Methodological Aspects of Organizing Scientific Research on Financial Risk Management in the Field of Intellectual Property (Наукові праці Міжрегіональної Академії управління персоналом Економічні науки)

Summary

• The article studies the theoretical and methodological foundations and the staged organization of scientific research on financial risk management for intellectual property.p.1
• The authors argue that classical methodology based on the normal distribution is inadequate for assessing rare events with catastrophic consequences.p.1
• A conceptual framework with five levels is presented, from the methodological level to a target block on financial stability and loss minimization.p.5
• Research design requires a justified benchmark and backtesting of models on real historical data.p.4
• Machine learning models face the black box problem and data bias, which can scale modelling errors.p.7
• A verification system built on data cleaning, algorithmic transparency and backtesting, supported by explainable AI, is proposed.p.8

Key insights

1. Complex stochastic models create a time lag that conflicts with the need for fast strategic decisions.p.1
2. A unified approach should be replaced by a classification of financial risks by type of intellectual property: patents, trademarks, copyright and know-how.p.1
3. Any error in forming the data sample distorts the final modelling results.p.4
4. Backtesting on real historical data confirms the adequacy of the tools and sets limits to their practical use.p.4
5. Quantitative methods listed include Value at Risk, Expected Shortfall, stress testing and Monte Carlo simulation.p.5
6. The black box problem of neural networks deprives risk managers of the ability to verify decision logic.p.7
7. Explainable AI helps break complex model decisions into understandable influencing factors.p.8