2009 Headlines

Retail Sales Linked To Financial Market Returns

New study by Vishal Gaur provides an improved model for equity analysts when forecasting annual retail sales

January 14, 2009 | Ithaca, NY | Researchers have developed an accurate way to forecast annual retail sales using a model which incorporates data from financial market returns. Authors Vishal Gaur,

Vishal Gaur

associate professor of operations management at the Johnson School at Cornell University, Nikolay Osadchiy of the Stern School of Business at New York University, and Sridhar Seshadri of the McCombs School of Business at the University of Texas, also examined the usefulness of financial market returns for operational decision-making and risk management.

The new study focuses on firm-level, dollar-denominated annual sales realized by retailers and correlates this data with trade-oriented financial market indicators. After examining the sales forecast error via the forecast terms used by equity analysts, financial market returns during the forecast terms, and various volatility parameters, the researchers establish the association between sales forecast error and market return for various retail segments. The correlation is expressed as a numerical coefficient which averaged 0.17 across the numerous retail segments and peaked as high as 0.83 for office supplies stores to insignificant or negative for discount and bargain stores such as Wal-Mart and Family Dollar.

"There is ample evidence that end-consumer demand is correlated with broad financial market indicators," Gaur says. "Our research shows that forecast updates using financial market returns improve forecast accuracy when combined with the forecasts by equity analysts."

The results are also classified according to discretionary and everyday purchases. Segments in the discretionary group such as jewelry, home improvement, consumer electronics and apparel were shown to have higher correlation coefficients on average than the everyday purchase group which consists of discounters and wholesale clubs. Moreover, the researchers discovered that the correlation coefficient is higher for high-margin retailers and increases with the terms of the sales forecast.

"Beyond forecast updating, our model can be applied for risk management, calculating the value of decision postponement and planning other operational decisions," Gaur says. "Recent research has examined financial hedging based on the correlation between demand and the price of traded assets. The correlation information from our model can be used to inform such decisions and shows that its value varies depending on the retail segment, gross margin of a firm and time horizon."