Dr. David Blanchett, Head of Retirement Research at PGIM DC Solutions, has been recently active on LinkedIn, advocating his 2022 paper “Redefining the Optimal Retirement Income Strategy.” In this paper, Dr. Blanchett suggests several changes to traditional Monte Carlo models currently used by many financial advisors. These changes include:
- Inclusion of a dynamic process, similar conceptually to the Actuarial Approach advocated in this website, to determine how retiree spending should be adjusted from year to year, and
- Development of a better metric for evaluating scenario results than the traditional “probability of success” metric.
Instead of developing the probability of success for a household with a spending goal of $X per year for a fixed period of years, Dr. Blanchett’s model develops a “Goal Completion Score” for a household with a spending goal of $X per year (with Y% considered “Needs” and (100%-Y%) considered “Wants”) for a fixed period of years, but with such household agreeing to increase or decrease future spending in a future year “t” based on their Funded Status at time t (the present value of their assets at time t divided by the present value of their spending liabilities at time t) and other adjustment rules (algorithms). It is anticipated that the spending adjustments “baked into” Dr. Blanchett’s Monte Carlo model would affect household “wants” spending first.
Dr. Blanchett’s dynamic algorithm for changing spending in retirement from year to year is similar to the approach anticipated by the Actuarial Approach outlined in this website in that both approaches anticipate:
- an annual comparison of the present value of household assets and household spending liabilities to determine the household’s Funded Status (which Dr. Blanchett calls the Funded Ratio),
- establishment of separate funding buckets and asset/liability comparisons for essential and discretionary spending (which Dr. Blanchett calls needs and wants), and
- mapping of assets/investments to the two buckets using Liability Driven Investing (LDI) theories.
Dr. Blanchett’s dynamic spending approach is simpler (for calculation simplicity purposes) than the Actuarial Approach (Actuarial Financial Planner model and annual valuation process) for a number of reasons, including
- It doesn’t anticipate non-recurring expenses (e.g., long-term care, new car purchases, family assistance, etc.)
- Needs and wants are assumed to be constant percentages of total spending each year
- It uses a fixed period of years for the household lifetime with no adjustments upon the first anticipated death within the couple
- It doesn’t anticipate different rates of future increases for different types of assets or expense liabilities
Our quick assessment of Dr. Blanchett’s proposed changes
Readers of our blog know that we aren’t big fans of Monte Carlo modeling when it comes to ongoing financial planning during retirement. Generally, these models are more consistent with one-and-done (static) types of analyses and can be quite useful for facilitating certain decisions, such as developing or changing an investment strategy. Having said that, we believe the changes proposed by Dr. Blanchett are a step in the right direction. We might even go so far as to agree that his approach could provide value at the onset of a household’s retirement by quantifying the range and likelihood of future spending adjustments or other risks (and would be even better if it used our calculated Funded Status measure). However, since his approach has spending adjustments based on future calculations of household Funded Status baked in, we wonder how useful his approach would be for years subsequent to the initial application when the household is expected to simply apply a spending adjustment algorithm (like our Funding Status guardrails) to determine spending for that future year.
Conclusion
We believe that the process used to adjust spending from year to year is more important than the model used to calculate a Household’s Funded Status or project other results. While our Actuarial Financial Planner (AFP) may not be the most sophisticated model around, it is more robust in determining Funded Status than the approach anticipated by Dr. Blanchett. In addition, we believe periodically stress-testing the assumptions used in the AFP can provide users as good, if not better, risk information as may be obtained with Monte Carlo modeling (either traditional or Dr. Blanchett’s improved version). And, for fans of Occam’s Razor out there, the Actuarial Approach provides a much simpler and easier to understand solution.
