资源描述
09:15 28/3/2019 RFS-OP-REVF180132.tex Page: 2062 20622106 How Valuable Is FinTech Innovation? Mark A. Chen J. Mack Robinson College of Business, Georgia State University Qinxi Wu Hankamer School of Business, Baylor University Baozhong Yang J. Mack Robinson College of Business, Georgia State University We provide large-scale evidence on the occurrence and value of FinTech innovation. Using data on patent filings from 2003 to 2017, we apply machine learning to identify and classify innovations by their underlying technologies. We find that most FinTech innovations yield substantial value to innovators, with blockchain being particularly valuable. For the overall financialsector,internetofthings(IoT),robo-advising,andblockchainarethemostvaluable innovation types. Innovations affect financial industries more negatively when they involve disruptive technologies from nonfinancial startups, but market leaders that invest heavily in their own innovation can avoid much of the negative value effect. (JEL G14, G20, G29, G39) Received May 31, 2017; editorial decision September 30, 2018 by Editor Andrew Karolyi. In recent years, the rapid advance of FinTech, or financial technology, has attracted considerable attention within the finance industry. Many observers havewelcomedtheriseofFinTech,claimingthatnewlyemergingtechnologies have the potential to radically transform financial services by making transactions less expensive, more convenient, and more secure. 1 Worldwide, external funding for FinTech development has been rising quickly. During the first half of 2018, global investment in FinTech companies totaled $57 billion, a striking increase from $38.1 billion for all of 2017 (KPMG 2018b). We are grateful to Itay Goldstein, Wei Jiang, and Andrew Karolyi (the Sponsoring Editors for the RFS FinTech Initiative); Paul Tetlock (the RFS FinTech Workshop discussant); and two anonymous referees for valuable guidanceandsuggestions.WealsoappreciatehelpfulfeedbackfromVikasAgarwal,JerryHoberg,JamesHodson, Andreas Hoepner, Lixin Huang, Dalida Kadyrzhanova, Omesh Kini, Laura Xiaolei Liu, and Xiaoyan Zhang and participants at the 2017 RFS FinTech Workshop, the 2018 RFS FinTech Conference, and the 2018 Western Economic Association International Conference, and seminar participants at Georgia State University, Peking University, and Tsinghua University. Yuan Gao, Justin Lee, Han Ma, An Nguyen, Honglin Ren, and Liang Zhang provided capable research assistance. Send correspondence to Mark A. Chen, J. Mack Robinson College of Business, Georgia State University, 35 Broad St. NW, Atlanta, GA 30303; telephone: 404-413-7339. E-mail: machengsu.edu. 1 See, for example, The Economist (2015), McKinsey Kogan et al. 2017; Mann 2018). 4 In Section 1, we define these seven technology categories and provide illustrative examples of each. 2063 Downloaded from by Renmin University user on 28 November 2019 09:15 28/3/2019 RFS-OP-REVF180132.tex Page: 2064 20622106 The Review of Financial Studies / v 32 n 5 2019 mobiletransactionshaveexperiencedthemosttotalinnovationoverthesample period. Blockchain is currently the smallest but fastest-growing category of FinTech innovation. To explore the implications of FinTech innovation, we develop a new methodology for estimating the value of a patent filing to one or more publicly traded firms. Our valuation approach is based on observed stock market responses to USPTO disclosures of patent filings. Importantly, our approach accounts for the markets anticipation of different types of patent filings made by different filers. 5 We start by estimating innovation arrival intensities via Poisson regression models that account for factors such as technology type, time effects, a patent filers prior innovation experience, and filer fixed effects. Then, for each patent filing, we combine its predicted count intensity with a firms stock price movements to infer the underlying value of the innovation to the firm. Using our valuation approach, we examine how much firms in the financial services sector stand to gain from their own FinTech innovations. The calculations show that a FinTech innovations private value (i.e., the value accruing to the innovator) is typically large and positive. For instance, in 2017 dollars, the median private value of a FinTech innovation is about $46.7 million, which is much higher than the median private value of $3.1 million for other financial innovations. Overall, the FinTech innovation types that are most valuable to innovators are blockchain, cybersecurity, and robo-advising. When wecontrolinmultivariateregressionsfortime-varyingfirmcharacteristics,firm fixed effects, and patent quality measures, we find that blockchain and robo- advising emerge as the most valuable types, thus underscoring the economic importance of these new segments of the FinTech space. WeextendourvaluationmethodtostudyhowFinTechinnovationsaffectthe financial services sector and its key component industries: banking, payment processing, brokerage, asset management, and insurance. 6 Our calculations show that, for the financial sector as a whole, the typical FinTech innovation brings positive value. The most valuable innovation types are IoT, robo- advising, and blockchain, with median value impacts of $24.5 billion, $15.5 billion, and $8.1 billion, respectively (2017 dollars). Nonetheless, we find substantial variation in value across different technology-industry pairings. Among mobile transaction innovations, for example, the median value impact is negative for the banking industry but positive for the payments industry. What explains the wide cross-sectional variation in the value effects of FinTech innovation? We argue that the value effects of an innovation are driven by two key factors: (1) how inherently disruptive is the underlying technology; 5 Throughout, we use the term “filer” to refer to the original assignee of a patent application filing. 6 Because our method requires the use of stock-price data, any inferences or conclusions we draw about an innovations value impact to an industry must be tempered by the fact that we cannot directly measure value effects for privately held firms within the industry. 2064 Downloaded from by Renmin University user on 28 November 2019 09:15 28/3/2019 RFS-OP-REVF180132.tex Page: 2065 20622106 How Valuable Is FinTech Innovation? and (2) whether the innovator poses a competitive entry threat to the industry. To study this issue, we use technological spillovers emanating from individual, nonfirm inventors to construct a data-based measure of “disruptiveness” across technology-industry pairs. Consistent with theories of disruptive innovation (e.g., Christensen 1997; Christensen and Raynor 2003; Downes and Nunes 2013), we find that a FinTech innovation tends to destroy significantly more industry value when its underlying technology is disruptive and when it originates from a young, nonfinancial firm (“FinTech startup”). Next, we examine how FinTech innovation affects value from the viewpoint of individual incumbent firms, that is, market-share leaders and their rivals. Theoretical considerations suggest that disruptive innovation by potential entrants can be especially harmful to an industrys market leaders, which are sluggish in adapting to change and focused on existing customers (Tang 1988; Christensen and Raynor 2003). On the other hand, industry-wide disruption mightbeadvantageoustomarketleadersbecause,comparedtorivals,theyhave largerscaleeconomiesandmorefinancialresourceswithwhichtoinnovatenew lines of business (Dasgupta and Stiglitz 1980; Scherer 1980; Blundell, Griffith, and van Reenen 1999; Czarnitzki, Etro, and Kraft 2014). Our empirical tests support the latter prediction and also suggest that market leaders ability to avoid harm from disruptive outside innovation is strongly linked to the amount of resources that they devote to their own research and development (R Hall, Jaffe, andTrajtenberg2005;KortumandLerner2000;Lerner2009;Bravetal.2018). While this literature has delivered valuable insights about corporate patenting and innovation in general, much of the research relies on patent grant data and thus cannot fully capture the FinTech innovation activity that has occurred over just the past few years. 7 By focusing on patent applications and utilizing the BDSS data, we can mitigate the data truncation problems inherent in relying on patent grants and thus provide a more complete picture of very recent trends and patterns in FinTech innovation. Second, our work builds upon a stream of research that uses stock price data to study the value of innovation (see, e.g., Pakes 1985; Austin 1993; Hall, Jaffe, and Trajtenberg 2005; Nicholas 2008; Kogan et al. 2017). The methods that we develop extend this literature by recognizing the count-based nature of innovation events over time, thus permitting more precise estimates of the true value impact of such events. More generally, our approach of combining stock price reactions with predicted Poisson arrival intensities could be useful 7 Empirical studies document that, for observed patent grants, the average time between application and granting is about 2 years (e.g., Hall, Jaffe, and Trajtenberg 2005; Seru 2014; Cornaggia et al. 2015). Some applications remain under review for much longer than 2 years and thus may not be captured by currently available grants data. 2065 Downloaded from by Renmin University user on 28 November 2019 09:15 28/3/2019 RFS-OP-REVF180132.tex Page: 2066 20622106 The Review of Financial Studies / v 32 n 5 2019 for studying other types of recurring, partially anticipated phenomena, such as revisions to analyst estimates, sequences of corporate news releases, or waves of mergers or bankruptcies. Third, our findings contribute to the finance, strategy, and economics literatures that explore the role of innovation in shaping industry competition. Theoretical research has modeled how innovation from outside of an industry can harm or benefit incumbent firms (see, e.g., Lieberman and Montgomery 1988;HendersonandCockburn1996;Christensen1997;Adner2012)andhow incumbents can use their own innovation to protect themselves from outside threats (see Dasgupta and Stiglitz 1980; Gilbert and Newbery 1982; Aghion et al. 2001; Aghion and Griffith 2005). Testing such theories is challenging because of the difficulty of obtaining a large data sample of competitive threats from innovation. Our work employs a new data set and provides systematic evidence of how innovation by potential entrants can affect individual firms within an industry. Finally, our approach to identifying and classifying FinTech patent filings contributes to the literature that applies textual analysis and machine learning to finance and economics. Researchers have used text-based methods to study newsarticles,onlineforumpostings,corporatefilings,andanalystreports(e.g., Antweiler and Frank 2004; Tetlock, Saar-Tsechansky, and Macskassy 2008; Hanley and Hoberg 2010; Loughran and McDonald 2011; Jegadeesh and Wu 2013;HobergandPhillips2016;Bellstam,Bhagat,andCookson2017;Manela and Moreira 2017). Other work studies the application of machine-learning methods to economics (e.g., Kleinberg et al. 2015; Glaeser et al. 2016; Naik, Raskar, and Hidalgo 2016; Athey and Wager 2018; Athey, Tibshirani, and Wager 2018). A number of the machine-learning algorithms that we use for text classification appear to be new to the finance domain and can be applied to study a broad set of questions relating to patent filings, legal documents, media stories, and other textual data. 1. Categories of FinTech What is FinTech? Although FinTech can be broadly defined as any technology that enables or enhances the provision of financial services, such a definition is oflimiteduseforempiricallyidentifyingandclassifyingreal-worldFinTech.To proceedwithouranalysis,wethereforerequireatypologythat(1)distinguishes innovation within the FinTech space from other types of financial or scientific innovation;and(2)articulatesthekeytechnologicaldifferencesamongdifferent instances of FinTech innovation. We begin with the premise that FinTech ultimately consists of the set of recently developed digital computing technologies that have been appliedor that will likely be applied in the futureto financial services. Then, based on a 2066 Downloaded from by Renmin University user on 28 November 2019 09:15 28/3/2019 RFS-OP-REVF180132.tex Page: 2067 20622106 How Valuable Is FinTech Innovation? Table 1 Categories of FinTech Category definition Key technologies Real-world examples Cybersecurity: Hardware or software used to protect financial privacy or safeguard against electronic theft or fraud Encryption, tokenization, authentication, biometrics Diebold iris-scanning ATM, Mastercard Biometric Card, USAA face recognition login, Experian CreditLock Mobile transactions: Technologies that facilitate payments via mobile wireless devices, such as smartphones, tablets, and wearables Smartphone wallets, digital wallets, near-field communication Apple Pay, Android Pay, PayPal Mobile Express Checkout, Venmo, Square Data analytics: Technologies and algorithms that facilitate the analysis of transactions data or consumer financial data Big data, cloud computing, artificial intelligence, machine learning Equifax NeuroDecision credit scoring, JPMorgan Contract Intelligence (COiN), Bloomberg Social Sentiment Analytics Blockchain: Distributed ledger technologies with a primary application to financial services Cryptocurrency, proof-of-work, smart contracts, directed acyclic graphs Bitcoin, Ripple Payment Network, Visa B2B Connect, Nasdaq Linq asset trading platform Peer-to-peer (P2P): Software, systems, or platforms that facilitate consumer-to-consumer financial transactions Crowdfunding, P2P lending, customer-to-customer payments GoFundMe, Kickstarter, Lending Club, Prosper Marketplace, Zelle Robo-advising: Computer systems or programs that provide automated investment advice to customers or portfolio managers Artificial intelligence, big data, machine learning Betterment, E-Trade Core Portfolios, Schwab Intelligent Portfolios, Vanguard Personal Advisor Services Internet of things (IoT): Technologies relating to smart devices that gather data in real time and communicate via the internet Smart devices, near-field communication, wireless sensor networks, actuators UnitedHealthcare Motion F.I.T. tracker, Nationwide SmartRide telematics, Travelers Insurance smart home sensors This table shows a proposed typology of FinTech. The definitions, technologies, and examples listed are based on the authors reading of news articles, industry reports, and surveys. general reading of various articles and reports, 8 we formulate a broad typology of FinTech comprising seven categories: cybersecurity, mobile transactions, data analytics, blockchain, peer-to-peer (P2P), robo-advising, and IoT. Table 1 provide
展开阅读全文