Guided Spending Rates: Rethinking “Safe” Initial Withdrawal Rates
Guest Expert: David Blanchett, PhD, CFA, CFP®, PGIM DC Solutions
Date:
Webinar Replay Description
David Blanchett presented a fresh approach to estimating optimal portfolio withdrawal rates by incorporating spending flexibility and an outcomes metric that better reflected retiree sentiment.
Comments
A few questions asked by listeners with David’s reponses:
Bruce Z.asked:
"How about sharing with clients their probability of success and providing a range rather than a single figure? For example, if they have a 70% probability of success, showing them a 60-80% range?"
David's Answer: While this might be better than sharing a precise value (since it suggests estimation error), it could actually make things worse if the client gravitates to the lower end of the range (e.g., in the above example if the actual probability is 70%, but they are told 60%-80% , they may think it’s like 61%). I really think it’s best to not provide values whenever possible!
Stephanie M. asked:
"How would he incorporate 'sequencing risk'? More specifically, how does he handle modeling of poor returns early in retirement?"
David's Answer: sequence risk is going to be explicitly incorporated into any kind of stochastic (Monte Carlo) model, so definitely included!
Sivaparvati T. asked:
"Majority of my clients are high net worth clients with multiple sources of income. How would he modify the projections and risk modeling for them?"
David's Answer: Not sure I would change that much. I think longevity is going to be a bigger issue with high net worth clients, so making sure they have enough lifetime income is going to be essential to make sure they are comfortable spending.
Mike M. asked:
"On slide 36, what would happen to the guided spend amount if there were a large market decline?"
David's Answer: In theory, they would probably increase. If the market goes down, odds are the expected returns on assets are going to increase. That being said, the value of the portfolio would have declined, so it would be a higher percentage of a lower value, and these would likely offset to some extent.
Mike M. also asked:
"What are your thoughts about using a time segmentation or bucket strategy compared to a probability-based strategy?"
David's Answer: I think time segmentation is a great behavioral way to get retirees comfortable taking market risk. Note, I definitely don’t think there’s any kind of alpha benefit associated with segmenting a portfolio into buckets versus just using a single portfolio, but I do think to the extent using buckets makes a client more comfortable taking risk (more beta) would increase expected returns and likely improve expected outcomes.
Robert D. asked:
"Please discuss the idea of a much higher withdrawal rate in the early years of retirement when the retiree is more active."
David's Answer: it’s all about a “trade” when we think about withdrawal rates. If you take out 2%, there’s a really, really good chance you’re going to have a big balance later in retirement. Alternatively, if you take out 10%, there’s a pretty good chance you’ll have to cut back later. The key is understanding these trade-offs, especially in the context of how people usually spend (tends to decline in real terms during retirement, on average) and the fact most Americans have a decent amount of guaranteed lifetime income (like Social Security) that provides an important base should the portfolio become depleted. If I say the right initial withdrawal rate is closer to 5%, on average, for some folks they could do 8% knowing that on average it may work out, but if the markets move against them they’ll have to cut back on spending.
Gail asked:
"What tools are available for his theories?"
David's Answer: I don’t think there are that many. I think IncomeLab is a good tool that does dynamic withdrawals. I’m guessing there’s more coming in the future!
Bruce Z.asked:
"How about sharing with clients their probability of success and providing a range rather than a single figure? For example, if they have a 70% probability of success, showing them a 60-80% range?"
David's Answer: While this might be better than sharing a precise value (since it suggests estimation error), it could actually make things worse if the client gravitates to the lower end of the range (e.g., in the above example if the actual probability is 70%, but they are told 60%-80% , they may think it’s like 61%). I really think it’s best to not provide values whenever possible!
Stephanie M. asked:
"How would he incorporate 'sequencing risk'? More specifically, how does he handle modeling of poor returns early in retirement?"
David's Answer: sequence risk is going to be explicitly incorporated into any kind of stochastic (Monte Carlo) model, so definitely included!
Sivaparvati T. asked:
"Majority of my clients are high net worth clients with multiple sources of income. How would he modify the projections and risk modeling for them?"
David's Answer: Not sure I would change that much. I think longevity is going to be a bigger issue with high net worth clients, so making sure they have enough lifetime income is going to be essential to make sure they are comfortable spending.
Mike M. asked:
"On slide 36, what would happen to the guided spend amount if there were a large market decline?"
David's Answer: In theory, they would probably increase. If the market goes down, odds are the expected returns on assets are going to increase. That being said, the value of the portfolio would have declined, so it would be a higher percentage of a lower value, and these would likely offset to some extent.
Mike M. also asked:
"What are your thoughts about using a time segmentation or bucket strategy compared to a probability-based strategy?"
David's Answer: I think time segmentation is a great behavioral way to get retirees comfortable taking market risk. Note, I definitely don’t think there’s any kind of alpha benefit associated with segmenting a portfolio into buckets versus just using a single portfolio, but I do think to the extent using buckets makes a client more comfortable taking risk (more beta) would increase expected returns and likely improve expected outcomes.
Robert D. asked:
"Please discuss the idea of a much higher withdrawal rate in the early years of retirement when the retiree is more active."
David's Answer: it’s all about a “trade” when we think about withdrawal rates. If you take out 2%, there’s a really, really good chance you’re going to have a big balance later in retirement. Alternatively, if you take out 10%, there’s a pretty good chance you’ll have to cut back later. The key is understanding these trade-offs, especially in the context of how people usually spend (tends to decline in real terms during retirement, on average) and the fact most Americans have a decent amount of guaranteed lifetime income (like Social Security) that provides an important base should the portfolio become depleted. If I say the right initial withdrawal rate is closer to 5%, on average, for some folks they could do 8% knowing that on average it may work out, but if the markets move against them they’ll have to cut back on spending.
Gail asked:
"What tools are available for his theories?"
David's Answer: I don’t think there are that many. I think IncomeLab is a good tool that does dynamic withdrawals. I’m guessing there’s more coming in the future!
Guided Spending Rates: Rethinking “Safe” Initial Withdrawal Rates 05-02-2025
Attendees Comments:
Bruce Z.asked:
"How about sharing with clients their probability of success and providing a range rather than a single figure? For example, if they have a 70% probability of success, showing them a 60-80% range?"
David's Answer: While this might be better than sharing a precise value (since it suggests estimation error), it could actually make things worse if the client gravitates to the lower end of the range (e.g., in the above example if the actual probability is 70%, but they are told 60%-80% , they may think it’s like 61%). I really think it’s best to not provide values whenever possible!
Stephanie M. asked:
"How would he incorporate 'sequencing risk'? More specifically, how does he handle modeling of poor returns early in retirement?"
David's Answer: sequence risk is going to be explicitly incorporated into any kind of stochastic (Monte Carlo) model, so definitely included!
Sivaparvati T. asked:
"Majority of my clients are high net worth clients with multiple sources of income. How would he modify the projections and risk modeling for them?"
David's Answer: Not sure I would change that much. I think longevity is going to be a bigger issue with high net worth clients, so making sure they have enough lifetime income is going to be essential to make sure they are comfortable spending.
Mike M. asked:
"On slide 36, what would happen to the guided spend amount if there were a large market decline?"
David's Answer: In theory, they would probably increase. If the market goes down, odds are the expected returns on assets are going to increase. That being said, the value of the portfolio would have declined, so it would be a higher percentage of a lower value, and these would likely offset to some extent.
Mike M. also asked:
"What are your thoughts about using a time segmentation or bucket strategy compared to a probability-based strategy?"
David's Answer: I think time segmentation is a great behavioral way to get retirees comfortable taking market risk. Note, I definitely don’t think there’s any kind of alpha benefit associated with segmenting a portfolio into buckets versus just using a single portfolio, but I do think to the extent using buckets makes a client more comfortable taking risk (more beta) would increase expected returns and likely improve expected outcomes.
Robert D. asked:
"Please discuss the idea of a much higher withdrawal rate in the early years of retirement when the retiree is more active."
David's Answer: it’s all about a “trade” when we think about withdrawal rates. If you take out 2%, there’s a really, really good chance you’re going to have a big balance later in retirement. Alternatively, if you take out 10%, there’s a pretty good chance you’ll have to cut back later. The key is understanding these trade-offs, especially in the context of how people usually spend (tends to decline in real terms during retirement, on average) and the fact most Americans have a decent amount of guaranteed lifetime income (like Social Security) that provides an important base should the portfolio become depleted. If I say the right initial withdrawal rate is closer to 5%, on average, for some folks they could do 8% knowing that on average it may work out, but if the markets move against them they’ll have to cut back on spending.
Gail asked:
"What tools are available for his theories?"
David's Answer: I don’t think there are that many. I think IncomeLab is a good tool that does dynamic withdrawals. I’m guessing there’s more coming in the future!