Retail Sales Ground Truthing

By Peter Perrault

Intern, Index of Economic Activity for Humboldt County

August 2002

A diverse sample of 10 Humboldt County retailers is surveyed each month in order to create the retail sales sector of Index of Economic Activity for Humboldt County. The Index provides the only monthly source of economic information on retail sales in Humboldt County. In contrast, the California Board of Equalization tracks all taxable retail sales in Humboldt County, and thus provides much more comprehensive retail sales data. Unfortunately the retail sales data from the Board of Equalization is more than a year old by the time it is finally released. Thus there is a tradeoff between timely and current monthly retail sales information from the Index, reflecting performance by a relatively small sample of retailers, and old but comprehensive data from the Board of Equalization. The quality of the retail sales index depends upon how accurately it tracks trends in overall retail sales in Humboldt County. Consequently this report addresses the question of how accurately the retail sales index from the Index of Economic Activity reflects trends in overall retail sales.

To answer this question, I have conducted statistical and graphical comparisons of real (inflation adjusted) retail sales in Humboldt County (based on figures provided by the California State Board of Equalization) and the (non-seasonally adjusted) values from our retail sales index. As will be shown below, the results of this study reveal a strong correlation between trends in the retail sales index from the Index of Economic Activity and trends in total taxable retail sales in Humboldt County. Before discussing these results, however, I will briefly explain the analytical methods used in this study.

First, the quarterly retail sales figures provided by the California Board of Equalization were first adjusted for inflation in order to make them comparable to the retail sales index from the Index of Economic Activity, which is derived from inflation-adjusted retail sales from our sample of reporting retailers. We adjust both data series using the Consumer Price Index (CPI) for All Urban Consumers, put out by the U.S. Department of Labor, Bureau of Labor Statistics. The adjustment put the nominal sales figures into real dollar amounts. The CPI uses 1982-1984 as its base year. Second, I converted the monthly retail sales index from the Index of Economic Activity to quarterly values in order to make them comparable to the quarterly taxable retail sales data reported by the Board of Equalization. I then computed a Pearson’s correlation coefficient between the retail sales index values and the real County retail sales figures.

The correlation coefficient for the two data series is 0.704. This means that 70.4% of the trend in County retail sales can be explained by the retail sales index of the Index of Economic Activity. Testing for significance (using a t-test) I found that with over 99% confidence we can reject the null hypothesis that there is no statistical significance between the two sets of numbers. Overall a 70 percent correlation can be considered to be a strong correlation, indicating that the retail sales index performs strongly as an indicator of overall retail sales trends in Humboldt County.

The accompanying diagram is used to compare the trend in the retail sales index of the Index of Economic Activity with the trend in total (inflation-adjusted) Humboldt County retail sales. Within the four and a half years for which the data were available, one can see that the retail sales index performs best as an indicator for the first two quarters of any given year. County retail sales increase sharply in QI (Jan-Mar) and QII (Apr-Jun), and the retail sales index indicates this trend with remarkable accuracy. Interestingly, the retail sales index tends to reach a peak in QII and decline thereafter, while total County retail sales peak in QIII (Jul-Sep) and decline thereafter. We do not have an explanation for why our sample of retailers experience peak sales activity in QII rather than QIII of each year. Consequently the retail sales index serves as a leading economic indicator of the trend in overall County retail sales in the second half of each year.

Tourism and travel make an important contribution to retail sales in Humboldt County, and tourism is a very seasonal activity in our area due to the rainy season. One can clearly see the seasonal nature of retail sales in the accompanying diagram.