He Loves Me, He Loves Me Not A Forecast of U.S. Jewelry Sales Alex Gates Ling-Ching Hsu Shih-Hao Lee Hui Liang Mateusz Tracz Grant Volk June 1, 2010
Purpose Examine trends in jewelry sales in the United States Create a model capable of forecasting future U.S. jewelry sales Forecast U.S. jewelry sales for one year
Data Data obtained from U.S. Census Bureau website ( Data is monthly U.S. Jewelry Sales from January, 1992 through March, 2010
Jewelry Sales Data (JSALES)
JSALES Correlogram
Pre-Whitened JSALES
SDLNJSALES
SDLNJSALES – Unit Root Test
SDLNJSALES AR(1)
SDLNJSALES ARMA(1,1)
SDLNJSALES – Final ARIMA Model: AR(1) MA(1) MA(12) MA(14) MA(16)
SDLNJSALES - Correlogram
SDLNJSALES Serial Correlation Test
SDLNJSALES Correlogram of Residuals Squared and ARCH Test
SDLNJSALES - ARCH/GARCH
Actual, Fitted, Residual
Histogram of Residuals Non-normal, but single-peaked
Correlograms Standardized ResidualsResiduals Squared
ARCH Test for Heteroskedasticity
Conclusion: Accept the model!
Forecasting SDLNJSALES for Previous 12 Months
Forecasting SDLNJSALES for Next 12 Months
Recolored Forecast of JSALES
Detailed Graph
JSALES Data and Forecast Actual values of U.S. Jewelry Sales from January, 2009 through March, 2010 Forecasted values from April, 2010 through March, 2011
Final Conclusions Forecasted recovery of Jewelry Sales for 2010 holiday season 10.35% increase predicted for December 2010 sales over 2009 levels Upper bound of 95% confidence level predicts that, if everything goes very well, sales could be their highest since 1992
Final Conclusions