As prospective retirees struggle to figure out how much money they need to accumulate in order to retire, a key assumption is what anticipated spending will be in retirement. After all, the more spending that must be supported, the more assets that may be necessary (in addition to other income sources) to support that spending.
Historically, a popular "rule of thumb" was to assume a replacement ratio of 70% to 80% in retirement, although in recent years this guidance has been lambasted by planners who suggest that client lifestyles tend to remain steady in retirement (or even increase in some cases), not decrease.
Yet in reality, it appears that planners have been mis-applying the replacement ratio research, which is based on a percentage of pre-retirement income, not pre-retirement spending! As a result, it turns out the 70% replacement ratio for moderately affluent clients may be remarkably accurate, and in fact could be too high for some wealth clients!
The inspiration for today's blog post are some of the recent conversations I've been having with fellow planners regarding reasonable spending assumptions in retirement, after my blog post from February about whether clients tend to reduce their spending in their later retirement years. During several of the conversations, the idea that clients may cut their spending in their later retirement years was compared to the 'unrealistic' assumption that clients cut their spending at the beginning of their retirement years from their pre-retirement spending, as embodied by the "replacement ratio" rule of thumb. Yet as I had to repeatedly point out, the replacement ratio research is about measuring income, not spending!
The Income Replacement Ratio Is Really A Spending Replacement Ratio!
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%.
Some would suggest that the aforementioned number might even be a bit high, as even maintaining the same general lifestyle in retirement might not entail the exact same expenditures across the board. For instance, without work, there tend to be fewer lunches eaten out, fewer miles driven (affecting everything from gas costs to car maintenance and repairs to the pricing on auto insurance), and less clothing replacement for work clothes (especially suits, where applicable!). Accordingly, the same home lifestyle might still cost less than it did during the working years. But assuming that perhaps a few other expenses go up as well - as there is more time to fill in retirement - the baseline is still approximately a 78% replacement ratio.
And in fact, the 78% income replacement ratio is exactly what the research has shown, in an ongoing retirement replacement ratio research study by Aon Consulting for the past 20 years, based on data from the Consumer Expenditure Survey! Notably, the replacement ratio is higher for those with lower incomes (who under our progressive tax structure, pay less in taxes, and therefore have less of a drop from pre-retirement gross income to post-retirement spending), and falls for those with higher incomes, although the study only goes up to a $90,000/year pre-retirement income level.
Lower Income Replacement Ratios As Income Rises
For many financial planning clients, who have incomes of $100,000+/year and face higher tax rates, and who also rely more heavily on personal assets for retirement and therefore have higher pre-retirement savings, the replacement ratio is likely some amount less than 78%, simply because there is a higher proportion of savings and taxes that don't need to be replaced in retirement. On the other hand, at these income levels, some additional replacement income becomes necessary for taxes in retirement, so even clients facing very high tax rates can only have their replacement ratios drop so far. On the other hand, for some clients that need to engage in significant savings in the final peak income years to make up for prior undersaving, it's possible the replacement ratio could be dramatically lower than 70%; in some extreme cases, clients save upwards of 20% or more of income (in addition to paying 20%-30% in taxes), leading to potential replacement ratios of only 50%.
Because of these variances in tax rates and especially savings rates for higher income and more affluent clients, it may be difficult to set a very clear and consistent replacement ratio rate (or at the least, it would likely need to vary by income, as it does at lower income levels in the Aon study). Nonetheless, some amount less than the 78% for those with $60,000 of pre-retirement income, such as 70% for those who are more affluent - seems a reasonable estimate.
But the bottom line is that, aside from especially high tax rates for very affluent clients, or exceptionally high saving rates for certain clients trying to complete the final pre-retirement stretch, the 70% replacement ratio (or 80% for clients of more moderate means) is likely a fairly good estimate of retirement needs - simply because it is likely a good estimate of take-home pay during the working years that is already being spend! Accordingly, as long as the replacement ratio is applied correctly, as a replacement ratio of income that constitutes level pre- and post-retirement spending, it can be a reasonable guideline for clients.
So what do you think? Have you ever talked about replacement ratios with clients? How much of your clients' pre-retirement income is actually spent in the final years before retirement? Do your clients spend about 70%-80% of their pre-retirement income? Does that mean a 70%-80% replacement ratio for income needs in retirement - to be fulfilled by some combination of personal assets and fixed income sources - is accurate for most of your clients?
(Editor's Note: This blog post was featured in the 12th edition of the Carnival of Retirement on Finance Product Reviews.)
Gregory Rogers says
Excellent commentary, as always. With a planning client, I would never feel comfortable using the 70-80% replacement ratio in a planning discussion, unless the topic is “rules of thumb”, or it’s a conversation to set up a more detailed/customized (and hopefully more accurate) approach.
I would think the planner needs to assess the current level of spending (at some level of detail that both the planner and the client are comfortable with), and then do some detailed work on what spending categories will change over time. This can get very subjective with younger clients, or those with younger kids, as the range of potential changes is wide. But it gets a bit easier with those closer to retirement. Potential income growth for younger clients may be extreme as well, so the “70% of what number?” question adds another challenge with the rule of thumb.
As long as the planner has a good handle on the likely changes in expense categories that the client can’t forsee (but we should know, based on experience, and much of the research you and others cite in your posts), I think that he/she can avoid relying on ratios that may or may not reflect the reality of each individual client.
As I don’t have a real good idea of how much I will actually spend in retirement, I looked at my spending (range, conservative to moderate) based on income minus savings and taxes (ignoring for now the payments for healthcare, as those will continue). I found my range was 68-79%. To me, that seem “close enough” to drive my savings total expectations prior to retiring. (Then of course, there’s the “How long I’ll live” question that muddies everything up, too, but staying conservative seems an appropriate approach.)
Don Heffernan says
I understand using income replacement estimates for middle income earners who don’t save but it seems fruitless for financial planning clients with larger, often dual, incomes. It is becoming very common for higher wage earners to save at 20% and higher rates. They also may be paying high taxes, mortgage payments and college tuition payments through most of their earning years. For such clients it seems to me that the only sensible way to plan for retirement is to carefully evaluate spending, not income.
Michael Kitces says
Ironically, then, the implication of your point is that 70% might be too HIGH of a replacement ratio for clients who are bigger savers?
Does that mean a replacement ratio approach can be effective as a “conservative baseline” that we modify from there?
Joe Alfonso, CFP says
I agree that the key issue is projected spending in retirement. Typically, clients nearing retirement have achieved the lifestyle they would like to maintain in retirement so the issue is to ensure that their savings, social security and pensions support that level of spend going forward. Factoring in projected taxes is important, including adjusting for the elimination of FICA taxes.
Rather than relying on a rule of thumb, I feel that looking at actual client spending in the pre-retirement years yields a more accurate assessment of the target savings needed for a successful retirement.
Michael Kitces says
I don’t disagree with the need for analyzing client-specific situations, but I’m still struck that often, we spend a LOT of time trying to narrow down this spending number… and basically find that if we just took gross income and subtracted taxes and savings, we’d be “awfully close”.
That doesn’t necessarily mean we need to just abandon trying for greater accuracy, but I have to wonder how much time we’re spending trying to get more precise about something that doesn’t need to be THAT precise anyway.
If it turns out that my client’s replacement ratio would have really been 71.147285% with a detailed analysis, I’m not sure the actual retirement advice would have been any different?
Or to generalize more broadly – how do we more effectively identify situations that might be materially different than a replacement ratio approach?
Retirement income replacement should = 100% of whatever the client is spending before retirement. If they’re saving 20% of their income and spending 70% (spending means everything, including taxes), then of course the replacement ratio would be 70%. Duh.
Michael Kitces says
Indeed, except in reality many advisors and academics have lambasted the rule without a clear understanding of the exact point you just made.
Thus, the purpose of the post. 🙂
The big cost that is being overlooked is health insurance costs – the part paid by the employer – and increased medical expenses.
Christine Parker says
Thank you, Michael. I always enjoy reading your posts. However, I’m unclear about the example here: $60,000/yr income before retirement is reduced to $47,000/yr of income in retirement, especially if the retiree is distributing from tax-deferred accounts such as a traditional IRA because they are still paying income taxes in retirement. If the retiree is distributing retirement income from tax-free Roth IRA’s, your example to give up “roughly $11,000/year” seems to be more applicable. Can you help me understand your point?
Michael Kitces says
The point is simply this: 70% of pre-retirement INCOME is a good estimate of post-retirement SPENDING needs.
The point is simply to answer the question: how much do your clients SPEND in retirement? And can it be answered without taking a microscope to their weekly, monthly, and annual cash flows, which can be very arduous – if not impossible – for both the client and the advisor.
How you generate those spending needs, the associated strategies, and the tax consequences of those strategies, is a separate (although obviously also relevant) issue.