Once a puzzle, always a puzzle: Reading Housing Data

Posted by anamaria

Fri Oct 30th, 2009 08:40 AM

Share |

We couldn't help but take note of today's article in the Real Deal that notes the downside of giving too much credence to national housing data.

"With the glut of housing data and statistics available, it's difficult to know which figures give the most accurate representation of home sales and prices. In Manhattan, the disparity between national housing figures, such as average home price and sales, and city numbers can be particularly noticeable. Rather than one national market, there are, in reality, many mini-markets to evaluate, according to broker Douglas Heddings, president of the Manhattan-based Heddings Property Group at Charles Rutenberg Realty. Heddings told Fox Business News that it's unwise for both homebuyers and mortgage lenders to rely on monthly national data to determine housing trends. The data "can be incredibly confusing to the buying and selling public," Heddings said."

Let's take a peek at the various aspects of reading data, and ways to avoid the pitfalls in trying to digest it.

Seasonality: It's not news that real estate is highly seasonal. This means buyers buy in the spring and fall, renters lease in the summer and most activity is dead in peak winter months, each and every year for the most part. To adequately analyze housing data, you need to compare numbers to those of the same "season" last year. This is why month to month comparisons fail to see the big picture. Rather than waiting a whole year to compare data as it is generated, researchers "seasonally adjust" data to make it more useful and relevant, smoothing it out over the course of the year. Pay attention to the numbers quoted: seasonally adjusted data is reported as "SA", and not seasonally adjusted data is reported as "NSA". The trick is knowing which is which and how to read it. In a period of high seasonal volume, the adjusted numbers will be lower than the not adjusted, and vice versa, precisely due to this smoothing out process. Reading that SA housing starts are up by 20%, for example, doesn't mean that starts themselves are up by that much; rather that they beat the expectations of the smoothed out numbers we would have seen had we ignored seasonal influences. Understsand the nature of the numbers you are reading, SA or NSA, and read analyses through those respective lens.

Margin of error: New home sales data comes out monthly, only to be revised up or down a some time later (same goes for unemployment figures, jobless claims, home prices, etc.). Needless to say, when the margin of error % is greater than the actual reported change in sales, the released figure becomes meaningless. Since the markets are forward looking, few people actually look back to see the revised numbers, relying purely on the first-reported estimates. Compare the margin of error with the degree of change being reported to gauge how meaningful the data really is, and don't neglect revisions.

Trend numbers are so last year: Trend numbers imply a linearity of sorts. One could look at prices in February versus May, for example, draw a straight line and conclude the degree of movement (falsely assuming the data reflects the same 1-bed that sold for in February for $600k is now selling for $550k). What such trends neglect is the actual shift in inventory from month to month or quarter to quarter. The key question is: Is there a seasonal difference in actual market inventory, what does it look like and how significant is it? Observe the changing inventory of what you are comparing as a backdrop against which to analyze the data.

Beware of sequential reporting: Take month-on-month and quarter-on-quarter data analysis with a grain of salt, as it neglects the very seasonality we've been discussing. Of course Q2 will be busier than Q1, for example; this happens every year. This is why researchers primarily use seasonally adjusted numbers versus not seasonally adjusted data. Year on year comparisons (y-o-y) provide a more accurate perspective on market activity. Do not make decisions or enter negotiations relying solely on quarter-on-quarter data.

Year on year imperfections: While Y-o-Y data is the gold standard, even it is imperfect. Great examples can be found on the Lower East Side and Midtown East, where a plethora of new condo developments have significantly skewed year-on-year sales numbers upwards based on luxury inventory which previously did not exist. Neighborhoods have evolved significantly over the last few years and will continue to change over time. Analyze year on year data with an understanding of neighborhood-specific developments.

To tidy up all of these points and wrap'em with a ribbon, not so long ago, we came across a WSJ article mentioning that NYC housing prices were flat, only to add in a small caption that the NY data did not include co-ops and condos. Reader beware. Don't take headlines at face value, particularly national headlines. While there are nation-wide, macro dynamics at work, real estate has and always will be a local game, with all the pros and cons that come with that.

So, while Noah and company at UrbanDigs will still provide real time analysis on changing trends in the Manhattan residential marketplace ("Expect Significant Quarter-to-Quarter Improvements"), their will always be a caution tag attached.