At the crossroads of corporate strategy and finance lies valuation1. Primarily, it helps in determining the resale value, potential mergers and acquisitions or sourcing funding for the firm. Since it needs in-depth analysis by professional services, advisory houses or investment banks, it also enables a better understanding of the company’s assets. However, valuation comes at an exorbitant cost and is not free from human bias that can radically change its results.
In this article we explore the possibilities of applying unsupervised machine learning (ML) to optimize valuation and to reduce human errors.
A springboard for machine learning assisted valuation
The market approach is chosen by investment professionals and financial analysts, due to its ease of calculation and availability of data and it is the most common method for valuation. In this approach, analysts identify peer companies for the firm in question and by using their data, a certain multiple is calculated (such as the P/E, EV/EBITDA and others).
Valuation requires collection and cross-referencing of a large amount of data requiring hours of manual work by analysts who can be limited by their opinions or experiences. This creates the need for a better system.
The data ranges from financial statements, stock and debt price movements, industry analysis, market environment and other factors left to the discretion of the financial analysts, in order to identify the best peers to be used for valuation. Moreover, the results of the valuation can vary due to factors like individual competencies.
A fresh outlook to valuation with unsupervised machine learning
Unsupervised ML models are typically used to identify patterns across datasets and cluster different observations into groups. This fits the objectives of peer analysis. AI excels at analyzing feature variables mentioned above, as well as the general strategic direction of the firms, in a quick and optimal manner. The process is complete in a matter of minutes and it removes the potential for human errors. Moreover, valuation using ML is economically viable, as the process can be reproduced at a marginal cost.
The need to reduce manual analysis, human errors and associated costs opens unprecedented avenues for ML assisted valuation.
In the future, far from replacing financial analysts, ML will augment their role by facilitating better access to information and reducing effort. By bringing together individual expertise and accuracy of ML, a new landscape for corporate valuation will be created.
Source: Infosys Consulting: AI powered peer selection for multiples valuation
Benefits of machine learning assisted valuation
Apart from the benefits mentioned above, applying unsupervised ML to valuation can have a number of benefits:
- Setting correct benchmarks through NLP: Immediate identification of peer firms and their relevant multiples by considering the future potential of the firm, the financial performance and the stock and debt movements. The first is captured by incorporating the management reports, earnings call transcripts and the strategic direction. This is achieved thanks to the natural language processing (NLP) of the annual and quarterly reports and the press coverage. The financial statements, as well as their historical movements are good predictors of the company’s overall performance and shed light on the management performance. Stock and debt price movements capture the investor sentiment and preferences, as well as anticipate future funding costs.
- Peer Strategy Analysis & Peer Benchmarks: Apart from the valuation, the model output provides opportunities for investors, investment banks and the management to assess the performance against peers and gain valuable insights.
- Peer influencing features & segment drivers: ML Identifies influencing factors in order to cluster peers. These factors help in understanding segment drivers and focus areas to improve valuation.
- Continuous learning & update: Thanks to marginal costs and ease of repeating the process for firms, to empirically validate their growth assumptions, and test and compare against competition
Through years of experience in Financial Services and the AI technology deeply seated into the DNA of our “Next in Banking” unit, Infosys Consulting has built the expertise required to deliver the model depicted above.
Want to learn more? Get in touch with our experts.
Gevorg Karapetyan
Senior Consultant
Gevorg has more than 8 years of experience in digital transformation and Consulting. He currently focuses on AI-powered digital banking platforms for our German Financial Services practice. He specializes in strategy and digital product design leveraging AI/machine learning and automation. He complements our financial projects with strategic approach development including our proprietary 32E innovation and design approach and supports our innovation hub team with use cases for “Next in Banking”. He holds an MSc. from the Goethe University in Frankfurt and a Licence from IAE Jean Moulin Lyon 3 University.
Daniela Rothley
Associate Partner, Infosys Consulting
Daniela has more than 18 years of experience in the banking industry and heads up our capital markets business in Germany. She joined Infosys Consulting in 2017 from Sopra Steria Consulting where she was a business unit manager and prior to that, she was a director at SQS. Daniela started her career as a banker, working for LBBW and Dresdner Kleinwort Wasserstein, before moving to consultancy.
Elif Ercan
Principal , Infosys Consulting
Elif has more than 10 years of experience in capital markets and leads the front office, treasury as well as the future technologies area of our capital markets business in our German Financial Services practice. Over the course of her career, Elif has had extensive experience in pricing interest rate derivatives, implementing regulatory requirements, and leading teams in strategic product development for the front office and treasury business. She specializes in advising financial institutions in defining their digitization strategy leveraging AI/ML and automation.
