|Friday, February 24|
|PS2 Poster Session 2 and Refreshments||
Fri, Feb 24, 5:15 PM - 6:30 PM
Conference Center AB
Enhancing Monthly Retail Holiday Effect Methodology Through Daily Data (303450)*Rebecca Jean Hutchinson, U.S. Census Bureau
Keywords: Big Data, Seasonal Adjustment, Retail Trade, Economic Indicators, Holiday Adjustment, Credit Card Data
The U.S. Census Bureau is researching the potential of leveraging Big Data to enhance methodologies used in its economic indicator programs. One example of this is using daily credit/debit/prepaid gift card transaction data to identify effects of holidays on our Monthly Retail Trade Survey. A holiday may influence economic activity not only on the day that it falls but also before and/or after it. Moving holidays--like Easter--can influence monthly activity differently depending on when they fall. Current holiday effect methodology relies on subject matter expertise and seasonal model fit comparisons. This methodology has been improved using daily data from FirstData that capture 58 billion transactions annually. We now have the ability to find evidence for/against a larger variety of holidays or pseudo-holiday events including Super Bowl Sunday, Easter Sunday, Ramadan, Chinese New Year, and Cyber Monday as well as more granularity of sales movement around holidays. As a result of this work, we have implemented an Easter Sunday effect in our seasonal adjustment methodology. This poster describes holiday research completed to date and other applications utilizing daily data.