The December 2010 Advance Monthly Sales for Retail Trade and Food Services Report for November was released this morning. Here is the opening paragraph of the release:
The U.S. Census Bureau announced today that advance estimates of U.S. retail and food services sales for November, adjusted for seasonal variation and holiday and trading-day differences, but not for price changes, were $378.7 billion, an increase of 0.8 percent (±0.5%) from the previous month, and 7.7 percent (±0.7%) above November 2009. Total sales for the September through November 2010 period were up 7.8 percent (±0.5%) from the same period a year ago. The September to October 2010 percent change was revised from +1.2 percent (±0.5%) to +1.7 percent (±0.2%).
The chart below shows the complete data series from 1992, when the US Census Bureau began tracking the data. I’ve highlighted recessions and the approximate range of two major economic episodes that have impacted consumer attitudes. The Tech Crash (NASDAQ:QQQQ) that began in the spring of 2000 had little impact on consumption. The Financial Crisis of 2008 has had a major impact. In fact, the latest sales number is 0.3% below the all-time high of November 2007, the month before the Great Recession began.
Here is the same chart with two trendlines added. These are linear regressions computed with the Excel Growth function.
The green trendline is a regression through the entire data series. The latest sales figure is 9.2% below the green line end point.
The blue line is a regression through the end of 2007 and extrapolated to the present. Thus, the blue line excludes the impact of the Financial Crisis. The latest sales figure is 16.0% below the blue line end point.
We normally evaluate monthly data on a month-over-month or year-over-year basis. The November 0.8% increase over October and 7.7% over November 2009 look quite encouraging. But a snapshot of the larger historical context illustrates the devastating impact of the Financial Crisis on the US economy.
Here is a version of the chart with a logarithmic vertical axis and exponential regression trendlines. The trendlines bisect the data as in the linear version, but the linear regressions on a log y-axis are straight lines rather than curved.
Doug Short Ph.d is the author of dshort.com.