For an Accurate Picture of the Market, Case-Shiller’s Non-Seasonally Adjusted Monthly Numbers Won’t Cut It
Ok, I’ve officially had it. First, the Case-Shiller index becomes increasingly seasonal and volatile because of foreclosure re-sales. Second, because seasonality is not constant and increases in magnitude because of the foreclosure crisis, Case-Shiller now instructs us all to stop looking at the seasonally adjusted numbers and focus instead on annualized changes in the non-seasonally adjusted index.
But, people being what we are and craving more immediate feedback on our environment, many are now routinely citing monthly change in the non-seasonally adjusted numbers. The result of this has created the perception that home prices increased in April relative to March of this year. While I hate to be the harbinger of bad news here, I can assure you that home prices did not increase in the Composite-20 markets between March and April.
My thoughts about our current situation are as follows:
1. Take a look at Figure 1 which shows the monthly change in the non-seasonally adjusted Case-Shiller Composite-20 home price index. I’ve marked the points for the past four April numbers. Even before the housing recession, there was a healthy dose of seasonality in the Case-Shiller index. But the seasonality really took off during the recession when the index was influenced not just by seasonal price trends (homes selling for more in the spring and for less in the winter) but by the proportion of monthly sales that are foreclosure re-sales (where the proportion is lower in the spring which, again, boosts prices and higher in the winter which lowers prices).
2. If you want to talk about the Case-Shiller index, please talk about year-over-year numbers. If you must talk about monthly numbers and still insist on using Case-Shiller, at least talk about the seasonally adjusted monthly numbers. I know S&P has cautioned about doing this because the seasonal adjustment no longer removes all the seasonality, but they’ve really left you few good choices (other than confining yourself to only talking about annual numbers). They didn’t give guidance with the expectation that people would spurn the seasonally adjusted monthly numbers in favor of the non-seasonally adjusted monthly numbers.
3. For April, home prices on this seasonally adjusted index are down 0.1% between March and April for the Composite-20 markets. When using these numbers, it’s probably a really good idea to at least mention that this seasonal adjustment still doesn’t eliminate all the seasonality. Specifically, it removes the seasonality associated with homes selling for higher prices in spring than in winter but does not remove the seasonality associated with there being fewer foreclosure re-sales proportionally in the spring versus winter.
4. Figure 2 below provides the comparison of the monthly change in the Zillow Home Value Index (Composite 20 markets) and the seasonally adjusted Composite-20 Case-Shiller index. As noted exhaustively in the past, Case-Shiller is more negative than the ZHVI from mid-2007 through 2009 and is now more positive. As also noted, Zillow believes that the current more positive view of Case-Shiller is directly attributable to a large volume of foreclosure re-sales that initially sold at substantial discounts now being sold again for prices closer to market value. Zillow distinctly views the appreciation of such a sale not as organic market appreciation but rather attributable to the well-known discount associated with foreclosure re-sales. If you take a distressed sale bought from a bank at a 30% discount that sells again three years later for 20% more than the REO sale (i.e., no distressed discount but the market has fallen 10% over those three years) as evidence that typical home prices are increasing, then Case-Shiller is the index for you. If you don’t think this sale tells you anything about what typical non-foreclosure prices are doing in your area, then you might want to take a closer look at the Zillow Home Value Index.
