Abstract
We survey 231 undergraduate students to analyze how retail investors’ characteristics influence their decision to use a robo-advisor. Characteristics such as the willingness to take risk, extraversion, and optimism are significant in univariate tests but the willingness to take risk and the internal locus of control are significant in multivariate settings. Participants who use the robo-advisor invest more and are more likely to also invest on their own in both stocks and bonds. We also find statistically significant differences between participants who exclusively use the robo-advisor for investments in stocks and bonds and participants who use the robo-advisor and invest some money in stocks and bonds on their own.
Similar content being viewed by others
Avoid common mistakes on your manuscript.
1 Introduction
In this study, we analyze how retail investors’ individual characteristics influence their decision to invest through a robo-advisor. The most popular robo-advisors for investors propose an investment portfolio that typically consists of stock and bond index funds (Horn and Oehler 2020).Footnote 1
One of the potential advantages of robo-advisors over more traditional investment advice is their 24/7 availability. This increased flexibility regarding time engenders lower consultation costs that makes robo-advisors interesting to low net-wealth individuals (D’Acunto and Rossi 2020; Horn et al. 2020; Rossi and Utkus 2020b). Moreover, robo-advisors can help investors who might not have private information build well-diversified portfolios with little behavioral biases that are typical of human advisors (see Linnainmaa et al. 2020, Foerster et al. 2017, Inderst and Ottaviani 2009). Robo-advisors can mitigate biases because they follow predefined rules (see D’Acunto et al. 2019).
The sum of all robo-advisors’ assets under management in the year 2020 was only 1.07 trillion USD.Footnote 2 This is a small amount when compared to the assets under management of the largest asset managers such as BlackRock (8.7 trillion USD), Vanguard (7.2 trillion USD), and State Street (3.5 trillion USD). Furthermore, at least 15 asset management firms had more assets under management individually than all robo-advisors combined.Footnote 3 Beyond the recency of robo-advising (Fisch et al. 2019), explanations for the relatively low assets under management and low inflows remain unclear. Hohenberger et al. (2019) find that retail investors who associate the use of a robo-advisor with joy are more likely to use it. Rossi and Utkus (2020a) show that one main reason to avoid robo-advisors is the lack of possibilities for personal communication and a lack of trust. Specifically, investors who trust their human financial advisor are very unlikely to use a robo-advisor.
We conducted a questionnaire-based survey of 231 undergraduate business students at a German medium-sized university in November 2019 to analyze the influence of individual characteristics on the decision to use robo-advising. Our reasoning is that there is a higher probability that young adults will use a robo-advisor than older adults. Young adults are less subject to algorithm aversion (Rossi and Utkus 2020a). They have no relationship or a less established and less trust-based relationship with a human advisor. Moreover, young adults are a very important target group for robo-advisors. A survey of young adults is thus also useful for the practice of robo-advisory.
In the survey we use validated questionnaire items to measure the characteristics that potentially influence investment decisions such as gender, financial knowledge, statistical knowledge, risk attitude, cognitive reflection, Big Five personality factors, locus of control, PANAS, and trust in institutions and persons other than a human financial advisor (see the overviews in, e.g., Oehler et al. 2018a, Oehler et al. 2020, Oehler et al. 2021, and Oehler and Wendt 2018). The literature shows that the majority of these characteristics also influence technology acceptance (Behrenbruch et al. 2013; see also Rossi and Utkus 2020a on the influence of algorithm aversion). Usually, more conscientious, less neurotic people with more technology experience, confidence, and trust in technology are more likely to use a new technology (Barnett et al. 2015; Lee et al. 2017).
Our participants receive a fixed participation fee that mitigates any incentive to gamble. Further, we do not provide a concrete investment suggestion in order to sidestep participants’ confirmation bias (see Alemanni et al. 2020). This approach also eliminates the potential for a hypothetical bias as described in Cummings et al. (1997). However, the participants’ answers are not financially binding in any case. The potential downside of using a survey approach is its external validity. Our findings only apply to likely users of robo-advisors and not to the entire population. But Oehler et al. (2018b) perform a survey of German undergraduate students on their financial portfolio choices, and Oehler et al. (2018c) use a comparable subset from the representative survey data of the German central bank. The results of both suggest that our findings have broader relevance.
Even young adults might perceive robo-advisors as less trustworthy than a human advisor. We hypothesize that the young adults in our survey with individual characteristics such as higher levels of extraversion, openness to new experiences, and optimism are more likely to be less risk averse and to use a robo-advisor. Presumably, not everyone among the individuals who give the robo-advisor a chance will go all-in and only use it for investing. Therefore, we allow participants to invest only a share or nothing at all of their investable wealth and to invest in risk-free to relatively risky asset classes that match their risk attitude.
Based on tests of equality, we find that participants who are willing to use the robo-advisor are more willing to take financial risks, are more extraverted, are more optimistic, and are less pessimistic than participants who are not willing to use the robo-advisor.Footnote 4 These are key variables used in the literature and our results are in line with its findings (Dohmen et al. 2011, Oehler et al. 2018a, Puri and Robinson 2007; see also Kaustia et al. 2019 for an overview). When integrating all characteristics into a logit regression, the participants’ willingness to take financial risks remains statistically significant but extraversion, optimism, and pessimism do not. Instead, participants with a lower internal locus of control are significantly more likely to use the robo-advisor (although the locus of control is not significant in the univariate analysis). Moreover, participants who are willing to use the robo-advisor invest a larger amount in stocks and bonds than participants who are not willing to use the robo-advisor, which is in line with the idea that those willing to use the robo-advisor are more willing to take financial risks.
We also find statistically significant differences between participants who exclusively use the robo-advisor for investments in stocks and bonds and participants who use the robo-advisor and invest some additional money on their own. The latter group invests more in stocks and bonds in total. However, the amount that they invest via the robo-advisor is less than the amounts invested by participants who exclusively use the robo-advisor. The participants who use the robo-advisor and invest in stocks and bonds on their own have higher levels of perceived financial knowledge and experience than participants who exclusively use the robo-advisor. These results are further support for the findings in the literature that more experienced and literate investors are more likely to invest higher amounts and shares of their wealth in risky assets and establish better diversified portfolios (van Rooij et al. 2011; Graham et al. 2009; von Gaudecker 2015).
However, we observe an inverse relation between the perceived statistical knowledge and the likelihood to invest additionally in stocks and bonds when using a robo-advisor. A multinomial regression supports our univariate findings for the effect of statistical knowledge and shows that participants who use the robo-advisor and invest in stocks and bonds on their own are less optimistic but more willing to take financial risks. Further, they have a lower internal locus of control than participants who exclusively invest with the robo-advisor. However, the results of our analysis do not indicate that any of the characteristics actually influence the difference in the amount that these two groups of participants invest with the robo-advisor.
Our findings provide implications for researchers and for providers of investment services and financial advice. For example, research on retail investors’ investment decisions and on financial intermediation should consider that not only individuals’ risk attitude in the financial domain but also that other individual characteristics might significantly influence decisions on which asset classes to invest in and at the same time which investment advisor to use.
The paper is structured as follows: We describe our survey design and our empirical methods in the next section. In Sect. 3, we present the empirical results. We discuss our findings and conclude in Sect. 4.
2 Design and Methodology
2.1 Survey Design
The participants received information gathered from the homepage of a real robo-advisor. This information covered an explanation of the robo-advisor’s basic functionality, past returns, costs, risks, and advantages.Footnote 5 Thereafter, the participants were given 1,000 Euros to invest. Other studies have used investable amounts that range from a few Euros (see Alemanni et al. 2020) to 10,000 Euros (see Nosic and Weber 2010). We chose an amount that presumably best reflected what students usually would handle. Furthermore, Oehler et al. (2018c) measured students’ risk attitude in a theoretical lottery design with different Euro amounts and showed that students’ answers reflected a reasonable degree of risk aversion; that is, students did not act as if they perceived the investment task as gambling with play money when the theoretical amount at stake was 500 Euros or more. Therefore, our choice seemed plausible given students’ lifestyles and earning compacity.
The investment task was structured as follows: First, participants indicated whether they were willing to use the robo-advisor to invest (a share of) the 1,000 Euros or not. If participants were willing to use the robo-advisor, they stated the amount of money that they would invest out of the 1,000 Euros with the robo-advisor. In addition, they indicated the allocation to stocks and bonds (on a scale of 10%/90%, 20%/80%, and so on) that the robo-advisor should follow for the portfolio. One condition was that our participants could only invest up to the 1,000 Euros because we did not allow them to borrow more money. For the amount that they did not invest with the robo-advisor, participants were asked to state the amounts that they had invested in liquidity provisions, fixed-term deposits, bond ETFs, stock ETFs, real estate funds, bond mutual funds, stock mutual fonds, individual bonds and stocks, and/or another asset class. Hence, participants could design portfolios on a continuum ranging from almost risk-free in liquid assets to fairly risky (with or without the robo-advisor) in individual stocks or even options or certificates.
Additionally, the questionnaire included items to capture the individual characteristics of the participants. An overview of the variables used in the empirical analysis, how they are measured, and their place in the literature is presented in Table 1.
Table 2 displays the descriptive statistics. Although the students in our sample might not be a representative cross-section of the population, they should represent a sample of potential users of robo-investors as they (1) should be aware of the existence of robo-advisors; (2) should show relatively good financial capability (see Oehler et al. 2018b, see also Hohenberger et al. 2019 for the importance of perceived investment capabilities); (3) likely do not have a well-established and trust-based relationship with a financial advisor; and (4) are less likely to be algorithm averse (Rossi and Utkus 2020a).
Participants received the questionnaire in the German language; see the Internet Appendix for an English translation of the questionnaire.
2.2 Empirical Analysis
We first provide univariate tests of equality (t-tests) to analyze the differences in the characteristics and asset allocation of participants who are willing to use the robo-advisor (robo-users) versus those who are not (robo-avoiders).
Then, we run a logit regression to analyze the determinants of the decision to use a robo-advisor with the following model:
The dependent variable equals one if the participant is willing to use the robo-advisor and zero otherwise. Table 1 has the descriptions of the explanatory variables and vectors of variablesFootnote 6 that enter the regression analysis.
Robo-users do not have to use the robo-advisor exclusively but are also allowed to invest any remaining money on their own. Therefore, we also perform a series of t-tests between two subsamples within the group of robo-users. Specifically, we analyze the differences in asset allocations and individual characteristics between participants who only used the robo-advisor (robo-only-investors) versus participants who used the robo-advisor for some of the money while investing the remaining amount on their own (robo-plus-investors). Furthermore, we perform a multinomial logistic regression using the following model:
ROBO_USAGEi covers three outcomes: no use of the robo-advisor, that is, robo-avoiders; only use the robo-advisor for investing, that is, robo-only-investors; and to use the robo-advisor for some of the money in combination with investing the remaining amount on their own, that is, robo-plus-investors. The first outcome serves as the reference point in our analysis. Using Eq. (2) with SOLELY_USING_ROBOi as the dependent variable, we also apply a logit regression analysis to the subsample of robo-users to discover the characteristics that drive the decision to use the robo-advisor solely or in combination with their own investments in stocks and bonds.
To be able to uncover the determinants of the amount that participants would invest with the robo-advisor and the stock ratio they would choose for this investment, we need to account for the fact that not all participants use the robo-advisor. Therefore, we applied two approaches. First, we used a Tobit regression approach on the full sample. Second, we used an OLS regression to approximate a truncated regression only on the subsample of robo-users, that is, only including responses from participants that actually invest an amount using the robo-advisor that is different from zero.Footnote 7 The underlying model is as follows:
For the dependent variable, ROBO_INVESTMENTi, we use either the amount of money that participant i is willing to invest with the robo-advisor, INVEST_WITH_ROBO_AMOUNTi, or the stock ratio that he or she chooses for the investment with the robo-advisor, INVEST_WITH_ROBO_STOCK_RATIO i.
3 Results
3.1 Characteristics and Asset Allocation of Robo-users vs. Robo-avoiders
We report the mean values of the characteristics of the participants who are willing to invest with the robo-advisor (robo-users) and of the participants who are not willing to invest with the robo-advisor (robo-avoiders) in Table 3. Of the participants, 53% are willing to invest on average 524 Euros (out of 1,000 Euros) with the robo-advisor. This usage shows that our sample contains a high percentage of participants who do not preclude using a robo-advisor a priori. The percentage supports our approach of using young adults as participants. The results of the t-tests show that robo-users show a significantly higher willingness to take financial risks (4.89 vs. 4.28), a higher degree of extraversion (3.49 vs. 3.23), more optimism (11.44 vs. 10.79), and less pessimism (7.10 vs. 7.67) than robo-avoiders. This difference is statistically significant at the 5% level for each of these characteristics.
The results of the logit regression in Table 4 show that participants with a higher willingness to take financial risks, that is, a lower degree of risk aversion, are more likely to use a robo-advisor. The corresponding coefficient is 0.26, and the influence is statistically significant at the 1% level. This finding is in line with the t-test results. In contrast to the t-tests, however, the results of the regression analysis do not show a significant influence from the participants’ degree of extraversion, optimism, and pessimism. The latter effect is not driven by multicollinearity between risk attitude, extraversion, optimism, and pessimism. The untabulated results of a correlation analysis indicate no significant correlation between the risk attitude and the remaining three factors (however, the latter ones are correlated with each other with statistical significance at the 1‰ level). Moreover, with statistical significance at the 1% level, the coefficient of -1.03 for INTLOCUS indicates higher chances that participants with a lower internal locus of control use a robo-advisor, that is, that people who attribute outcomes of events (e.g., investment outcomes) less to their own control are more likely to transfer the portfolio management to the robo-advisor. The results of the logit regression indicate a good model fit as 68.4% of the estimates are correct. A post hoc test for statistical power shows high power for the regression and that the results are not driven by multicollinearity.Footnote 8
The asset allocations of robo-users and of robo-avoiders as presented in Table 5 again show robo-avoiders’ lower willingness to take financial risk. All differences in the asset allocations are statistically significant at the 1‰ level. Robo-users on average invest 773 Euros in stocks and bonds (including the 524 Euros they invest using the robo-advisor). Robo-avoiders only invest 387 Euros in stocks and bonds. Since robo-avoiders still invest some money in stocks and bonds, not using the robo-advisor may indicate an active choice against the robo-advisor and not a choice against risky investments per se.
As an interim conclusion, we can state that participants with a lower degree of risk aversion are more likely to use a robo-advisor. Our findings so far indicate that investors with a higher willingness to take financial risk have the opportunity to benefit from delegating investing to a robo-advisor as it provides them with the chance to establish a well-diversified portfolio and to potentially mitigate the negative effect of their own behavioral biases.
3.2 Univariate Differences Between Robo-only- and Robo-plus-investors
Table 6 presents the mean amounts of money invested in the different asset classes by robo-only-investors and robo-plus-investors. Among the robo-users, 40 participants (33%) only invest with the robo-advisor (robo-only-investors) while 82 (67%) of the participants combine investing with the robo-advisor and investing in individual stocks, bonds, and index and mutual funds (robo-plus-investors). The t-tests on the asset allocations of these two groups of participants show that robo-only-investors invest a significantly smaller amount in stocks and bonds than robo-plus-investors (632 vs. 843 Euros). However, the robo-plus-investors invest a significantly lower amount via the robo-advisor (471 Euros). Furthermore, robo-only-investors and robo-plus-investors choose a significantly different stock ratio for their investments with a statistical significance at the 2‰ level. While robo-only-investors allocate 51% of their investment amount in stocks, robo-plus-investors choose a stock ratio of 63% for investment with the robo-advisor. For the remaining investments without the robo-advisor, robo-plus-investors choose an even higher stock ratio of 74%. A paired-sample t-test between the stock ratio of investments with and without the robo-advisor of robo-plus-investors shows a statistically significant difference at the 1% level.
Besides their asset allocation, robo-only-investors and robo-plus-investors also differ in their characteristics. The mean values of the characteristics of robo-only-investors and robo-plus-investors and the corresponding t-test results are presented in Table 7. Robo-plus-investors have better financial knowledge (4.20 vs. 3.03, significant at the 1% level) and the greater financial experience (0.24 vs. 0.05, significant at the 1‰ level). On the other hand, robo-only-investors have more knowledge of statistics (4.58 vs. 4.21) with statistical significance at the 5% level. Since studies have shown that statistical knowledge, investment experience, and financial knowledge are positively related to investment performance (see Campbell et al. 2014; Corgnet et al. 2018; Graham et al. (2009), von Gaudecker (2015), Nicolosi et al. 2009, Seru et al. 2010), our results do not indicate that robo-only-investors have better investment skills than robo-plus-investors or vice versa.
3.3 Multivariate Results
We provide the results of multinomial logistic regressions to identify the characteristics that drive the decision to become a robo-only-investor or a robo-plus-investor relative to a robo-avoider in Table 8. The results show a positive exponentiated coefficient of 2.77 for the effect of statistical knowledge on the likelihood of being a robo-only-investor at the 5‰ level of statistical significance (1.21 and not statistically significant for robo-plus-investors). The positive relationship between the degree of optimism and being a robo-only-investor is weak with an exponentiated coefficient of 1.45 but still significant at the 5% level (0.93 and not statistically significant for robo-plus-investors). Robo-plus-investors, on the other hand, show a higher willingness to take financial risks with an exponentiated coefficient of 1.38 with statistical significance at the 5‰ level and a lower internal locus of control with a coefficient of 0.36Footnote 9 that is statistically significant at the 1% level. In contrast, robo-only-investors are not associated with a statistically significant higher willingness to take financial risks (exponentiated coefficient of 1.27) or a lower internal locus of control (exponentiated coefficient of 0.30). The regression analysis has strong statistical power and is not driven by multicollinearity issues.
We perform a further logit regression for the subsample of robo-users with a binary dependent variable that indicates whether participants are robo-only-investors or robo-plus-investors. The detailed results are presented in Table 10 of the Appendix. Among the robo-users, the participants with better statistical knowledge and a higher degree of optimism are more likely to only use the robo-advisor. We also find that participants with a higher degree of pessimism are also more likely to be robo-only-investors than robo-plus-investors and that is statistically significant at the 1% level. However, the test for statistical power shows that the regression only has marginal statistical power and should therefore be treated with caution. Hence, it may only serve as mild support for the results in Tables 7 and 8, which is why the detailed results are only presented in the Appendix.
In the regression analysis that uses model (3), we tried to identify the determinants of the amount that participants invest when using the robo-advisor and the stock ratio that they choose when investing with the robo-advisor. We leave the results untabulated as they only marginally add to our findings.Footnote 10 When only using characteristics as independent variables in Tobit regressions on the full sample, the results support our previous findings regarding the determinants of using a robo-advisor; participants with a higher willingness to take financial risks and a lower internal locus of control invest a higher amount with the robo-advisor and choose a higher stock ratio for their investments. In particular, the coefficients for RISK_ATTITUDE and INTLOCUS in the regression with the invested amount (stock ratio) as the dependent variable are 75 and -326. (7.8 and -29.5). Moreover, participants with a higher positive affect invest a higher amount with the robo-advisor (coefficient of 117). When we add a dummy variable that indicates whether a participant is a robo-only-investor or a robo-plus-investor, the significance of participants’ willingness to take financial risks, internal locus of control, and a positive affect stay stable or even slightly increase while the explanatory power of the model with the invested amount (stock ratio) doubles to a Pseudo R2 of 0.06 (0.06). These findings are in line with those reported in Tables 4 and 8. However, the significance of these relations fully disappears when only using an OLS to approximate model (3) for the subsample of the 122 robo-users. The only relations left, with statistical significance at the 5% level and coefficients of -113 and 56, are that participants with a lower internal locus of control and higher positive affect invest a higher amount with the robo-advisor.
Assessing investors’ capacity and preferences regarding the risk and return of an investment is crucial for robo-advisors in order to propose a suitable stock ratio. When the dependent variable in the OLS on the subsample of robo-users is the stock ratio chosen for the investment through the robo-advisor, then none of the characteristics has a significant influence. Hence, the characteristics neither directly determine the amount invested with the robo-advisor nor the stock ratio chosen for this investment. This lack of an effect indicates that only using the characteristics is not sufficient to derive the risk-return-preferences and/or the risk and return of the investors regarding the different asset classes. Consequently, it is unlikely that robo-advisors would be able to provide a suitable portfolio proposal when solely using these characteristics at hand.
Overall, we conclude that the participants’ characteristics have a significant influence on the decision to (additionally) use a robo-advisor. However, further research is needed to better identify the factors that determine the amount invested with the robo-advisor or the stock ratio chosen for this investment.
4 Discussion and Conclusions
Our study focuses on younger adults as a target group for robo-advisors to analyze the influence of a broad range of individual characteristics on young retail investors’ decision to use a robo-advisor. Beyond the question of whether or not they use the robo-advisor, we also analyze the effect of the individual characteristics on the amount they are willing to invest with the robo-advisor. We use a questionnaire-based survey among 231 undergraduate business students at a German medium-sized university and a series of tests of equality and regression analyses.
Our findings support the hypothesis that less risk-averse retail investors are more likely to use a robo-advisor. We also find that the characteristics of extraversion, optimism, and pessimism are significant in univariate but not in multivariate analyses. Instead, participants with a lower internal locus of control are more likely to use the robo-advisor. In line with the idea that robo-users are less risk averse, we find that they invest a larger amount in stocks and bonds than robo-avoiders.
Thus, our findings do not provide convincing support for the idea that several widely used personality traits or specific trust components matter. While personality traits and trust affect actual investment decision-making (as shown by, e.g., Guiso et al. 2008; Oehler et al. 2018a), those effects indicate that they do not systematically affect the choice of the robo-advisor. However, the trust might also indicate that our participants do not (yet) transfer it from other contexts as measured in our study to the still relatively new robo-advisors.
When further analyzing the subsample of robo-users, we find that investors who use the robo-advisor and invest in risky assets on their own have better financial knowledge and more experience. Therefore, these investors might be either overconfident regarding their skills at picking assets or they might actually have better investment skills based on their experience (Seru et al. 2010).
Investors who solely use the robo-advisor for investments in risky assets have better statistical knowledge and are more optimistic. This finding indicates that they might be actively avoiding risky investments other than those with the robo-advisor as they can assess the limited potential for an enhancement in portfolio performance when they are already invested in a diversified portfolio with the robo-advisor (Oehler and Wanger 2020).
Our findings provide implications for researchers and practitioners alike. Researchers should be aware that not only retail investors’ risk attitude in the financial domain but also other (sometimes related) personal characteristics such as locus of control, life orientation, and statistical knowledge might significantly influence their decisions on which asset classes to invest in and which investment service provider to use. Since new, particularly purely digital, financial service providers have arisen in recent years, further research should focus on the characteristics that influence the long-term relation between retail investors and these new service providers, including the likelihood of changing them.
In addition, more analysis is necessary to determine the impact of individual characteristics on the amount of money an investor is willing to invest using the robo-advisor and on different types of portfolios beyond a simple stock–bond mix. It might be a promising path to investigate the latent factors that affect investors in the process of investment decision making (see, e.g., Yang 2013). Further research should also focus on other groups of potential users of robo-advisors and on further robo-advisors.
Providers of digital financial advice and portfolio management should consider further developing their business model to cater to the needs of investors with higher degrees of risk aversion. For instance, robo-advisors might wish to additionally offer advice on insurance products and real estate finance that are perceived as less risky by retail investors.
Change history
03 August 2022
Missing Open Access funding information has been added in the Funding Note.
Notes
Examples for robo-advisors include: Betterment (USA), MoneyFarm (UK), Nutmeg (UK), Quirion (Germany), Scalable Capital (Germany), Schwab Intelligent Portfolios (USA), Wealthfront (USA). Some robo-advisors also provide recommendations regarding individual stocks (see, e.g., D’Acunto et al. 2019). We do not analyze this type of robo-advising.
See https://www.blackrock.com/corporate/literature/annual-report/blackrock-2020-annual-report.pdf, https://about.vanguard.com/who-we-are/fast-facts/, https://annualreport.statestreet.com/Y2020/default.aspx, https://www.statista.com/statistics/322452/largest-asset-managers-worldwide-by-value-of-assets/.
Optimism and pessimism are not considered as the two poles of one scale but as two independent factors (see Glaesmer et al. 2008).
We did not pick a robo-advisor with particularly good or bad past performance. See the return and risk information that respondents received in the Internet Appendix (p. 7 of the questionnaire).
Vectors of variables and vectors of the corresponding coefficients are in bold print.
See Breen (1996) on truncated regressions.
The only variables with a variance inflation factor (VIF) higher than 10 are the pairs LN(LOTTERY_RISKY)/LN(LOTTERY_UNCERTAIN) and EXPERIENCE_STOCK_MARKET/EXPERIENCE_STOCK_MARKET_SATISFIED which is not surprising given that the latter variable is an interacted variable that includes the previous one. The VIFs for the regression models that exclude LN(LOTTERY_RISKY) and EXPERIENCE_STOCK_MARKET do not exceed 2.2. The corresponding regression coefficients are almost identical to the ones reported in Table 4. In particular, the regression coefficients for the two significant characteristics RISK_ATTITUDE and INTLOCUS do not change considerably (.26 vs. .25 and -1.03 vs. -1.03), see Appendix, Table 9 for the detailed results.
Please consider that exponentiated coefficients – reflecting odds ratios – are provided in Table 8; the non-exponentiated coefficient of INTLOCUS is -1.03.
The table is available from the authors upon request.
References
Abreau M, Mendes V (2012) Information, overconfidence and trading: Do the sources of information matter? J Econ Psychol 33(4):868–881
Alemanni B, Angelovski A, Di Cagno D, Galliera A, Linciano N, Marazzi F, Soccorso P (2020) Do investors rely on robots? Evidence from an experimental study. CONSOB Fintech Series No. 7, available at SSRN: https://ssrn.com/abstract=3697232. Accessed 10/02/2021
Barberis N, Thaler R (2003) A Survey of behavioral finance. In: Konstantinides G, Harris M, Stulz R (ed) Handbook of the Economics of Finance 1B, pp 1053–1123
Barnett T, Pearson A, Pearson R, Kellermanns F (2015) Five-factor model personality traits as predictors of perceived and actual usage of technology. Eur J Inf Syst 24(4):374–390
Behrenbruch K, Söllner M, Leimeister J, Schmidt L (2013) Understanding Diversity – The Impact of Personality on Technology Acceptance. Hum Comput Interact – INTERACT 2013: 306–313
Breen R (1996) Regression models: censored, sample-selected, or truncated data. Sage, Thousand Oaks
Camerer C, Hogarth R (1999) The effects of financial incentives in experiments: a review and capital-labor production framework. J Risk Uncertain 19(1–3):7–42
Campbell J, Ramadorai T, Ranish B (2014) Getting Better or Feeling Better? How Equity Investors Respond to Investment Experience. NBER Working Paper No. 20000
Corgnet B, Desantis M, Porter D (2018) What Makes a Good Trader? On the Role of Intuition and Reflection on Trader Performance. J Finance 73(3):1113–1137
Cummings RG, Elliott S, Harrison G, Murphy J (1997) Are hypothetical referenda incentive compatible? J Political Economy 105(3):609–621
D’Acunto F, Rossi A (2020) Robo-Advising. CESifo Working Paper No. 8225, available at SSRN: https://ssrn.com/abstract=3578259. Accessed 10/02/2021
D’Acunto F, Prabhala N, Rossi A (2019) The Promises and Pitfalls of Robo-Advising. Rev Financ Stud 32(5):1983–2020
Deaves R, Lüders E, Luo G (2009) An Experimental Test of the Impact of Overconfidence and Gender on Trading Activity. Rev Financ 13(3):555–575
Dohmen T, Falk A, Huffman D, Sunde U, Schupp J, Wagner GG (2011) Individual Risk Attitudes: Measurement, Determinants and Behavioral Consequences. J Eur Econ Assoc 9(3):522–550
Fehr E, Fischbacher U, von Rosenbladt B, Schupp J, Wagner GG (2002) A Nation-Wide Laboratory Schmollers Jahrb 122:519–542
Fisch J, Labouré M, Turner J (2019) The Emergence of the Robo-Advisor. In: Agnew J, Mitchell O (ed) The Disruptive Impact of FinTech on Retirement Systems, pp 13–37
Foerster S, Linnainmaa J, Melzer B, Previtero A (2017) Retail financial advice: Does one size fit all? J Finance 72(4):1441–1482
Frederick S (2005) Cognitive Reflection and Decision Making. J Econ Perspect 19(4):25–42
Glaesmer H, Hoyer J, Klotsche J, Herzberg P (2008) Die deutsche Version des Life-Orientation-Tests (LOT-R) zum dispositionellen Optimismus und Pessimismus. Zeitschrift Für Gesundheitspsychologie 16(1):26–31
Glaser M, Langer T, Reynders J, Weber M (2007) Scale Dependence of Overconfidence in Stock Market Volatility Forecasts, Working Paper
Graham JR, Harvey CR, Huang H (2009) Investor Competence, Trading Frequency, and Home Bias. Manag Science 55(7):1094–1106
Guiso L, Sapienza P, Zingales L (2008) Trusting the Stock Market. J. Finance 63(6):2557–2600
Hohenberger C, Lee C, Coughlin J (2019) Acceptance of robo-advisors: Effects of financial experience, affective reactions, and self-enhancement motives. Financial Plan. Rev. 2(2):e1047
Horn M, Oehler A, Wendt S (2020) FinTech for Consumers and Retail Investors: Opportunities and Risks of Digital Payment and Investment Services. In: Walker T, Gramlich D, Bitar M, Fardnia P (ed) Ecological, Societal, and Technological Risks and the Financial Sector, pp 405–421
Horn M, Oehler A (2020) Automated Portfolio Rebalancing: Automatic Erosion of Investment Performance? J Asset Manag 21(6):489–505
Inderst R, Ottaviani M (2009) Misselling through Agents. American Economic Review 99(3):883–908
Kaustia M, Conlin A, Luotonen N (2019) What drives the heterogeneity in portfolio choice? The role of institutional, traditional, and behavioral factors. Working Paper available at SSRN: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2845963. Accessed 10/02/2021
Kovaleva A, Beierlein C, Kemper C, Rammstedt B (2012) Eine Kurzskala zur Messung von Kontrollüberzeugung: Die Skala Internale-Externale- Kontrollüberzeugung-4 (IE-4). Gesis Working Papers, Mannheim
Krohne H, Egloff B, Kohlmann C, Tausch A (1996) Untersuchung mit einer deutschen Version der „Positive and Negative Affect Schedule“ (PANAS). Diagnostica 42(2):139–156
Lee C, Ward C, Raue M, D'Ambrosio L, Coughlin JF (2017) Age differences in acceptance of self-driving cars: A survey of perceptions and attitudes. International Conference on Human Aspects of IT for the Aged Population
Linnainmaa J, Melzer B, Previtero A (2020) The Misguided Beliefs of Financial Advisors. J Finance forthcoming
Nicolosi G, Peng L, Zhu N (2009) Do Individual Investors Learn from Their Trading Experience. J Financ Markets 12(2):317–336
Nosic A, Weber M (2010) How Riskily Do I Invest? The Role of Risk Attitudes, Risk Perceptions, and Overconfidence. Decis Anal 7:282–301
Oehler A, Wendt S (2018) Trust and Financial Services: The Impact of Increasing Digitalisation and the Financial Crisis. In: Sigurjónsson TO, Schwarzkopf D, Bryant M (eds) The Return of Trust? Institutions and the Public after the Icelandic Financial Crisis, Emerald, pp 195–211
Oehler A, Horn M, Wedlich F (2018a) Young adults’ subjective and objective risk attitude in financial decision making: Evidence from the lab and the field. Rev Behav Finance 10(3):274–294
Oehler A, Horn M, Wendt S, Reisch L, Walker T (2018b) Young adults and their finances: An international comparative study on applied financial literacy. Econ Notes 47(2–3):305–330
Oehler A, Wendt S, Wedlich F, Horn M (2018c) Investors’ Personality Influences Investment Decisions: Experimental Evidence on Extraversion and Neuroticism. J Behav Finance 19(1):30–48
Oehler A, Wanger HP (2020) Household Portfolio Optimization with XTFs? An Empirical Study Using the SHS-base. Res. Int. Bus. Finance, forthcoming. https://doi.org/10.1016/j.ribaf.2019.101103
Oehler A, Wedlich F, Wendt S, Horn M (2020) Does personality drive price bubbles? Stud Econ Finance 38(3):619-639
Oehler A, Horn M, Wendt S (2021) Information Illusion: Different Amounts of Information and Stock Price Estimates. Working Paper accepted at Annual Meeting of American Economic Association, January 3–5, 2021
Primi C, Morsanyi K, Chiesi F, Donati M, Hamilton J (2016) The Development and Testing of a New Version of the Cognitive Reflection Test Applying Item Response Theory (IRT). J Behav Decis Mak 29(5):453–469
Puri M, Robinson D (2007) Optimism and economic choice. J Financial Economics 86(1):71–99
Rammstedt B, Kemper C, Klein M, Beierlein C, Kovaleva A (2012) Eine kurze Skala zur Messung der fünf Dimensionen der Persönlichkeit: Big-Five-Inventory-10 (BFI-10). Gesis Working Papers 2012/23
Rossi A, Utkus S (2020a) The Needs and Wants in Financial Advice: Human versus Robo-advising. Working Paper available at SSRN: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3759041. Accessed 10/02/2021
Rossi A, Utkus S (2020b) Who Benefits from Robo-advising? Evidence from Machine Learning. Working Paper available at SSRN: https://doi.org/10.2139/ssrn
Seru A, Shumway T, Stoffman N (2010) Learning by Trading Rev Financ Stud 23(2):705–739
SOEP (2003) SOEP 2003 – Erweiterter Pretestbericht zum Befragungsjahr 2003 (Welle 20) des Sozio-oekonomischen Panels − Fragebogen und Verhaltensexperiment
Statista (2020) Robo-Advisors worldwide, available at https://www.statista.com/outlook/337/100/robo-advisors/worldwide. Accessed 10/02/2021
Van Rooij M, Lusardi A, Alessie R (2011) Financial Literacy and stock market participation. J Financial Economics 101(2):449–472
Von Gaudecker HM (2015) How Does Household Portfolio Diversification Vary with Financial Literacy and Financial Advice. J Financ 70(2):489–506
Yang A (2013) Decision Making for Individual Investors: A Measurement of Latent Difficulties. J Financ Serv Res 44:303–329
Funding
Open Access funding enabled and organized by Projekt DEAL.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflicts of interest/Competing interests
None.
Availability of data and material
Corresponding author provides data upon request.
Code availability
Not applicable.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
We would like to thank Haluk Ünal (Editor-in-Chief), Nagpurnanand Prabhala (Co-Editor), an anonymous referee, participants of the International Conference on Operations Research 2021, participants of the Behavioural Finance Working Group Conference 2021, and participants of the Psychonomic Society 2020 Annual Meeting for valuable comments and suggestions. A previous version of the paper has been presented under the title “Do Robo-Advisors Address Less Risk-Averse People?”. All remaining errors are our own.
Appendices
Appendix
Appendix
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Oehler, A., Horn, M. & Wendt, S. Investor Characteristics and their Impact on the Decision to use a Robo-advisor. J Financ Serv Res 62, 91–125 (2022). https://doi.org/10.1007/s10693-021-00367-8
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10693-021-00367-8