Unpacking the Inverted U-shape between Regional AI and Business Performance

AbstractThis study examines how regional artificial intelligence (AI) influences firms’ business performance from the viewpoint of economic geography. We employ the instrumental variable method to analyze 3633 American listed companies. We find the “regional AI and business performance” relationship appears in an inverted U-shape. By applying the plausible instrumental variable method, our robustness check suggests that our findings are reliable. Theoretically, our paper enriches current regional AI studies with firm-level evidence; practically, our paper sheds light on how to make firm location decisions in the AI era.Keywords: Artificial intelligencefirm performancereturn on equityAmericaJEL CLASSIFICATION CODES: R11L10 Disclosure statementNo potential conflict of interest was reported by the author(s).Notes1 According to preliminary math using the equation: y = β0+β1X + β2X2, when β2 is significant, then the (inverted) U-shape exists. β0 and β1 can change the position of the (inverted) U-shape in the coordinate axis, but it is β2 that determines the (inverted) U-shape. Therefore, “β1 is insignificant but β2 is significant” never means that the (inverted) U-shape does NOT exist or means that it is an increasing (or decreasing) return.2 We would like to thank our two anonymous reviewers for bringing this to our attention.Additional informationFundingThis paper is financed by the Humanity and Social Science Foundation (grant number: 22YJCZH117), the National Social Science Funds of China (grant number: 22CJL013) and the Innovative Team Project of Guangdong Universities (grant number: 2019WCXTD008).