Could the Government Take the Bite Out of RMDs?

Affluent retirees love to complain about required minimum distributions, and it's easy to understand why: No one likes a requirement to take money they do not need out of their 401(k) and IRA accounts. That's particularly true for retirees who are bumped into higher marginal tax rates due to these withdrawals, especially when these RMDs increase the percentage of Social Security benefits that are subject to taxation.

My Morningstar colleagues have written about this over the years and have some strategies for how to minimize the RMD bite. Even with these strategies, there is no way to avoid the fact that RMDs are the other side of the tax-deferred account bargain. Savers enjoyed tax benefits when they contributed and as their accounts (hopefully) grew, in retirement the government would like to collect deferred part of the tax.

But people in government have been rethinking this bargain; the executive and legislative branches have been floating ideas to reduce the RMD bite recently.

The changes that are the most likely to happen are the least likely to be meaningful enough for retirees to even notice. However, a legislative proposal from Sen. Rob Portman and Sen. Ben Cardin would dramatically change RMDs for retirees over the next decade, although it might also lead to future RMD shocks.

First, let's quickly review the RMD rules to see what might need adjusting. When the IRS introduced changes to RMDs for the 2003 tax year, they were a real improvement over the status quo because they dramatically simplified the RMD calculation. Whereas RMD amounts had been based on the actual joint life expectancy of the account owner and his or her spouse, the new rules directed most retirees to simply look up their age in the Uniform Lifetime Table to determine how much they needed to take out. (People with a spouse 10 or more years younger use a different table.) The IRS designed the table to spread 401(k) and IRA withdrawals out over a couple's lifetime, based on mortality data from the year 2000.

One obvious thing to do would be to adjust the mortality assumptions that drive these RMDs, because they are going on 20 years old. In fact, President Donald Trump directed the Treasury Department to do just that, and when the government reopens, it will likely propose some adjustments. However, while life expectancy has increased, my back-of-the-envelope calculations reveal that such a change would not make much difference.

The reason an adjustment, while modestly helpful for retirees, would not make a big difference is that the current rules already make a very generous assumption that a spousal beneficiary is 10 years younger than the current account holder, regardless of his or her actual age. This means the RMDs are already based on long life expectancies for at least one member of a couple. A slight increase in these factors for somewhat increased life expectancy since 2000 would only slightly reduce the percentage retirees need to take out of their accounts. Not only that, the actual benefit to retirees would just be their marginal tax rate multiplied by the decrease in the RMD, or just a fraction of a fraction.

For example, the required distribution at age 70 1/2 might fall about .19% based on my adjustment to the mortality table. Retirees with a well-funded, $1 million IRA would need to withdraw about $1,900 less and would save just $418 in taxes, assuming they were in the 22% bracket. Certainly, that's better than nothing, but it's not a huge change even for a fairly affluent retiree.

In contrast, Portman and Cardin have proposed legislation that would make a big change to RMDs by pushing them to age 72 in 2023 and then to age 75 in 2029, up from age 70 1/2 today. At first glance this seems like it would be great for retirees because it would help them preserve their assets for longer and shield them from taking RMDs during market downturns. However, it would also mean that when RMDs start, they would be a much higher percentage of a retiree's account than they are today because they would spread the remaining distributions over a shorter expected lifespan.

This change would also benefit more affluent retirees because they have longer life expectancies, tend not to spend IRAs at all before RMDs kick in, and have higher retirement account balances that are more likely to push them into a higher tax bracket when they begin taking their RMD. While doing so, this proposal would cost the government quite a bit in forgone revenue as the new age phased in, which might make it unpalatable.

Policymakers may ultimately conclude that given Roth accounts and conversions, they have already given savers and new retirees some flexibility regarding RMDs. Roth accounts, of course, do not have an RMD requirement because people contribute to a Roth with aftertax dollars. It is true that the cohort of people who are already retired did not have the chance to use Roths in a meaningful way because they were not created until 1997 and took some time to gain in popularity; however, proposals that phase in a later RMD will not benefit this current cohort anyway.

The bottom line is that RMDs really are the flip side of the tax deferral that savers enjoy before they reach retirement (or at least old age), and there's not much policymakers can do to make that obligation disappear without forgoing a lot of revenue. All they can do is shift the tax collection slightly. If it is any comfort, the taxes paid on withdrawals are at the effective tax rate, which usually lower than the marginal tax rate retirees avoided by making contributions to tax-deferred accounts during their working lives. And savers and retirees can manage their RMDs by using Roth vehicles along with traditional tax-deferred accounts.