US Demographics and Housing

Expectations of a robust housing recovery are not well supported by US demographics.

From Bloomberg News (February 8, 2012): Chief Executive Officer Jamie Dimon told investors and analysts in a January conference call that housing is “getting closer” to a bottom. “We’re going to add 3 million Americans every year for the next 10 years. That’s 30 million Americans who need 13 million dwellings,” he said.

Mr. Dimon’s estimate looks too optimistic on two accounts.  First, the US population will more likely grow by 23 to 26 million in the next 10 years.  And second, the demographic bracket which includes over 80% of home buyers will grow at a much lower rate.

The US population grew by 2.8 million people in the sixteen months from April 2010 to July 2011, but over 700,000 of these 2.8 million were new immigrants.  In the remainder of this decade, the population will grow by an average of 2.6 million people a year, assuming a more typical 1 million new immigrants per year.  My estimates are derived from data compiled by the Center for Disease Control (which tracks life expectancy among other things) and by the US Census. In the next decade 2020-29, the population will grow by an average 2.3 million per year and in the following decade by 0.9 million per year, again assuming 1 million new immigrants per year.  Note that without immigration, the population would shrink in 2030-39 and in 2040-49, as I argued in America Heading Towards Zero Population Growth?.

It is not enough to tabulate the number of new Americans (newborns or immigrants) to estimate the likely impact on housing.  We have to also look at the likelihood that they are in an age segment and in an economic bracket that will lead them to spend a fair amount of money on a home, regardless of whether they are renting or buying.  An ideal scenario would be a large increase in the population segment of young and middle-aged Americans who can be characterized as middle-class or richer.  But this scenario is unlikely in the next decade.

Of the 2.6 million annual addition to the population in the present decade, 4.2 million will be from new births and 1 million from new immigrants.  And because there will be 2.6 million deaths, the sum total of these three figures comes to a 2.6 million (4.2 + 1 – 2.6 = 2.6) addition.

The US Census recently disclosed that foreign-born households are made up on average of 3.4 people, more than the 2.5 people average for native-born households. Assuming new immigrants are in similarly-sized households, one million immigrants would absorb about 294,000 dwellings per year. Incidentally, there were only 700,000 immigrants in 2011, equivalent to a demand for 206,000 units.

Outside of immigration, the annual growth in population in the present decade will be 1.6 million.  Babies don’t buy homes but growing families with new babies may do so because they need more space.  In this case however, they are vacating one home and moving into another, resulting in a net demand of zero in terms of units (say trading a 3 bedroom for a 4 bedroom home) and a subdued net positive demand in dollar terms (say trading a $250,000 home for a $300,000 home).  At the other end of the age spectrum, older people who pass away often leave a home vacant, adding to the supply of existing homes for sale.  The demand for upgrades is debatable but the supply from deaths is certain, resulting in a murky overall supply-demand situation.

A better way to look at the demand is to estimate the size of the population of people who are likely home buyers.  75% to 80% of home buyers are in the 30 to 60 age bracket (see footnote). The remaining 20% to 25% of home buyers are aged less than 30 or more than 60 and are generally buyers of smaller homes. In terms of dollar value, the 30-60 bracket could therefore represent as much as 90%+ of the residential real estate market. This bracket now has about 125 million people but it has stopped growing (excluding new immigration) in 2005 and will not resume its growth until after 2020.  If the overall pool of likely home buyers is not growing, then people are just trading homes amongst themselves without any significant net gain in overall demand for new homes. Upsizing by growing families could be largely matched by downsizing from older couples whose children have left home.

Therefore if you assume no demand from existing residents and demand of 294,000 homes from new immigrants, the total net demand in the next ten years would be around 3 million units, far fewer than the 13 million estimated by Mr. Dimon. Total home construction will probably exceed this 3 million figure because of some demolition of older homes and because supply and demand are often not in the same location.  Florida or Texas may see net demand while some other regions see net supply.

Some attention has been given to other measures of demand such as housing affordability and the rate of household formation. The National Association of Realtors said in March that the housing affordability index rose above 200 for the first time since recordkeeping began in 1970. An elevated index is an indication of high affordability.  But in any transaction, there is the ability to buy and the need or desire to buy. The affordability index shows that the ability to buy is strong but it is not indicative of the need to buy. If people are already in homes and there are few incremental buyers, housing affordability is not an indication that demand will soon return. If the index is high because demand is muted and interest rates are low, it does not follow that greater demand will inevitably return. That would only be true in the context of favorable demographics such as we have had in previous recoveries.

As to household formation, it is still low but seems to have bounced in the first quarter.  It declined after 2007 because of a fall in the divorce rate and because more young people stayed under their parents’ roofs.  The youth unemployment rate has recently been improving, raising expectations that household formation will also recover. But even during the good times, the young demographic contributes a relatively small demand in terms of units and an even smaller demand in terms of value (since they tend to live in smaller and less expensive housing). Furthermore, counting the young people who enter the house-buying age bracket and neglecting to count the older people who simultaneously leave this bracket only presents half of a full picture.

Outside new immigration, there will be little demand for housing from demographics until the end of this decade.  When the 30-60 age bracket starts growing again in the 2020s, housing demand will grow again but it will be weaker than in 1985-2005 and will have to contend with the new supply of homes vacated by rising numbers of dying or downsizing baby boomers.

As to the exchange-listed home builders, two factors may work in their favor despite the poor demographic backdrop. One is that real estate is famously local, which means that, with sufficient research, they may be able to focus on areas of the country which are growing even in a time when the overall market is not.  For example, some parts of Florida or Texas may see more demand than supply while some regions in the Midwest or Northeast see more supply than demand. The second factor is the trend towards urbanization with more people choosing to live in cities instead of suburbs. In this vein, several home builders are shifting some of their focus to multi-family developments.

The main risk to these forecasts arises from the difficulty in assessing the impact of illegal immigration. It clearly contributes to some incremental demand, perhaps not in the early years after a person’s arrival, but certainly after a few years. But here again, the data is mixed because, as noted by the Pew Research Center, the inflow from Mexico has recently reversed for the first time.

Footnote: The National Association of Realtors publishes every year a Profile of Home Buyers and Sellers. The Profile is compiled from responses to a survey which the NAR mails to tens of thousands of consumers who purchased a home during the period under study, usually July to June.  For 2010, the NAR mailed 111,004 surveys and received 8,449 responses, or 7.9%, and for 2011, it mailed 80,099 surveys and received 5,708 responses, or 7.3%. It is not clear whether these samples are large and diverse enough to be representative of all home buyers. It is possible, or perhaps even probable, that the profile of survey respondents differs markedly from that of home buyers in general. Notwithstanding the methodology, the 2011 Profile shows a marked jump in the average age of home buyers from 39 in 2010 to 45 in 2011.  This jump is directly related to the expiration of the first-time home buyer tax credit which was still on offer in 2010 (and 2008-09) but not in 2011.  Although the age distribution of buyers in 2010 showed a large percentage in their twenties, we view this as a discrepancy caused by the tax credit. In addition, it is very likely that the current historically very low mortgage rates have skewed the age distribution towards younger age groups.  We are comfortable with our estimate that, under normalized circumstances, 75 to 80% of home buyers are in the 30 to 60 age bracket.