AP Again Pretends That 'Housing Starts' Equals 'Housing Construction,' This Time Falsely Claiming a 60%, 3-1/2 Year Improvement

Once again, a reporter from the Associated Press, aka the Administration's Press, has told a major fib about the situation in the new-home construction industry, thereby vastly exaggerating its degree of improvement -- claiming a 60% surge during the past nearly 3-1/2 years when it has been 15% at most.

Today's figures from the Census Bureau on housing starts weren't terrible, but they surely weren't cause for major optimism -- except at the AP, where Martin Crutsinger cited "steady progress in the housing recovery" and committed the same serious mistake other AP writers have made (examples here, here, and here), namely pretending that the term "housing starts" has the same meaning as "home construction."

Total seasonally adjusted starts, which were predicted to be 767,000, instead came in at 750,000. That was up by 2.3% over July's downwardly revised 733,000, and a tad below June's 754,000. The single-family component had a better month, as its 535,000 starts were up by 5.5% over July but only 0.8% above June.

Now let's look at how Crutsinger reported it:

U.S. builders started work on more homes in August, driven by the fastest pace of single-family home construction in more than two years. The increase points to steady progress in the housing recovery.

The Commerce Department said Wednesday that construction of homes and apartments rose 2.3 percent to a seasonally adjusted annual rate of 750,000 last month. That's up from 733,000 in July, which was revised lower from last month's initial estimate.

Single-family housing starts rose 5.5 percent to an annual rate of 535,000 homes, the best pace since April 2010.

... The rate of home construction has risen nearly 60 percent since hitting a recession low of 478,000 in April 2009.

Crutsinger is definitely wrong in his characterization of the single-family situation, and egregiously wrong in his claim that "the rate of home construction has risen nearly 60%" since April 2009.

Single-family home construction is NOT occurring at "the fastest pace ... in more than two years." Starts are, but "construction" isn't:


"Home construction" has three elements, which answer the following questions: "What did you begin working on?" (starts); "What are you already working on?" (under construction); and "What did you finish?" (completions).

As seen above, while single-family starts are indeed at their highest seasonally adjusted level since April 2010, the number of units under construction is the same as it was in November 2010 (21 months ago), and completions are below, if barely, where they were just four months ago.

If one wants to identify where gradual improvement began, it would be fair to see when starts finally began consistently outnumbering completions. That first occurred, after almost four years of going in the opposite direction, in November 2011, only ten months ago.

What isn't fair or accurate is claiming, as Crusinger did, that August represented "the fastest pace of single-family home construction in more than two years." It wasn't.

Looking at the same elements above, Crustinger's claim that "the rate of home construction has risen nearly 60 percent since ... April 2009" is absurdly and scandalously wrong. As I stated when AP reporter Christopher Rugaber made a similar error in May: "I would call that assertion 'horse manure,' but that would be unfair to equine excrement."

The truth:

  • Overall starts -- the annual rate of groundbreaking -- are actually up 57% (750,000 compared to 478,000).
  • Units under construction in August were 492,000, or 27% below the April 2009 level of 677,000.
  • Units completed, at a seasonally adjusted 689,000 in August, were the best since June 2010, but were 19% below the April 2009 figure of 846,000.

The "rate" or degree of change in all three items listed is relevant to determining the change in the "the rate of home construction." It sure as heck isn't limited to what has happened with housing starts.

The best way to get a handle on the value and direction of "home construction" activity is to look at the "residential construction" component of real gross domestic product. Though that number includes home remodeling,  its GDP (in 2005 chained dollars) came in at an annualized $359.7 billion in the second quarter of 2012. That's up 7.8% from the $333.7 billion seen in the second quarter of 2009, which of course contains Crutsinger's April 2009 reference point. Even acknowledging a bit of improvement during the current quarter and that April 2009 was probably worse than the two months which followed, there's no way the new-home component of residential construction is up 60% during the past 40 months.  I would give it 15% at best (it's probably more like 10%, but I'm in a generous mood tonight), meaning that Crutsinger misled his readers by a factor of at least four.

Over a year ago, someone who is now a former AP reporter tried to claim in a conversation that treating "home construction" as identical to "housing starts" is okay because the average reader, listener, or viewer doesn't really understand what "housing starts" are. Give me a flippin' break. Equating the terms is misleading and deceptive. How much more obvious can it be?

What I wrote in connection with Rugaber's error in May holds with Crutsinger's narrative, namely that it's "the kind you would expect a news organization seeing the above data to withdraw." But as I also wrote then: "Readers here know that's not going to happen."

Cross-posted at BizzyBlog.com.

Economy Business Coverage Housing Media Bias Debate Bias by Omission Labeling Wire Services/Media Companies Associated Press Christopher Rugaber Martin Crutsinger