By Hao ZHANG, Associate Professor in Accounting
University of Macau
The proposal from Beijing to establish a stock exchange denominated in RMB in Macau has generated lively discussions, both inside Macau and within the GHM (Guangdong-Hong Kong-Macau) Greater Bay Area, about the future direction of financial development in Macau.
Alongside this is the on-going discussion about the need to diversify the gaming-led Macau economy through organic introductions of innovative industries and technologies such as Artificial Intelligence (AI) and Big Data Analytics into the city. In this column I argue that these two lines of discussion, far from being separate, are in fact synergic. The link lies in the highly promising ability of AI analytics to enhance the underlying information infrastructure that is required for a functional modern financial market.
A viable modern financial market is predicated on a functional underlying information infrastructure that can enable investors and other stakeholders to make informed and efficient investment-related decisions. Specifically, the underlying information infrastructure would need to be able to: (1) process and translate all available firm-relevant data (raw information) for a period of time into descriptive information readily available for further analysis (addressing the “what is this?” question); and (2) transform the descriptive information through further analysis into predictive or useful intelligence or knowledge for a firm, a sector, and the market overall (addressing the “what is likely to happen?” question). Of course, the resultant predictive intelligence/knowledge can be put through yet further prescriptive analyses tailor-made for individual investors or stakeholders to facilitate individual decision-making (addressing the “what should I do now?” question).
The nature of firm-relevant data (raw information) is such that the data are always multimodal in form, including the traditional numerical mode (actual earnings, earnings forecasts, periodic debt obligations, etc.), the textual mode (earnings-related textual explanations, clarifications, forecasts, analyses by the firm in question, the media and information intermediaries, or online forums, etc.), and the audio and visual modes (audio or video records of media interviews, public speeches, round-table discussions of senior management, earnings-conference calls, vlogs by individuals and those representing or affiliated with different organizations, etc.). In particular, data in non-numerical modes (textual, audio, visual) are conventionally known as unstructured data, in part because they cannot be readily indexed for comparisons across firms and over time. With the advent of the internet and wireless technology, the amount of unstructured raw information in non-numerical modes available to the public has grown exponentially in volume at any one point in time. Needless to say, the cost of processing, aggregating, and translating this increasingly vast amount of relevant but unstructured raw data at any one time into descriptive information and then transforming it into useful intelligence/knowledge is non-trivial and can be daunting.
Economists starting as early as F.A. Hayek (and in the days when financial information was construed as in the numerical mode only) have realized that the process of aggregating and transforming voluminous raw information into predictive and prescriptive intelligence is not a costless exercise. The solution proposed by economists (for example, S.J. Grossman and J.E. Stiglitz) is in the form of a cost-benefit tradeoff:
process more raw information only if the additional benefit of processing it exceeds the additional cost. While the cost-benefit tradeoff approach may well be fine at the individual level, it is a different matter for the information infrastructure of a financial market as whole. To the extent that the additional cost of transforming more relevant raw information into useful intelligence can be reduced by, for example, technological advances, the informative function of the information infrastructure of any financial market is further enhanced and the financial market is made more competitive and vibrant. In other words, a viable financial market is best supported by an underlying information structure which is inclusive of relevant innovative technologies.
The encouraging news is that recent advances in AI Analytics are highly promising in reducing the cost of processing and transforming relevant data, particularly unstructured Big Data in non-numerical modes. Specifically, data in non-numerical modes (textual, audio, visual) can be quantified or indexed alongside more traditional numerical data for comparative analysis across firms and/or over time. The AI methods of natural language processing (NLP), machine learning (ML), and deep learning (DL) are increasingly mature and when supported by modern computing power, feasibly operational. Armed with these methods and using them with acumen, we can process and analyze a vast amount of multimodal, unstructured financial data at any time with reasonable ease and speed as well as within moderate and decreasing costs. The fusion of relevant but unstructured Big Data in different modes into a coherent whole in order to enhance the overall informativeness of available data has become a hot academic research topic in accounting, finance, and computer science. It is now possible to provide real-time or near real-time predictive analyses based on latest available data on a firm, a sector, or the market.
Thus, the logic that the development of a financial market goes hand in hand with the development of innovative industries in Macau is straightforward. A functional financial market requires the infrastructure support of an AI-lead information industry and the process of financial development induces more organic developments in both financial and innovative industries. Since the Covid-19 experience in the last year or so has made it clear that Macau needs to diversify its gaming-led economy, coordinated deliberations and planning of Macau SAR government policies for developing both the financial and information industries warrant serious and urgent consideration.