The present investigation aims at modeling the sales response to advertising and, in the process, sheds some light on the sales-advertising relationship subject, which has been at the center of a decades-long controversy due to its inherent complexities. We studied three Colorado Lottery games, Lotto, Powerball, and Scratch, over a four-year period of operation. To synthesize a model that appropriately described the sales-advertising behavior of each one of these games, we addressed three fundamental questions driving the modeling process itself: 1. Is there a relationship between sales and advertising? 2. If such relationship exists, is there an advertising "carryover effect" on sales? And, 3. What is the shape of the sales-advertising relationship? We put forward two general-response models (Current Effects and Koyck's) in combination with eight functional forms (one linear and seven nonlinear forms) to address the above questions and test the respective hypotheses. Employing the available time series data corresponding to game sales, game advertising expenditures, state population, state unemployment rate, and jackpot (for the relevant games), we performed the respective regression analyses. We, then, evaluated the posited relationships and selected the best predictive model for each game, when statistical evidence supported a significant sales-advertising association. Using this final model, we addressed the three research questions at the core of this study. The results of this investigation suggested the existence of a significant positive and nonlinear (concave-downwards) Scratch sales-advertising relationship. No sales-advertising association was found for the Lotto or Powerball games. The data analyzed did not seem to support either the advertising "carryover effect" on sales on any of the games studied. From the theoretical point of view, these findings extend prior empirical research that has generally assumed, for simplification purposes, a linear sales-advertising relationship with its corresponding consequences. From the practical perspective, this study highlights advertising’s contribution to sales, which can help debunk mistaken beliefs frequently stigmatizing advertising as a resource-spending function and quell the long-established skepticism about its financial accountability. / text

I examine whether sales news and cash flow news are more informative in valuation when they are consistent with earnings news. Prior studies show that earnings response coefficient (ERC) is larger when the earnings surprise is consistent in sign with the sales surprise or the cash flow surprise because the consistency suggests higher earnings persistence. I provide new evidence that indicator consistency increases sales response coefficient and cash flow response coefficient as well as ERC. This consistency effect for sales and cash flow cannot be explained by the standard persistence argument from prior studies. I propose a new argument that can explain the consistency effects for all three indicators. I posit that investors perceive consistent indicators to each be more precise and thus rely more on each indicator. Under this precision argument, I predict and show that indicator consistency is particularly useful when there is high uncertainty about indicator precision. / Business Administration/Accounting