The 70% Replacement Rate In Retirement Is Rubbish
Originally published in Forbes.
Michael Kitces’ 2012 article, “In Defense of the 70% Replacement Rate,” examines the validity of the personal financial planning industry’s 70-80% replacement-rate target. Kitces points out that the income-replacement ratio is really a spending replacement ratio – the ratio of post-retirement spending to pre-retirement income.
Kitces provides this illustration:
For instance, assume a couple of moderate means that earns $60,000/year in the years before retirement. Such a couple gives up approximately 18% of their income to taxes (roughly $11,000/year), between Federal and state taxes (after deductions, which lowers their effective rate), and the employee share of FICA taxes. In addition, if we assume savings roughly consistent with the national savings rate over time – e.g., about 3%, or $2,000/year – we find that the couple’s take-home pay that was being spent was about $47,000/year. Which means if the couple wants to continue their current standard of living in retirement, they need to replace $47,000/year of pre-retirement spending in their retirement years… and generating $47,000/year of income in retirement, after generating $60,000/year of income before retirement, is a replacement ratio of 78%.
I’m a big fan of Kitces’s blog. He’s by far one of the top personal finance writers in the country. But I need to take him to the woodshed on the replacement rate.
I checked Kitces’s calculation by running his hypothetical couple through my company’s life-cycle financial planning tool, MaxiFi. MaxiFi determines precisely what the household should save (and, therefore, spend) each year rather than simply assume something like a 3% saving rate. MaxiFi’s annual spending suggestions are designed to keep the household’s living standard per household member stable through time.
I considered an age-50 couple with no children, with both spouses earning $30,000. I also assumed they would both continue to earn the $30,000 in today’s dollars through age 65 and have maximum ages of life of 100. In addition, I assumed they both began work at 20 earning $15,000, with these nominal earnings growing smoothly to $30,000 in the current year. And I assumed the couple owns a $200,000 house with a 20-year, $100,000 mortgage, including monthly payments of $600 and other annual housings costs of $3,000. Finally, I assumed the couple has $25,000 in regular assets and $100,000 each in a 401(k) account to which they and their employer each contribute 3% of pay and on which the couple earns a 2.5% real return.
In the base case, I have the couple retire at 65, and take their Social Security and retirement account withdrawals at 66. Their optimal savings rate at 50 is 4.11%, not the 3% that Kitces assumed.
The ratio of the program’s age-66 total spending (not including federal and state taxes or Medicare Part B premiums) to age-65 total income is 91.5% – higher than Kitces’ 78% replacement rate. Of course, 91.5% is a long way from 70%, the figure Kites is defending. Moreover, the correct replacement rate is hypersensitive to the couple’s precise circumstances, effectively making the 70% "rule of thumb” essentially useless.
Some examples …
If the couple has no retirement account assets, its replacement rate is 68.2%, which is miles below 91.5%. If each spouse’s initial retirement accounts are three times larger, the replacement rate jumps to 134.4%. If the couple’s maximum ages of life are 85, not 100, the replacement rate is 110.1%. If the spouses and their employers contribute 5% to their 401(k)s, not 3%, the replacement rate is 102.5%. If each spouse earns $20,000, not $30,000, the replacement rate is 135.1%. If each spouse earns $60,000, not $30,000, the replacement rate is 62.3%. If the spouses earn 1.5%, not 2.5% real, on their retirement account, the replacement rate is 82.6%. Finally, if the two spouses earn 3.5% real on their retirement accounts, the replacement rate is 92.7%.
Now assuming an advisor or a household on its own could correctly guess the couple’s correct replacement rate (which is truly impossible), what does the replacement rate mean operationally? That is, how does knowing the replacement rate help us decide what our client or we as individuals should accumulate in regular assets by retirement age?
The answer is not much.
One might think that with replacement-rate spending targets ranging from 68.2% to 135.1%, the couple should accumulate a lot of assets, which would throw off a lot of income to cover these seemingly high spending requirements. The answer in seven of the eight cases referenced above is that targeted retirement assets should be zero or very close to zero. In the event the couple has no 401(k) accounts, they should accumulate $205,043 (in today’s dollars) in regular assets by 65. The reason that the seven cases entail little or no asset holdings as of age 65 is that the couple’s Social Security benefits and retirement account withdrawals suffice to cover the couple’s annual spending, tax and Medicare Part B premium obligations.
As Kitces suggest, many if not most practitioners interpret the replacement rate rule of thumb as an income replacement rate. Interpreted in that light, this couple's replacement rate is not the vaunted 70%, nor the often higher 85% figure bandied about. Instead, it's zero or essentially zero in seven of the eight cases considered. There is, as indicated, no lost income to be replaced since the couple, except in one case, has enough Social Security and retirement account withdrawals to support and sustain their old-age spending. Even in the one case where significant asset accumulation is required, the income-replacement rate is extremely low and heads to zero over time.
A different operational question is what should the couple spend, in total, at age 50 under each situation. The answer ranges from $42,880 to $62,728. The retirement replacement rate in the former case is 135.1% and 91.3% in the latter case. No planner or household can, on their own, figure out the correct replacement rate at retirement age let alone convert the correct replacement rate into the correct current spending and saving recommendations.
I've been pushing economics-based personal financial planning for a quarter century when I started my software company. I believed then and continue to believe that conventional financial planning is asking the wrong questions, using the wrong methods, and is wholly unable to provide appropriate personal financial advice.
Economic Matters - The podcast is hosted by Laurence Kotlikoff and moderated by Alex Kotlikoff.
Laurence Kotlikoff is a Boston University Economist, a NY Times Best Selling Author, President of maxifi.com, and Author of Money Magic.
Stay Connected: Facebook, Twitter, LinkedIn
Hi, Michael's ok with a little back and forth. Pls call me at 617 834-2148 and explain what's complex and unhelpful. I can change it online in two secs. Unless one shows quantitatively the problem, it's just hot air. But let's talk. best, Larry
So, despite the introductory kudos to Kitces and lengthy analysis by you, gotta say that this article by you looked a lot like a private pissing contest between the two of you, and I found it unnecessarily complex and unhelpful. If you want to promote Maxifi, (which I use and do think is a very helpful planner) I suggest you bring down your commentary to a level that is more accessible to the rank and file and focus on the final outcome of your analysis, which I think (despite the obtuse approach) is that rules of thumb, including Kitces', are just that, and every economic situation calls for it's own specific evaluation of many factors. And then list the main ones. Or something. Not this.