Which unrecognized intangible assets drive financial performance in leading global technology firms from 2008 to 2021, and how does ML-driven feature selection validate their strategic importance?

Purpose: Leveraging Machine Learning’s growing role alongside econometrics for high-dimensional data, this Master’s Thesis investigates and validates the influence of unrecognized intangible assets on financial performance in leading global technology firms. Design/methodology/approach: A multi-method framework combines Ordinary Least Squares (OLS) fixed-effects regressions with ML-driven feature selection (Bootstrap Lasso and Random Forest) to scrutinize contemporaneous and lagged impacts of proxies (brand value, employee satisfaction, R&D intensity, patent claims, SG&A intensity, SEC filing sentiment) on firm performance (Tobin’s Q, Price-to-Earnings P/E, Total Shareholder Return TSR) for 10 leading global technology firms from 2008–2021. Findings: Empirical findings unequivocally establish Brand Value as the paramount positive driver of market valuation (ln(Tobin’s Q)) and shareholder returns (ln(TSR)), with Employee Satisfaction as a fundamental positive predictor. Conversely, SG&A Intensity demonstrates a robust negative association, suggesting market interpretation of inefficiency rather than strategic investment. Disclosure Sentiment from SEC filings also often presented a paradoxical negative link, particularly with ln(Tobin’s Q). Traditional innovation metrics (R&D Intensity, Patent Claims) showed a muted direct impact. ML techniques confirmed significant non-linearities in these relationships. Originality/value: While ML’s application to unrecognized intangibles is nascent, this study contributes a data-driven hierarchy of these assets in the technology sector. It pioneers a methodological triangulation (OLS, Lasso, RF) for robust feature selection and validation, offering a granular perspective on how these nonphysical assets translate into financial outcomes and revealing complexities in their market interpretation. To the author’s knowledge, this is one of the early comprehensive studies to systematically rank such a breadth of unrecognized intangibles and validate their varied impacts using this combined OLS-Lasso- Random Forest approach in the context of leading global technology firms.