This past spring has been the season of home builders' discontent. A hoped-for recovery in housing markets is little in evidence. Builders are erring on the side of caution. The publics keep yanking their previous earnings guidance and remain mum about what the rest of the year will look like for the simple reason that they just can't tell what is ahead.

Eric Belsky The speed with which new-home demand dried up and inventories of unsold vacant homes skyrocketed caught us by surprise. Even more, it has shaken builder confidence in the long-run outlook for housing. Preferring to be surprised on the upside, many question the ballyhooed consensus that demand for new housing (satisfied by single-family and multifamily completions as well as manufactured home placements) over the 2005 through 2014 period will total upwards of 19 million.

Trust in more aggressive forecasts of housing activity has eroded even more. In 2003, a group of chief economists of housing and mortgage finance trade associations and firms theorized an upper estimate by pegging average annual new-home demand from 2004 through 2013 at 2.17 million in its high scenario–above even the level of production in 2005 and 2006. That now seems too aggressive, given the battering the market has taken from that level of production.

There is reason to think it is overly rosy, as it assumes that immigration will run at an annual rate of 1.7 million. Still, in late 2006, the NAHB projected that new-home demand would likely total 20 million from 2006 through 2015. That's well ahead of the 18.1 million from 1995 to 2004.


Is it time to declare the long-term outlook a bust? Don't count out just yet the fundamentals that will likely drive the market higher over the next 10 years. Although it is difficult to project total demand for new homes, there is still reason for optimism that when summed up in 2014, the period from 2005 to 2014 will far surpass total housing completions plus manufactured home placements from the decade earlier.

A cause for optimism: household growth is projected to accelerate by 14.6 million during 2005 to 2015, thanks to strong immigration. This projection assumes immigration of about 1.2 million per year– the rate the nation has achieved this decade.

What's more, our housing stock will be older on average than last decade. Logically, older homes are more likely to get abandoned or demolished than newer homes, so new-home demand to replace scrapped units and units lost to disaster is expected to increase. Also, trends in some of the largest metropolitan areas toward living closer to urban cores will likely accelerate teardowns, since housing in inner-ring suburbs is typically older, smaller, and no longer fits the underlying land value nor meets the needs of people who desire to live there.

As for changes in the number of vacant units, it is clear that we entered this period with a significant overhang of vacant units for rent or sale that ballooned as a result of completions in 2005 and 2006. But with starts falling well below what appears to be the long-run average in 2007, and the same likely to occur in 2008, that overhang will diminish. Underlying demand from increases in sustainable vacancies should compare favorably with the earlier decade because of expanded second-home demand from aging baby boomers and additional vacancies to accommodate the normal turnover of housing as a larger base of households move.

Thus, with household growth projected to increase by about 2 million, replacement demand likely to grow, and demand for vacancies and second homes up because of a larger, older, and wealthier household base, production should come in higher than in 1995 to 2004, despite overbuilding that occurred late in that period and several years of substandard growth likely to unfold to make up for well-above trend completions in 2005 and 2006.

–Eric Belsky is executive director at the Joint Center for Housing Studies at Harvard University.