Under-diversification And The Role Of Recency And Probability
Matching
Uri Ben Ziona, Ido Erevb, Ernan Haruvyc , Tal Shavitd
a Department of Economics, Ben-Gurion University, Beer-Sheva, Israel.
b Faculty of Industrial Engineering and Management, Technion, Haifa, Israel.
c School of Management, University of Texas at Dallas.
d Department of Economics and Management, Open University of Israel.
Motivation
Empirical observation show that many many individual investors hold fewer individual individual investors hold fewer individual stocksstocks than necessary to eliminate idiosyncratic risk (e.g., Blume and Friend, 1975; Statman, 1987; Kelly, 1995; Odean, 1999; Polkovnichenko, 2004; Goetzmann and Kumar, 2004)
The current research
Three experiments are presented that compare alternative explanations to the coexistence of risk aversion and under-diversification in investment decisions.
The participants were asked to select one of several assets under two feedback conditions. In each case, one asset was a weighted combination of the other assets, allowing for lower volatility.
We Examine whether under-diversification is a result of momentum trading behavior (e.g., Grinblatt, 1995) we refer as the “recency hypothesis”.
Under momentum investment strategies, investors buy
past winning stocks and sell past losing stocks.
Thus, momentum trading is an example of a tendency to
rely on recent outcomes (Barron & Erev, 2003).
The Experiments
*Experiment 1 focused on a simplified investment task.
*In each of the 100 trials, participants were asked to choose one of three assets: A, B and C.
*The participants were 80 subjects in their second or third year of undergraduate studies that had taken at least one course in statistics.
*The experiments took place at a computer Laboratory at Ben-Gurion University, and lasted approximately half an hour.
*The participants were informed that they would be asked to invest 100 experimental tokens in one of the assets in each trial.
*To provide concrete incentives, subjects were told that the return in tokens in each round would be converted into NIS at a rate of 1 NIS for 200 tokens.
*They were also told that their profit from previous trials could not be used for reinvestment in the assets.
The treatments
We conducted two experiments including 40 subjects each.
The participants in each experiment were divided into two equal groups (of 20 subjects): Full and Limited information.
The groups differed with respect to the feedback provided after each trial.
After each trial in the Full Information Full Information groupgroup the participants saw their return and earning from the asset they chose and the forgone payoffs (the return of the other assets).
In the Limited Information groupLimited Information group the feedback was limited to the obtained payoff.
The Assets
The assets’ returns are constructed on the basis of two independent variables:
UU was drawn from uniform distributions in each trial in the range 0% to 100%.
εε was drawn from uniform distributions in each trial in the range -5% to 5% .
Assets
The correlation between assets A and B is –0.977. Asset C’s return was always between assets A and B returns.
Asset Asset return Mean return Standard
deviation.
A RA = U 50% 29.3%
B RB = (75%-0.5*U)+ ε 50% 14.92%
M RM = 0.5* RA +0.5* RB 50% 7.46%
ResultsChoice Frequencies in Experiment 1
Group Asset A Asset B Asset M
Full Information 46.65% 39.70% 13.65%
Limited Information 33.15% 27.45% 39.4%
Average Relative Frequencies –Experiment 1.
C-C0.082385959
Limited Information 1
05
10152025303540455055
1-20 21-40 41-60 61-80 81-100
Period
per
cen
tA
B
M
Full Information1
05
10152025303540455055
1-20 21-40 41-60 61-80 81-100
Period
per
cen
tA
B
M
in the Full Information condition, on average, subjects chose asset M less frequently than in the Limited Information condition (t =4.9, p < 0.01).
The difference, predicted under the contingent recency and probability matching hypothesis.
Subjects have no preference to asset M.
Experiment 2Experiment 2 was conducted to examine the robustness of the results of experiment 1 in settings where the less volatile asset has a higher expected return.
We replace asset M of experiment 1 with asset M+, which has 2% higher return.The results were the same as in experiment 1.
Experiment 3
To compare the laboratory results with a more realistic investment scenario, we recruited upper-division (juniors and seniors) economics students and MBA business students from two schools in Israel to participate in an online investment experiment. The subjects had all taken courses in finance and statistics
Participants were asked to log in to an investment web site and to make investment choices among five Vanguard mutual funds:
(1) Vanguard Total Stock Market Index Fund.(2) Vanguard Total Bond Market Index Fund.(3) Vanguard European Stock Index Fund.(4) Vanguard Pacific Stock Index Fund. (5) Vanguard Target Retirement 2015 Fund.
A composite of the other funds with the following percentages: 51.2% in Total Stock Market Fund, 36.0% in Total Bond Market Fund, 7.4% in European Stock Index Fund, 3.4% in Pacific Stock Index Fund and 2.0% in other.
To complete the experiment, participants had to make 30 separate investment decisions on 30 different days.
Each day they logged in they received feedback regarding the previous decision.
Full information treatment, and Limited information treatment.
Subjects received payment upon completion of the experiment for the returns they made in the experiment
RESULTSThe best performing fund was the European
Stock Index Fund (average daily return of 0.12%), followed by the Total Stock Market Index Fund (0.11%) and the Pacific Stock Index Fund (0.07%).
The most volatile by a large difference was the Pacific Stock Index Fund (std. dev of 0.82% compared to the second highest std. dev. of 0.66% for the European stock index fund).
In the limited information condition, the Total Stock Index and European Stock Index were most popular.
In the full information condition, the Total Stock Index and Pacific Stock Index were most popular.
In the limited information condition the volatile Pacific Index Fund was unpopular in the fourth place in terms of demand (17%), whereas in the full information condition it was the most popular (39%), particularly following strong performances
Summary and Conclusions
Previous studies of investment decisions highlight two robust but apparently inconsistent behavioral tendencies: Investors tend to exhibit strong risk aversion, but they also tend to prefer underdiversified portfolios.
The results demonstrate under- diversification that can be described a byproduct of a contingent recency effect and of probability matching.
It is important to emphasize that the current research does not prove that under-diversification in the stock market is driven by a contingent recency effect and/or probability matching.
The main contribution of the current research is the demonstration that under-diversification can be a product of these robust psychological principles.
