Investor Activity and Performance

Abstract

This dissertation, which examine different aspects of investor activity and performance, were made possible by a unique dataset covering all investors' transactions in the Norwegian stock market over more than two decades. I spent the better part of a year cleaning data and consulting with market participants and the data vendor to understand the data's intricacies and formulate relevant research questions. While processing half a billion transactions was an extremely demanding task, this novel market-wide dataset enabled both the investigation of new questions and a holistic reexamination of existing theories about market participants.

In the first paper, “Investor Activity and Wealth Transfers”, I investigate the evolution of investor activity and the consequential wealth transfer between investor categories. I document excessive activity across investor categories with significant downward-sloping trends in both activities and wealth transfers over time. This study employs a novel performance attribution model decomposing transactions into market timing, stock selection and short-term trading. The method avoids the caveats of statistical estimations by directly calculating activity and wealth transfers from the detailed data. The first paper's comprehensive analysis of all investor categories and market activity evolution over time makes novel contributions across multiple research streams, enriching our understanding of market dynamics. Its descriptive approach offers fresh perspectives that complement traditional academic frameworks. Following coverage in the leading national financial newspaper, the paper garnered much interest from market practitioners, reaching the top 10% most viewed papers on SSRN within weeks of publication. This enthusiastic response demonstrates how bridging theoretical research with practical market insights can create particularly valuable contributions to both academia and industry.

In the second paper, “Investor Category Trading Imbalances and Return Predictions When Predictions Balance Out”, I investigate an apparent paradox in the literature regarding how different investors' trading imbalances predict stock returns. The research offers three main contributions to the field: It first demonstrates that institutional and retail trading patterns have opposing relationships with future returns - institutional buying generally signals positive returns over periods up to 250 days, while retail buying signals negative returns of comparable size. Second, the paper explains why previous studies may find conflicting results, particularly when they examine incomplete investor groups or use metrics that are not zero-sum across all investors. Finally, by analyzing four distinct institutional investor categories, the research challenges existing theories about liquidity provision, finding that neither retail investors nor dealer banks serve as liquidity providers to other institutions seeking rapid trade execution.

In the third paper, "Six combinations of disposition effects: Investor heterogeneity and the effect of aggregation", I investigate a well-researched topic, the disposition effect, among mutual funds. I find that even the most sophisticated professional investors exhibit this bias in their trading patterns. The average mutual fund shows two distinct tendencies: a propensity to sell winners over losers (the classical disposition effect) and a propensity to sell stocks with high absolute returns (the absolute or V-shaped disposition effect). Through detailed transaction data analysis, I demonstrate that aggregation obscures significant heterogeneity in selling behavior, and that the V-shaped disposition effect prevalent among mutual funds significantly detracts from selling performance.