Navin, Jordan, authorWeiler, Stephan, advisorAnderson, Aaron, committee memberPena, Anita, committee memberKroll, Stephan, committee memberMushinski, David, committee member2019-09-102019-09-102019https://hdl.handle.net/10217/197263Wildlife-aircraft conflict poses a substantial economic and safety threat in the United States (US). Dolbeer, Wright, Weller, Anderson, and Begier (2014) estimates direct costs related to wildlife strikes burdened the US economy by approximately $157 million annually between 1990 and 2014. In 1995, the Federal Aviation Administration (FAA) collaborated on a project with the United States Department of Agriculture's (USDA) Wildlife Services to investigate the magnitude and nature of the wildlife strike problem, ultimately resulting in the creation of the National Wildlife Strike Database (NWSD). However, reporting strikes (and associated information, such as repair costs) to the NWSD is not mandatory, and information used to calculate economic damage estimates from wildlife strikes in the US relies on voluntarily reported cost data. This dissertation focuses on the direct costs of wildlife strikes in the US and the associated disclosure behaviors of large domestic American airlines. Chapter 1 investigates the relationship between the likelihood of voluntary repair cost disclosure after a wildlife-strike event by such airlines and market competitiveness and idiosyncratic firm profits. Results show changes in competitiveness and profitability impact the voluntary disclosure of wildlife-strike repair costs by major US airlines to the NWSD. Chapter 2 similarly examines airline voluntary disclosure accuracy, employing emerging methods from economics and accounting literature that test the accuracy of self-reported data based on a statistical property exhibited by large datasets, known as Benford's Law (de Marchi & Hamilton, 2006; Dumas & Devine, 2000; Nigrini, 1996; Zahran, Iverson, Weiler, & Underwood, 2014). Analogous to Chapter 1, findings indicate the accuracy of repair costs American air carriers report to the NWSD is linked to market competition and profits. Chapter 3 relates to developing a method for interpolating missing repair costs in the NWSD using machine learning techniques. Results show that a neural network outperforms both linear regression and random forest models when predicting out-of-sample data, and furthermore, interpolating missing costs in the NWSD with a neural network delivers an average annual estimate of the direct costs of wildlife strikes in the US that is approximately $75 million, significantly less than prior estimates. Specifically, the neural network approach yields estimates $19 and $82 million lower, respectively, than when using mean cost assignment and Dolbeer et al. (2014)'s reported estimate derived using a variation of the same method.born digitaldoctoral dissertationsengCopyright and other restrictions may apply. User is responsible for compliance with all applicable laws. For information about copyright law, please see https://libguides.colostate.edu/copyright.Text