Monday, August 04, 2014

Goldman Sachs' Hui on House Price Seasonal Adjustments

by Bill McBride on 8/04/2014 01:03:00 PM

A few excerpts from some analysis by Goldman's Hui Shan: "Are investors getting too pessimistic about US housing"

This month’s housing data showed weakness in both price and activity measures. On the price side, market consensus expected 0.3% month-over-month growth in the S&P Case-Shiller 20-City Composite. The actual print was -0.3%, the first decline since January 2012 when house prices bottomed. ...

In our previous research, we showed that seasonal adjustments have become a tricky business for housing owing to heightened distressed sales over the past few years (see “Enough about the weather: Let’s talk seasonal adjustments”, Mortgage Analyst, February 28, 2014). In the US housing market, more people buy and sell homes in the summer than in the winter because of the school calendar. A thinner market in the winter implies that sellers typically accept lower prices to close transactions. As a result, seasonal factors would nudge prices up in the winter and down in the summer to arrive at the seasonally adjusted series. Normally, such adjustments are within the range of plus/minus one percentage point.

The recent housing downturn is the most severe in US history since the Great Depression, featuring a wave of distressed sales. According to the National Association of Realtors, at the worst of the crisis, nearly half of home sales were distressed. Because distressed sales take place throughout the year while non-distressed sales are more concentrated in the summer, distressed sales account for a larger share of total sales in the winter than in the summer. Also, because distressed properties tend to transact at a significant discount relative to non-distressed properties, they drag down observed transaction prices and the effect is more pronounced in the winter (when the distressed share is high) than in the summer (when the distressed share is low). This amplifies the normal seasonal pattern. In recent years, seasonal adjustments have expanded from the normal “plus/minus one percentage point” to “plus/minus three percentage points”.

As the housing market recovers and the share of distressed sales drops, the true underlying seasonal pattern is normalizing. However, seasonal factors are usually derived using data from the past 5-7 years. In other words, we are using the amplified seasonal factors to adjust more muted seasonal patterns, meaning pushing up house prices too much in the winter and also pushing down house prices too much in the summer. This is consistent with what we have seen over the past year: the S&P Case-Shiller house price index beat consensus expectations at the beginning of the year and it was surprisingly weak for the May reading. This seasonal adjustment problem is likely to persist in coming months, causing the next few months’ house price prints to appear softer than they really are.
emphasis added
CR Note: This is similar to the argument I made last week A few comments on the Seasonal Pattern for House Prices. The seasonal adjustment should normalize over the next few years.