Characteristics of a Sound Goals-Based Investing Method
We believe goals-based investing is the future of wealth management because it holistically solves the investor’s main challenge of optimizing assets to efficiently fund lifetime goals. There are different approaches to goals-based investing. But regardless of the approach, goals-based investing should be built on first principles and foundational research. The purpose of this article is to propose that any sound goals-based asset allocation method should contain the following four characteristics:
- An anchor to portfolio theory
- Incorporation of liabilities (goals) and time
- An intuitive definition of risk preference
- Integration of recourse decision-making and adaptive trade-offs
Portfolio Theory
A significant body of published academic research is available to guide developers of goals-based methods. Portfolio theory is important to understanding asset allocation methods, underlying assumptions, and key issues. A lack of anchoring to portfolio theory leads to a greater risk of suboptimal outcomes and unforeseen consequences. This is crucial because asset allocation is the primary driver of portfolio return and risk.
The early financial economics literature formed a foundation in our understanding of asset allocation. This literature includes the mean-variance theory of Markowitz (1952) and the capital asset pricing model (CAPM) of Sharpe (1964). But neither of these models incorporates goals and time. A basic goals-based method can be constructed from Markowitz’s mean-variance theory or Sharpe’s CAPM, but related methods intentionally bring goals and time into the optimal asset allocation solution.
When considering time horizon in asset allocation, one must be aware of the important work of Samuelson (1969). Samuelson disproved time diversification under the assumption of independent asset returns. This means that investors with the same return-to-risk preference will have the same static asset allocation regardless of time horizon. Building on Samuelson’s work, Bodie et al. (1992) showed that asset allocation can change through time when labor income (human capital) is taken into account. Just three conditions justify a dynamic asset allocation based on time horizon: changing human capital, mean-reverting risky assets (i.e., returns are not independent), and the changing character of liabilities with time. In contrast, most asset allocation glide paths that support retirement or college savings were not constructed in consideration of these technical issues.
Contemporary intertemporal CAPM theory considers liabilities and time (Waring and Whitney 2009; Cochrane 2014). It provides a solid theoretical foundation for goals-based asset allocation, and it can be adapted to accommodate multiple goals. Another goals-based approach related to Markowitz’s mean-variance theory incorporates multiple goals with unique risk preferences, using shortfall probability as the definition of risk (Das et al. 2010).
Goals and Time
An investor’s assets should serve a purpose —to fund a lifetime of financial goals. Goals are liabilities on a lifetime balance sheet. At the highest level, goals include consumption and gifts. If assets serve the purpose of funding lifetime goals, then optimal lifetime asset allocation should be goals-based and multi-period. From this perspective, maximizing return per unit of risk is not a goal but a means to achieving a goal.
Not all goals are the same. Behavioral economists have argued that investors view their portfolios to comprise different underlying mental account subportfolios, with each subportfolio having its own purpose (Thaler 1985; Shefrin and Statman 2000). Mental accounts can include goals such as retirement, education, and bequests. Different goals can have different priorities ranging from high-priority and near-term to aspirational and long-term. The risk profile of each subportfolio that funds its respective goal should be risk-aligned. Fortunately, Sharpe et al. (1999) and Das et al. (2010) show that the total portfolio is mean-variance efficient when each mental account subportfolio also resides on the efficient frontier.1
Standard mean-variance theory is single- period optimal (typically one year), whereas goals are multi-period and funded over a lifetime. The timing of goals, their magnitude, and risk preferences can all affect asset allocation through time. Optimal lifetime asset allocation requires explicitly incorporating goals and time. For many investors, a lifetime of annual-consumption needs represents their largest and most important goal. Nontradable assets such as human capital and pensions occur at different times of the lifecycle and naturally fund part of this consumption. These nontradable assets need to be incorporated into the optimal asset allocation solution.
An Intuitive Definition of Risk Preference
Risk is multi-dimensional. Risk is volatility. It is tail risk. It is the permanent loss of capital. Risk is a failure to meet financial goals. In reality, these definitions are all closely related. Although standard deviation and conditional value-at-risk (CVaR) are excellent measures of risk for the statistically informed, they are not intuitive for most private investors. This opens the door for misalignment between portfolio selection and the investor’s true risk aversion. There are more-intuitive ways to express risk preference that can be translated into the language of portfolio theory.
Loss aversion is the behavioral tendency to prefer avoiding losses over acquiring gains. This behavior probably is related to the marginal utility of wealth. A diversified multi-asset-class portfolio should offer an approximately symmetrical return distribution. Under this condition, a rational investor would consider risk to be the variance around the expected mean return. However, loss aversion suggests the investor weighs the negative returns more heavily than what is implied by the variance around the expected mean return.
Loss aversion has been incorporated in goals-based investing to provide a more intuitive definition of risk preference for portfolio selection. For example, high-priority goals can be aligned with the intertemporal risk-free asset (or a close proxy) to guarantee funding of those goals when they arrive in time. In another goals-based approach, portfolios that fund discrete goals can be selected using shortfall probability as the definition of risk. Shortfall probability is the probability that a portfolio will not achieve the required return to meet a goal threshold. Historical stress tests (and subsequent recovery) and simulation can help capture and communicate risk in relation to goals in tangible ways, so that private investors can more precisely select portfolios that are aligned with their goals and risk preferences.
We note there are issues to using shortfall probability as the definition of risk for portfolio selection in a multi-period framework. The main criticism relates to the arguments of Samuelson (1963, 1969) on the law of large numbers and time diversification when returns are independent, where a high average return can contribute to a lower perceived risk to goal funding (i.e., a lower probability of shortfall).
Continue reading this article now. Download this article PLUS two additional articles on alternative investing from IMCA’s Investments & Wealth Monitor now. Peter Mladina is director of portfolio research for wealth management at Northern Trust. He earned a BA in economics from the University of California, Los Angeles, and an MBA from Edinburgh Business School in Britain. Contact him at [email protected].
© Investment Management Consultant Association (IMCA)
