APPROACHES TO ASSET ALLOCATION

Learning Outcomes

• compare the investment objectives of asset-only, liability-relative, and goals-based asset allocation approaches

• contrast concepts of risk relevant to asset-only, liability-relative, and goals-based asset allocation approaches

ChatGPT Summary

https://chat.openai.com/c/e58c5cdd-1e7b-4bb6-9e6f-d11691247bc2

The document you've provided outlines the fundamental concepts and distinctions between three primary approaches to asset allocation: asset-only, liability-relative, and goals-based. Here's a concise explanation of each, focusing on their investment objectives, contrasting concepts of risk, and their relevance in financial planning and investment management.

Asset-Only Asset Allocation

• Objective: Maximizes the Sharpe ratio (return per unit of risk) of the portfolio by focusing solely on the assets' expected returns, volatilities, and correlations without considering liabilities or specific goals.

• Risk Concept: Utilizes volatility (standard deviation of portfolio returns) as the primary risk measure. It emphasizes the risk-return trade-off inherent in different asset class combinations.

• Application: Suitable for investors or institutions where liabilities or specific financial goals are undefined or deemed less critical, such as some sovereign wealth funds, foundations, and endowments.

Liability-Relative Asset Allocation

• Objective: Aims to fund liabilities by choosing an asset allocation that covers both current and future obligations. This approach may use surplus optimization (assets minus liabilities) or construct a liability-hedging portfolio alongside a return-seeking portfolio.

• Risk Concept: Focuses on shortfall risk—the risk of having insufficient assets to meet liabilities when due. Risk measures may include the volatility of the surplus and sensitivities of assets and liabilities to financial conditions like interest rates.

• Application: Predominantly used by entities with defined obligations, such as pension funds, insurers, and banks, where the penalty for not meeting liabilities is high.

Goals-Based Asset Allocation

• Objective: Aligns asset allocation with individual or familial goals, each associated with specific cash flows, time horizons, and required probabilities of success. This approach segments the portfolio into sub-portfolios, each dedicated to achieving a particular goal.

• Risk Concept: Concerns the probability of failing to meet specified goals. It employs a more personalized risk assessment that considers the unique aspects of each goal, such as time horizon and required success probability.

• Application: Primarily for individual investors, allowing for a more tailored investment strategy that caters to varying degrees of risk tolerance and different financial aspirations, from maintaining lifestyle to leaving a bequest.

Contrasting Risk Concepts

• Asset-Only: Risk is seen in terms of portfolio volatility and the trade-off between risk and return.

• Liability-Relative: Risk involves not just volatility but also the capacity to meet liabilities, focusing on shortfall risk and the matching of asset and liability characteristics.

• Goals-Based: Risk is personalized, related to the likelihood of not achieving specific, defined goals.

Investment Practice Implications

Each approach has its place in investment practice, depending on the investor's specific circumstances, objectives, and constraints. Asset-only approaches offer a straightforward framework for maximizing returns for a given level of risk, suitable for contexts where liabilities or specific goals are not a primary concern. Liability-relative approaches provide a rigorous method for institutions or individuals with defined liabilities to ensure these obligations can be met, emphasizing the matching of asset and liability characteristics. Goals-based asset allocation offers the most personalized framework, aligning investment strategies directly with individual goals, making it particularly relevant for personal financial planning.

Understanding these approaches helps investors and financial planners develop more effective asset allocation strategies that align with specific investment objectives and risk tolerances.

We can identify three broad approaches to asset allocation: (1) asset-only, (2) liability-relative, and (3) goals-based. These are decision-making frameworks that take account of or emphasize different aspects of the investment problem.

Asset-only approaches to asset allocation focus solely on the asset side of the investor’s balance sheet. Liabilities are not explicitly modeled. Mean–variance optimization (MVO) is the most familiar and deeply studied asset-only approach. MVO considers only the expected returns, risks, and correlations of the asset classes in the opportunity set. In contrast, liability-relative and goals-based approaches explicitly account for the liabilities side of the economic balance sheet, dedicating assets to meet, respectively, legal liabilities and quasi-liabilities (other needs that are not strictly liabilities but are treated as such) or goals.

Liability-relative approaches to asset allocation choose an asset allocation in relation to the objective of funding liabilities. The phrase “funding of liabilities” means to provide for the money to pay liabilities when they come due. An example is surplus optimization: mean–variance optimization applied to surplus (defined as the value of the investor’s assets minus the present value of the investor’s liabilities). In modeling, liabilities might be represented by a short position in a bond or series of bonds matched to the present value and duration of the liabilities. Another approach involves constructing a liability-hedging portfolio focused on funding liabilities and, for any remaining balance of assets, a risky-asset portfolio (so called because it is risky or riskier in relation to liabilities—often also called a “return-seeking portfolio” because it explicitly seeks return above and beyond the liability benchmark). Liability-driven investing (LDI) is an investment industry term that generally encompasses asset allocation that is focused on funding an investor’s liabilities. Related fixed-income techniques are covered in the fixed-income sequence under liability-based mandates.

All approaches to asset allocation can be said to address goals. In investment practice and literature, however, the term “goals based” has come be widely associated with a particular type of approach to asset allocation and investing.

Goals-based approaches to asset allocation, as discussed here, are used primarily for individuals and families, involve specifying asset allocations for sub-portfolios, each of which is aligned to specified goals ranging from supporting lifestyle needs to aspirational. Each goal is associated with regular, irregular, or bulleted cash flows; a distinct time horizon; and a risk tolerance level expressed as a required probability of achieving the goal.7 For example, a middle-aged individual might specify a goal of maintaining his current lifestyle and require a high level of confidence that this goal will be attained. That same individual might express a goal of leaving a bequest to his alma mater. This would be a very long-term goal and might have a low required probability. Each goal is assigned to its own sub-portfolio, and an asset allocation strategy specific to that sub-portfolio is derived. The sum of all sub-portfolio asset allocations results in an overall strategic asset allocation for the total portfolio. Goals-based investing (GBI) is an investment industry term that encompasses the asset allocation focused on addressing an investor’s goals.

Institutions and Goals-Based Asset Allocation

Asset segmentation as practiced by some life insurers has some similarities to goals-based investing. Asset segmentation involves notionally or actually segmenting general account assets into sub-portfolios associated with specific lines of business or blocks of liabilities. On one hand, such an approach may be distinguished from goals-based asset allocation for individual investors in being motivated by competitive concerns (to facilitate offering competitive crediting rates on groups of contracts) rather than behavioral ones. On the other hand, Fraser and Jennings (2006) described a behaviorally motivated goals-based approach to asset allocation for foundations and endowments. Following their approach, components of an overall appropriate mean–variance optimal portfolio are allocated to time-based sub-portfolios such that uncomfortably novel or risky positions for the entity’s governing body are made acceptable by being placed in longer-term sub-portfolios.

Although any asset allocation approach that considers the liabilities side of the economic balance sheet might be termed “liability relative,” there are several important distinctions between liabilities for an institutional investor and goals for an individual investor. These distinctions have meaningful implications for asset allocation:8

• Liabilities of institutional investors are legal obligations or debts, whereas goals, such as meeting lifestyle or aspirational objectives, are not. Failing to meet them does not trigger similar consequences.

• Whereas institutional liabilities, such as life insurer obligations or pension benefit obligations, are uniform in nature (all of a single type), an individual’s goals may be many and varied.

• Liabilities of institutional investors of a given type (e.g., the pension benefits owed to retirees) are often numerous and so, through averaging, may often be forecast with confidence. In contrast, individual goals are not subject to the law of large numbers and averaging. Contrast an estimate of expected death benefits payable for a group of life insurance policies against an individual’s uncertainty about the resources needed in retirement: For a 65-year-old individual, the number of remaining years of life is very uncertain, but insurers can estimate the average for a group of 65-year-olds with some precision.

Liability-Relative and Goals-Based Approaches to Investing

Various perspectives exist concerning the relationship between liability-relative and goals-based approaches to investing. Professor Lionel Martellini summarizes one perspective in the following three statements:9

1. Goals-based investing is related to a new paradigm that advocates more granular and investor-centric investment solutions.

2. This new investment solutions paradigm translates into goals-based investing (GBI) approaches in individual money management, in which investors’ problems can be summarized in terms of their goals, and it translates into liability-driven investing (LDI) approaches in institutional money management, where the investors’ liability is treated as a proxy for their goal.

3. GBI and LDI are therefore related, but each of these approaches has its own specific characteristics. For example, GBI implies the capacity to help individual investors identify a hierarchical list of goals, with a distinction between different types of goals (affordable versus non affordable, essential versus aspirational, etc.) for which no exact counterpart exists in institutional money management.

Relevant Objectives

All three of the asset allocation approaches listed here seek to make optimal use of the amount of risk that the asset owner is comfortable bearing to achieve stated investment objectives, although they generally define risk differently. Exhibit 5 summarizes typical objectives.

Exhibit 5:

Asset Allocation Approaches: Investment Objective

Asset Allocation Approach	Relation to Economic Balance Sheet	Typical Objective	Typical Uses and Asset Owner Types
Asset only	Does not explicitly model liabilities or goals	Maximize Sharpe ratio for acceptable level of volatility	Liabilities or goals not defined and/or simplicity is important Some foundations, endowments Sovereign wealth funds Individual investors
Liability relative	Models legal and quasi-liabilities	Fund liabilities and invest excess assets for growth	Penalty for not meeting liabilities high Banks Defined benefit pensions Insurers
Goals based	Models goals	Achieve goals with specified required probabilities of success	Individual investors

In a mean–variance asset-only approach, the objective is to maximize expected portfolio return per unit of portfolio volatility over some time horizon, consistent with the investor’s tolerance for risk and consistent with any constraints stated in the IPS. A portfolio’s Sharpe ratio is a characteristic metric for evaluating portfolios in an asset-only mean–variance approach.

The basic objective of a liability-relative asset allocation approach is to ensure payment of liabilities when they are due.

A goals-based approach is similar to a liability-relative approach in that it also seeks to ensure that there are sufficient assets to meet the desired payouts. In goals-based approaches, however, goals are generally associated with individual sub-portfolios, and an asset allocation is designed for each sub-portfolio that reflects the time horizon and required probability of success such that the sum of the sub-portfolios addresses the totality of goals satisfactorily.

Relevant Risk Concepts

Asset-only approaches focus on asset class risk and effective combinations of asset classes. The baseline asset-only approach, mean–variance optimization, uses volatility (standard deviation) of portfolio return as a primary measure of risk, which is a function of component asset class volatilities and the correlations of asset class returns. A mean–variance asset allocation can also incorporate other risk sensitivities, including risk relative to benchmarks and downside risk. Risk relative to benchmarks is usually measured by tracking risk (tracking error). Downside risk can be represented in various ways, including semi-variance, peak-to-trough maximum drawdown, and measures that focus on the extreme (tail) segment of the downside, such as value at risk.

Mean–variance results, although often the starting point for understanding portfolio risk, are regularly augmented by Monte Carlo simulation. By providing information about how an asset allocation performs when one or more variables are changed—for example, to values representing conditions of financial market stress—simulation helps complete the picture of risk, including downside and tail risk. Insights from simulation can then be incorporated as refinements to the asset allocation.

Liability-relative approaches focus on the risk of having insufficient assets to pay obligations when due, which is a kind of shortfall risk. Other risk concerns include the volatility of contributions needed to fund liabilities. Risk in a liability-relative context is generally underpinned by the differences between asset and liability characteristics (e.g., their relative size, their interest rate sensitivity, their sensitivity to inflation).

Goals-based approaches are concerned with the risk of failing to achieve goals.10 The risk limits can be quantified as the maximum acceptable probability of not achieving a goal.11 The plural in “liabilities” and “goals” underscores that these risks are generally related to multiple future points in time. Overall portfolio risk is thus the weighted sum of the risks associated with each goal.

Generally, a given statistical risk measure may be relevant in any of the three approaches. For example, standard deviation can be used to assess overall portfolio volatility in asset-only approaches, and it may be used to measure surplus volatility (the volatility of the difference between the values of assets and liabilities) or the volatility of the funded ratio (the ratio of the values of assets and liabilities) in liability-relative asset allocation.