While Capital Economics is forecasting housing starts to surge by almost 40% over the next few years, reaching 1.5 million by the end of 2017, they’re not so sure the industry will see a return to pre-recession 2006 levels.

It’s possible, write Ed Stansfield and Andrew Hunter in a client note, that starts could rise even more rapidly in the short term as builders make up for under-building in recent years.

“But there are good reasons why we are unlikely to see starts regain their pre-recession highs of more than 2 million in the foreseeable future,” they write.

They argue that the near-term prospects for the homebuilding recovery are increasingly bright. After some weather-related difficulties early this year, housing starts recently rebounded to a seven-year high.

“We expect that strong jobs growth combined with low mortgage rates and an easing in lending standards will support a steady strengthening in demand,” they say. “That will help to push starts above 1.2 million by the end of this year, and to 1.5 million by end-2017.”

Few argue against the fact that the surge in homebuilding in the run-up to the recession was unsustainable, resulting in housing completions consistently outstripping household formation and a sharp increase in the vacancy rate.

But there is evidence of a downward trend in this measure over the past 50 years, or more precisely a shift to a new, lower, long-term average in starts per capita since 1990.

“This reflects a major demographic shift in the form of a reduction in the average number of people per household. This trend meant that, throughout the 60s and 70s, an increasing share of the population required a home of their own,” they say. “The high levels of homebuilding during this period, both in absolute and per capita terms, were therefore justified by strong rates of household formation.”

This means that the high levels of homebuilding throughout the 1960s and 70s were, at least partially, justified by stronger demand for housing, they say. As the size of the average household declined, a higher share of the population needed a home of their own. More housing units per capita were therefore built to meet this higher demand.

“The balance of evidence suggests that housing starts per capita have settled at an average level which is materially lower than what was typical in previous decades,” they write. “Accordingly, the 2006 highs of more than 2 million starts, or more than 6 starts per 1000 people, were excessive relative to new long-run norms. And unless there is another significant demographic or cultural shift on the horizon, there is no longer scope for a sustainable rise in starts back to these levels in the foreseeable future.”