Week 9 - Investor Behaviour Flashcards
Study of individual investors
Household finance
Data on household finance
- Survey of consumer finances
- Brokerage data
- Finnish Data
- Swedish Data
- 401k data
Facts about individual investors
- Non-Participation
- Buying high, selling low
- Life-cycle behaviour
- Under-diversification
- Preference for active management
- Stock-picking performance
- Selling behaviour
- Buying behaviour
Non-participation
- Basic rational model without frictions predicts that people will participate in risky assets
- surprisingly, a large fraction of the population, even wealthy people, doe not participate in the stock market
- worse in earlier years (28% in 1984, 48% for those >100,000 in liquid assets)
Non-Participation facts
- Participation is positively related to education, income and wealth
Cross country participation trends
- As traditional gender role attitudes increase in a country, the share of direct participation in stock market decreases
Rational explanations for Non-Participation
- Transaction costs
Buying high, selling low pehnomenon
- Evidence that some groups of investors buy high and sell low
- Increase exposure to a risky asset class in advance of low returns
- Decrease exposure in advance of high returns
Evidence on Buy high, sell low
- Dollar weighted average return on mutual funds is lower than the time-weighted average by 1.5% per year
Life Cycle Behaviour of risky asset exposure
Most rational models predict that the share of financial wealth allocated to risky assets will decline over life cycle
- Empirical evidence is broadly consistent with this
- Empirical level of risky asset holdings is lower than predicted
Under-Diversification
- Basic rational model without frictions predicts that people will holds diversified portfolios
- Empirically, many households appear to be under-diversified
Under-Diversification Type 1
Type 1: Home Bias
- Households allocate a large fraction of equity holdings to domestic equity
- Eg. in 1989, in US, UK, and Japan - 94, 82 and 98% of equity holdings were in domestic equity
Under-Diversification Type 2
Local Bias
- Households allocate a large fraction of their domestic equity holdings to stocks headquartered locally
- Using Odean Brokerage data, Ivkovich and Weisbenner show that 31% of stocks are located within 250 miles of investor, v 12% of all stocks
- Average distance of Portfolio is 917 miles, average distance to all stocks: 1225 miles
Under-Diversification Type 3
Concentrated holdings of individual stocks
- Underdiversified definition: more than 50% of equity exposure in brokerage account with less than 10 stocks
- in 1989, approximately 30% are undiversified
- 2001, 14% are undiversified
Under-Diversification Type 4
Large Holdings of own-company stock
- in 401K retirement plans, people allocate large amount to the stock of their own company
- out of 154 firms in the S&P500 studied: 90 are cash-match firms, 64 are stock-match
- Across all plans, people invest 23% of discretionary contributions in company stock - 18% are cash-match, 29% stock-match
Actively-managed funds Facts
- Average net return of actively managed funds appears to be below that of index funds
- In basic rational model without frictions, average net returns of active and index funds should be the same
- Poor performance of active funds suggests that these funds may be too popular
Actively-managed funds fund flows
- fund flows are a positive function of past alpha - relatively little persistence in fund alphas
- The basic rational model predicts that fund flows should not predict alpha - no fund characteristic should predict alpha
- However, several variable do eg. active share, fund size, manager education
Overall: fund flows may not be fully rational
Stock-Picking performance
Study: 66,000 households through large, discount brokerage firm from 1991 to 1996
Findings: net returns of typical households is below range of benchmarks:
- Household’s beginning of the year portfolio
- Value-weighted market return
- CAPM benchmark
- 3-Factor Benchmark
Selling Behaviour
- Individuals have a greater propensity to sell a stock trading at a gain relative to purchase price than one trading at a loss - disposition effect
Selling Behaviour Disposition effect
- On any day on which an investor sells a stock in his portfolio, put each stock in his portfolio into one of four buckets: realised gain, paper gain, realised loss, paper loss
- Count up number of items in each bucket across all investors and over full sample
- Compute: proportion of gains realised (PGR) & proportion of losses realised (PLR)
- Disposition effect: PGR > PLR
Selling Behaviour Rational explanations
- Information
- Liquidity
- Taxes
- Rebalancing
Selling Behaviour Behavioural Explanations
- Prospect Theory
- Mean-Reverting beliefs
- Cognitive Dissonance
- Realisation Utility
Disposition effect across different groups of investors
Stronger for less sophisticated investors:
- More sophisticated investors exhibit the effect less
- Mutual fund managers exhibit it, but less than individuals
- Households also exhibit a disposition effect in the real estate market
Selling Propensity v returns since purchase
Selling propensity has a V-Shaped relationship with returns since purchase
- Selling propensity also depends on a stock’s rank in the investor’s portfolio
Buying Propensity
- Barber and Odean (2009) find that individual’s buying propensity is a positive function of a stock’s past returns
Institutional Investors
- Basic Frictionless model with skilled managers and rational investors predicts that mutual fund’s gross alpha will be positive on average and net alpha will be zero on average
- In practice, gross alpha is positive but not substantially so