Web Appendix. Need for Cognitive Closure, Risk Aversion, Uncertainty Changes, and Their Effects on Investment Decisions.

Transcription

1 Web Appendix Need for Cognitive Closure, Risk Aversion, Uncertainty Changes, and Their Effects on Investment Decisions David Disatnik and Yael Steinhart Web Appendix A STUDY 1 Design We randomly assigned each participant to one of four conditions in a 2 (noise: with or without) 2 (uncertainty change indicator: decrease or increase) experimental design. Stimuli In what follows we present the screens that participants saw during the experiment. For clarity of presentation the text below includes a title for each screen; in the experiments participants did not see these titles. Screen 1: General instructions You are about to take part in a survey that explores consumers attitudes and perceptions toward investment decisions. Please read the instructions carefully and respond to the questions that follow. There are no right or wrong answers just respond to the best of your ability. The survey is anonymous and conducted for research purposes only. We would like to thank you in advance for your cooperation. Screen 2: The risk-aversion items Please rate how much you agree with the following statements: 1 strongly disagree strongly agree I would rather be safe than sorry. I want to be sure before I purchase anything. I avoid risky things. Screen 3: Various explanations The terms risk and reward are strongly related, and when one increases so does the other. That is, when it comes to financial products, if a product has a relatively high risk/reward level, it means that by investing in this product one can either make a lot of

2 money or lose a lot of money. In contrast, investing in a product with a relatively low risk/reward level can result in either modest gains or modest losses. In this survey, we focus on financial products (portfolios) that consist of bonds (which have low risk/reward levels) and stocks (which have high risk/reward levels). Thus, for example, a portfolio that consists of 10% bonds and 90% stocks is considered as a financial product with a relatively high risk/reward level. In addition, when comparing two portfolios each consisting of stocks and bonds the portfolio with the larger portion of bonds is considered safer. The noise manipulation [Participants in the with noise condition] Screen 4: The noise Prior studies indicate that consumers usually engage in investment decisions in a noisy environment. Therefore, we have included noise within this research, which will appear while you make your financial decisions. For that purpose, please make sure that your speakers work and keep them at their maximal level from now on until the end of the experiment. Screen 5: Noise validation Participants heard a voice saying the name Nicolas Sarkozy. Immediately afterwards they were told: Please report the name that you just heard: a) Angela Merkel; b) Barack Obama; c) Nicolas Sarkozy d) David Cameron Those who marked Nicolas Sarkozy heard from that point on a dull beep tone (1000 Hz). We screened out those who did not mark the correct answer. [Participants in the without noise condition] Participants in this condition did not see Screens 4 and 5 and were not exposed to the noise throughout the experiment. Screen 6: Choosing the initial investment portfolio (all participants) Assuming you had the opportunity to choose your optimal allocation between stocks and bonds, what would be your choice (given that the proportions of stocks and bonds have to sum to 100%)? Screen 7: Choosing the updated investment portfolio [Participants in the increase in market uncertainty condition] You chose an investment allocation consisting of XX% bonds and XX% stocks [each participant saw the allocation s/he had chosen in the previous screen]. Now, assume that new information (that was not known before) has become available and triggered an increase in uncertainty in the financial markets. Given the new information, you are now offered an opportunity to either preserve the allocation between stocks and bonds you have chosen, or

3 change it. That is, you can either choose to keep the allocation of the portfolio as is, or choose a new allocation of bonds and stocks. In any case please report your choice: [Participants in the decrease in market uncertainty condition] You chose an investment allocation consisting of XX% bonds and XX% stocks [each participant saw the allocation s/he had chosen in the previous screen]. Now, assume that new information (that was not known before) has become available and triggered a decrease in uncertainty in the financial markets. Given the new information, you are now offered an opportunity to either preserve the allocation between stocks and bonds you have chosen, or change it. That is, you can either choose to keep the allocation of the portfolio as is, or choose a new allocation of bonds and stocks. In any case please report your choice: Screen 8: Additional items [Participants in the with noise condition] At this stage, we stopped the noise and told participants: You will no longer hear the noise. Please rate your agreement with the following statements: 1 strongly disagree strongly agree I found the noise to be pleasant. I completed the survey patiently. [Participants in the without noise condition] 1 strongly disagree strongly agree Please rate your agreement with the following statement: I completed the survey patiently. 1 STUDY 2 Design All participants were exposed to the same stimuli that we describe below. Stimuli In what follows we present the screens that participants saw during the experiment. For clarity of presentation the text below includes a title for each screen; in the experiments participants did not see these titles. 1 A t-test analysis indicates that, on average, participants who were exposed to the noise found it to be significantly unpleasant (M = 1.42 vs. M mid-scale = 4; t (107) = , p <.001). We also obtain that, on average, participants who were exposed to the noise reported completing the task less patiently compared with those who were not exposed to the noise (M = 5.11 vs. M = 5.75, respectively, t (240) = -2.96, p <.01). We do not report these tests in the paper, since we present additional checks for the noise manipulation in Web Appendix D.

4 Screen 1: General instructions You are about to take part in a survey that explores consumers attitudes and perceptions toward investment decisions. Please read the instructions carefully and respond to the questions that follow. There are no right or wrong answers just respond to the best of your ability. The survey is anonymous and conducted for research purposes only. We would like to thank you in advance for your cooperation. Screen 2: Compensation We assigned to you an initial amount of five dollars. At the end of the experiment the amount you will receive will be the outcome of the investment decisions you will make during this survey. If, for example, you make a 10% profit, you will receive $5.50, whereas if you lose 10%, you will receive $4.50. Screen 3: The risk-aversion items Please rate how much you agree with the following statements: 1 strongly disagree strongly agree I would rather be safe than sorry. I want to be sure before I purchase anything. I avoid risky things. Screen 4: The NFCC items. Please rate how much you agree with the following statements: 1 strongly disagree strongly agree I don t like situations that are uncertain. I dislike questions which could be answered in many different ways. I find that a well ordered life with regular hours suits my temperament. I feel uncomfortable when I don t understand the reason why an event occurred in my life. I feel irritated when one person disagrees with what everyone else in a group believes. I don t like to go into a situation without knowing what I can expect from it. When I have made a decision, I feel relieved. When I am confronted with a problem, I m dying to reach a solution very quickly. I would quickly become impatient and irritated if I would not find a solution to a problem immediately. I don t like to be with people who are capable of unexpected actions. I dislike it when a person s statement could mean many different things. I find that establishing a consistent routine enables me to enjoy life more. I enjoy having a clear and structured mode of life. I do not usually consult many different opinions before forming my own view. I dislike unpredictable situations. Screen 5: Various explanations The terms risk and reward are strongly related, and when one increases so does the other. That is, when it comes to financial products, if a product has a relatively high risk/reward level, it means that by investing in this product one can either make a lot of

5 money or lose a lot of money. In contrast, investing in a product with a relatively low risk/reward level can result in either modest gains or modest losses. In this survey, we focus on financial products (portfolios) that consist of bonds (which have low risk/reward levels) and stocks (which have high risk/reward levels). Thus, for example, a portfolio that consists of 10% bonds and 90% stocks is considered as a financial product with a relatively high risk/reward level. In addition, when comparing two portfolios each consisting of stocks and bonds the portfolio with the larger portion of bonds is considered safer. Screen 6: Choosing the initial investment portfolio Please allocate your five dollars between stocks and bonds (given that the proportions of stocks and bonds have to sum to 100%)? [Unlike the case in all our other studies, this was the end of the first part of the experiment. After more than a day, we went back to the participants.] Screen 7: Choosing the updated investment portfolio The other day you chose an investment allocation consisting of XX% bonds and XX% stocks [each participant saw the allocation s/he had chosen in the first part of the experiment]. Now, assume that new information (that was not known before) has become available and triggered an increase in uncertainty in the financial markets. Given the new information, you are now offered an opportunity to either preserve the allocation between stocks and bonds you have chosen, or change it. That is, you can either choose to keep the allocation of the portfolio as is, or choose a new allocation of bonds and stocks. In any case please report your choice: Screen 8: Paying the participants (we determined arbitrarily that bonds had gone up by 2% and stocks had gone down by 5%) Since your initial investment decision, bonds have gone up by 2% and stocks have gone down by 5%. As a result, you are entitled to XX dollars [each participant was paid in accordance with his/her updated allocation between stocks and bonds]. STUDY 3 Design We randomly assigned each participant to one of four conditions in a 2 (time limit: with or without) 2 (type of initial allocation: internal or external) experimental design. Stimuli In what follows we present the screens that participants saw during the experiment. For clarity of presentation the text below includes a title for each screen; in the experiments participants did not see these titles.

6 Screen 1: General instructions You are about to take part in a survey that explores consumers attitudes and perceptions toward investment decisions. Please read the instructions carefully and respond to the questions that follow. There are no right or wrong answers just respond to the best of your ability. The survey is anonymous and conducted for research purposes only. We would like to thank you in advance for your cooperation. Screen 2: The risk-aversion items Please rate how much you agree with the following statements: 1 strongly disagree strongly agree I would rather be safe than sorry. I want to be sure before I purchase anything. I avoid risky things. Screen 3: Various explanations The terms risk and reward are strongly related, and when one increases so does the other. That is, when it comes to financial products, if a product has a relatively high risk/reward level, it means that by investing in this product one can either make a lot of money or lose a lot of money. In contrast, investing in a product with a relatively low risk/reward level can result in either modest gains or modest losses. In this survey, we focus on financial products (portfolios) that consist of bonds (which have low risk/reward levels) and stocks (which have high risk/reward levels). Thus, for example, a portfolio that consists of 10% bonds and 90% stocks is considered as a financial product with a relatively high risk/reward level. In addition, when comparing two portfolios each consisting of stocks and bonds the portfolio with the larger portion of bonds is considered safer. The time-limit manipulation [Participants in the with time limit condition] Screen 4: Time limit It is not necessary to spend a great deal of time pondering the questions you are about to see. In fact, please answer the questions as quickly as possible (while still providing an accurate answer) without taking a lot of time to ponder each one simply provide your initial, honest opinion. In the next screens, when response latencies exceeded five seconds, participants in the withtime-limit condition were presented with the phrase, Please respond faster. [Participants in the without time limit condition] Participants in this condition did not see Screen 4 and were not exposed to the reminder Please respond faster in the next screens.

7 Screen 5: Choosing the initial investment portfolio [Participants in the internal initial allocation condition] Assuming you had the opportunity to choose your optimal allocation between stocks and bonds, what would be your choice (given that the proportions of stocks and bonds have to sum to 100%)? [Participants in the external initial allocation condition] Imagine that unexpectedly you inherited from a distant relative a portfolio, which consists of 60% bonds and 40% stocks. Screen 6: Choosing the updated investment portfolio [Participants in the internal initial allocation condition] You chose an investment allocation consisting of XX% bonds and XX% stocks [each participant saw the allocation s/he had chosen in the previous screen]. Now, assume that new information (that was not known before) has become available and triggered an increase in uncertainty in the financial markets. Given the new information, you are now offered an opportunity to either preserve the allocation between stocks and bonds you have chosen, or change it. That is, you can either choose to keep the allocation of the portfolio as is, or choose a new allocation of bonds and stocks. In any case please report your choice: [Participants in the external initial allocation condition] You inherited an investment portfolio consisting of 60% bonds and 40% stocks. Now, assume that new information (that was not known before) has become available and triggered an increase in uncertainty in the financial markets. Given the new information, you are now offered an opportunity to either preserve the allocation between stocks and bonds you inherited, or change it. That is, you can either choose to keep the allocation of the portfolio as is, or choose a new allocation of bonds and stocks. In any case please report your choice: STUDY 4 Design We randomly assigned each participant to one of two conditions: no clear guidance and clear guidance. Stimuli In what follows we present the screens that participants saw during the experiment. For clarity of presentation the text below includes a title for each screen; in the experiments participants did not see these titles.

8 Screen 1: General instructions You are about to take part in a survey which explores consumers attitudes and perceptions toward investment decisions. Please read the instructions carefully and respond to the questions that follow. There are no right or wrong answers just respond to the best of your ability. The survey is anonymous and conducted for research purposes only. We would like to thank you in advance for your cooperation. Screen 2: The risk-aversion items Please rate how much you agree with the following statements: 1 strongly disagree strongly agree I would rather be safe than sorry. I want to be sure before I purchase anything. I avoid risky things. Screen 3: Various explanations The terms risk and reward are strongly related, and when one increases so does the other. That is, when it comes to financial products, if a product has a relatively high risk/reward level, it means that by investing in this product one can either make a lot of money or lose a lot of money. In contrast, investing in a product with a relatively low risk/reward level can result in either modest gains or modest losses. In this survey, we focus on financial products (portfolios) which consist of bonds (that have low risk/reward levels) and stocks (that have high risk/reward levels). Thus, for example, a portfolio that consists of 10% bonds and 90% stocks is considered as a financial product with a relatively high risk/reward level. In addition, when comparing two portfolios each consisting of stocks and bonds the portfolio with the larger portion of bonds is considered safer. The noise manipulation [All participants were exposed to the noise] Screen 4: The noise Prior studies indicate that consumers usually engage in investment decisions in a noisy environment. Therefore, we have included noise within this research, which would appear while you make your financial decisions. For that purpose, please make sure that your speakers work and keep them at their maximal level from now on until the end of the experiment. Screen 5: Noise validation Participants heard a voice saying the name Nicolas Sarkozy. Immediately afterwards they were told: Please report the name that you just heard: a) Angela Merkel; b) Barack Obama; c) Nicolas Sarkozy d) David Cameron Those who marked Nicolas Sarkozy heard from that point on a dull beep tone (1000 Hz). We screened out those who did not mark the correct answer.

9 Screen 6: Choosing the initial investment portfolio Assuming you had the opportunity to choose your optimal allocation between stocks and bonds, what would be your choice (given that the proportions of stocks and bonds have to sum to 100%)? Screen 7: Choosing the updated investment portfolio [Participants in the no clear guidance condition] You chose an investment allocation consisting of XX% bonds and XX% stocks [each participant saw the allocation s/he had chosen in the previous screen]. Now, assume that new information (that was not known before) has become available and triggered an increase in uncertainty in the financial markets. Given the new information, you are now offered an opportunity to either preserve the allocation between stocks and bonds you have chosen, or change it. That is, you can either choose to keep the allocation of the portfolio as is, or choose a new allocation of bonds and stocks. In any case please report your choice: [Participants in the clear guidance condition] You chose an investment allocation consisting of XX% bonds and XX% stocks [each participant saw the allocation s/he had chosen in the previous screen]. Now, assume that new information (that was not known before) has become available and triggered an increase in uncertainty in the financial markets. As a result, investment experts recommend increasing the portion of bonds in investment portfolios. Given the new information, you are now offered an opportunity to either preserve the allocation between stocks and bonds you have chosen, or change it. That is, you can either choose to keep the allocation of the portfolio as is, or choose a new allocation of bonds and stocks. In any case please report your choice:

10 Web Appendix B PRE-TEST OF THE EFFECT OF THE NOISE AND TIME-LIMIT MANIPULATIONS ON NFCC (STUDIES 1 AND 3) The pre-test was conducted among 144 participants, who were each randomly assigned to one of three conditions: (a) with noise, using a noise manipulation similar to the one used in Study 1; (b) with time limit, using a time manipulation similar to the one used in Study 3; and (c) without noise and without time limit (control condition). To confirm that the noise and the time limit caused high levels of NFCC, we adopted (with slight changes to reflect the tasks in our experiments) the five items that Kardes et al. (2004) used to assess NFCC: I was worried about choosing the updated portfolio allocation suboptimally (reverse scored [R]); I continued to think about the updated portfolio allocation I chose, even after I provided my answer (R); I struggled with the task of choosing the updated portfolio allocation (R); The answer of how to update the portfolio allocation came to me quickly ; I disliked the task of choosing the updated portfolio allocation because it was confusing (R). Participants were asked to report on a 7-point scale their extent of agreement with each of the above items (1 = extremely disagree, 7 = extremely agree), and for each one of the participants, we calculated the average score of agreement with the above items. The average score of agreement of participants who were exposed to the noise was on average significantly higher than that of participants in the control condition (M = 4.66 vs. M = 4.12, respectively, t (95) = 2.70, p <.01). Likewise, the average score of agreement of participants who were exposed to the time limit was on average significantly higher than that of participants in the control condition (M = 4.57 vs. M = 4.12, respectively, t (94) = 2.19, p <.05). Thus, we conclude that both the noise manipulation and the time-limit manipulation cause high levels of NFCC. PRE-TEST OF AWARENESS OF THE TIME LIMIT (STUDY 3) The pre-test was conducted among 111 participants, who were each randomly assigned to one of two conditions with time limit or without time limit using a time manipulation similar to the one used in Study 3. Participants were asked to report on a 7- point scale how quickly they had completed the survey (1 = very slowly, 7 = very fast). Participants who were exposed to the time limit reported, on average, completing the survey significantly faster compared with participants who were not exposed to the time limit (M = 4.41 vs. M = 1.98, respectively, t (109) = 8.15, p <.001). PRE-TEST OF OWNERSHIP PERCEPTIONS (STUDY 3) The pre-test was conducted among 42 participants, who were each randomly assigned to one of four conditions in a 2 (time limit: with or without, using a time manipulation similar to the one used in Study 3) 2 (type of initial allocation: internal or external) experimental design. The results indicate that participants in the external-initial-

11 allocation condition did not differ from participants in the internal-initial-allocation condition in terms of their ownership perception (M external = 4.90 vs. M internal = 4.86, respectively, t (41) =.22, p >.1). In addition, we told pre-test participants that new information about an increase in market uncertainty had become available and asked them to report the likelihood that they would rely on this information, if they had the option to remain with their previously chosen/inherited investment allocations or to change them. We obtained a significant interaction in terms of the likelihood of relying on the new information (F (1,40) = 3.97, p =.05). Among participants in the with-time-limit condition, those in the externalinitial-allocation condition reported a higher likelihood to rely on the new information than did those in the internal-initial-allocation condition (M external = 5.75 vs. M internal = 5.05, respectively, t (19) = 2.01, p =.06). In contrast, among participants in the without-time-limit condition, individuals did not differ in the extent to which they expected to rely on the new information (M external = 4.77 vs. M internal = 4.86, respectively, t (21) = -.44, p >.1).

12 Web Appendix C RESULTS OF STUDY 1 WITHOUT AND WITH A THREE-WAY INTERACTION In Table W1 below, we present the results of Study 1 without and with the three-way interaction of risk aversion, NFCC, and the uncertainty change indicator. We can see that the coefficient of the three-way interaction is not statistically significant (β = -1.14; t (233) = -.49, p >.1). This implies that, in the model with the three-way interaction, the interaction of any two of the variables risk aversion, NFCC, and the uncertainty change indicator does not significantly vary across different levels of the third variable. This is in line with the version of the model that does not include the three-way interaction at all. Table W1 also suggests that the coefficients of risk aversion, NFCC, and the uncertainty change indicator and their respective significance levels, as well as the intercept and the coefficient of the initial portion of bonds (and their respective significance levels) are very similar between the two models (the coefficients of the two-way interaction terms represent simple interactions in the model with the three-way interaction, as opposed to interactions in the model without the three-way interaction, and therefore it is less interesting to compare between them). In the model with the three-way interaction, an exploration of the effect of risk aversion on the change in portion of bonds in all four possible conditions of NFCC and the uncertainty change indicator generated the following results. In line with H1, when participants were told about an increase in market uncertainty, the slope of risk aversion was not significant under the high NFCC condition (β = -.21; t (233) = -.15, p >.1). In contrast, the slope of risk aversion was significant and positive under the low NFCC condition (β = 1.87; t (233) = 2.02, p <.05). Likewise, when participants were told about a decrease in market uncertainty, we also obtained results that were consistent with H1. The slope of risk aversion was not significant under the high NFCC condition (β = -.63; t (233) = -.53, p >.1). In contrast, the slope of risk aversion was significant and positive under the low NFCC condition (β = 2.59; t (233) = 2.46, p <.05). As Table W2 below shows, these results are virtually the same as those obtained when the model without the three-way interaction is employed. Hence, whether the three-way interaction is included or not and for both an increase and a decrease in uncertainty, individuals high in NFCC do not update their portfolios in a manner that corresponds to their risk-aversion levels, whereas individuals low in NFCC do.

13 TABLE W1 Regression Results Without and With a Three-Way Interaction (Study 1) Change in portion of bonds Without a three-way interaction Intercept *** (2.07) Risk aversion 2.06 ** (.85) NFCC * (1.86) Uncertainty change indicator *** (1.74) Initial portion of bonds -.14 *** (.03) Risk aversion NFCC ** (1.16) With a three-way interaction *** (2.09) 1.87 ** (.93) * (1.86) *** (1.75) -.14 *** (.03) (1.71) Risk aversion Uncertainty change indicator NFCC Uncertainty change indicator Risk aversion NFCC Uncertainty change indicator.31 (1.12) 2.61 (2.60).72 (1.40) 2.66 (2.61) (2.34) R Overall F 7.30 *** 6.40 *** df 7, 234 8, 233 Notes: In the two regressions reported in the table, the change in portion of bonds is the dependent variable. As for the predictors, risk aversion is a continuous variable that takes values from 1 to 7 and was mean-centered; NFCC is a dummy variable (1 high NFCC, 0 low NFCC); the uncertainty change indicator is a dummy variable (1 decrease in uncertainty, 0 increase in uncertainty); and the initial portion of bonds is a continuous variable that takes values from 0 to 100. Entries in the table represent unstandardized coefficients. Standard errors are reported in parentheses. * p <.1, ** p <.05, *** p <.01.

14 TABLE W2 The Effect of Risk Aversion on the Change in Portion of Bonds in the Four Possible Conditions of NFCC and the Uncertainty Change Indicator (Study 1) Without a three-way interaction High NFCC, Increase in market uncertainty With a three-way interaction Low NFCC, Increase in market uncertainty 2.06 ** 1.87 ** High NFCC, Decrease in market uncertainty Low NFCC, Decrease in market uncertainty 2.36 ** 2.59 ** Notes: * p <.1, ** p <.05

15 Web Appendix D TABLE W3 Floodlight Analysis Change in Portion of Bonds (Study 2) NFCC Effect SE LLCI ULCI t (45) p Notes: The table presents the effect of risk aversion on the change in portion of bonds for specific values of NFCC. The first column shows the NFCC levels. The second column shows the effect of risk aversion on the change in the portion of bonds. The third column shows the standard errors. The fourth column shows the low level of the confidence interval and the fifth column shows the high level of the confidence interval. The sixth column shows the values of the t-test and the seventh column shows their respective p-values. The Johnson-Neyman points are obtained at NFCC = 3.57 and NFCC = 6.65 (p =.05).