Update on Household Formations

Written by Richard Hokenson 

The most recent Current Population Survey (CPS) estimate of net household formations is disappointing. They estimate an increase of only 405,000 for the 12 months ending March 2017 (see Chart 1). It is the smallest gain since 2010 and only the second time in the past 7 years that net household formations (demand) were less than housing completions (supply). The slight increase reported in the CPS is surprising because the estimate from the Housing Vacancy Survey (HVS) showed a much larger gain (see Chart 2).


Although it is not possible to calculate annual net household formations by age, one can proxy age-specific developments by calculating the headship rate which is defined as the number of household heads divided the population for specific age groups. For example, the headship rate for persons aged 15-24 is the number of households headed by someone aged 15-24 divided by the number of persons aged 15-24. The age-specific headship rates are displayed in Charts 3-12. The following are the most important results:

  • The headship rate for persons aged 15-24 (persons born 1993-2002) dipped only modestly during the Global Financial Crisis and has remained stable since then as almost all of them entered the job market following the crisis (see Chart 3).

  • The rate of household formations by millennials continues to be pressured. The headship rate for persons aged 25-34 keeps on declining and has more than reversed its post-crisis bounce (see Chart 4). Decomposition by 5-year age groups shows that for persons aged 25-29 years old, there was no post-crisis bounce and their rate of household formations drifts ever lower (see Chart 5). This is the most disadvantaged segment of millennials as they began looking for employment just as the financial crisis was full blown. Older millennials (persons aged 30-34 years old) had shown a postcrisis recovery but that has now fully reversed (see Chart 6). The lack of recovery in the rate of household formation for either group despite robust employment growth is disconcerting.

  • Gen X has fared much better (persons born 1963-1982). The headship rate for persons aged 35-44 years old was only slightly affected by the Global Financial Crisis and remains stable (see Chart 7). The headship rate for persons aged 45-54 reveals a slight negative drift (see Chart 8). This is not negative as it mostly reflects more persons becoming 55 years old (exits from this age group) relative to the number turning 45 years old (entrants to this age group).

    • The headship rate for younger baby boomers, i.e. persons aged 55-64 (born 1953-1962) is very volatile in a relatively tight range (see Chart 9).
  • The downward drift in the headship rate for persons aged 65 years old and over is most likely a reflection of improved longevity (see Chart 10). There are more households where both partners are still alive, e.g. husband and wife. In the past one spouse would have died – the effect is that the number of households (the numerator in the calculation of headship remains the same but the population (the denominator) is smaller which would boost the headship rate. Now the denominator is larger which depresses the headship rate. That phenomenon is true for persons aged 65-74 years old (see Chart 11) and is especially evident for persons aged 75 years old and over (see Chart 12).

Investment Implications:

It is disconcerting that the headship rate for millennials continues to drift lower. It is a bit puzzling since job growth has been robust both for persons aged 25-29 (see Chart 13) and persons aged 30-34 (see Chart 14). While the downside might be that housing-related developments are not as strong as they could be, the upside is that this group’s income is highly discretionary. Their living expenses, especially rent, are minimal. Investment in residential structures may not add as much to GDP but consumer outlays will.


This update was researched and written by Richard Hokenson, as of December 7 2017