by Calculated Risk on 3/26/2012 10:33:00 PM

All data is lagged, but some data is lagged more than others.

In times of economic stress, I tend to watch the high frequency data closely: initial weekly unemployment claims, monthly manufacturing surveys, and consumer sentiment. The “high frequency” data is lagged, but the lag is usually just a week or two.

Most of the time I focus on the monthly employment report, GDP, housing starts, new home sales and retail sales. The lag for most of this data is several weeks. As an example, the BLS reference period contains the 12th of the month, so the report is lagged a few weeks by the time it is released. The housing starts and new home sales data released last week were for February, so the lag is also a few weeks after the end of the month. The advance estimate of quarterly GDP is released several weeks after the end of the quarter.

But sometimes the lag can be much longer. Tomorrow morning the January Case-Shiller house price index will be released. This is actually a three month average for house sales recorded in November, December and January. (Update: April 24: S&P obtains the data when recorded, but uses closing dates, not recording dates for the price index).

But remember that the purchase agreement for a house that closed in November was probably signed in September or early October. So some portion of the Case-Shiller index will be for contract prices 6 or even 7 months ago!

Other house price indexes do a little better. CoreLogic uses a weighted 3 month average with the most recent month weighted the most – and they will release their February index next week, almost a month ahead of Case-Shiller. The LPS house price index is for just one month (not an average) and uses only closings (not recordings like other indexes that can add an additional lag).

But the key point is that the Case-Shiller index will not catch the inflection point for house prices until well after the event happens. Just something to remember ...