Pub. 4 2016 Issue 3

Issue 3. 2016 13 headline says it is. Well just to throw another curve ball, guess what. That headline number you just heard reported is going to change anyway! That’s right, nearly every major source of hous- ing data revises their figures, often multiple times and sometimes by almost comically large margins. There have been times when some in the media have caught on to this, the New York Times noted that “in the 10 months from April 2008 through January 2009, the current estimate is that the economy lost 1.9 million more jobs than the initial reports indicated.” 1.9 million jobs is bigger than the average annual employment gain since 1940, so this is not an insignificant figure. That was the Bureau of Labor Statistics, but there are plenty of other examples also. Late last year the Bureau of Economic Analysis (bea.gov) revised the U.S. savings rate two months in a row, from 5.6% to 5.0%, and then from 5.0% to 4.4%. That’s over 20% of revision in 2 months, and represented $160 billion in savings that we thought we had in our bank accounts disappearing (interestingly, the resulting upward revision to Gross Domestic Product suggested this unforeseen spending was mostly directed toward health care costs!). But it’s not only the government that faces these challeng- es. The National Association of Realtors had what could only be described as a massive data mea culpa in 2011 as reported by CNN Money, “NAR said it plans to downwardly revise sales of previously-owned homes going back to 2007 during the release of its next existing home sales report on Dec. 21 …. While NAR hasn’t revealed exactly how big the revision to home sales will be, the agency’s chief economist Lawrence Yun said the decrease will be “meaningful.”” Turns out NAR had been base-lining a lot of their data sample on information out of the 2000 Census. Unfortunately, the nature of the U.S. housing market changed a LOT over the following decade, and they fell into a statistical pit of quicksand. #4 What is Single Family Anyway? If someone asked you what a single family house was, what would you say? Most people in the housing industry would probably talk about owning the land, mentioning terms like fee-simple, or detached, maybe with an image in mind of a yard, possibly throw in a tree with a tire swing and a golden retriever. But interestingly, the U.S. Department of Commerce doesn’t quite see it that way, and when we hear about data that mentions single family homes from the government, it can mislead us if we don’t think about it from their perspective. The Census provides a good description, “The single-family statistics include fully detached, semi-detached (semi-attached, side-by-side), row houses, and townhouses. In the case of attached units, each must be separated from the adjacent unit by a ground-to-roof wall in order to be classified as a single-family structure. Also, these units must not share heating/air-conditioning systems or utilities, such as water supply, power supply, or sewage disposal lines.” Note no refer- ence of land ownership, it doesn’t have to be fully attached, and certainly no golden retrievers! For those interested in new home statistics this is important, as apart from stacked condominiums, the vast majority of the new for-sale housing market is going to be considered single family by our government. That means the remainder in multi-family is largely going to be rentals, with a minority of stacked for-sale condos that partially contribute. This also will vary from market-to-market, since Miami is going to have a lot more condominiums than Omaha. So the next time you hear about Single Family permits, sales, or housing starts, keep in mind that it’s a pretty big net being cast. #5 Can’t Always See the Trees in the Forest This title sounds like a mixed up metaphor but it’s the reality when it comes to a lot of national housing data series. Remem- ber our earlier example about Census New Home Sales? Our friends at the Census provide a lot of great data, but they don’t count every single new home sales contract nationwide. Instead, they rely on a survey of permit activity. There are 20,000 permit issuing places nationwide that send in their information to the Census Bureau. But the data that gets sent in is fairly limited, and doesn’t indicate why the structure is being built or if they are planning on selling it. So the Census has 900 Survey of Construc- tion offices that they communicate directly with to find out a bit more information. For new home sales, they survey 1 out of 50 single family permits (remember single family is a big tent for the Census!) to find out if the home has been sold. So, let’s think about that, 1 out of 50 permits in 900 out of 20,000 permit issu- ing places. That’s not exactly a huge sample size given the results will be posted for the entire nation. Because of this, the Census only releases information on New Home Sales (and a number of other series, like Housing Starts) on the regional level; they can’t  Housing Data Mistakes — continued on page 14