Tomorrow we will release retail sales data for November amid a steady drumbeat of news stories about holiday shopping. These stories beg the question of just how important sales are during the holiday season -- for overall consumer spending as well as for specialty stores. The short answer is that total retail sales (excluding motor vehicle sales) tend to creep up about 3 percent every year between September and October (bearing in mind that September tends to be in a bit of a post-summer lull). The upward creep continues at a bit over 4 percent in November, and then sales spike 17 percent in December.1 Putting these results in a different light, in 2010 we saw that the average household spent $497 more in December than in November, and $121 higher than in November than October. You can sign up here to get an email [1] with the most up-to-date 2011 data as soon as it is available. Not surprisingly, this surge in spending during the holiday season doesn’t affect all types of stores equally. Below we’ll explore the importance of holiday sales, how we arrive at the numbers and for which types of stores the holiday shopping season is more important.
One way to answer the questions of the importance of holiday sales is to look at the seasonal adjustment factors used by the Census Bureau’s Advance Monthly Retail Sales Report [2]. Bear with me here. Seasonal adjustment factors, estimated using sophisticated statistical techniques, measure the extent to which sales “usually” change month-to-month by looking at the past. When the Census Bureau reports, for example, the headline number that retail sales increased 0.5 percent in October from September, that estimate takes into account what usually happens in October. So if sales in October usually increase 2.0 percent, and if sales actually increase 2.5 percent, then Census reports a 0.5 percent gain on a seasonally adjusted basis (most other economic indicators that the government releases are seasonally adjusted, including GDP, employment, personal income, imports and exports of goods and services). Removing seasonal effects makes it easier to assess the underlying trend in the data.
The figure below shows total retail sales excluding motor vehicles on a seasonally adjusted basis (red line) and on a non-seasonally adjusted basis (black line) – notice how much smoother that seasonally adjusted series is relative to the non-seasonally adjusted series. It’s important to note that seasonal adjustment giveth and taketh away equally, so although the gain in sales in December are lowered, sales in January are raised (note that the black line spikes by about 17% in December compared with the seasonally adjusted data for the same period). Over the course of a year, the seasonal adjustments downward and upward tend to cancel each other out (or come very close). These seasonal adjustment factors imply that total retail sales are $497 higher per household in December than in November and $121 higher in November than October. The average household spent $3,287 in December; $2,790 in November, and $2,669 in October.
So, looking at the seasonal adjustment factors tells us what usually happens to sales on a month-to-month basis. What retail sectors typically have large spikes in sales at the end of the year? Two of the most pronounced are jewelry stores and electronic shopping and mail order houses. Jewelry stores make 36 percent of their yearly sales from October through December whereas electronic shopping and mail order houses make 30 percent of their sales for the year during the last quarter. The figures below show the seasonally unadjusted and adjusted figures for these two types of stores. Other shops that have large industries in retail sales where seasonal effects are large include , clothing, electronic and appliance, sporting, hobby, book and music; and department stores.
One important thing to keep in mind about these numbers is that total consumer spending includes more than just retail sales, and retail sales (which approximate consumer spending on goods) comprise about 30 percent of total personal consumption, according to the Bureau of Economic Analysis (consumption is similar to spending, but we’ll leave that discussion for another day). The other major, non-retail sales based categories of personal consumption, such as rent, health care and education are less likely to be strongly influenced by holiday season spending sprees.
~Mark Doms, Chief Economist, U.S. Department of Commerce
December 9, 2011
1. These results are from 2010. Each month Census revises certain seasonal adjustment factors. Also, Census does an annual seasonal adjustment review at the end of April that extends back to January 1996.
Are you on Twitter? We are! Follow us at EconChiefGov [3]!
