Overview

The Center for Economic and Management Research (CEMR) Price College of Business, The University of Oklahoma produces the Oklahoma Monthly Retail Sales Series containing monthly estimates of retail sales for Oklahoma, the Oklahoma City, Tulsa and Lawton Metropolitan Statistical Areas and 48 selected cities in Oklahoma.

Estimates published for the State and metro areas are available by kind of business categories. Estimates published for selected cites are total retail sales only. The series runs from March 1988 to current for the state, Lawton Metropolitan Statistical Area and all selected cities. The series runs from March 1999 to current for the Oklahoma City and Tulsa Metropolitan Statistical Areas .

Definitions

Kind of Business Classification Establishments are assigned industry classifications according to their primary types of economic activity.  Those establishments engaged primarily in selling merchandise for personal or household consumption are designated as part of the Major Industry Group "Retail Trade" and are assigned an industry classification code ranging between 52 and 59 according to definitions provided in the 1987 Standard Industrial Classification manual. 

Aggregations for the Oklahoma Monthly Retail Trade Series are made as follows:  

Data Sources

The series is based on sales tax collection data provided by the Business Tax Division, Oklahoma Tax Commission (OTC). The data provided includes tax receipts (monthly sales tax collections) by establishment, type of business, and geographic location.

Data Limitations

Several issues should be considered when using the information provided in this series.

1)  The information provided by OTC is that it includes not only current month tax collections (resulting from actual sales in that month) but also collections from any late reports or penalty payments from previous months. This has the unwanted effect of introducing month-to-month variations in tax collections not associated with actual sales from the business month. The method used to minimize this variation is discussed below. 

2)  Automobile sales, which are not subject to sales tax in Oklahoma, are not included.

3)  Caution should be exercised in interpreting changes in the series for any particular area.  A change in the series simply means that the level of retail sales has changed. It does not necessarily imply changes in other economic indicators. For example, sales may decline even though individuals in a city may be earning as much as they were previously. Individuals may simply  be growing more cautious in their expenditures. However, this is the kind of situation that bears watching because, in general, sales tend to fluctuate less than personal income.

4)  It should also be noted that changes in the series can reflect changes in regional spending patterns rather than a general deterioration in economic conditions. For example, a new highway which makes it easier to shop in a larger city may reduce local purchases even though income in the region remains the same.

Despite these qualifiers, the information presented in this report should prove useful for persons interested in economic performance. The data are the most timely available, and, when used in conjunction with other economic indicators, provide an informative picture of local economic conditions.

Methodology

All sales tax collections from non-retail establishments (primarily wholesale taxes) are eliminated from the data set provided by OTC.  Sales tax collections are next sorted according to their business classifications using Standard Industrial Classification (SIC) Codes.

Sales subject to sales tax are then estimated by dividing collections by the appropriate area's sales tax rate. For example, suppose a city raised $10,000 in sales tax collections in a particular month. If the city's sales tax rate is 2.0 percent, then the dollar value of the sales subject to city sales tax would have equaled $500,000 during that month. 

Seasonal Adjustment

Since month-to-month changes in retail sales are not meaningful unless the normal seasonal variation is removed, our preliminary estimates are next “seasonally adjusted." This permits examination of the "real" month-to-month changes unaffected by normal seasonal trends.

An example of normal seasonal variation is the increase in sales associated with the Christmas holiday. Assume the average rate of increase in General Merchandise sales between November and December was 20 percent over the last five years. Next assume that the change in General Merchandise between November 1994 and December 1994 was also 20 percent. The seasonally adjusted data for December 1994 would show no increase for December 1994 because the seasonal trend--in this case the Christmas holiday season--has been removed.

CEMR uses a statistical process commonly referred to as the "X-11" program to seasonally adjusted retail sales estimates. The program was developed at the U.S. Bureau of the Census. 

Central Averaging Following this procedure, an effort is made to remove variation in the data produced by tax collections in a given month not resulting from actual sales in that month, i.e., late payments, penalty payments, etc. This is accomplished using a statistical technique known called Centered Averaging. 

Past series of retail estimates were "benchmarked" to the Census of Retail Trade. In other words, the Census was used as a base year and percent changes in sales tax collections were used to move that base figure forward in time. It was decided to abandon this method after close investigation revealed that the method used by the Census to assign business classifications was inconsistent with the method employed by the Oklahoma Tax Commission.

Once these steps are completed, the categories are next aggregated either a Total Durable or Total Non-durable category.

Gasoline Sales Estimates Sales of gasoline and motor fuels are not subject to sales tax but rather an excise tax. Monthly gasoline sales are then estimated by dividing the excise tax collection data by the excise tax rate.

Revisions

New seasonal factors are computed annually for the entire series to adjust  for seasonal variations as well as trading day and holiday differences. The seasonal adjustment program used to generate the seasonal factors  is X-11.

Contacts

For additional information on the Oklahoma Monthly Retail Trade Series, contact:

John McCraw
Assistant Director
Center for Economic and Management Research
Michael F. Price College of Business
The University of Oklahoma
307 W. Brooks
Norman, Oklahoma 73019
405-325-5864
mailto:jmccraw@ou.edu

Back to top