Research & Policy Briefs From the World Bank Chile Center and Malaysia Hub No. 21, February 2019 Robo-Advisors: Investing through Machines Facundo Abraham Sergio L. Schmukler José Tessada Investing through online automated platforms, known as robo-advisors, is increasingly popular. Robo-advisors expand access to wealth management services by making it easier and less costly to open investments accounts and receive financial advice, as well as plan and automate investment decisions. However, the rise of robo-advisors requires consumers to understand the limitations of these services and to get proper financial education. Policy makers need to grapple with the impact of robo-advisors on the overall financial system, as well as reassess their regulatory and supervisory practices. The Rise of Robo-Advisors Robo-advisors are starting to grow in other parts of the world, too. In Europe, there are currently over 70 The financial industry is continuously adopting new robo-advisors, with 5 of them managing more than €100 technologies to deliver financial services in cheaper and million (Burnmark 2017). Emerging economies have also more efficient ways. The adoption of these technologies witnessed the emergence of their own robo-advisors. For particularly deepened after the 2007−08 global financial example, the number of robo-advisors is growing fast in crisis, when tighter regulations on traditional banks and Asia, driven by an emerging middle class and high developments in computer science increased incentives to technological connectivity (Forbes 2017). Robo-advisors are develop non-bank, technology-based financial companies already present in China (mainland); Hong Kong SAR, China; (IFC 2017). Some examples of technological innovations in India; Japan; Singapore; Thailand; and Vietnam, among finance include ATMs (automated teller machines), mobile other economies. Other emerging regions also have payments, and trade finance using blockchain. Now, robo-advisors, but their presence is limited so far. For technological disruption has reached the realm of wealth instance, there are only 6 robo-advisors in Africa and Latin management services, where automated financial advisors, America altogether (Burnmark 2017). Robo-advisors are known as robo-advisors, are starting to compete with expected to continue expanding around the world in years human advisors. to come. Some projections even forecast that robo-advisors will manage around 10 percent of global investment assets Conceived as a low-cost alternative to traditional human by 2020 (Business Insider 2017). advisors, robo-advisors are online platforms that use algorithms to automatically build and manage clients’ Even though they are labelled “advisors,” robo-advisors portfolios. Though robo-advisors started as fintech start-ups typically provide services that go beyond simple advisory in the aftermath of the global financial crisis, they have services to encompass comprehensive portfolio grown in popularity in recent years, particularly as more management services that allow individuals to plan and traditional financial institutions have started to offer their delegate their investment decisions. For example, in 12 out own robo-advisory services. For example, Charles Schwab, of 15 surveyed economies, robo-advisors offered both one of the largest brokerage firms in the United States, advisory and management services (IOSCO 2016). In launched a robo-advisory service, Intelligent Portfolios, in addition to portfolio allocations, the services provided can 2015. Also in 2015, BlackRock, the world’s largest asset include portfolio rebalancing and tax management. manager, acquired the robo-advisory company Future Advisor. Robo-advisory has also caught the attention of bank How Do Robo-Advisors Work? giants, such as Bank of America and Wells Fargo, which have recently started to offer their own automated advisory To help with investment decisions, robo-advisors start by services. defining the investment strategy of each individual based on his/her investment goals and risk profile. Robo-advisors ask The United States is, by far, the leading market for potential clients about the purpose of the investment and robo-advisors. As of 2017, it had more robo-advisors than the time horizon. Robo-advisors offer investment strategies any other economy in the world (about 200) and captured for a variety of goals, including retirement, saving for large 57 percent of all investments in robo-advisors (Burnmark expenditures, establishing a rainy day fund, or generating a 2017; CBInsights 2017). The estimated value of assets stream of income to cover expenses. These questions are managed by robo-advisors in the United States exceeded complemented with objective and subjective questions that US$400 billion in 2018 and is anticipated to grow at an evaluate a client’s willingness and capacity to tolerate risk. average annual rate of 31 percent, reaching almost US$1.5 Objective risk metrics can include a client’s income and trillion by 2023 (Figure 1). Currently, the largest years to retirement. Subjective questions ask, for example, robo-advisors in terms of assets under management are how the client would react to a market decline and how Vanguard (US$112 billion), followed by Intelligent Portfolios comfortable he/she is with fluctuations in the market (Lam (US$33 billion) and Betterment (US$14 billion) (Figure 2). 2016). To keep costs low and the process simple, clients’ Affiliation: Abraham and Schmukler: Development Research Group, the World Bank. Tessada: Pontificia Universidad Catolica de Chile (PUC & FinanceUC). E-mail addresses: fabraham@worldbank.org, sschmukler@worldbank.org, jtessada@uc.cl Acknowledgement: We received very useful comments from Ana Maria Aviles, Norman Loayza, and Luis Servén, as well as helpful edits from Nancy Morrison. The World Bank Chile Research and Development Center, the FCI group at the Malaysia Global Knowledge and Research Hub, and the KCP and SRP programs provided financial support for this brief. Objective and disclaimer: Research & Policy Briefs synthesize existing research and data to shed light on a useful and interesting question for policy debate. Research & Policy Briefs carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions are entirely those of the authors. They do not necessarily represent the views of the World Bank Group, its Executive Directors, or the governments they represent.- Global Knowledge & Research Hub in Malaysia Robo-Advisors: Investing through Machines Figure 1. Projected Assets Managed by Robo-Advisors in the United States, 2018−23 The value of assets under management by robo-advisors is expected to more than triple between 2018 and 2023. 1,600 1,486 1,405 1,400 1,266 1,200 1,049 1,000 US$billion 800 750 600 426 400 200 0 2018 2019 2020 2021 2022 2023 Source: Statista 2019. assessments are conducted using standard online short traded funds (ETFs) and index funds. These investment questionnaires. instruments tend to follow a basket of securities or an index (investing in all the securities included in the index and in Based on these two dimensions, robo-advisors use the same proportion as the index). As a result, by acquiring automated algorithms to make recommendations on how only a few funds, investors can achieve a “market portfolio” to allocate funds across different types of assets. In most while minimizing trading costs. Moreover, by passively cases, these algorithms are based on modern portfolio holding funds, investors do not need to engage in active theory (Bjernes and Vukovic 2017). This portfolio selection monitoring and trading, reducing trading costs even further. framework, introduced by Harry Markowitz (1952), is the In practice, robo-advisors seem to follow a conservative most popular model of asset allocation and states that the approach, offering funds that have wide coverage, long optimal portfolio is one that maximizes the expected return operating history, market liquidity, and good performance given a level of risk tolerance or, alternatively, minimizes risk over time (Deutsche Bank 2017; Phoon and Koh 2018). given a level of expected return. Robo-advisors construct portfolios with higher risk by increasing the ratio of equity to The entire process of using a robo-advisor (from opening bonds and, within each type of instrument, investing in an account to monitoring and readjusting the portfolio) can riskier assets (for example, moving away from government be performed online with no human interaction. Several to municipal bonds or from U.S. to emerging market stocks). robo-advisors are fully automated. These types of Portfolio optimization is adjusted by taking into account robo-advisors are usually less costly, and thus are oriented investment goals and the desired risk level. For example, for toward the mass market. Nevertheless, other robo-advisors a given level of risk, asset allocation will be different if the offer a hybrid system where clients have human interaction, goal is to generate income for expenses or saving for the albeit limited. For example, only a certain number of long term. contacts are allowed or they can occur only via the Internet, not in person. These robo-advisors charge higher fees, but In addition to recommending an initial allocation of are still cheaper than traditional human advisors. funds, algorithms can be designed to continuously monitor portfolios and detect deviations from the targeted risk. Benefits and Limitations of Robo-Advisors Whenever deviations are identified, the portfolio is automatically rebalanced. For example, the value of equity The use of robo-advisors for management wealth services might increase faster than the value of bonds over time, can provide several advantages over traditional services increasing the share of the portfolio invested in equity. that rely on human advisors. One appeal of using Because a higher share of equity increases risks, the robo-advisors is that they are easily accessible. Instead of portfolio might be rebalanced by selling equity. Moreover, having to set an appointment with an advisor and attend a the portfolio can be automatically rebalanced to reduce meeting at a physical location, robo-advisors offer clients risks as time goes by and as the time when the funds are the possibility of obtaining financial advice and managing needed approaches. Rebalancing can also occur when the investments at any time, from anywhere with an Internet investor changes his/her risk tolerance or investment goals. connection. A corollary of modern portfolio theory is that by Robo-advisors can also reduce the costs of financial diversifying assets, investors can reduce risk without advice. In contrast to human advisory firms, robo-advisors sacrificing expected returns. To achieve diversification at a can save on fixed costs, such as the salaries of expensive low cost, robo-advisors mainly offer to invest in exchange financial advisors or the maintenance of physical offices. As 2 Research & Policy Brief No.21 Figure 2. Largest Robo-Advisors in the United States, 2018 Vanguard dominates the robo-advisor industry in the United States. 120 112 100 80 US$billion 60 40 33 20 14 10 8 0 Vanguard Intelligent Portfolios Betterment Wealthfront Personal Capital (Charles Schwab) Assets under Management Source: Ortner 2018. a result, they can reduce minimum investment and other financial firms not because they are the cheapest requirements. For instance, Bank of America requires but because they receive higher commissions from them US$25,000 to open an account with a private financial (Fein 2015). Moreover, algorithms are inevitably advisor, but only US$5,000 to open an account with their programmed by humans, so biases could be introduced robo-advisor. Some robo-advisors, such as Betterment, do during their design, consciously or unconsciously. For not require a minimum investment at all. In addition, example, robo-advisors could recommend that clients hold robo-advisors can charge lower fees than human advisors. a relatively large share of their investment in cash to then For example, a fully automated robo-advisor can charge a re-invest for profit (Vox 2016). fee as low as 0.25 percent of assets managed, whereas the Whereas robo-advisors can increase accessibility and fees for traditional human advisors are no less than 0.75 affordability of wealth management services, they can also percent and can even reach 1.5 percent (EY 2015). On top of entail costs. Although straightforward and time-saving, administrative fees, financial advisors usually charge fees for robo-advisors might not be able to know clients as well as trades performed, which robo-advisors typically minimize human advisors do through multiple interactions, tailored by following passive investing. questions, and closer relationships. “One-size-fit-all” Using robo-advisors can yield additional savings in the questionnaires might be too simple and narrow to provide a form of “tax harvesting.” This is the practice of selling assets complete overview of a client’s financial situation and that experience a loss and using the proceeds to buy an his/her needs. Furthermore, these questionnaires assume asset with similar risk (keeping the same risk profile of the that individuals with a similar risk profile would provide the portfolio). Recording a loss decreases capital gains, reducing same answers to the same subjective questions, which taxable income. Performing tax harvesting can be complex: might not necessarily be true (Deutsche Bank 2017). it involves identifying harvesting opportunities in a portfolio Robo-advisors also lack other important aspects of a with several assets, finding suitable substitutes, and client-advisor relation, such as helping clients define their performing multiple trading, among other tasks. financial goals, counseling during market downturns, or Robo-advisors can perform tax harvesting more efficiently dealing with possible changes in their lives (Accenture and frequently than human advisors. 2015). Furthermore, limited risk-assessment might not provide Robo-advisors can also help reduce some of the a complete overview of a client’s overall financial condition. behavioral biases that are common in financial advisory. Robo-advisors might not ask about a client’s other Human advisors can be subjective, favor products for which investments (such as pension funds and real estate), future they receive commission, have a limited capacity to monitor expenses, potential liabilities, spouse’s financial condition, several assets simultaneously, and focus on domestic or insurances purchased, among other information (FINRA securities, among other biases. Thus, by transferring the 2016). If robo-advisors act on partial information, they decision-making process from humans to automated might not provide optimal recommendations. algorithms, robo-advisors can mitigate some of these biases. Nevertheless, even when robo-advisors are used, Robo-advisors can also lead to consumer some biases might still be present. For example, similar to disengagement. In other words, because the entire process human advisors, robo-advisors might use certain brokers is automatic, consumers might not make efforts to 3 Robo-Advisors: Investing through Machines understand how the service works, or even continuously (Baker and Dellaert 2018). Regulatory agencies around the monitor their investments. This issue is particularly relevant world have already started to think about how to when robo-advisors are offered to individuals with relatively adequately adopt robo-advisors and have issued guidelines, lower wealth who might have no experience with reports, and opinions on this issue. Regulators have stated investment products (OECD 2017). that they would have to develop new skills to supervise robo-advisors effectively. For example, they would need to Because robo-advisors are relatively new, their business have the technical capacity to assess robo-advisors’ models have not been tested in the long term and under financial stress. Thus, it remains unclear to what extent algorithms. Similarly, regulators would need to understand consumers will be protected in case a robo-advisor how automated profiling of clients works. Regulators have company fails. Some jurisdictions have taken steps to also emphasized the importance of consumer education. protect consumers. For example, robo-advisors in the Prospective clients need to have enough information to United States are required to be members of the Securities understand how robo-advisors operate and whether or not Investor Protection Corporation (SIPC), which provides they are suitable to their needs. Regulatory organizations insurance for up to US$ 500,000 per customer in case of have also raised concerns related to cybersecurity and data bankruptcy of a member firm. privacy, among other issues (ESAs 2015; FINRA 2016; IOSCO 2016). The Future of Robo-Advisors The robo-advisory industry is still at an early stage and, Because of their low cost and easy accessibility, as such, few studies have analyzed its different impacts on robo-advisors have the potential to promote more the financial system, including on asset markets. As sophisticated investment practices within a population not robo-advisors expand and more information on them used to having access to financial advisors. Robo-advisors becomes available, more analyses could try to shed light on could be particularly attractive to certain groups such as these issues. For example, future work could analyze who households with relatively lower income or younger uses robo-advisors, and thus to what extent they are individuals, who might not invest because their investable contributing to better financial decisions by a wider array of funds are too small, are located far from urban centers, or investors. Further analyses could also focus on whether simply feel intimidated by human advisors. Increased human and robo-advisors are substitutes or complements participation in capital markets, would offer these (either catering to different population segments or individuals new ways to save for retirement, rainy days, or individuals using both at the same time), as well as on how any other purpose. At the same time, robo-advisors might the profit margins and cost structures compare between benefit individuals that already have investments. Not only the two. To the extent that robo-advisors can be accessed can their costs be reduced, but thanks to the enhanced by any individual from any location, it would also be computational power and (at least in theory) objectivity, interesting to study whether robo-advising activity tends to robo-advisors can design more efficient portfolios be concentrated in a few economies and accessed by compared to humans. In fact, there is empirical evidence investors from all over the world, or economies have their that the use of robo-advisors can be associated with higher own robo-advisors catered to their own domestic investors. diversification and less behavioral biases (D’Acunto, It would be useful to analyze if the same robo-advisor gives Prabhala, and Rossi 2018). consumers in different economies tailored Proper regulation and supervision will be a key recommendations based on local products and determinant of the success of robo-advisors. Policy makers environments (such as tax codes), or instead provides would benefit from establishing good practices that standard international products. Further insights on these guarantee that robo-advisors are objective and transparent, topics would help to better understand the true potential and provide advice appropriate to each client’s needs and pitfalls of robo-advisors. References Accenture. 2015. The Rise of Robo-Advice: Changing the Concept of Wealth FINRA. 2016. Report on Digital Investment Advice. Management. Forbes. 2017. “Why Are Robo-Advisors on the Rise in Asia?” November 15. Baker, T., and B. Dellaert. 2018. “Regulating Robo Advice Across the Financial Services Industry.” Faculty Scholarship, University of Pennsylvania Law School. IFC (International Finance Corporation). 2017. “Digital Financial Services: Challenges and Opportunities for Emerging Market Banks.” EMCompass, Note 42. Bjernes, L., and A. Vukovic. 2017. “Automated Advice: A Portfolio Management Perspective on Robo-Advisors.” Master’s thesis, Norwegian University of IOSCO (International Organization of Securities Commissions). 2016. Update to Science and Technology. the Report on the IOSCO Automated Advice Tools Survey. Burnmark, 2017. Digital Wealth. Lam, J. 2016. “Robo-Advisors: A Portfolio Management Perspective.” Paper Business Insider. 2017. “Robo Advisors: Online Financial Advisors that Fit in Your presented to the Distinction in the Major thesis, Yale College. Pocket.” January 4. Markowitz, H. 1952. “Portfolio Selection.” The Journal of Finance 7 (1): 77−91. CBInsights. 2017 “A Wealth Tech World: Mapping Robo-Advisors Around the Globe.” April 21. OECD (Organisation for Economic Co-operation and Development). 2017. Robo-Advice for Pensions. D’Acunto, F., N. Prabhala, and A. Rossi. 2018. “The Promises and Pitfalls of Robo-Advising.” CESifo Working Paper 6907, CESifo Group Munich. Ortner, J. 2018. “Questions Still Remain if Robo Advising Is Right for You.” Seeking Deutsche Bank. 2017. “Robo-Advice − A True Innovation in Asset Management.” Alpha, Nov 12. EU Monitor Global Financial Markets. Phoon, K., and F. Koh. 2018. “Robo-Advisors and Wealth Management.” The ESAs (European Supervisory Authorities). 2015. Joint Committee Discussion Journal of Alternative Investments 20 (3): 79−94. Paper: On Automation in Financial Advice. Statista. 2019. “Robo-Advisors.” https://www.statista.com/outlook/337/109/robo- EY. 2015. Advice Goes Virtual: How New Digital Investment Services Are Changing advisors/united-states. the Wealth Management Landscape. Vox. 2016. “Charles Schwab Offers Free Investment Advice − But There's a Big Catch.” http://documents.worldbank.org/curated/en/795161508437504068/Corporate-debt-maturity-in-developing-countries-sources-of-long-and-short-termism Fein, M. 2015 “Robo-Advisors: A Closer Look.” Banking & Insurance Journal 7 (174). May 24. Global Knowledge & Research Hub in Malaysia