Beware the Crystal Ball

Journal of Financial Planning; June 2014


Jonathan Guyton, CFP®, is principal of Cornerstone Wealth Advisors Inc., a holistic financial planning and wealth management firm in Edina, Minnesota. He is a researcher, mentor, author, and frequent national speaker on retirement planning and asset distribution strategies.

No doubt you get this client question at least once a week: “So, what is the market going to do the rest of the year?” If you’re like me, you have several well-worn responses. Some may even be laced with humor to make your point. “Well, we just happen to keep our crystal ball right over there in that credenza,” I often say while looking over my shoulder at a certain cabinet door. A short pause accompanied by a smile usually completes the message I wish to send.

Financial planners have mostly learned not to fall for the question. We realize the harm we can cause if we are wrong in either direction. As planner and fellow Journal columnist Ross Levin once wrote, “The more confidence we have in our forecasts, the more likely we are to be wrong about them.

Even if we have thoughtful views on the subject, we frame them in the context that our clients’ ability to achieve their goals neither depends on nor requires us to make such predictions. Yet, I am concerned that as a profession we are doing just that. Unfortunately, I see signs that subtly (and perhaps unwittingly) promote the view that before you can offer advice, you must first render a prediction about the future. 

Don’t fall for it. At the very least, keep your eyes wide open for those moments when you are being tempted to do so, then proceed with caution in light of that awareness.

Before describing what such temptations look like, I should describe what I consider to be the three frames of reference that financial planners can use when analyzing a client’s situation and making recommendations: backward-looking, forward-looking, and present situation.

Backward-Looking Perspective

Simply put, a backward-looking approach makes the key assumption that the past is a reasonable proxy for the future. What worked even in the most challenging moments of the past (sufficient savings rates, return and risk assumptions, safe withdrawal rates, etc.) will work for clients in the future. This approach has its advantages. It is grounded in fact. It seems both prudent and defensible. And by definition, it avoids having to decide whether “this time, it’s different.”

One of its chief disadvantages, in my view, is its inability to assess—soon enough, at least— when a client might be in trouble relative to their goals. In its most rigid form, “in trouble” simply does not compute, because the adviser’s recommendations have always worked before.

Its other disadvantage (to clients, at least) is that it usually tends toward advice most appropriate for a worst-case scenario—only when the conditions (market valuation level, inflation rate, propensity of client to under-save or over-spend, etc.) replicate the highest historical barriers to success. You might be surprised how many planners always recommend the same asset allocation or withdrawal rate for a retiree.

Forward-Looking Perspective

If a historically based advisory perspective seems unequal to the task or inappropriate for the present conditions (if one finds oneself saying, “This time, however, it is different”), then we can easily find ourselves in a forward-looking frame of reference, peering into the future; trying to imagine it. We soon find ourselves in prediction mode, whether or not we claim to have a crystal ball behind a certain cabinet door in our conference room.

How might this happen? Clearly, there are always more than enough prognosticators ready to share the rationales behind their predictions. At least one widely used provider of financial planning software uses its own assumptions about the future—returns, inflation, volatility, etc.—as capital market inputs, thereby shifting its results from being historically based. Its stochastic calculation engine, which by definition already models outcomes under conditions that vary from the historical norm, also contains an additional variable: the impact of the ‘leanings’ implicit in its current assumptions. Such results, and the advice planners render on their basis, are in part dependent on the accuracy of its predictions of how, this time, it’s different.

My point is not that this is necessarily harmful. Skillful predictions that accurately anticipate how the future will differ from the past will make any advice more valuable. The opposite, of course, is also true. In the wake of the Great Recession’s bear market, how often were predictions made that future returns would now be noticeably lower than historical returns—even as the Journal of Financial Planning published Scott Leonard’s historically based paper, “U.S. Equity Returns after Major Market Crashes,” in October 2009 citing strong evidence of just the opposite? Guess which, thus far, is the more accurate.

Instead, my point is that in a forward-looking approach, our skill at predicting how the future will differ from the past is made into a significant determinant of our clients’ success, especially because such predicted differences are usually assumed to exist throughout our clients’ financial lives. Such an approach forces thoughtful financial planners to first ask whether they agree with the predicted differences before they can determine whether to base their advice on it.

Research has started to test this hypothesis under several sets of capital market expectations, with one usually being long-term, historically based data and the other(s) illustrating different assumptions about the future. There can be a real advantage to this, because a hypothesis that works under a variety of conditions is inherently more appealing.

However, an unintended consequence is to require a practitioner—at least a practitioner who wants to use its findings as a basis for her advice—to first choose the future capital market assumptions on which she will base that advice. And the differences, or volatility, in that advice can be quite significant to a client. 

Consider the situation of the new retiree at the start of 2009. His portfolio value declined 25 percent in the fall 2008 market meltdown to $900,000. How much can (should) he withdraw in yearly distributions assuming a 50 percent to 60 percent equity allocation, a 90 percent success rate, a yearly inflation-adjusted increase, and no other adjustments in any subsequent year? According to Bengen’s 1996 historically based analysis, $40,500 (4.5 percent). Or, according to a widely used planning software’s below-norm real return assumptions from that time, $29,700 (3.3 percent). Based on Kitces’ 2008 historically based approach that sets the initial withdrawal rate based on the starting market valuation, $45,000 (5.0 percent). What a difference!

The first answer said, “The future will basically resemble the past.” The second one was based on the prediction that the future will be very different from the past. Of course, its answer would have undoubtedly been different under a different prediction. The third said, “Where we are today, relative to past history, matters.” A forward-looking approach, particularly one using auto-pilot analytics and centered on a predicted future, adds volatility and uncertainty to our advice.

Policy-Based Approach

The third approach has no need to ask whether the future will be different from the past. It relies on regular observations of present conditions and circumstances. At regular intervals, it evaluates whether or not the present situation warrants any adjustments to what was previously advised. This is the policy-based approach. In the words of planners Elissa Buie and Dave Yeske, (in their July 2006 Journal article “Policy-Based Financial Planning Provides Touchstone in a Turbulent World”) a policy is broad enough to encompass new or unexpected events, yet specific enough to minimize doubt as to what action to take in response.

A policy-based approach is dynamic. It expects to make mid-course adjustments. (Clearly, in any area where a client cannot or will not accept such flexibility, such adjustments must not be presumed.) It is based on underlying analytics that assume today’s adjustment will be re-tested for possible further adjustment in another year, and again a year later. It does not need to ask if or how the future might differ from the past. And if the future does significantly differ from the past, the policy-based “action taken in response,” adjustments modify the plan to the trajectory it should have taken could this particular future have only been predicted.

Policies are not new to financial planning. Policy-based approaches to investment management have existed for several decades. In distribution planning, the two areas most powerfully affected when governed by dynamic policies are the portfolio’s equity allocation and the annual withdrawal amount. In the last 10 years, the impact of such policy-based approaches has been persuasively demonstrated in articles by Bill Bengen, William Klinger, David Blanchett, Michael Kitces, Larry Frank, David Zolt, Wade Pfau, Neera Gupta, and others.

How much of a difference can a policy-based approach make? Let’s revisit that 2009 retiree who wants to know his sustainable withdrawal amount after seeing his balanced, well-diversified portfolio decline 25 percent or $300,000 to $900,000. Using the 2009 dynamic allocation policies of Kitces, his sustainable withdrawal amount is $46,800 (5.2 percent). Based on the 2006 dynamic withdrawal policies of Guyton/Klinger, it’s $49,500 (5.5 percent). However, a careful reading and combined application of these policy sets strongly suggests it to be $54,000 (6.0 percent), if not slightly higher.

So, the next time you find yourself wondering if you need to make a crystal-ball prediction before you can soundly advise your client, just say no. The risk is too great, the potential opportunity cost too high, and the payoff too small. The present, and the evidence-based policies that take advantage of it, can tell you all you need to know.