Census Bureau Takes a Small Step in Better Describing the Structure of the Modern Family -- but More Can Be Done

by George Masnick
Fellow

On November 25 the Census Bureau released its latest package of tables describing American families and living arrangements. These tables highlight the growing complexity of living arrangements among children—and the challenges that demographers and housing analysts face in charting changing household composition.

Since 2007 these tables have included a breakdown of family groups that identify couples who were not legally married but were joint parents of at least one minor child in the household.  This change reflects the trend for families to increasingly be started by the birth of a child rather than by marriage. Over 85 percent of births to teens are out of wedlock, as are over 60 percent of births to 20-24 year olds and over 30 percent of births to 25-29 year olds. Among those in their 20s co-residence of the parents is usually the norm, but in many cases, marriage does not take place for several years, and may never take place, certainly if the couple splits up.  Prior to 2007, these particular family groups were lumped into the category of “other families” with either a male or female reference person as head.  It was impossible under this old definition to distinguish in the tabulated data when unmarried family groups contained joint parents.

Many who referred to the older data assumed (incorrectly) that if adults in such family groups were not “currently married,” then the child or children were living in a “single”-parent household.  The implication was that unmarried two-parent households would behave more like one-parent households than like married couples across a wide range of issues of importance for public policy, including housing consumption.

The magnitude of the numbers of two-parent families under the old and new definitions can be seen in Exhibit 1.  While only about 7 percent of two-parent families are not married, that number is up from 5 percent in 2007. (Click exhibits to enlarge.)

  

Source: Current Population Survey March and annual Social and Economic Supplement, 2012 and earlier, http://www.census.gov/hhes/families/data/families.html. Table FM-2; http://www.census.gov/hhes/families/data/cps2013.html

In 2013 about 76 percent of all parents of minor children were married. Among young adults with minor children, however, the share that is currently married is much lower than this average (Exhibit 2). Only 43 percent of such parents under the age of 25 are married, as are just 65 percent of parents age 25-29.  The higher shares of older parents of minor children that are married reflect both the lower share of births to unmarried women when these parents were younger as well as the tendency for people to marry later. Whether today’s younger cohorts of parents will carry forward higher percentages of unmarried two-parent and “single”-parent living arrangements when they reach middle age remains to be seen.  I have put the word “single” in parentheses because it refers to legal marital status only, and these parents may well be partnered. 

Source: Census Bureau tabulations of 2013 CPS/ASEC  

When the parents of minor children are broken down by race/ethnicity we can see quite a large amount of variability in marriage/living arrangements (Exhibit 3).  The largest discrepancy is between Blacks and Asians.  Only 51 percent of Black parents are currently married compared to 89 percent of Asian parents of minor children. The share of non-Hispanic White parents of minor children who are married is almost 82 percent. Fully 42 percent of Black parents are in “single” parent living arrangements compared to only 9 percent of Asians. We would like to be able to identify the degree to which these differences are accounted for by differences in age of parents and by nativity status, but the data in the Census Bureau’s releases do not allow us to fully do this.  The data are especially silent when attempting to determine the presence of non-parent adults in the “single” parent category.

Source: Census Bureau tabulationsof 2013 CPS/ASEC  

While identifying joint-parent unmarried couples as a separate category is a step forward, especially among parents in their 20s, a further breakdown of the data is still needed to better describe the modern family.  Married couples consist of persons in their first marriage and those who have been remarried.  If we are now identifying unmarried parents that are both the biological parent of at least one minor child in the household, shouldn’t we also identify married couples where only one parent is the biological parent of any child?  Some “single” parents are living with a partner to whom they are not married, who for all intents and purposes are helping to support the family and acting like a parent.  Some “single” parents are living with non-partner adults who also might be playing parental roles.  Many children are in “joint custody” households.  These “blended” and “extended” living arrangements are all very much part of the modern family, but cannot be readily identified in the Census data, especially by age cohort.

The next steps that the Census Bureau can take to present a better picture of the modern family seem straightforward.  Marital status could include a category “remarried,” and married couples should be further broken down by marriages in which one or both partners are remarried. Unmarried parents of minor children could be broken down by those living with a partner and those not. Among those not living with a partner, the presence or absence of other adults could be identified.  Minor children in married couple living arrangements could be identified as the biological child of both parents or as a stepchild of one parent.  And minor children in the household could be identified as living exclusively in the household or regularly spending some of their time in another household.

Generational differences in living arrangements at the onset of family formation, and the extent to which these differences persist as cohorts age, are key descriptors of the modern family. Therefore, many of the CPS tables should provide the age of the reference parent as a variable that is cross tabulated against other variables.  It would also be helpful if these new tables are produced separately by race/Hispanic origin of the reference parent.  This detailed breakdown by age and race/Hispanic origin will stretch the CPS data quite thin, to be sure, but the user can always aggregate up to gain robustness.

Finally, a few comments about the sharp decline since 2007 in Exhibit 1 in the number of two-parent families with minor children. This decline is certainly related to the effects of the Great Recession.  One reason for the decline is that immigration fell sharply in 2006 and has just begun to recover. Immigrant women have higher fertility than native born and experienced the greatest fertility decline during the economic down turn. These are trends consistent with the poor economic conditions that have affected young adults most severely.  Immigrants also have a much higher share of births to married couples compared to native born (76.4 percent versus 61.2 percent), and the decline in immigration during the Great Recession thus contributed to the recent rise in the share of all births that are to unmarried women.

It is normal that during a recession, both marriages and births are postponed.  A recovery in marriages would be expected to lag the recovery in the economy to allow for some planning of the event.  Meanwhile, both the decline and the recovery in births should each lag the trend in the economy by a year or more.  Although year-to-year instability in the CPS series is often the result of simple random variability, perhaps the upturn in 2013 in the number of families with minor children is further evidence that the economic recovery has begun in earnest. 

Childless Households Have Become the Norm

by George Masnick
Fellow

In 1960 almost half of all households were families with children under 18.  Since then, the number has fallen to under 30 percent (Figure 1).  By definition, the declining share of family households with children exists because households without children have increased more rapidly (Figure 2).  There are many reasons for this trend: delayed age at marriage and later age at childbearing, smaller family sizes, higher divorce rates, and more couples choosing not to have children (Table 1).  The changes in each of these measures over the last few decades are quite striking. In 1960 the median age at first marriage was 22.8 for men and 20.3 for women, compared to 28.6 and 26.6 in 2012.  The share of households with four or more people in 1960 was over 40 percent, falling to just under 23 percent in 2012.  Women who were 25 in 1960 ended their childbearing years in the mid 1980s with only 8.5 percent of them remaining childless. Women born in 1960 finished childbearing in 2010 with nearly twice as many of them childless (16.3 percent). In 1960, only 13 percent of all households were single persons, but by 2012 that percentage had risen to 28. All of these trends result in households having fewer children and fewer households having any children at all. (Click charts to enlarge.)

Source: Current Population Survey March and annual Social and Economic Supplement, 2012 and earlier. Table FM-1.  Minor children numbers from Census Bureau's population estimates for July 1 of each year.

Source: Current Population Survey March and annual Social and Economic Supplement, 2012and earlier. Table FM-1

Source: Census Bureau Current Population Survey historical tables.

The interesting aspect of this long-term trend is that it continued in spite of the strong upswing in the sheer number of American children, which grew after 1990 (also Figure 1).  That increase is due to the largest baby boomers having their own children (the echo boom) and to childbearing by the flood of immigrants who arrived between 1985 and 2005.  (Note that in 2012, fully 87.5 percent of children under the age of 18 who have an immigrant parent were themselves born in this country.) 

To be sure, baby boomer and immigrant childbearing did increase the actual number of households with children.  For example, the number of households with children under the age of 18 increased from 33.3 million in 1985 to 38.6 million in 2012. This 5.3 million increase was far less than the 11.3 million increase in total number of children in the population over this period because many households with children contained two or more children under the age of 18.  More importantly, however, the increase in households without children surpassed the 5.3 million growth of households with children by a considerable margin.

Two key reasons for the recent increase in childless households have been the aging of the population and increasing longevity. The large baby boom generation (age 45-64 in 2010) is now entering the empty nest stage (at least regarding children under 18). Between 2002 and 2012, households with at least one child, headed by today’s 45-64 year old cohort, declined by 12.3 million. There are still 11.5 million 45-64 year old headed households with children, and most will become households without children over the next decade.  Furthermore, empty nest households headed by those over the age of 65 are surviving longer and longer, making it likely that the trend in the decline of households with children will continue well into the future.

Significantly, the decline in the number of households with children accelerated after 2007.  Much of the decline can be explained by the sharp drop in the number of births. Annual births rose from just over 4 million in 2001 to over 4.3 million in 2007, the highest on historical record, but then fell to just below 4 million in 2011.  The total fertility rate (births per 1000 women age 15-44) fell from 69.5  (a 17 year high) to 64.4, a decline of 7.3 percent over this same period. Both the decline in births and the drop in the fertility rate are linked to the decline in immigration that followed the Great Recession. Because newly arrived immigrants are concentrated in the childbearing ages, and because immigrants have higher fertility than the native born, the loss of immigrants has had a disproportional effect on declining fertility.  The effect of the Great Recession on lowering fertility among the native born is also of importance, but this decline could be temporary.  The echo boom generation began to turn 25 in 2010, and has most of its childbearing years yet ahead of it. A return to higher levels of immigration and/or a rebound in fertility could reverse the decline in number of births and ease the long-term decline in the share of households with children, but will not likely reverse it.

A Surge in Hispanic Household Growth? The Challenge of Interpreting Short-Term Trends in Datasets that are Occasionally Adjusted

by Dan McCue
Research Manager

Interpreting year-to-year changes in annual surveys from the Census Bureau can be a tricky business, especially around decennial censuses.  Because it is the largest and most comprehensive count of the population, after each new decennial census is released, the smaller but more frequently issued surveys available from the Census Bureau, such as the Current Population Survey (CPS), Housing Vacancy Survey (HVS) and American Housing Survey (AHS), are updated, or “re-benchmarked” based on the findings from the new decennial census.  Prior to this, these surveys were controlled to extrapolations based off of the prior decennial census. While it is inevitable that ten years of extrapolation can lead controls to drift off course, failing to recognize when and how datasets are re-benchmarked to correct for this drift can lead to misinterpretations about short-term trends.  The danger is that the re-benchmarking adjustment can be misinterpreted as an actual trend that occurred in a single month or year, rather than what it really is: a discontinuity in the data due to an adjustment made to correct the net sum of ten years of extrapolation errors that had accumulated in the dataset since the last decennial census.

Take for instance, the following data overview in a recent online article:

"The latest U.S. census figures, for June, show year-over-year Hispanic homeownership increased by 7.3 percent, from 6.2 million to 6.7 million. For black-owned households during the same time, the numbers dipped by 1.3 percent, from 6.3 million to 6.2 million. Likewise, whites' homeownership also saw a slight decrease of about 1 percent, from 58.4 million to 57.8 million." - National Journal

On its face, this data leads us to conclude that the number of Hispanic homeowners surged from June 2011 to June 2012, while at the same time the number of homeowners among both blacks and whites dropped significantly, and therefore without growth in Hispanic homeownership the overall number of homeowners in the US would have dropped significantly over the past 12 months instead of growing slightly as was reported.

However, the Census Bureau’s Housing Vacancy Survey (HVS) showed that both Hispanic and non-Hispanic homeownership rates dropped during the June 2011 to June 2012 period, a time wherein Hispanics also suffered higher than average unemployment rates. At first glance, the divergence in the two reports is puzzling. However, on the Census Bureau’s HVS website, there is a short but significant sentence under the “Changes in 2012” section of the Source and Accuracy of Estimates web page:   

“Beginning in the first quarter 2012, the population controls reflect the results of the 2010 decennial census.”  - HVS Source and Accuracy of Estimates

This note is important, because the distribution of occupied households by tenure, race, and ethnicity of households is based on these population controls.  Therefore, any changes in the number of homeowners by race and ethnicity that spans across the first quarter of 2012 is also incorporating change due to the shift in the distribution of households by age, race, and tenure that occurred with the re-benchmarking of the survey..

The adjustment to Hispanic households due to the re-benchmarking appears to be significant. Looking at the Hispanic share of households in HVS before and after Q1 of 2012, we can see that the re-benchmarking in that quarter led to a significant upwards adjustment that forms a discontinuity in this series (Figure 1).  The existence of a discontinuity is corroborated by data from the Current Population Survey, which re-benchmarked to the 2010 Census in 2011. The CPS Table Creator allows us to see the impact of the re-benchmarking directly by comparing the Hispanic share of households in 2011 under both 2000 and 2010 Census weights.  It shows that the 2010 census weights raise the Hispanic share of households a full percentage point, from 11 to 12 percent, compared to the 2000 census weights.  In short, this all suggests that results from the 2010 Census found that the 2000 Census-based population extrapolations had been underestimating Hispanic household growth in the 2000s, and therefore these household counts needed to be shifted upwards in 2012 as a correction.

Figure 1:  The Shift to 2010-Based Population Controls in Q1 of 2012 in the HVS Coincides with an Apparent Discontinuity in the Hispanic Share of Householders

Source: JCHS tabulations of US Census Bureau, Housing Vacancy Survey data.

With the change in population controls in the HVS in Q1 of 2012, the amount to which the shift in the distribution of households towards Hispanic households was underestimated incrementally over the last ten years gets corrected all at once, and gets attributed as change measured between Q4 of 2011 and Q1 of 2012.  And as we see in Figure 2, the quarterly change recorded in Q1 of 2012 has a huge influence over our view of the recent trend in household and homeownership growth by Hispanic ethnicity. 

Figure 2: Concurrent with the Switch to Census 2010-Based Population Controls, The First Quarter of 2012  Has a Large Influence on the Recent Trend in Hispanic and Non-Hispanic Household Growth 

Source: JCHS Tabulations of the 1995-2011 AHS

Without the ability to compare alternative HVS household counts for Q1 of 2012 under both 2000- and 2010-based population controls, it is difficult to determine exactly how much of the change in Hispanic and non-Hispanic households and homeowners in 2011 to 2012 was due to the re-benchmarking and how much was due to actual change measurable in the survey.  We refrain from presenting alternative scenarios here, but because the quarter is such an outlier, most assumptions to smooth or discount that quarter of data would conclude with much lower Hispanic household and homeowner growth and much higher growth among non-Hispanics over the past year.  

Different Data Sources Tell Different Stories About Declining Geographic Mobility

by George Masnick
Fellow

The Census Bureau recently released its usual extensive Current Population Survey (CPS)-based package of tables on geographic mobility for 2011-12.  A new feature of this release is a series of historical charts, one of which is reproduced below. Examining just the last decade, mobility rates took a sharp turn downward in 2007-08, with most of the decline occurring in moves between states (Figure 1).  This sharp decline has been the impetus for many stories about the decline in the geographic mobility rate and its implications for housing (see this example).  Most have assumed that the mobility decline was caused by the Great Recession: with reduced job opportunities across the country, there was less inducement to change residence in search of employment, particularly among young adults who were unable to leave the parental nest. Some have asserted that the loss in housing values and tight mortgage lending have “locked in” owners who otherwise would like to move, especially those owners who are now under water on their mortgages.

Figure 1

Source: U.S. Census Bureau, Current Population Survey, 1948-2012, select years.

Yet some recent efforts to scrutinize mobility rate trends and associations have raised doubts about the basic facts. Research at the Minneapolis Fed suggests Interstate Migration Has Fallen Less Than You Think.  Another series of papers have debated the strength of the association between negative equity and reduced mobility. Findings published in 2010 that owners with negative equity are one-third less mobile were challenged as largely a result of the authors dropping some negative-equity homeowners' moves from the data.   The challenge received a rebuttal that was lukewarm at best.

But a more fundamental question is whether there was indeed as sharp a decline in mobility in the late 2000s as the CPS data in Figure 1 suggests. Since 2006 the Census Bureau’s American Community Survey (ACS) has also provided annual estimates of mobility rates that are consistently higher than those of the CPS and suggest a different trend (Figure 2). One factor contributing to higher ACS rates could be that, starting in 2006, the ACS included in its sample the more mobile institutional population. In contrast, the CPS sample excludes most people that live in group settings such as correctional facilities, military barracks, and college dormitories.  The Census Bureau has recalculated the ACS mobility rate based only on the population living in households for 2006 through 2009.  These modified ACS rates plotted in Figure 2 are significantly lower than those with the group quarters population included, are in line with the long-term more gradual decline in the pre-2000 CPS trend, and definitely do not show as sharp a decline around 2007.

Figure 2

Source: Current Population Survey (CPS) and American Community Survey (ACS) published tables.  The Census Bureau has recalculated the ACS mobility rate based on population living in households for 2006 through 2009 (www.census.gov/prod/2011pubs/p20-565.pdf).  The ACS did not cover the entire U.S. until 2005.

The differences between the ACS and CPS mobility rates in Figure 2 are supported by additional analyses of inter-county and inter-state migration trends from these two sources. This research also shows that the ACS levels and trends are mirrored almost exactly by migration rates from IRS data, further adding credence to the ACS. Mobility rates of household heads calculated from the American Housing Survey (AHS) also closely follow the levels from both the ACS and the IRS data.  The persistently lower rates of mobility in the CPS since 2000 are not well understood, but might be explained by the CPS data being collected primarily by a telephone survey that might not fully reflect the recent growth of cell phone-only households - assuming that those households contain persons that are among the most mobile.  (The ACS is primarily a mail survey, IRS data are from filed tax returns, and the AHS follows the occupants of particular housing units over time.)

If there is a story in the ACS trend, aside from one of gradual decline over the long-term, it is that the period immediately leading up to the Great Recession was one of above-trend mobility.  More people were moving than might have been expected during the peak of the housing boom. IRS migration trends in the analysis cited above support this story as well.  The bursting of the housing bubble has mostly just returned geographic mobility rates to their long-term trend.  The long-term decline in mobility is likely due to a host of broad social, economic, and demographic trends: the aging of the population; delays in the transition to adulthood; the increase in dual-career households; the changing race/Hispanic origin of the population; more working from home; more homogenized employment opportunities across different locations; the increase in long-distance commuting patterns; etc.  Absent another housing boom, we should expect near-term mobility rates to continue to gradually decline.

Update on Joint Center Household Projections

by Dan McCue
Research Manager

In December 2012, the Census Bureau released a new set of national population projections that both incorporates data from the 2010 decennial US Census and updates the projection methodology.  These new projections are lower than those issued in 2008 due to reduced assumptions for both net immigration levels and fertility rates – the effects of which are only partially offset by lower assumptions used for mortality rates.  In the near term, the lower immigration assumptions account for over 80 percent of the population growth adjustment.  For example, in years 2015-2025, net annual immigration levels in the new projections average 898,000 (nearly 40 percent lower than in the 2008 series) and net additions from natural increases average 1.7 million (lower by about 6.5 percent).  Overall, total population growth projected for this period is now 6.8 million less than that which was called for in the 2008 projections.

What does such a large downward adjustment to projected population growth imply about expectations for future household growth?  According to the methodology used in the JCHS’s2010 household growth projections, the new 2012 Census Bureau population projections lead to a new projected baseline household growth trend that is directly between the low- and high- scenarios projected in 2010.

Figure 1: How the New 2012 Census Bureau Population Projections Impact the 2010 JCHS Household Growth Projections

This is because the 2010 JCHS household growth projections considered the 2008 Census Bureau population projections as its high-growth scenario, while the 2010 JCHS low-growth scenario cut immigration levels to half of those in the 2008 Census series resulting in population growth even more modest than the new 2012 population projections. Figure 2 shows the differences in adult population and growth between the three different scenarios.

Figure 2: Population Growth in the 2012 Census Population Projections is within the High- and Low- Assumptions used in the 2010 JCHS Household Growth Projections

Immigration remains the greatest source of volatility in population growth projections and population growth volatility is still a major driver of household growth volatility. Therefore, future household growth projections will require a range of sensitivity for potential future immigration levels. 

In addition to the new Census population projections, another major consideration within the household growth projection is the rate at which people form independent households (or the headship rate).  The 2010 JCHS household growth projections were based on headship rates that prevailed during 2007-09. Since then there has been some falloff in these rates as the weak economic recovery has been associated with increases in families doubling up and fewer young adults living on their own.  However, assuming headship rates don’t experience a substantial additional falloff, projected household growth levels over the next decade under the 2010 JCHS methodology are not nearly as sensitive to differences in headship rates as they are to differences in population growth projections as shown above.  Applying the rates from 2007-9, as used in the 2010 JCHS household growth projections, results in an estimate of 12.8 million additional households over the 2012-2022 period.  When the more recent 2011 ACS household counts are used instead for headship rates, the estimate falls 3.8 percent to 12.4 million households, which amounts to 40,000 fewer households per year.  A very similar estimate of 12.5 million results from using headship rates based on household counts in the 2010-2012 CPS.  In short, depending on the headship rate assumption used, the updated Census population projections suggest that household growth over the next ten years should fall in the range of between 12.4 and 12.8 million, which represents an increase of between 6 and 9 percent over our previous low series estimate.