For now, we only have data up through mid 2014, but we can get a sense of the overall trend.
(These numbers are from the Colorado Demography Office and are based on population and household data. The total household number is based on occupied housing unit data, and the housing unit data originates with the US Commerce Dept., and includes both single-family and multifamily.)
In 2014, new household formation and new housing units were fairly even matched.We can see that both increased by about 1.25 percent. Specifically, housing units increased by 1.2 percent while households increased by 1.3 percent (for Colorado Springs):
Not all years are as evenly, matched, though. We can see back in 2010 there was a lot more household growth than there was new unit construction. Back in 2003, there was a lot more housing construction than there was new household growth.
Not surprisingly this reflect what we might have suspected all along — namely that during the housing boom, housing construction was outpacing new household formation, while it has been the opposite since the financial crisis.
To look at this a different way, let's look at the difference between the total number of new units and the total number of new households.
In this graph, for example, we see that in 2014, there were 94 more households formed than there were new units constructed. By contrast, in 2003, there were 6498 fewer households formed than there were new housing units.
Since 2008, this points toward a tight housing market, and will likely mean low vacancy rates, since we know from this that only a small portion of new units have been multifamily (for Colorado Springs):
