New Research On Social Security Claiming Strategies
I’ve written numerous articles over the years about strategies for maximizing your Social Security (SS) benefits. But the number one decision all retirees face is when to start taking payments. Which is better: to start as early as possible or to delay benefits for as long as you can?
There will always be claimants who need the money now or who have medical conditions suggesting an early death. For them the choice is clear: start as soon as possible. For everyone else it’s more complicated. Those who do not trust the federal government’s commitment to continue to fund SS and/or to keep payments in line with inflation will start as soon as possible and collect as much as they can before benefits are taken away or the SS trust fund goes bankrupt. Those who recognize SS as government insurance providing lifetime support for retirees regardless of longevity will choose to wait as long as possible in order to maximize longer-term inflation-adjusted SS monthly payments in case they live an exceptionally long life. Neither of these choices requires any sort of complex analysis.
The third group is more interesting. They view SS as an investment opportunity, turning the decision into an analytical maximization exercise (i.e. which choice will generate the greatest lifetime income). They not only use the breakeven age in their calculations but additionally attempt to take into account the return on investment (ROI) of the earlier SS payments (assuming they don’t need to use the money at the time) as well as other factors such as future inflation rates and tax rates.
Wade Pfau and Steve Parrish at The American College of Financial Services recently completed two case studies to attempt to quantify which starting age (62, 67/FRA, or 70) results in the greatest legacy value for retirees dying at age 95. In both cases they created a financial plan based on an investment portfolio involving multiple account types (taxable, IRA, and Roth) and a spending model which is highest during the earlier retirement years but with spikes in later years due to long-term care and other health issues. For each starting age they simulated 34-year results from three allocation models (25%, 50%, and 75% large-cap stock allocations with ten-year constant-maturity treasuries making up the fixed income portion) drawn from historical data going back to 1871. Taxes, IRMAA, and SS benefits were based on 2022 rules. If SS benefits and RMDs from the IRA exceeded the expenses and taxes in any given year, the surplus was added as new savings to the taxable account.
Case 1 was a single individual with a $1 million portfolio. Pfau & Parrish found that for all allocation models, claiming at age 67 resulted in greater after-tax post-death portfolio valuations as compared to claiming at age 62. For the 25% stock allocation level the later claiming was better 94% of the time, 82% of the time for the 50% allocation, and 67% for the higher-risk portfolios (75% stocks). Interestingly, delaying until age 70 was about 5 percentage points less beneficial than claiming at age 67 with all allocations, but still significantly better than starting SS at age 62.
The second case was a married couple (same age) with $3 million in assets. Again, claiming at age 67 was superior to doing it at age 62 (96% of the time for the 25% stock allocation, 76% for the 50% allocation model, and 62% for the 75% stock model). Even more interesting, delaying until age 70 provided the greatest legacy value for the married couple utilizing a 50% allocation.
The researchers’ conclusions: it’s possible to achieve a better outcome by claiming early, but it is not common. In fact, beating the benefit of delaying SS would require a high tolerance for risk together with an aggressive asset allocation as well as a significant amount of discretionary wealth.
The study can be found in the January 2023 issue of the Journal of Financial Planning.
FYI here’s my summary (not analysis) of ALL the reasons I’ve come across to start or defer:
This is as brief as I could reasonably make it, yet still rather long, demonstrating that there are a LOT of possibly relevant personal variables!