Well, that’s not supposed to be how supply/demand works. Demand has clearly tapered off, and we know there continues to be ample supply. So, how do we reconcile the two opposing trends?
Here’s the picture…
And the problem is…
There are two problems: the data used is different, and the timing of the data is different.
- Existing home sales – The primary data is from the National Association of Realtors. It collects closing sale information from associated firms. So, for the incentive program, a sale is counted when it closes, meaning the effect can be scattered through to the final September 30 closing date.
- New home sales – The primary data is from the Census Bureau. It collects sales agreement and deposit information. So, for the incentive program, a sale is counted when the buyer signs the initial papers. (Now we can see why those sales dropped off more quickly than did the existing ones. Unfortunately, because new homes are a small proportion of total sales, they are not a reliable measure of overall activity.)
- Home prices – The best data is from S&P/Case-Shiller. They use “matched pairs” of prices. That is, they measure the price change for the same house over time. However, their data is triple-delayed. First, they report on a two-month lag (the July report was for May, not June like the sales reports). Second, they only use recorded deed information, which means the sale must be closed and the sale must be recorded and publicly available. Third, even though they report monthly, each month’s report is a 3-month moving average. For example, the May report includes March, April and May price information. Clearly, then, the May data is still being affected by the incentive program. Their methodology also means we will never get a clear demarcation of price effect.
So… There is no reason to believe demand/supply factors in the housing market have stopped working. In time, we will have a better, albeit fuzzy, picture of the incentive program’s overall effectiveness and the subsequent activity.
The three driving considerations (location, location and location) remain in effect for real estate. Even the best US data is a mixture of regional, city and local issues and trends. The S&P Case-Shiller data for each of its 20 cities provides a good view of this fact:
The S&P/Case-Shiller data above is seasonally-adjusted. There has been some discussion about whether that is appropriate. I addressed the issue in “The Wall Street Journal Gets House Price Trend Wrong – Interprets Case-Shiller Data Incorrectly” (April 29). Here is the latest pricing data showing that seasonally-adjusting still accomplishes the purpose of extracting seasonal volatility seen every year.
Disclosure: No positions