Kitces: Preparing for Lower Long-Term Returns

[Gone4Good]

I note that for a median valuation portfolio, the safe withdrawal rate will be higher than 4%.

I think that is probably true, at least from a historic perspective.

I also think statements like "my withdrawal strategy is 95% successful" is only applicable if you're starting point is solidly within the middle of the data set.

 

[samclem]

I thought he was talking about 33% drop from a bit above today's levels (we are under 24% now), not 60%?

You are talking more about undershooting?

Sorry, I'm not following you. I was trying to say that if we are at a CAPE of approx 24X now and the median is about 16X, then if we had a plummet in equity values of 60% (from today's prices), we'd be at a CAPE of about 10. My stocks worth $500K today would (after the drop) be worth $200K. But, because of the new low CAPE of 10, G4G's method would have us "recalibrating" their value to 16/10 (or 1.6) = $320K. We'd take the "standard" 4% of that (or 5% if we follow your suggestion of using the median rather than the 4% "covers the worst case" WR). So, we'd be withdrawing $320K x .04 = $12.8K or $320K x .05 = $16K. In the real world, if my portfolio has gone from $500K to just $200K, and I withdraw $16K, that's 8% of what I've got, 8% of my beaten-down equity shares that I'm hoping may mean-revert to something close to their former value (or at least to the median 16x value) to avert a future, final dip of my portfolio to FIRECalc's dreaded X-axis.
Selling more (%age-wise) when share prices are in the dumpster seems counterintuitive. What seems logical is to shift allocations and increase bond allocations when equity prices are high, and then buy more equities (selling bonds to do it) when their prices are low (by historic standards). As far as withdrawals--keep 'em at a constant percentage (taken from year-end values). That seems like a good fit to my emotional preferences (tighten the belt in tough times) and to the requirement to preserve the future growth potential of the portfolio (don't cash out those beaten-down shares at a >higher< rate when the history of CAPE tells us they are ripe for a comeback in the near future).

 

[Gone4Good]

If we had a very big decline in equity prices (thus, a lower PE10 and lower valuations), Gone4Good's method would require (or at least allow) for a much higher withdrawal percentage that year--maybe 8% or more if we got down to CAPEs of about 10. Emotionally, after a drop in my stock portfolio of more than 60%, I don't think it would be easy (or even possible), to withdraw that much of the beaten-down equities unless it were truly a necessity (i.e. needed for the most Spartan essential spending). I believe in the utility of the CAPE, and I believe in mean reversion, but that would be quite a leap.

Some of that happens naturally. The portfolio declines, withdrawals stay the same, so the withdrawal percentage goes up in a down market. The worse the market, the larger the increase in WR.

Now if you've provisioned for that decline in portfolio value, then it's not that big of a deal. Your WR goes from 3% to 4% or 4.5% and that still feels OK.

Would anyone really increase their WR beyond that because return expectations have increased? Doubtful. But they might just feel comfortable enough with their financial situation to rebalance back into equities while in the teeth of the bear just like the unemotional FIRECalc bot assumes you will.

Some of us might even use the opportunity to increase our equity allocations beyond simple rebalancing.

 

[audreyh1]

Sorry, I'm not following you. I was trying to say that if we are at a CAPE of approx 24X now and the median is about 16X, then if we had a plummet in equity values of 60% (from today's prices), we'd be at a CAPE of about 10. My stocks worth $500K today would (after the drop) be worth $200K. But, because of the new low CAPE of 10, G4G's method would have us "recalibrating" their value to 16/10 (or 1.6) = $320K. We'd take the "standard" 4% of that (or 5% if we follow your suggestion of using the median rather than the 4% "covers the worst case" WR). So, we'd be withdrawing $320K x .04 = $12.8K or $320K x .05 = $16K. In the real world, if my portfolio has gone from $500K to just $200K, and I withdraw $16K, that's 8% of what I've got, 8% of my beaten-down equity shares that I'm hoping may mean-revert to something close to their former value (or at least to the median 16x value) to avert a future, final dip of my portfolio to FIRECalc's dreaded X-axis.
Selling more (%age-wise) when share prices are in the dumpster seems counterintuitive. What seems logical is to shift allocations and increase bond allocations when equity prices are high, and then buy more equities (selling bonds to do it) when their prices are low (by historic standards). As far as withdrawal rates--keep 'em at a constant percentage (taken form year-end values). That seems like a good fit to my emotional preferences (tighten the belt in tough times) and to the requirement to preserve the future growth potential of the portfolio (don't cash out those beaten-down shares at a >higher< rate when the history of CAPE tells us they are ripe for a comeback in the near future).

But that's how the system works with the % of initial portfolio value method. If the portfolio drops a lot right a year after you retire, your withdrawal in a given year be much higher percentage wise.

If you have several years of increasing portfolio value before the drop, then it might just be reverting back to where you started.

I didn't understand why you were using the scenario of the CAPE dropping to 10.

 

[Gone4Good]

I didn't understand why you were using the scenario of the CAPE dropping to 10.

I think he was trying to apply the logic in reverse.

I'd restate it this way, if we were currently at a CAPE of 10x, my suggestion is that a "safe" withdrawl rate would be something higher than 4%.

I think the confusion came in when trying to apply that logic to a portfolio that has been hit by a down market and suggesting it implies higher dollar withdrawals. I'm not sure it works that way.

If I was drawing 3% of a 100% equity portfolio valued at 24x, that same draw would be 4.7% if equities declined to 16x and would be a 7.4% draw if equities declined to 10x.

So yes, the approach says you can draw ~7% at 10x. It just doesn't envision you actually increasing the dollar amount of your withdrawals.

 

[samclem]

My approach is to discount current markets back to a level that represents the median valuation of the data set used to originally calculate your SWR. If you're using FIRECalc and PE-10, that would mean discounting the current equity market down to a valuation of about 16x from it's current 24x (a 33% haircut).

I'd also do the same thing with the bond market. 10-year treasuries are yielding 1.74% versus a median of 3.89%. Using an average bond market duration of about 6 years, that 200bp lower current yield results in a price discount of about 12%.

So assuming a 1MM portfolio and a 50/50 asset allocation, I'd mark the $500,000 in stocks down by 33% to $333,000 and the bond allocation down by 12% to $440,000.

My resulting $773,000 portfolio puts me right in the middle of the valuations used in our historic data set. Applying a 4% SWR, I get a withdrawal of about $31,000.

That means my undiscounted portfolio can support a withdrawal rate of 3.1%. Said another way, a 3.1% withdrawal today is equivalent to a 4% withdrawal at median valuations.

Okay,I see where I got off course, You aren't proposing doing this adjustment every year to adjust withdrawal rates based on changing valuations. This is a calculation you do one time at the start of retirement to determine the WR to be used forever.

Some of that happens naturally. The portfolio declines, withdrawals stay the same, so the withdrawal percentage goes up in a down market. The worse the market, the larger the increase in WR.

Yes, it doesn't work that way for those of us taking a %age of year end portfolio values. Now, though, I understand your explanation.

 

[Gone4Good]

You aren't proposing doing this adjustment every year to adjust withdrawal rates based on changing valuations.

Nope. It's just a sanity check that helps me keep these paper portolio gains in perspective.

We're not always as rich as we think we are. And that's never more true than after a seven year bull market.

 

[CaliforniaMan]

There is really something fishy about this. The best predictability is at 15 years. Really? I don't buy it. More likely this is an artifact of a limited data set. If we just look at a plot of the CAPE over time we see three near thirty year "cycles". Approximately 1901 to 1929, then from 1929 or 1936 to 1966, and then to 2000 or so. For whatever the cause, these quasi-periodic cycles would make it appear that there is some roughly 15 year predictability. I have wondered about this for some time and thought I might delve into the statistics at some point, but so far too lazy. Wish someone would.

This supposed 15 year predictability could be simply an artifact of these approximately 30 year "cycles," which to me seems far more likely than actually being able to predict something about the economy 15 years out.

 

[Gone4Good]

Since I use % of remaining portfolio as my withdrawal rate, I can't really project lowering the withdrawal % now in anticipation of a reversion to mean 10 years from now.

I would have to have a dynamic withdrawal % that dropped when valuations were high, went to a median level when valuations seemed to be median, and raised when valuations were low.

I haven't thought hard about a % of portfolio withdrawal method, but it seems to me a similar concept would apply.

All I'm doing, after all, is calibrating my initial withdrawal rate to the median market valuation. I'd want to do the same thing with my initial withdrawal rate for a percentage of portfolio approach too.

Say you want $40K in living expenses and have a portfolio currently valued at $1MM. If you started drawing 4% today and the market is really over valued you'll face long periods where you're withdrawal amounts are well below what you originally wanted to spend.

One way to guard against that is to mark down your current portfolio to a level reflecting "normal" valuations. If I use my previous calculation to arrive at a "fair value" for my $1MM portfolio of $773,000, then I might want to start with a 5.2% withdrawal rate rather than a 4% rate to make sure I'm still getting my $40K in income if markets mean revert.

Maybe I'm comfortable drawing 5.2% or maybe I'm not. Either way it seems to me this is a reasonable way to calibrate one's initial withdrawal amounts in an over valued market.

 

[Gatordoc50]

I think that is probably true, at least from a historic perspective.



I also think statements like "my withdrawal strategy is 95% successful" is only applicable if you're starting point is solidly within the middle of the data set.

+1
This is just common sense but I appreciate the way you have expressed it mathematically. Very helpful.

 

[Gone4Good]

This supposed 15 year predictability could be simply an artifact of these approximately 30 year "cycles," which to me seems far more likely than actually being able to predict something about the economy 15 years out.

I agree that it looks like we have a bunch of ~30 year valuation cycles, but that is from peak to peak (1901 to 1929, 1929 to 1966, 1966 to 2000).

The peak to trough, and trough to peak periods are closer to 15 years in duration. It is those periods where valuation meaningfully impacts investment returns.

For example, if we have a period where valuations start at a peak of 25x and end at a peak of 25x three decades later, your investment return over those three decades isn't influenced by changes in valuation at all. Both starting and ending valuations are the same. But your returns are hugely influenced in between the peaks when valuations may have declined form 25x to 14x and then rose again from 14x back to 25x.

So assuming a 30 year peak to peak cycle, valuations will be a more important driver of investment returns over some period shorter than 30 years.

 

[audreyh1]

I haven't thought hard about a % of portfolio withdrawal method, but it seems to me a similar concept would apply.

All I'm doing, after all, is calibrating my initial withdrawal rate to the median market valuation. I'd want to do the same thing with my initial withdrawal rate for a percentage of portfolio approach too.

Say you want $40K in living expenses and have a portfolio currently valued at $1MM. If you started drawing 4% today and the market is really over valued you'll face long periods where you're withdrawal amounts are well below what you originally wanted to spend.

One way to guard against that is to mark down your current portfolio to a level reflecting "normal" valuations. If I use my previous calculation to arrive at a "fair value" for my $1MM portfolio of $773,000, then I might want to start with a 5.2% withdrawal rate rather than a 4% rate to make sure I'm still getting my $40K in income if markets mean revert.

Maybe I'm comfortable drawing 5.2% or maybe I'm not. Either way it seems to me this is a reasonable way to calibrate one's initial withdrawal amounts in an over valued market.

I think I've kind of done this by having a ~3% withdrawal rate. The $ amount withdrawn has increased as the portfolio has gone up, and it will drop when the portfolio drops. Even if the portfolio were to drop 25%, it would still be a 3% withdrawal. Obviously income would drop 25%, but I can handle that as currently withdrawn income (after taxes) exceeds our budget by more than that.

I like that our excess income has built up over several years to have funds available even if our income is drastically cut. I have not been willing to reinvest excess, but keep it in short-term funds.

 

[Chuckanut]

So, maybe one of those variable withdrawal rate formulas is a practical answer to all of this confusion? One that buffers both the upside and downside withdrawal levels?

 

[Gone4Good]

I think I've kind of done this by having a ~3% withdrawal rate. The amount withdrawn has increased as the portfolio has gone up, and it will drop when the portfolio drops. Even if the portfolio were to drop 25%, it would still be a 3% withdrawal. Obviously income would drop 25%, but I can handle that as currently withdrawn income exceeds our budget by more than that.

I like that our excess income has built up over several years to have funds available even if our income is drastically cut. I have not been willing to reinvest excess, but keep it in short-term funds.

That makes perfect sense.

It'd be tough to implement if you retire directly into a bear market without the benefit of having some years of excess earnings accumulate. That's why it might be better for folks to start by drawing more money than they need right out of the gate to protect against a year-1 market collapse.

 

[audreyh1]

That makes perfect sense.

It'd be tough to implement if you retire directly into a bear market without the benefit of having some years of excess earnings accumulate. That's why it might be better for folks to start by drawing more money than they need right out of the gate to protect against a year-1 market collapse.

That's why I've been adamant about withdrawing the planned percentage even though the portfolio has increased a great deal in the past few years, and even when we don't spend it all in the same year. Because I know that any excess that is reinvested in the portfolio might go "poof" in a downdraft the next year. Theoretically I shouldn't need to reinvest the excess, because I'm already withdrawing at a low percentage, plus I'm prepared to take less income in $ when my portfolio does drop.

 

[Lsbcal]

That's why I've been adamant about withdrawing the planned percentage even though the portfolio has increased a great deal in the past few years, and even when we don't spend it all in the same year. Because I know that any excess that is reinvested in the portfolio might go "poof" in a downdraft the next year. Theoretically I shouldn't need to reinvest the excess, because I'm already withdrawing at a low percentage, plus I'm prepared to take less income in $ when my portfolio does drop.

This is part of the idea behind the VPW tool. One sets the AA and Last Withdrawal Age. Then one looks at some downturn's sequence of returns in really bad retirement years like 1929 (bear market + deflation) or 1968 (bear market + inflation).

The relatively high withdrawal rates (for us) can be put aside in a "poof" fund of unspent reserves. If the bad sequence shows up, the "poof" fund can be used to supplement spending until market recovery years which have always occurred ... so far. Our "poof" fund is in short term investment grade bonds right now.

 

[Gone4Good]

That's why I've been adamant about withdrawing the planned percentage even though the portfolio has increased a great deal in the past few years, and even when we don't spend it all in the same year. Because I know that any excess that is reinvested in the portfolio might go "poof" in a downdraft the next year. Theoretically I shouldn't need to reinvest the excess, because I'm already withdrawing at a low percentage, plus I'm prepared to take less income in $ when my portfolio does drop.

I like it!

In good years your portfolio naturally tilts toward more conservative investments and in down years it naturally tilts a bit more towards a riskier allocation.

It also has a built in warning system that lets you know when it's time to cut spending.

And, it has the unambiguous virtue of being 100% fail-proof, provided you don't fail to cut your spending as required to keep withdrawals within the stated fixed percentage.

Lots to like.

 

[samclem]

I like that our excess income has built up over several years to have funds available even if our income is drastically cut. I have not been willing to reinvest excess, but keep it in short-term funds.

I find this appealing, too. But, like a "buckets" approach, keeping the unspent "extra" money fenced in ST bonds (or cash) for eventual spending can be expected to produce lower returns than leaving the "excess" in the major pot to be rebalanced and buy more equities as they decline in price during "down" years. Maybe there's a way to do both: Leave the unspent surplus in the "big pot" being rebalanced (just like everything else, and for the same reason: Higher volatility-adjust returns due to rebalancing), but track (in your books) the value of what was left in, adjust its value each year by the same rate as the portfolio as a whole, and know it is available for withdrawal during lean years without violating the letter or spirit of the withdrawal rules set up on day one. It's "extra money" that you were "due" in previous years.

 

[CaliforniaMan]

I agree that it looks like we have a bunch of ~30 year valuation cycles, but that is from peak to peak (1901 to 1929, 1929 to 1966, 1966 to 2000).

The peak to trough, and trough to peak periods are closer to 15 years in duration. It is those periods where valuation meaningfully impacts investment returns.

For example, if we have a period where valuations start at a peak of 25x and end at a peak of 25x three decades later, your investment return over those three decades isn't influenced by changes in valuation at all. Both starting and ending valuations are the same. But your returns are hugely influenced in between the peaks when valuations may have declined form 25x to 14x and then rose again from 14x back to 25x.

So assuming a 30 year peak to peak cycle, valuations will be a more important driver of investment returns over some period shorter than 30 years.

I didn't mean to suggest that I think there actually are 30 year economic cycles, only that for whatever chance reason they occurred in the past in this time series, and that due to these "cycles" (note the quotes) the 15 year correlations found are probably spurious and without meaning.

I think we may be reading something into the data that is not there. We think we see a 15 year out correlation when actually it can be completely explained by these meaningless past "cycles."

 

[Gone4Good]

I think we may be reading something into the data that is not there. We think we see a 15 year out correlation when actually it can be completely explained by these meaningless past "cycles."

Absolutely. We basically have something like 4 cycles, so not a robust data set from which to forecast.

And considering that it's been 15 years since the last cycle peaked in 2000 . . . "Yay, that means we're at trough valuations now." Not sure anyone believes that.

 

[audreyh1]

This is part of the idea behind the VPW tool. One sets the AA and Last Withdrawal Age. Then one looks at some downturn's sequence of returns in really bad retirement years like 1929 (bear market + deflation) or 1968 (bear market + inflation).

The relatively high withdrawal rates (for us) can be put aside in a "poof" fund of unspent reserves. If the bad sequence shows up, the "poof" fund can be used to supplement spending until market recovery years which have always occurred ... so far. Our "poof" fund is in short term investment grade bonds right now.

That's funny - a poof fund! It's more like an anti-poof fund.

Personally, I've been thinking of it as my war chest.

 

[audreyh1]

I like it!

In good years your portfolio naturally tilts toward more conservative investments and in down years it naturally tilts a bit more towards a riskier allocation.

It also has a built in warning system that lets you know when it's time to cut spending.

And, it has the unambiguous virtue of being 100% fail-proof, provided you don't fail to cut your spending as required to keep withdrawals within the stated fixed percentage.

Lots to like.

It's really simple too. All I have to do is calculate 3% of the Dec 31 portfolio value each year and remove that amount. I rebalance afterwards. That's it.

By the tilt I assume you mean selling from equities when they run up, and buying more equities when they correct - but any rebalancing does that.

The fail-safe - well that has to be qualified a bit. A few bad years in a row mean a combination of market drops and withdrawals. This can lead to a much lower portfolio and thus a drastic cut in income. But if you have excess funds set aside, you should be able to get through it without too much pain, and even have the fortitude to rebalance.

 

[Lsbcal]

I find this appealing, too. But, like a "buckets" approach, keeping the unspent "extra" money fenced in ST bonds (or cash) for eventual spending can be expected to produce lower returns than leaving the "excess" in the major pot to be rebalanced and buy more equities as they decline in price during "down" years. Maybe there's a way to do both: Leave the unspent surplus in the "big pot" being rebalanced (just like everything else, and for the same reason: Higher volatility-adjust returns due to rebalancing), but track (in your books) the value of what was left in, adjust its value each year by the same rate as the portfolio as a whole, and know it is available for withdrawal during lean years without violating the letter or spirit of the withdrawal rules set up on day one. It's "extra money" that you were "due" in previous years.

If I understand you correctly wouldn't this be exposing your designated spending money to the same portfolio risk as all the other dollars? For example, suppose we are in a 1968 sequence then the table below shows a $1M portfolio for a 50/50 AA and 3.4% spending. By year 8 the real spending amount (red in withdrawal column) is reduced a lot. This is a good year to take more spending money from our set-aside pot. But if that pot is exposed to the market, it would itself be down about 40% in real term (599990/1000000).

Probably the best way to avoid the set-aside lowering portfolio returns is to spend and enjoy your yearly allocation. Don't let that grow too big as it indicates (perhaps?) a failure of planning for fun. Does this make sense? I'm not entirely sure myself.

 

[Lsbcal]

That's funny - a poof fund! It's more like an anti-poof fund.

Personally, I've been thinking of it as my war chest.

Yes, a better name.

 

[audreyh1]

I find this appealing, too. But, like a "buckets" approach, keeping the unspent "extra" money fenced in ST bonds (or cash) for eventual spending can be expected to produce lower returns than leaving the "excess" in the major pot to be rebalanced and buy more equities as they decline in price during "down" years. Maybe there's a way to do both: Leave the unspent surplus in the "big pot" being rebalanced (just like everything else, and for the same reason: Higher volatility-adjust returns due to rebalancing), but track (in your books) the value of what was left in, adjust its value each year by the same rate as the portfolio as a whole, and know it is available for withdrawal during lean years without violating the letter or spirit of the withdrawal rules set up on day one. It's "extra money" that you were "due" in previous years.

I still have plenty of fixed income in the big pot for rebalancing. I don't see that as a problem or a constraint.

Reinvesting the excess will increase the very long-term returns, but I'm more interested in surviving the next 10 years in reasonable comfort and not so much having a high remainder value. I really don't want to subject withdrawn funds to market risk.

When CDs and high-yield savings accounts are yielding almost as much or more than short-term bond funds, or even intermediate treasuries for that matter, with no credit or interest rate risk, you aren't really being penalized on those short-term investments either.