian-American women, but have dropped greatly for other Hispanic women, black women and Native American women.

Multiple births are up almost 50% since 1990, with a disproportionate share of that growth among triplet and higher order births, and among non-Hispanic whites, both of which indicate an impact from fertility treatments.

In terms of lifetime births per woman the trends are as follows (by ethnicity)

1980 - 1.8
1990 - 2.0
2000 - 2.1
2004 - 2.1

1980 - 2.2
1990 - 2.5
2000 - 2.1
2004 - 2.0

American Indian, Eskimo, Aleut:
1980 - 2.2
1990 - 2.2
2000 - 1.8
2004 - 1.7

Asian or Pacific Islander:
1980 - 2.0
1990 - 2.0
2000 - 1.9
2004 - 1.9

1980 - NA
1990 - 3.0
2000 - 2.7
2004 - 2.9

Since 2002, black fertility has been lower than white fertility, reserving a long term trend.

As of 2006, the higher a woman's family income the more likely it is that she is childless, so the trend towards high income families with large families is offset matched by the large number of high income families that are childless. Of the poorest women 39.4% are childless, of the most affluent (75,000+ of family income) 47.9% are childless.

This isn't to say that there couldn't be economic effects. The very highest income families in America have dramatically increased their wealth in the past decade or so, capturing the lion's share of the nation's growth in GNP. This dramatic rise in prosperity and the feeling of a lack of economic constraints caused by this seemingly unlimited improvement in economic circumstances no doubt has helped the highest income Americans feel more able to afford larger families.

It is also worth noting that if the changes observed in the last couple of years prove to be lasting, they are a return to a historical norm, rather than a departure from it. The last two hundred years have been a population genetics mystery. For most of human history, those at the top of the socio-economic pyramid have had more children who lived to maturity, while those at the bottom have had fewer. This is what a population geneticist would expect: those with desirable traits used to secure better living conditions generally have more descendants; those at a disadvantage who live in substandard conditions generally have fewer descendants.

The economics of industrialized societies, however, have dramatically reduced fertility rates in every developing society. Improved means of contraception have made the transition particularly dramatic in countries like Mexico which have seen the average number of children per woman per lifetime drop by about four children per lifetime in a mere generation. Widespread higher education combined with increased gender equality and assortive mating has made the effect on fertility strongest for the upper middle class. But the latest trend is a return to longer run historical trends.

One wonders if government policy won't impose what we have seen in China, which is mandatory limits on fertility for most people.

Then again, today's paper also brought news of Russia which, in the face of dramatic post-Soviet population declines as life expectancy has dropped by a decade or two, and birth rates have cratered. Russia has responded with public service advertisements extolling the benefits of large families and large economic benefits for new mothers that are particularly large for women having second or later children. These efforts have helped produce a fifteen year high baby boom in the country.

Overall birth rates in the U.S. hit a record low in 2002, and have rebounded only slightly since then. The recent peak was in 1990, which had the highest birthrate for the period from 1972 to the present. The birth rate declined by roughly a third from 1960 to 1972, presumably as a result of improved contraception, continued to decline until 1976, rebounded slightly until 1990, declined slowly until 2002, and then has slowly rebounded again. Part of this change is also an echo of the baby boomers.

Infant mortality has declined almost every year for decades, dropping about 75% since 1960 and more than 25% since 1990.

Divorce rates have declined steadily since 1981 to reach the lowest level since 1970 in 2006, after rising from 1960 or earlier until then. But most or all of the decline in divorce rates since 1981 is attributable to a declining marriage rate over the same time period, which is at the lowest level since well before 1960.

No comments: