The Future Monetary System
After completing this reading, you should be able to: Identify and describe the... Read More
After completing this reading, you should be able to:
Artificial intelligence refers to the ability of machines to execute complicated tasks and train through experience on how to improve the performance of these tasks. On the other hand, machine learning is a type of AI that employs the idea that systems can learn (train) from data, identify patterns, and make decisions with minimal human intervention. It has an iterative aspect in that when models are exposed to new data, they can adapt independently.
Financial firms have the main aim of increasing revenue and saving costs. This has contributed significantly to the exponential growth of machine learning and artificial intelligence. In the following section, we look at supply and demand factors more closely to improve our understanding of the growing application of Fintech by financial institutions.
Supply factors that have driven the growth of Fintech in financial firms include:
The inclination to increase profitability propels the demand for AI and ML. The following are drivers for the growing adoption of AI and ML in the context of demand:
AI and ML have gained popularity in the financing industry over the last few years. Large amounts of client data are fed into the AI and ML algorithms to assess credit quality and price loan contracts. The following are some of the customer-focused uses of AI and ML in the financial sector:
Machine learning tools designed for credit scoring speed up lending decisions and limit incremental risk. Firms now rely on data not generally in their credit reports to perform credit quality analysis. Individually, lenders have turned to additional, unstructured, and semi-structured data sources such as social media activities to capture a nuanced view of creditworthiness. Applying ML algorithms to this group of new data enables assessment of qualitative factors such as willingness to pay and even consumption behavior.
Besides, the use of ML algorithms in credit scoring has led to greater access to credit. Traditional credit scoring models required a potential borrower to have a sufficient amount of historical credit information to be considered scorable. Without this information, the model could not generate a score leaving out potentially creditworthy borrowers without any means to build a credit history.
ML is used in the insurance industry to analyze complex data, thus lowering costs and improving profitability.
AI and ML applications provide improvements to the underwriting process and thus assist agents in sorting through big datasets that insurance companies have collected to identify and isolate cases that pose a higher risk, potentially reducing claims.
Moreover, insurance companies can use ML to improve the pricing or marketing of insurance products through the incorporation of real-time highly granular data, such as online shopping behavior and telematics. Therefore, accurate pricing of insurance could significantly change the insurance market as well as improve claims processing
Chatbots are virtual assistants that help customers complete transactions and solve simple problems. These automated programs use neuro-linguistic programming (NLP) to interact with clients either by text or voice. Financial firms use chatbots mostly in their mobile apps or social media. Chatbots provide balance information, give alerts, or answers simple customer questions. Cost savings and better communications with customers can both increase profitability.
Financial institutions employ AI and ML tools for several operational applications. The following are some of these applications:
Capital optimization is a traditional function in running a bank that relies heavily on mathematical approaches. AI and ML tools are built on the foundations of computing capabilities, big data, and mathematical concepts of optimization to increase the efficiency, accuracy, and speed of capital optimization.
Banks use ML to make sense of large, unstructured, and semi-structured datasets and to monitor the outputs of primary models. Unsupervised learning algorithms, for example, can be used to detect abnormal projections generated by stress-testing models. Additionally, these ML algorithms help model validators with ongoing monitoring of internal and regulatory stress-testing models by determining whether the stress-testing models are performing within acceptable tolerances or diverting from their original purpose.
AI and ML tools can also improve risk management. These tools assist banks in modeling their capital markets business for stress testing. They can document any bias concerning the selection of variables by using unsupervised learning methods to review large amounts of data, leading to better models with greater transparency.
AI and ML can be used hand in hand with the conventional market impact models. Firms can use AI to identify non-linear relationships to obtain more information from sparse historical models. On the other hand, ML can be used to create trading robots that can learn ways of reacting to market changes. Market impact analysis entails the evaluation of the effect of a firm’s trading on market prices. A more accurate estimation of the impact of trades on market prices is key to timing trades and minimizing trading execution costs for most firms, which can be achieved by the application of AI and ML tools.
AI and ML techniques are research areas for asset managers and trading firms. Additionally, some firms use ML to devise trading and investment strategies. The following are areas in which AL and ML can be used in trading and asset management:
Trading firms look up to AI and ML to use historical data to learn and improve their ability to sell to clients, with minimum human input. Analyzing past trading behavior, for example, can help anticipate a client’s next order. The current use of computer-mediated transactions helps generate large quantities of data that is required by ML tools to work effectively.
Additionally, AI and ML can be used to manage risk exposures. Firms can determine when members’ trading account positions may have increased risk profiles that may warrant intervention using exchanges that apply machine learning as a basis for modeling risk.
Lastly, AI and ML can also help these firms comply with trading regulations. A Regtech application of AI to trading may help firms meet pre-trade and post-trade transparency requirements for the non-equity markets.
In portfolio management, AI and ML tools are used to identify new indicators on price movements and make more effective use of the available large datasets and market research than with the current models. ML tools work to identify signals from data on which predictions relating to price volatility can be made in different time horizons to generate higher and uncorrelated returns.
Asset managers are also using AI capability supplied by third parties to build indicators. However, trading signals derived from AI and ML strategies may follow a decay function over time, with the use of more data and hence become less valuable for gaining an edge over the other investors.
AI and ML techniques are also used for regulatory compliance by the regulated institutions and authorities for supervision. Regtech is seen as a subset of Fintech that focuses on ensuring that regulatory compliance is more efficient and effective compared to the existing capabilities.
Suptech, on the other hand, is the use of these technologies by the public sector regulatory and supervisory bodies. In Suptech, the objective of AI and ML applications is to enhance the effectiveness and efficiency of supervision and surveillance.
Regtech can use ML together with neuro-linguistic programming (NLP) to analyze unstructured data and monitor behavior and communication of traders for transparency and market conduct. ML, combined with NLP, can interpret data inputs such as documents, spoken words, e-mails, instant messaging, among others. This, in turn, raises the issues of the boundaries for the employees’ surveillance policy.
NLP can be used by asset management firms to cope with new regulations. Firms could potentially leverage NLP and other ML tools to interpret the regulations into a universal language. These rules are then analyzed and codified for automation into the integrated risk and reporting systems that help the firms to comply with the regulations cost-effectively.
Additionally, AI and ML can be applied in knowing the identity of customers – know your customer (KYC). The KYC process is costly, strenuous, and highly duplicative across many services and firms. ML can be used in remote KYC of financial services firms to do identity and background pre-checks.
AI and ML methods may be used to automate macroprudential analysis and data quality assurance. The new reporting requirements across jurisdictions have led to a higher volume and frequency of reported data. Also, completing these reports in time requires more significant amounts of resources from financial institutions. AI is also used to improve data quality by identifying missing or erroneous values. In general, AI is being used to replace human functions using NLP, unstructured text analysis, and analysis of large data sets.
AI is used by central banks to assist with monetary policy assessments. Additionally, central banks are using AI for economic forecasting to obtain economic indicators such as inflation and to improve their understanding of the relationships among these indicators. Moreover, AI can be used to forecast GDP, unemployment, industrial production, tourism activity, and the business cycle.
Regulators are using AI for fraud detection and detailed analysis of the tremendous amount of market and trading data that is currently available to them. Such analysis is suitable for detecting suspicious transactions and patterns that should be further investigated. This allows supervisors to focus on higher-risk transactions. Investigating suspicious transactions is time-consuming and, in most cases, turns out to be false positives due to defensive filings by the regulated entities. ML is used to identify intricate patterns and pinpoint the suspicious transactions that may be more serious and need closer attention.
Since AI and ML can substantially improve the efficiency of information processing, leading to reduced information asymmetries, their applications can strengthen the information function of the financial system. The following are the potential benefits and risks that AI and ML poses to financial markets:
Saving costs, improving products, and making efficiency gains are the key drivers of the proliferating use of AI and ML. AI and ML allows market participants to collect and analyze information on a grander scale. This automation reduces costs and increases customer options. AI and ML tools also help market participants understand the relationship between the formulation of market prices and various factors such as sentiment analysis. Consequently, this helps reduce information asymmetries, thus contributing to the efficiency and stability of the markets.
Using AI and machine learning dramatically improves fraud detection, protection against cyberattacks, and risk management. Additionally, the application of AI and ML reduce trading costs. These tools enable market participants to adjust their trading strategies following a changing environment in a swift manner. This improves price discovery and reduces the overall transaction costs, thus increasing the profitability and capital position in the long run.
Increased use of AI and ML algorithms may lead to additional risks. It is blurred how the increased use of AI and ML may affect financial markets in response to the financial crisis. This is because the decision making of AI systems may not be easy to understand when employing it.
Additionally, some regulators who do not understand the working of AI systems may experience challenges in assessing and managing risks of their use, and finding specialists in these systems may be hard.
The application of AI in fraud detection and improving the supervision of firms and markets may reduce the potential for financial market instability and systemic risk. However, large providers of AI applications may be operating outside the existing regulatory framework, thus increasing systemic risk and the likelihood of financial market instability.
Finally, some ML and other AI models may fail in periods with higher volatility. This is because they have learned over a period of low volatility in the financial markets. This could lead to financial instability in case of a financial crisis.
After completing this reading, you should be able to: Identify and describe the... Read More
After completing this reading, you should be able to: Differentiate between the Basel... Read More
After completing this reading, you should be able to: Describe climate-related risk drivers... Read More