events

Jens Lausen

5th of November 2018

Jours fixes usually take place on the first monday of the month, starting at 5:00 p.m., in HoF E.01 / Deutsche Bank of the House of Finance (Campus Westend).

[Jens Lausen, E-Finance Lab]

Prediction of Financial Intermediary Misconduct based on Self-Disclosed Information

Financial intermediaries are essential for investors since they enable investors to participate in financial markets and exhibit strong influence not only on financial performance but also on wealth and life planning. With increased usage of the Internet and electronic communication, personal interaction between investors and financial intermediaries has been reduced, which impedes the process of building trust. At the same time, the financial crisis and widely noticed financial market manipulations have challenged investors’ confidence towards financial intermediaries and their trust in the financial system. Consequently, instruments are needed to increase trust and to ensure the proper functioning of the financial system.

Based on information manipulation theory, we argue that information disclosure of intermediaries committing misconduct differs from information disclosure of reliable market participants. Moreover, drawing on warranting theory, we propose that external verification of self-disclosed information provides additional value to identify misconduct. Following this rationale, we address the research question of whether self-disclosed information with varying levels of external verification can be used to predict financial intermediaries committing misconduct. We determine different feature sets that are supposed to enable investors and regulatory authorities to distinguish financial intermediaries that have committed financial misconduct from reliable ones.

We evaluate different predictive models that automatically classify financial intermediaries to identify whether they are likely to commit misconduct. Furthermore, we examine the economic relevance of our classifiers by means of an economic evaluation based on the financial compensation requested and paid. We find that self-disclosed information provided by financial intermediaries is valuable to detect financial intermediary misconduct. Specifically, classifiers additionally taking externally verified information into account achieve a promising classification performance and their application leads to considerable economic value for the society.