The blended risk score averages away the binding constraint

A risk questionnaire does two jobs with one output: it asks about time horizon, income needs, and how the client feels about a bad quarter, converts the answers into a score, and the score picks the portfolio. Kitces argued in a January 2017 Nerd's Eye View post that the conversion is where the trouble sits. The score averages two quantities, willingness to take investment risk and financial capacity to bear it, and averaging happens to be the operation that erases whichever of the two would have capped the allocation.

The tail cases are where that shows up: a household with almost no appetite for losses and ample assets, or one with a high stated tolerance and a thin balance sheet heading into a spending year. Pool the two dimensions and the strong score lifts the weak one, so the funded portfolio can be far too risky for the situation. Kitces is explicit that advisors arrive there unwittingly, and the fault line in his telling runs through the instrument rather than through the person using it. The low score, he writes, should have acted as a constraint on the investment policy statement, and did not.

His alternative is procedural: measure the two dimensions separately and score them on a two-dimensional scale in which each one both contributes to and limits the result, instead of collapsing both onto a single continuum. What that buys a portfolio is a ceiling that can be located. A composite score cannot tell an advisor whether the constraint was a stomach or a balance sheet. That distinction decides the recommendation, and no blend can supply it.

The tools Kitces points to are few: FinaMetrica and Riskalyze are described as standalone measures of what he calls pure risk tolerance, while Tolerisk was built to gather risk attitudes and basic financial goals separately and score them on two axes. For planners who run retirement projections for every household, the plan itself is the capacity measure, and the Monte Carlo probability of success is the number that carries it.

Choosing among those instruments is the easy part; the harder decision is a policy one and belongs in writing: which dimension gets to cap the allocation, and at what score. An advisory firm can license a two-dimensional assessment and still map money off a blended figure, which amounts to buying information and discarding the half that constrains the portfolio.

There is an operational reason the blend survives: one score maps to one model portfolio, one model maps to one rebalancing rule and one set of performance reports, and the workflow runs cleanly from there. Two scores force an advisor to resolve a conflict in the file, in front of the client, and document why the lower dimension won. That is more work than most questionnaire processes were built to absorb, and it is the work the post is asking for.

Capacity is the dimension that shifts, which matters more as households cross from saving to spending: willingness is a preference the client reports, capacity is a limit the plan computes, and that computation moves with markets, spending and time. The decumulation conversation starts with the reserve fund and the floor, not the rollover. The two-dimensional framework supplies the arithmetic for that position: the floor is the reason capacity deserves to cap the allocation, and the probability of success in the plan is where the floor appears as a number an advisor can point to.

Where the ceiling binds is a design choice with more than one answer: an advisor can honor a capacity cap in the equity weight, in the withdrawal rate, or in the share of the portfolio committed to a guaranteed floor. Each places the constraint somewhere different inside the account, and the client's stated tolerance does not decide which.

Willingness is not the disposable half of the pair. The client's gut is financial data, and a stated want is a clue, as this publication has argued about vacation homes; a household that says it can sit through a deep drawdown is describing something real about whether the relationship survives the next one. Collecting the preference is the easy half; the error is letting the questionnaire answer stand in for a limit the household may not have.

Kitces's own perch suggests where this convention travels if it travels at all. He is head of planning strategy at Focus Partners Wealth, a practice that per PWD's records spans $181.9 billion in regulatory assets across 220,585 accounts and 1,915 employees, and at Focus Partners Advisor Solutions, described in the post as a turnkey wealth management services provider to independent advisors. He co-founded XY Planning Network, AdvicePay, fpPathfinder and New Planner Recruiting, and hosted the Financial Advisor Success podcast. A scoring convention pushed through a turnkey provider and a planner network is likely to reach model portfolios faster than one argued only in journal pages.

The test is small and unglamorous: before the model is funded, an advisor should be able to name the dimension that binds and the score at which it binds. Where the number behind the portfolio is a blend, the constraint is an intention that did not make it onto the page, and a drawdown meeting is a bad place to introduce it.