۲ Corporate Reputation (CR) is undoubtedly an intangible asset 3 which provides a competitive advantage for firms (Rose & Thomsen, 2004; Hall, 1992). However, controversy arises when the discussion Q2 4 5 turns into how financial markets value that reputation. Some stud- 6 ies conclude that favorable reputations contribute to increase the 7 market value of firms (Black, Carnes, & Richardson, 2000; Stuebs & 8 Sun, 2011; Wang & Smith, 2008), while others reject this assertion 9 (Brammer, Brooks, & Pavelin, 2004, 2009). This contradictory set of 10 results motivates the search for new methodological perspectives, 11 different from those traditionally used (as multiple regressions, 12 MR), with the purpose of shedding some light on the controversy. 13 Our study uses Generalized Regression Neural Networks (GRNN) to 14 measure the relationship between CR and the firms’ market value. 15 MR has an important role in identifying signs and meanings of 16 variables, but the impact analysis of variables using GRNN takes into 17 account non-linearity, adding significant results to our research by 18 comparing both techniques. Since the two approaches are mutually 19 informative, our research is intended to shed light on the importance of CR to explain the market value of firms, providing both conceptual 20 and practical contributions. To the best of our knowledge, GRNN have 21 not been used to investigate the effects of CR in the value of compa- 22 nies, modeling procedures using neural networks are expected to be 23 more robust than the traditional MR, adjusted for potential nonlin- 24 earities between the variables under study (Pao, 2008).

The structure of the paper is organized as follows. After the intro- 26 duction, relevant literature on the topic and research hypotheses are 27 developed in Section 2. Section 3 presents research models and meth- 28 ods. Section 4 is dedicated to the data used and the selected sample, 29 and Section 5 the results obtained in the investigation. Finally, main 30 conclusions and future research suggestions are shown.