Do individual investors have asymmetric information based on work experience?
Motivation Trend towards increased investor autonomy how well do people perform as their own money managers? Individuals own a sizeable fraction of stock markets & can affect asset prices Important to understand individual behavior and biases in investment decisions
We link investment behavior and individual s expertise Focus on expertise obtained through work experience
Two main questions Do individuals overweigh their holdings of professionally close stocks? Do such (excess) holdings generate a superior returns?
Outline I. Introduction II. Theory III. Data and descriptives IV. Holdings of professionally close stocks V. Returns on professionally close stocks VI. Behavioral interpretation VII.Conclusion
II. Theory
Professionally close stocks: Standard finance theory Uninformed individuals: Returns on professionally close stocks tend to be correlated with future labor income -> shy away from such investments Informed individuals: - Get information through social or professional network or through more intelligent interpretation of data - Geologist in oil industry -> Invest in professionally close stocks if can obtain excess returns
Standard theory predicts Uninformed investors should shy away from professionally close stocks, while informed investors should invest in it. -> excess holdings of professionally close stocks should be associated with superior returns
Behavioral theory Heath & Tversky (1991): holding judged probability constant - people prefer to bet in a context where they feel knowledgeable or competent than in a context where they feel ignorant or incompetent. Huberman (2001): familiarity bias Individuals overweigh professionally close stocks because they are more familiar with such stocks? Would not imply superior returns Consistent with negative abnormal returns?
Predictions from theory Excess holdings of professionally close stocks can be explained by both rational and behavioral theories Different predictions with respect to returns
What we do Use unique Norwegian data to 1. Document to which extent individuals invest in professionally close stocks, based on matching of SIC codes -> Expertise bias 2. Several tests on whether such holdings are associated with superior returns
Prior literature Home bias: - Investors tend to hold securities inside their own country Local bias: - Investors overweigh holdings in stocks that are headquartered within 250 miles of home (Coval & Moskowitz,1999) Could be due to hedging reasons stock prices positively correlated with prices of local goods such as real estate -> Professionally close stocks are better environment to test rational versus behavioral theories Other: Cohen et al (2007) on networks of professional managers.
Can individual investors beat the market? Debate on whether individual investors sometimes can beat the market Odean (1998) finds significantly negative buy-sell returns using US data Other papers find return persistence for small group of investors
III. Data and summary statistics
Data Sample: All Norwegian individual investors on Oslo Stock Exchange 1994-2005. Data sources: Transaction record of all trades 1994-2005, obtained from the Norwegian Central Securities Depositary (VPS).Ticker and quantity. Identification of holdings Ticker prices from OSE and Borsprosjektet NHH. Returns on holdings For all listed companies SIC code (1994-2005) and daughter companies (1996-2001, 2005) Identify industry of a listed company Panel of education and labor market information for all Norwegian workers, covering 1986-2005. Includes SIC code of employer Expertise measure and identification of own-company stock
Own-company stock Important to identify because associated with tax-breaks and employer matching Including such holdings would blur our estimates of excess holdings and returns We can identify and exclude such holdings for 6 of the years. Rest of the years we make imputations. Results are robust to using only years where we have full hierarchy
Inclusion criteria Individual-years where at least NOK 5000 in holdings at the start of year Individuals that work in an industry with at least one listed company Stocks not held in own employer around 200 000 individuals
Descriptives (end of 2000)
IV. Expertise bias
Basic measures of expertise Expertise stock = stock whose SIC code matches the individual s SIC code of employment Report results mainly for the 2-digit mapping Expertise bias, two measures Actual holding in expertise stocks Holding of expertise stocks in excess of market portfolio
Expertise bias I (corrected for employer stock) mean Std.dev
Expertise bias II
Correlates of bias
V. Abnormal returns on professionally close stocks?
We are particularly interested in whether purchases in expertise stocks conveys asymmetric information
Empirical issues 1) Cross-sectional correlation - large number of investors but small number of stocks - returns highly correlated across stocks 2) Choice of benchmark returns 3) Power No universally accepted method to deal with these three issues
Methodologies Medium and long-term returns - Calendar time portfolio approach - Raw returns and risk adjusted according to four-factor model of Carhart, JF 1997 Short-term returns - Control-firm approach - Bootstrapped confidence intervals Compare expertise buy and expertise sell t
Calendar time portfolio approach Aggregates recent trades of a certain type into a single portfolio and analyzes the (1-month) returns of this portfolio Portfolio build-up period: 4, 12 and 24 months Analyzes (weighted) buy-sell returns If asymmetric information than should be positive
Control firm approach 1. Calculate actual average short-run returns on expertise buys 2. Calculate returns of fictitious portfolio consisting of buys in random stocks with similar marketcap/b2m values. 3. Repeat 2 until have a distribution of 1000 fictitious buy returns observations (draw figure) 4. Under the null hypothesis that the fictitious buy returns have the same returns distribution as the actual buy returns, we can test for abnormal returns by observing whether the actual returns lie in one of the tails of the distribution of returns for fictitious buys.
Control firm results
Comments Fictitious buy returns seem very large on short horizons Artifact of the way we have chosen the replacement groups? No. Get same result if use 1 group of replacement securities instead of 16 - the 1 group portfolio closely tracks the unweighted index return -> fictitious returns are high because small stocks do well on those dates. But expertise stocks do not perform well Small stocks do well on days with high individual investor trading activity for reasons that are unknown to us.
Summary of findings Individuals overweigh stocks to which they are professionally close by 7-15% No evidence that professionally close stocks give higher returns, in spite of poor hedging properties. Evidence of negative abnormal returns Behavioral theories gets more support. But which?
More analysis Education and expertise Local bias
Education Concivable that education also gives expertise Hard to match education to SIC codes Can ask whether complements
Local bias Is the expertise bias a rediscovery of local bias? - Could be the case if tightly correlated
VI. Which behavioral bias?
Which behavioral bias? Overconfidence - Too tight confidence intervals over future returns - Trade because disagree with market valuations Familiarity - exposed to certain stocks through e.g., workplace interaction Both are consistent with too much trading in expertise stocks and with more active trading overall Overconfidence more consistent with negative relation between returns and trading activity (trading costs unaccounted but would probably make significant) Overconfidence more consistent with negative abnormal returns?
Experts trading activity
Robustness Quantile regressions Value-weighted returns 5-digit SIC mapping
VII. Conclusion
Individuals overweigh stocks to which they are professionally close by 7-15% No evidence that professionally close stocks give higher returns, in spite of poor hedging properties Evidence on mistake of individual investors Overconfidence seems most plausible behavioral mechanism -> need for more theory
Feeds into the debate on whether individual investors sometimes can beat the market