There are many opinions on the best way to manage an individual’s investment portfolio to provide for retirement savings with some growth. Typical approaches involve a menu of “sized” portfolios that allocate between low-risk/low-return investments (e.g., highly rated bonds) and high-risk/high-return investments (e.g., stocks). The typical menu also provides an array of asset class mixes that the investor can select from based on subjective risk tolerance and age. We think there is a better “mousetrap,” one that can be used in a relatively simple form but that also can be refined for more sophisticated applications. This approach was proposed by Wilcox (2003) as “discretionary wealth.” A list of other articles on this topic is included at the end of this piece.

The discretionary wealth approach assumes that each investor has an implicit personal “balance sheet” composed of “liabilities” and “assets”; the investor’s situation is very similar to that of an insurance company writing annuities except that in this case, the investor self-insures and various assets and liabilities that are not relevant to the present and future investment portfolio are excluded.

The assets consist not only of the usual investment assets but also of implied assets, such as that portion of your human capital that is projected to be converted through savings to your investment portfolio, as well as noninvestment assets, such as a business or a spare vacation house that could be converted to investment assets. The total value of your investment assets will vary over time because of changes in market value as well as contributions to or withdrawals from savings.

The liabilities consist of any investment-related debt plus implied liabilities. In the typical case, the main implied liabilities are the present value of what the investor thinks he absolutely needs to live on (i.e., lifestyle maintenance) for the rest of his life, perhaps after retirement, or to fund essentials for other family members (e.g., college tuition). Other potential implied liabilities include taxes on unrealized capital gains, planned future charitable contributions, and so on. Imagine that you can simply buy an annuity that pays you an after-tax amount sufficient to meet your annual expenses for the rest of your life, adjusted for inflation. These expenses likely will not be fixed but will vary over your life cycle (single, married, parent, empty-nester, etc.). The fair value of this self-insured annuity is the liability that needs to be funded from your investment assets.

The net or excess of your assets over your liabilities is discretionary wealth (D-W). D-W is analogous to the surplus of an annuity-writing insurance company, the “equity” on its balance sheet. D-W is the amount you can afford to lose without suffering a shortfall disaster. In many cases, the main liability will be the implied one of the present, or time-discounted, value of the future consumption liabilities. In the simplest case where there are no implied assets and no investment liabilities,

Investment assets (A) = Planned consumption (C) + Discretionary wealth (D-W).

Looked at from the opposite perspective, what is particularly important is the ratio of the total investment assets to D-W. This relationship works like leverage and is essential for determining how much investment risk you should take relative to the available investment opportunities. It quantifies how changes in the investment portfolio (A) amplify changes in D-W: The effect on return and variance is proportional to A/D-W. For example, if A is three times D-W, a 5 percent change in A will mean a 15 percent change in D-W. This relationship is analogous to how debt leverage amplifies the return on equity, both positive and negative, in a business. The greater the leverage, the riskier the wealth of the shareholders, or in this framework the discretionary wealth. Note that the closely related ratio of C/A is similar to a conventional loan-to-value ratio and is just a rearrangement of the components A, C, and D-W.

For most people, annual lifestyle consumption increases from young adulthood through middle age and then decreases into retirement with a possible late-in-life upsurge for medical care. Each passing year decreases the remaining number of annuity payments needed to support your lifestyle. To figure out how much you need from investment assets to fund the lifestyle annuity, you have to make estimates for the following variables:

• expected return (net of fees and costs),
• risk (volatility),
• taxes,
• inflation,
• lifespan, and
• lifestyle consumption.

Because the future can never be completely known, it is wise for you to specify an additional reserve liability against the unknown, both with respect to future capital market returns and future consumption. Alternatively, you may use Monte Carlo simulation to address uncertainty (see the article by Dan diBartolemeo is this issue).

Here is a simple example of an implied balance sheet for someone who is now 60 years old, retired, and expects to spend $100,000 per year, in 2009 dollars, for the rest of her life. Actuarial life expectancy is fine for large groups, but it is safer for individuals to use the maximum plausible remaining lifespan so as not to risk being destitute in old age by outliving the actuarial tables.

Following are the parameters:

A straightforward time discount factor for future consumption is the AAA municipal rate yield adjusted by an estimated inflation rate, so the present value of the lifestyle annuity is about $3.28 million. If the investor’s total assets are $5.0 million, the net $1.72 million is D-W. We call the ratio of assets to D-W “L”; here, it is 5.0/1.72 = 2.91, which is high enough to call for a conservative investment approach.

L becomes important to understanding how risk and return can be related directly to D-W. L is usually greater than 1. Consequently, L usually amplifies the effect that changes in assets cause for changes in D-W, both as to return and volatility. It maps into the conventional parameters of risk and return. If the expected return on total investment assets is R, then the expected return on D-W is L ´ R. Similarly, if the variance of total investment assets is V, then the variance of D-W is L2 ´ V and the standard deviation is L ´ V1/2. This relationship also explains why volatility can materially negatively affect the growth of wealth when L is high, because it is the variance of returns that acts as a drag on the compounding of returns over time. The compounding return may be roughly approximated as the average return less half of the variance. Following is a simple example of how the leveraged effect works for an L of 3.

Shortfall constraints (i.e., avoiding disaster) are at the root of this approach. A total portfolio loss of 10 percent becomes magnified for an investor with an L of 5´ (meaning the investor cannot lose more than 20 percent of total assets without wiping out all D-W), because relative to D-W it is a 50 percent loss.

When assets are extremely large relative to the lifestyle annuity (e.g., for the mega-rich, where they spend very little relative to their total wealth — think of Warren Buffett), these relationships become asymptotic with the simple case of just maximizing long-term sustainable growth of the entire portfolio of assets. By contrast, when L is high, you have to become very conservative because as L increases, risk goes up much faster than expected return. By tracking L, you can adjust the riskiness of your investment assets to your current and remaining future total circumstances.

Just as an insurance company properly constructs its investment portfolio so as to grow its surplus while protecting against future claims, an individual will usually do better by taking into account the risks and opportunities for return of a widely diversified investment portfolio. Future consumption does not have to be exactly immunized with a matching asset or special fund. Most often, investors will be better off with a very diversified portfolio composed of multiple assets classes (when possible) such that the risk–return profile of the total portfolio addresses both the future consumption and the D-W components as a whole. Partitioning assets into two categories or buckets, one of low-risk and one of high-risk assets, may be easier but may overlook opportunities; also, in all likelihood, their relative proportions should change with changing circumstances.

If increases in discretionary wealth do not lead to increased spending, then following the discretionary wealth approach tends to maximize long-term median D-W and consequently total wealth.

The discretionary wealth framework has other advantages as well:

• It can be adapted fairly easily to nonnormal return distributions by directly incorporating higher moments, such as skewness and kurtosis.
• It can be adapted to incorporate more than just the one category of lifestyle consumption so that there is a hierarchy of spending priorities and shortfall constraints (e.g., maintaining a private jet may or may not be essential to one’s lifestyle). See Wilcox, Horvitz, and DiBartolomeo (2006, pp. 27–28).

SUGGESTED READING
Wilcox, J. 2003. “Harry Markowitz and the Discretionary Wealth Hypothesis.” Journal of Portfolio Management (Spring):58–65.

Wilcox, J. 2008. “The Impact of Uncertain Commitments.” Journal of Wealth Management (Winter):40–47.

Wilcox, J., J. Horvitz, and D. diBartolomeo. 2006. Investment Management for Taxable Private Investors. Charlottesville, VA: Research Foundation of CFA Institute.