Your Client’s Biggest Asset Isn’t in Their Portfolio

Your client’s largest asset is not in their portfolio. It is the income they have not yet earned.

That simple observation can upend a standard allocation conversation. A 30-year-old with $50,000 saved and $2 million in discounted future earnings does not really have a small balance sheet. They have a very large one, most of it tied up in future labor income. If that income is stable, bond-like, and only modestly correlated with the stock market, the client may already be far more conservative than their brokerage statement suggests.

Consider a 40-year-old business owner sitting across the table. He runs a manufacturing firm that nets $500,000 in a good year, has $1 million in a brokerage account, and just received a $250,000 payout from a former partnership. He wants to know how to invest it. His risk-tolerance questionnaire says “moderate.” The target-date fund in his 401(k) says 80/20. A model portfolio might place him somewhere in the middle.

None of those inputs asks the most important question about his financial life: What is the present value of the income he has not yet earned, and how risky is it?

That question sits at the center of lifecycle finance. Since the pioneering work of Robert Merton, lifecycle-finance theory has suggested that the relationship between financial wealth and human capital, not age alone, should be a primary determinant of portfolio allocation. If most of an investor’s wealth consists of future earnings that behave like a bond, the financial portfolio can rationally take more equity risk.

The challenge has always been implementation. Despite decades of academic support, advisers have lacked a practical way to incorporate human capital into portfolio decisions.

A new working paper by James Choi, Canyao Liu, and Pengcheng Liu, Practical Finance: An Approximate Solution to Lifecycle Portfolio Choice, helps close that gap.

The key takeaway for advisers is that portfolio allocation should be driven less by age-based rules of thumb and more by a client’s total economic balance sheet.

The authors show that their simplified approach captures nearly all the benefits of the theoretical optimum, with a welfare cost of less than 0.1% of lifetime consumption, compared with 2% to 4% for common age-based heuristics. In practical terms, the formula gets advisers surprisingly close to the theoretically optimal answer without requiring them to solve a full academic lifecycle model.

Human capital functions as an implicit asset on a client’s balance sheet. Valuing it requires a discount rate, and the discount rate for risky labor income is difficult to estimate because earnings can be disrupted by job loss, disability, business failure, or industry decline.

Choi et al. solve the full lifecycle problem numerically across thousands of parameter sets, then work backward to find age-varying discount rates. The result is a spreadsheet-friendly method that translates a client’s human capital into a personalized equity allocation.

The inputs are familiar: age, current wealth, expected wage path, and labor-income volatility. The output is an optimal equity allocation.

That matters because two investors of the same age and with the same portfolio value may have very different economic balance sheets. A tenured professor, a government employee, a commission-based salesperson, and a business owner may all be 40 years old with $1 million invested. But their future income streams do not carry the same risk, stability, or correlation with the market. A purely age-based allocation treats them as similar. A human-capital approach does not.

Permanent income, not last year’s income, drives allocation.

A business owner who had a rough year, but whose underlying economics remain sound, should not be treated like someone whose long-term trajectory has changed. Temporary income fluctuations have almost no effect on the optimal allocation. What matters is the volatility of the permanent component of income — the durable earning power that is expected to persist over time.

This is a distinction advisers often make instinctively in conversation but rarely formalize in the portfolio.

The model also suggests that many working-age investors may be underweight equities. In many cases, it pushes allocations to 100% equities during the accumulation years, even with conservative capital-market assumptions. It is not the return forecast doing the work. It is the sheer size of human capital relative to financial wealth.

The asymmetry is striking. At a risk aversion of four — a level the authors consider reasonable for many investors — holding zero equities for life costs 7.9% of lifetime welfare. Holding 100% equities costs just 0.56%.

In other words, the model is far more forgiving of holding too much equity during the accumulation years than of holding too little. For investors whose human capital is large, stable, and bond-like, the greater liability may not be equity exposure. It may be failing to take enough of it.

But when income is correlated with the market, the answer changes. A business owner whose revenue rises and falls with the economic cycle already carries implicit equity exposure through the business. That client should generally hold less stock than a government employee with identical financial wealth. The direction is intuitive; the formula’s contribution is putting a number on the adjustment.

The practical value of the paper is not that advisers should replace judgment with a formula. It is that they should ask better questions before assigning a portfolio. How large is the client’s human capital relative to their financial wealth? Is future income stable, cyclical, or uncertain? Does the client’s income behave more like a bond, an equity, or a concentrated private business interest? Is the client’s labor income correlated with the stock market? Has a recent income change altered the client’s permanent earning power, or was it merely temporary?

These questions move the allocation conversation beyond age, account balance, and a questionnaire score. They also help explain why two clients with the same risk-tolerance result may deserve very different portfolios.

The formula is useful, but it does not eliminate the need for judgment.

First, it treats risk aversion as a fixed, known input. In practice, it is neither. Choi himself has acknowledged that investors do not know their own risk aversion. The client who selected “aggressive” in February 2020 may be the same client calling to sell in March.

Second, the model treats gains and losses symmetrically. Advisers know that clients do not. Decades of behavioral research confirm what advisers see daily: losses are felt more intensely than equivalent gains. A retiree who enters a bear market in year one of distributions faces a fundamentally different problem than one who encounters the same drawdown in year fifteen of retirement. Early-sequence damage is largely irreversible. The formula has no mechanism for that asymmetry.

Third, the model solves for one simplified decision: equities versus a risk-free asset. Real portfolios are more complicated. They include intermediate bonds, TIPS, alternatives, tax constraints, asset location, spending needs, concentrated positions, and legacy goals.

These limitations do not make the formula irrelevant. They make the adviser more important.

The formula closes nearly all of the welfare gap between common rules of thumb and the theoretical optimum, for essentially zero computational cost. That is a meaningful advance for any practitioner willing to look past age as the primary input to allocation.

But the part the formula cannot reach is where advisory work has always lived. It cannot reconcile a client’s stated risk tolerance with their revealed risk tolerance. It cannot know whether a client will abandon an allocation during a bear market. It cannot distinguish between the portfolio that is optimal in theory and the portfolio a client can actually hold when theory collides with experience.

Human capital may be the missing variable in portfolio construction. Human behavior remains the missing variable in implementation. The formula can tell you what may be optimal. It cannot tell a client what they can live with. That gap is the work.