A Report On Digital Investment Advice by FINRA
Introduction
Technology has long played a central role in financial services innovation. It continues to do so today as many firms in the securities industry introduce new digital investment advice tools to assist in developing and managing investment portfolios. FINRA undertook a review of selected digital investment advice tools to assess these developments.
The observations and practices in this report are drawn from FINRA’s discussions with a range of financial services firms that provide or use digital investment advice tools, vendors and foreign securities regulators as well our regulatory experience. This report uses the term “financial services firms” to include both broker-dealers and investment advisers. The rules discussed in this report apply to broker-dealers. The effective practices we discuss are specifically intended for FINRA-registered firms, but may be valuable to financial professionals generally.
The adoption of digital investment advice tools has stimulated discussions about the role of financial professionals and the evolving relationship between financial intermediaries and their clients. What role will financial professionals play in conjunction with digital services in providing investment advice? To what degree will investors rely primarily on digital investment advice? How well can software know a client? Can the skill, knowledge and service provided by well-trained and ethical financial professionals be incorporated in software? Can that software provide sound personal advice, especially for clients with more complex advice needs?
Without venturing to answer these questions, what is clear is that the role technology plays in supporting investment advice to clients will increase at many securities firms. With that in mind, FINRA issues this report to remind broker-dealers of their obligations under FINRA rules as well as to share effective practices related to digital investment advice, including with respect to technology management, portfolio development and conflicts of interest mitigation. The report also raises considerations for investors in evaluating investment advice derived entirely or in part from digital investment advice tools.
This report does not create any new legal requirements or change any existing broker-dealer regulatory obligations. Throughout the report, we identify practices that we believe firms should consider and tailor to their business model.
A Note on Terminology
As used here, digital investment advice tools (also referred to as digital advice tools) support one or more of the following core activities in managing an investor’s portfolio: customer profiling, asset allocation, portfolio selection, trade execution, portfolio rebalancing, tax-loss harvesting and portfolio analysis. These investment advice tools can be broken down into two groups: tools that financial professionals use, referred to here as “financial professional-facing” tools, and tools that clients use, referred to here as “client-facing” tools. Client-facing tools that incorporate the first six activities–customer profiling through tax-loss harvesting–are frequently referred to as “robo advisors” or “robos.”
A Brief History of Digital Investment Advice
Financial professionals have used digital investment advice tools for years. These tools help financial professionals at each point in the value chain described above, for example, to develop an investor profile, to prepare proposals and sales materials, to develop an asset allocation or to recommend specific securities to an investor. Those recommendations may be for individual securities, a customized portfolio or a pre-packaged portfolio for investors with a given profile. In addition, digital tools can help develop recommendations to rebalance investors’ portfolios on a periodic basis or to support tax-loss harvesting. The tools financial professionals use may be developed by their firms, acquired from third-party vendors by their firm or, in some cases, acquired by the financial professionals themselves.
In the late 1990s, the landscape of investment tools available directly to investors began to expand. Some firms started to make asset allocation tools available online. The landscape expanded further in 2005, when NASD Interpretative Material (IM) 2210-6 became effective, allowing broker-dealers to make “investment analysis tools” available to investors. FINRA defined an investment analysis tool to be an “interactive technological tool that produces simulations and statistical analyses that present the likelihood of various investment outcomes if certain investments are made or certain investment strategies or styles are undertaken.”
Following the 2008 financial crisis, a number of new entrants began offering a wide range of digital financial tools directly to consumers, including investment advice tools. Many of these firms had their roots in the technology industry and brought new perspectives on the role of technology in financial services. The client-facing digital investment tools these firms developed offer aspects of the functionality previously only available to financial professionals. The degree of human involvement in client-facing tools varies substantially. Some firms rely on a purely digital interaction with clients while others provide optional or mandatory access to a financial professional.
In many cases, securities industry participants are responding with digital investment advice strategies of their own. Some participants are developing or acquiring client-facing investment advice tools while others are developing or acquiring financial professional-facing tools to enhance their ability to serve clients and compete more effectively. Some of these latter tools include advanced analytic tools–e.g., to assess customer risk tolerance or portfolio risk–and in some cases presentation interfaces that enable the financial professional to present information to clients online. Vendors frequently position these tools as providing the basis for financial professionals to conduct more in-depth, sophisticated discussions with their client.
Governance and Supervision
Governance and supervision of investment recommendations are recurring topics of FINRA guidance and are equally relevant to digital investment advice tools. We focus here on governance and supervision in two areas: 1) the algorithms that drive digital investment tools; and 2) the construction of client portfolios, including potential conflicts of interest that may arise in those portfolios.
Algorithms
Algorithms are core components of digital investment advice tools. They use various financial models and assumptions to translate data inputs into suggested actions at each step of the advice value chain. The methodology by which the algorithm translates inputs into outputs should reflect a firm’s approach to a particular task, e.g., profiling an investor, rebalancing an account or performing tax-loss harvesting. If an algorithm is poorly designed for its task or not correctly coded, it may produce results that deviate systematically from the intended output and that adversely affect many investors.
For this reason, it is essential that firms effectively govern and supervise the algorithms they use in digital-advice tools. At the most basic level, firms should assess whether an algorithm is consistent with the firm’s investment and analytic approaches. For example, a number of clientfacing digital investment advice tools are based on precepts from Modern Portfolio Theory and use a passive, index-based approach to investing based on the risk tolerance of the client, while others incorporate active management of investment portfolios. Not surprisingly, the outputs and investment advice from algorithms developed based on these approaches are likely to be different.
Even when client-facing digital advice tools take a similar approach to investing, implementation of methods for specific investing tasks, for example asset allocation, may produce very different results. Cerulli Associates compared the asset allocation for a notional 27-year-old investing for retirement across seven client-facing digital advice tools. Equity allocations ranged as high as 90 percent and as low as 51 percent; fixed income allocations ranged from 10 percent to 40 percent. (See Figure 2.) A Wall Street Journal analysis found similar disparities.
These examples highlight the importance of firms 1) understanding the methodological approaches embedded in the algorithms they use, including the assumptions underlying the potential scenarios on expected returns, and the biases or preferences that exist in those approaches and 2) assessing whether these methodological approaches reflect a firm’s desired approach. These considerations apply both to the internal development of digital advice tools and third-party digital advice tools that firms acquire or private-label.
A look at two other areas of digital investment advice—customer risk tolerance assessment and portfolio analysis—reinforces the need for broker-dealers to establish and implement effective governance and supervision of their digital investment advice tools. FINRA reviewed several tools designed to help financial professionals understand investors’ risk tolerance. In some cases, these tools also analyze the alignment of investors’ portfolios with their risk tolerance and propose conforming changes to bring the portfolio into alignment. These tools vary considerably in approach to these tasks. (See Observations on Practices beginning on page 6 for a discussion of some of these approaches.) Good governance involves understanding if the approach to assessing customer risk tolerance is consistent with the firm’s approach.
FINRA also reviewed tools to help financial professionals and their clients understand the impact of potential shocks to clients’ portfolios, for example from an oil price fall, a global recession or a geo-political crisis. Careful governance would include understanding the analytic approaches that are used in these tools, including the assumptions that are made, about the impact of the shock events on the correlations in various asset price movements, among other things.
Developing an understanding of the algorithms a tool uses would also include understanding the circumstances in which their use may be inappropriate. For example, applying a tax-loss harvesting algorithm to one account of a married client where both spouses have multiple investment accounts may be detrimental. Without a full view of the couple’s portfolio, the algorithm may generate unusable realized losses.
See full report below.
