Financial technology, commonly referred to as fintech, continues to transform daily life, from paytech to analyzing stock prices, and Oliver Yao, the new dean at the Alfred Lerner College of Business and Economics, has stated that the University of Delaware’s vision for collaboration on technical innovation was one of the reasons that drew him to the Lerner job.
Numerous Lerner College professors are currently conducting research in the fintech area, and presented their findings in November at the Lerner Faculty Research Showcase.
“The goal of today’s showcase is to share research ideas, and hopefully everyone in attendance will get some ideas of how to complement their research with other faculty here. Our goal is to conduct more research in the fintech area,” said Lerner Deputy Dean and Aramark Chaired Professor Sheryl Kline, whose remarks opened the showcase.
Lerner faculty members gave four-minute lightning talks during the session, with topics ranging from utilizing technology to analyze financial documents to interpretable AI for portfolio management. The research presented included working papers and accepted and published research.
Paul Laux, professor of finance, JP Morgan Chase Senior Faculty Fellow, also highlighted the upcoming spring Fintech and Financial Institutions Research Conference on April 19. The conference is sponsored by the FinTech Innovation Hub and the Philadelphia Federal Reserve. The goal is to provide added value to the connectivity between fintech organizations and the greater financial services system. It would be open to all FinTech Innovation Hub users and firms, and a platform for interaction between researchers and practitioners.
Below are summaries of the presentations:
Research on PayTech
Oliver Yao, dean and professor of MIS
Yao has had two papers recently published in Information Systems Research on facial recognition payment and mobile money. His research interests in paytech include adoption of emerging paytech, impacts of consumer behaviors, firm strategies and social issues by emerging paytech, and paytech in retail.
Yao concluded with his paytech predictions which included that wallets will become obsolete since people will stop carrying paper money, consumers will use multiple currencies daily, and cryptocurrencies will prevail but not necessarily Bitcoin.
Inverse Optimization
Ming Zhao, associate professor of operations management
Zhao first described forward thinking as inputting data into an algorithm or model and arriving at a decision, for instance the weight of different risks. Conversely, inverse optimization features the decision saved in the data set, and using that information to test that against the judgment researchers have. Inverse optimization is a framework not only for scenario analysis but also for systematically improving decision making.
Utilizing Technology to Analyze Financial Documents
Jing He, assistant professor of accounting
He shared her experience with utilizing technology to analyze SEC findings. Researchers can utilize EDGAR to view a multitude of financial reports including 8K (current reports), 10K (annual reports) and 10Q (quarterly reports), speeding up the research process to. He’s study constructed and validated 8K-based voluntary disclosure measures for a broad sample of firms, constructed concurrent mandatory disclosure methods, and conducted content analysis of the information within voluntary 8K items and exhibits.
Reference Point of Retail Investors
Michael Gelman, assistant professor of finance
Gelman and his fellow researchers found that the market portfolio serves as the reference point, while presenting a different market index results in the investors’ reference point changing to the new one. They also found that risk-taking is also a function of how they trade. If investors use an app, they conduct more risky trading than using a computer or placing a phone call to a banker.
Attenuating Investment Decision Volatility Under Interruptions: Embedded Mindfulness
Asli Basoglu associate professor of accounting and MIS
Basoglu’s research looked at investment decision banking, and various applications an investor can use to pursue their investment decisions. Her research shows that in a multitasking environment, instability in decisions causes risk taking to increase. Mindfulness applications play a role in increased resilience, leading to awareness, focus and therefore better choices.
Brand Personality and User Engagement on Social Media
Xiaoye Cheng, assistant professor of MIS
Cheng noted that traditionally, brand personality is often conveyed through advertisements, brand logo and user word-of-mouth. In social media context, brands engage with customers through posts and responses on their business pages. Brand personality traits such as openness, conscientiousness and extraversion result in increased user engagement through likes, shares and comments.
Chief Information Officers (CIOs) Joining Outside Boards: Impact on Their Own Firms’ Cybersecurity
John D’Arcy. professor of MIS
D’Arcy stated that firms are looking for cybersecurity expertise for their boards of directors, but because companies are having trouble finding these experts in house, they are looking to outside firms for potential CIO board candidates. His research looked at whether “lending” a CIO to an outside firm helped or hurt the lending firm’s cybersecurity.
His research found that the home firm had a significantly increased likelihood of a data breach when its CIO joined an outside board, but only when it did not have a Chief Information Security Officer on the top management team.
Privacy-Preserving Credit Risk Prediction with Alternative Data
Xiao Fang, professor of MIS and JPMorgan Chase Fellow
Fang’s fintech research includes financial document analysis, involving designing novel deep learning algorithms and methods that analyze firms’ financial documents such as 10k forms and patent filings to solve important problems such as industry classification, firm-industry assignment and patent valuation.
Semi-Parametric Probabilistic Programs for Bayesian Inference
Adam Fleishhacker, associate professor of operations management
Fleishhacker described a machine learning model called Bayesian Additive Regression Trees (BART) which is ranked highly in every causal analysis data competition. Relating BART to fintech, Fleishhacker explained that financial companies often need explainable models when making loan decisions. Banks must disclose to customers why they were approved or denied credit. He described using an interpretable model which banks can use to tell a consumer why they were denied, but also get some advantages from a BART model which can handle non-linearities and interactions in a way that a logistic regression cannot.
The language of (non) replicable social science
Michal Herzenstein, associate professor of marketing
Herzenstein stated that there is a replication crisis in social science, since manual replications are costly. She and her fellow researchers used 299 publicly available papers and found that text helps predict replicability in paper and study texts. Language of replicable studies is transparent and detailed, while language of nonreplicable studies is obfuscated, tells a good story and uses clout, weasel words and positivity.
Does High-Frequency Trading Cause Stock Prices to Deviate from Fundamental Values?
Michael Jung, associate professor of accounting
Jung and his fellow researchers defined high-frequency trading (HFT) as using computers and algorithms to automatically submit, route and cancel orders, and execute trades of stocks. The positive is that HFT enables fast, easy and low-cost transactions, while the negative is that it can create unintended volatility and trading based on noise.
Since in the age of HFT, it’s impossible for humans to outwit and out-speed computers, profits are lower and there is less incentive to devote time and energy to research. With less researching, trading and investing by fundamental investors, stock prices can deviate from fundamental values. The research implications were that undervalued stocks become more undervalued and overvalued stocks can become more overvalued, and technology in trading can crowd out the human approach to fundamental investing.
Hospitality Analytics and FinTech
Jing “Joy” Ma, assistant professor of hospitality and sport business management
Ma’s research focuses on unconstrained demand forecasting and accuracy measures. Although her work is industry specific, some insights are generalizable and could facilitate collaboration with fintech.
Barriers to Invest in NFTs: An Innovation Resistance Theory Perspective
Ahmad A. Rabaa’i, associate professor of MIS
Rabaa’i discussed his research investigating the barriers influencing investors’ resistance to investing in non-fungible tokens or NFTs using the innovation resistance theory (IRT). The most prominent functional barrier to investing in NFTs is the usage barrier, as investing in NFTs in a relatively novel concept, investors may feel more comfortable with traditional investments in commodities like money, oil, real estate and stocks. The least influential barrier is the value barrier – in comparison to the more conventional investment choices, respondents did not perceive the value of investing in NFTs.
Competitive Intelligence as a Function in FinTech Firms
Tom Tao, assistant professor of management
Tao’s research found that online job posting data can provide large amounts of real-time, reliable information. He pulled 12 years of historical data from a repository of job postings that provided a near-complete list of all CI-related positions advertised in the U.S. fintech industry from 2010-22.
Interpretable AI for Portfolio Management
Jiaheng Xie, assistant professor of MIS
Xie’s research deals with financial data, which is very hard to predict due to high-dimensional, noisy, non-linear dynamics. The objective is an interpretable deep-learning approach for stock mean-variance predictions. Xie is using a transformer-based double machine learning model which gives a general framework for estimating causal effects using machine learning and also confidence intervals for those estimates.