Should we adjust the SWR based on last All Time High?
Should we adjust the SWR based on last All Time High?
Hey guys!
The SWR is a measure for a worst case scenario: Which withdrawal rate has always worked historically. An investor can start investing at any time and assume that the SWR will protect him from running out of money, even from bad start dates, like October 2007.
Then the rate should also be counted from the current All Time High. For example in March 2009, the investor starting in October 2007 (which was the All Time High at March 2009) had a drawdown of 50%. But still, his safe withdrawal amount is based on his assets at October 2007. If he had 100k in 2007, he will still be able to take out 3k in 2009 (if we use 3% SWR). His wealth at this point is only 50k. A person who starts out in 2009 will have a withdrawal amount of ony 1.5k with the same SWR applied.
Isn't it then better to adjust the SWR based on the current All Time High?
If the market currently is at 80% of ATH (down 20%): 3% / 0.8 = 3.75%
If the market currently is at 50% of ATH: 3% / 0.5 = 6%
You can even subtract for the withdrawal amount taken out since the ATH.
What do you think?
The SWR is a measure for a worst case scenario: Which withdrawal rate has always worked historically. An investor can start investing at any time and assume that the SWR will protect him from running out of money, even from bad start dates, like October 2007.
Then the rate should also be counted from the current All Time High. For example in March 2009, the investor starting in October 2007 (which was the All Time High at March 2009) had a drawdown of 50%. But still, his safe withdrawal amount is based on his assets at October 2007. If he had 100k in 2007, he will still be able to take out 3k in 2009 (if we use 3% SWR). His wealth at this point is only 50k. A person who starts out in 2009 will have a withdrawal amount of ony 1.5k with the same SWR applied.
Isn't it then better to adjust the SWR based on the current All Time High?
If the market currently is at 80% of ATH (down 20%): 3% / 0.8 = 3.75%
If the market currently is at 50% of ATH: 3% / 0.5 = 6%
You can even subtract for the withdrawal amount taken out since the ATH.
What do you think?
Re: Should we adjust the SWR based on last All Time High?
I think your math is wrong. Don't you mean ? -
If the market currently is at 80% of ATH (down 20%): 3% * 0.8 = 2.4%
If the market currently is at 50% of ATH: 3% * 0.5 =1.5%
If your balance has fallen, you don't want your WR to increase.
It seems like you're complicating this more than needed. An easier method would be to just take 3% of whatever your given balance is at out each month. So if your balance has shrunk by half, a 3% withdrawl on that would also have shrank by 50%. You don't need to set your withdrawl in stone once you've reached your big day. Let it be flexible.
If the market currently is at 80% of ATH (down 20%): 3% * 0.8 = 2.4%
If the market currently is at 50% of ATH: 3% * 0.5 =1.5%
If your balance has fallen, you don't want your WR to increase.
It seems like you're complicating this more than needed. An easier method would be to just take 3% of whatever your given balance is at out each month. So if your balance has shrunk by half, a 3% withdrawl on that would also have shrank by 50%. You don't need to set your withdrawl in stone once you've reached your big day. Let it be flexible.
Re: Should we adjust the SWR based on last All Time High?
No, I mean the SWR from the current assets should be adjusted based on ATH. Example: If you had 50k in 2009 you can use 6% of this, 3k, as your withdrawal amount. It was safe to withdraw 3% of the assets in 2007, and then it should be safe to withdraw 6% when the market is down 50% in 2009.
But isn't the point with safe withdrawal that you should be able to take out that *amount* forever? If you have saved 100k and apply a 3% SWR, then you should be able to take out 3k every year forever? Otherwise it will be wrong to say you are "retired" when you hit 33 times your annual spending? If the market crashes you can only be able to withdraw half of the money you need...
But isn't the point with safe withdrawal that you should be able to take out that *amount* forever? If you have saved 100k and apply a 3% SWR, then you should be able to take out 3k every year forever? Otherwise it will be wrong to say you are "retired" when you hit 33 times your annual spending? If the market crashes you can only be able to withdraw half of the money you need...
Re: Should we adjust the SWR based on last All Time High?
Interested to see what others say on this. I am taking an income approach so have given little thought to living on capital.
If I took a strategy of withdrawing capital, I'd keep a large cash fund (1-3 years, or whatever the average correction lasts) and keep my WR steady at 3%, making up any shortfall with the cash (or equivalent).
If I took a strategy of withdrawing capital, I'd keep a large cash fund (1-3 years, or whatever the average correction lasts) and keep my WR steady at 3%, making up any shortfall with the cash (or equivalent).
-
- Site Admin
- Posts: 15907
- Joined: Fri Jun 28, 2013 8:38 pm
- Location: USA, Zone 5b, Koppen Dfa, Elev. 620ft, Walkscore 77
- Contact:
Re: Should we adjust the SWR based on last All Time High?
The SWR based on a Trinity type study (which is what people usually do when they say "SWR") is taken from the time of pulling the trigger. It is NOT adjusted later.
However, as a matter of practicality. It seems wise to adjust spending, at least downwards, to reduce failure risk. (If you also adjust it upwards, it's an entirely different animal and 4% means nothing anymore. In that case, maybe use real return rates (3%) or corporate earnings growth (6%) minus inflation (3%) thus also arriving at 3%).
Also, it's increasingly worth crusading a bit against the blissfully ignorant optimism underlying the 4% rule as it's currently used by many noobs in the FIRE space. Some frequently ignored facts/potential game-stoppers:
However, as a matter of practicality. It seems wise to adjust spending, at least downwards, to reduce failure risk. (If you also adjust it upwards, it's an entirely different animal and 4% means nothing anymore. In that case, maybe use real return rates (3%) or corporate earnings growth (6%) minus inflation (3%) thus also arriving at 3%).
Also, it's increasingly worth crusading a bit against the blissfully ignorant optimism underlying the 4% rule as it's currently used by many noobs in the FIRE space. Some frequently ignored facts/potential game-stoppers:
- 4% was based on a 30 year period ... not the 50-75+ year periods assumed by many extreme early retires.
- 4% was based on US market performance, but I also see people using it in other countries where similar studies yielded lower SWRs in almost all of them.
- 4% was based on a 70 year dataset running from 1925-1995. This period contains a couple of unique transitions which can not be repeated such as retail/household stock ownership going from <10% to over 50%. Interest rates declining to 0%. None of these "boosts" can repeat.
- 4% results do not include the 16 year period from 2001 which has been flat---and which is the most recent one and perhaps more reflective of the 21st century than the far past. If it wasn't by global central banks spending $17T in helicopter money (1.25 years of GDP over a period of 5 years!) to pump up [mainly] the US stock market ... things wouldn't look so nice now .. and the experience of the last 8 years would have been much uglier. As it was, a lot of the recent experience has resembled the dotcom hype moreso than what's historically normal, except this time it wasn't dotcom stocks but the entire index. At one time the CME was actually offering a trading discount to foreign national banks if they wanted to engage in currency wars^H^H^H^H^H^H^Hdiversify their holdings.
Re: Should we adjust the SWR based on last All Time High?
@liberty -- Yes, what you're proposing is technically fine. Play with this Retirement Spending calculator for a while and you'll get a feel for a few different spending methods. The one you're referring to is called "Baseline %". But I'll just point out that just because you can spend more doesn't mean you must or should. Be smart about it. A dominant theme here at ERE is how spending =/= happiness.
@Jacob -- yeah, withdrawal rates are a lot more complicated than people realize. FWIW, my latest Withdrawal Rates calculator accounts for your 1st and 4th point, and I'm in the process of working on #2. But you're absolutely correct that over-reliance on a single number is not the most robust best plan.
@Jacob -- yeah, withdrawal rates are a lot more complicated than people realize. FWIW, my latest Withdrawal Rates calculator accounts for your 1st and 4th point, and I'm in the process of working on #2. But you're absolutely correct that over-reliance on a single number is not the most robust best plan.
-
- Posts: 3845
- Joined: Wed Apr 02, 2014 7:46 pm
Re: Should we adjust the SWR based on last All Time High?
I don't exactly remember the numbers, but even looking at fairly coarse data there have been an average of something like 5-6 "All Time Highs" per decade in the 30 years I've been investing in stock mutual funds. So it's pretty much a non-event. It's a great marketing ploy to sell people alternative assets (usually gold or insurance products) because stated in the right voice it sounds ominous.
The value of an SWR to me is for use as a planning tool. There are many variable withdrawal strategies out there, but as jacob notes the original Trinity Study assumed a fixed withdrawal adjusted for inflation. That isn't a very practical tactic.
There are many reasons one might adjust the amount they spend from their assets. Life should dictate that more than today's stock prices versus whatever the most recent peak is.
I'd suggest considering the use a "notional" SWR (modified downward from the generic 4% prescription for the reasons jacob points out) just to get a sense of roughly where your accumulation targets are, after taking a hard look at what you're likely to spend in the future and to recognize there can be a lot of variation year-to-year. It happens that I appear to be heading for an average of about 2%/year from age 55-80. I don't think that's as conservative as some people might think. The percentage will vary significantly some years, but based on life needs/desires, not market prices.
The value of an SWR to me is for use as a planning tool. There are many variable withdrawal strategies out there, but as jacob notes the original Trinity Study assumed a fixed withdrawal adjusted for inflation. That isn't a very practical tactic.
There are many reasons one might adjust the amount they spend from their assets. Life should dictate that more than today's stock prices versus whatever the most recent peak is.
I'd suggest considering the use a "notional" SWR (modified downward from the generic 4% prescription for the reasons jacob points out) just to get a sense of roughly where your accumulation targets are, after taking a hard look at what you're likely to spend in the future and to recognize there can be a lot of variation year-to-year. It happens that I appear to be heading for an average of about 2%/year from age 55-80. I don't think that's as conservative as some people might think. The percentage will vary significantly some years, but based on life needs/desires, not market prices.
-
- Site Admin
- Posts: 15907
- Joined: Fri Jun 28, 2013 8:38 pm
- Location: USA, Zone 5b, Koppen Dfa, Elev. 620ft, Walkscore 77
- Contact:
Re: Should we adjust the SWR based on last All Time High?
@T9000 - You probably already did this ... but if not. Plot the following.
x-axis: CAPE10 in starting years.
y-axis: life expectancy in years (you'll prob need to log this scale .. or just cut it off at 100 years .. when people have died.)
So, for year = 1875 to 2017 {addscatterplotpoint cape10(year),lifexp}
I'd use WR as a parameter, like do a 3% plot, a 4% plot, a 10% plot.
Might be interesting. The macro guy investor types have done plots like that for present (ttm) P/E vs subsequent returns. They show definite signal. You could in principle do this for all your portfolio runs.
PS: I realize that representative CAPEs for a given era is a lagging indicator where other factors likely have a big influence. Same reason why GDP/market cap ratios vary from country to country.
x-axis: CAPE10 in starting years.
y-axis: life expectancy in years (you'll prob need to log this scale .. or just cut it off at 100 years .. when people have died.)
So, for year = 1875 to 2017 {addscatterplotpoint cape10(year),lifexp}
I'd use WR as a parameter, like do a 3% plot, a 4% plot, a 10% plot.
Might be interesting. The macro guy investor types have done plots like that for present (ttm) P/E vs subsequent returns. They show definite signal. You could in principle do this for all your portfolio runs.
PS: I realize that representative CAPEs for a given era is a lagging indicator where other factors likely have a big influence. Same reason why GDP/market cap ratios vary from country to country.
-
- Site Admin
- Posts: 15907
- Joined: Fri Jun 28, 2013 8:38 pm
- Location: USA, Zone 5b, Koppen Dfa, Elev. 620ft, Walkscore 77
- Contact:
Re: Should we adjust the SWR based on last All Time High?
I'll summarize my take in fewer words.
Trinity (so-called 4% rule) should serve as an [historic] reality-check [on the plan]. Not as the foundation for the plan.
Trinity (so-called 4% rule) should serve as an [historic] reality-check [on the plan]. Not as the foundation for the plan.
-
- Posts: 257
- Joined: Sat Oct 27, 2012 4:48 pm
Re: Should we adjust the SWR based on last All Time High?
there is no Hard and fast "take it to the bank" number. The smartest way I've seen was an old lady I know who said, "each year we look at how things went and decide how much we can spend next year." annual review in other words. The swr number is for planning to decide how big the pile needs to be to get started at early retirement. (IMHO)
Re: Should we adjust the SWR based on last All Time High?
The ATH calculation seems to presume that you are 100% in some kind of stock fund, or close to it. You would need a different calculation for every portfolio.
Re: Should we adjust the SWR based on last All Time High?
brute would be surprised if all humans didn't do this anyway. who decided in 2009 to spend the full 4% and not adjust a bit for portfolio performance? it would likely take iron will to actually withdraw 4% in that circumstance.tommytebco wrote: ↑Wed Apr 19, 2017 4:08 pmThe smartest way I've seen was an old lady I know who said, "each year we look at how things went and decide how much we can spend next year."
-
- Posts: 3180
- Joined: Thu Jul 14, 2011 3:09 am
Re: Should we adjust the SWR based on last All Time High?
Wait, so the OPs goal was to DOUBLE withdrawal during market lows? For... Safety?
If I HAD to withdraw in 2009, I would be looking to sell whatever was doing least badly, and making a series of smaller withdrawals, rather than a full year's withdrawal. A strategy that would have worked in 2009, but probably not as well had the market not recovered so quickly.
Flexibility, for the win. In spending. In assets. In life.
Yes, I just channeled Ego for a minute. I'm all better now.
If I HAD to withdraw in 2009, I would be looking to sell whatever was doing least badly, and making a series of smaller withdrawals, rather than a full year's withdrawal. A strategy that would have worked in 2009, but probably not as well had the market not recovered so quickly.
Flexibility, for the win. In spending. In assets. In life.
Yes, I just channeled Ego for a minute. I'm all better now.
Re: Should we adjust the SWR based on last All Time High?
There's no getting better. You have been infected and are now in quarantine.Riggerjack wrote: ↑Wed Apr 19, 2017 8:47 pmYes, I just channeled Ego for a minute. I'm all better now.
Re: Should we adjust the SWR based on last All Time High?
Well, my point is to use the same withrawal amount as on market highs. If you have 100k on a market high, the withdrawal amount is 3k. Then the amount you safely can take out should still be 3k when the market is 50% down and you have 50k remaining. This means that you at this point can use 6% of your current portfolio (3% of what you had before the crash).Riggerjack wrote: ↑Wed Apr 19, 2017 8:47 pmWait, so the OPs goal was to DOUBLE withdrawal during market lows? For... Safety?
Of course it's safer to take out only 3% of the current portfolio, but then there is no point with a SWR...
-
- Posts: 3845
- Joined: Wed Apr 02, 2014 7:46 pm
Re: Should we adjust the SWR based on last All Time High?
Okay, what you're describing there is more-or-less the Trinity Study methodology where you make your withdrawal calculation at retirement and stick with it (the dollar amount, not the %) going forward. Well, except you adjust it every year for inflation. One of the "goals" in their choice of methodology was a scheme where what you might call spending power of the withdrawals is held constant.liberty wrote: ↑Thu Apr 20, 2017 1:47 amWell, my point is to use the same withrawal amount as on market highs. If you have 100k on a market high, the withdrawal amount is 3k. Then the amount you safely can take out should still be 3k when the market is 50% down and you have 50k remaining. This means that you at this point can use 6% of your current portfolio (3% of what you had before the crash).Riggerjack wrote: ↑Wed Apr 19, 2017 8:47 pmWait, so the OPs goal was to DOUBLE withdrawal during market lows? For... Safety?
Of course it's safer to take out only 3% of the current portfolio, but then there is no point with a SWR...
Anyway, I misunderstood what you were getting at in the OP. I thought you were talking making ongoing adjustments to the % based on what the market was doing on a continual basis. Using the Trinity Methodology technically does that, but the % only comes into play at the onset of withdrawals.
-
- Posts: 3180
- Joined: Thu Jul 14, 2011 3:09 am
Re: Should we adjust the SWR based on last All Time High?
OK. This makes a bit of sense. Please read Jacob's post above, concerning weaknesses of Trinity. There are early retirement boards all over the Internet, and they will treat Trinity with greater or lesser versions of holy writ. Here, it is seen as an excellent starting point. Searching SWR will get you all the times we have discussed this.Well, my point is to use the same withrawal amount as on market highs. If you have 100k on a market high, the withdrawal amount is 3k. Then the amount you safely can take out should still be 3k when the market is 50% down and you have 50k remaining. This means that you at this point can use 6% of your current portfolio (3% of what you had before the crash).
Personally, I count 3.33%, with flexibility, and other income streams. But that's me. In my case I have a younger wife, who doesn't share my interest in finances. So adding a bit of margin, to ensure longer duration and a higher end balance, is prudent.
Returns may continue to match what we consider American norms as we enact policies that make us more like Western Europe, but I can't see why that would be likely. So, when looking at SWR in foreign markets, 4% doesn't seem so safe. Maybe it won't be safe here, either.
But then, I don't hate my job, or working. So my motivations may not match yours.
In the end, you will need to make your own choice of SWR. Have you looked at portfoliocharts.com yet?
Re: Should we adjust the SWR based on last All Time High?
Ah - I understand better now. You're looking backward to raise your initial spending, not forward to raise your future spending beyond inflation. My previous comment still applies, but there are a few caveats.
First, be careful about getting too granular looking for your ATH. Almost every withdrawal rates study I've seen uses annual data that filters out daily noise. If you start looking for every daily ATH, the same SWR may not apply.
Second, it's important to be clear on terms. Measuring your own investment account from its all time high is different than measuring the market from its all time high. SWRs are based on the assumption of regular ATHs in the market (aka "growth"). If you've never hit your savings goal, be very careful about using too much mental gymnastics to justify retiring anyway.
Finally, this is IMHO a good argument for selecting a portfolio before retirement that is not so susceptible to large drops. Not only will it negate worry about the markets, but it will also likely have a higher withdrawal rate anyway.
First, be careful about getting too granular looking for your ATH. Almost every withdrawal rates study I've seen uses annual data that filters out daily noise. If you start looking for every daily ATH, the same SWR may not apply.
Second, it's important to be clear on terms. Measuring your own investment account from its all time high is different than measuring the market from its all time high. SWRs are based on the assumption of regular ATHs in the market (aka "growth"). If you've never hit your savings goal, be very careful about using too much mental gymnastics to justify retiring anyway.
Finally, this is IMHO a good argument for selecting a portfolio before retirement that is not so susceptible to large drops. Not only will it negate worry about the markets, but it will also likely have a higher withdrawal rate anyway.
Re: Should we adjust the SWR based on last All Time High?
Yeah that's a great site! I have played around with it.Riggerjack wrote: ↑Thu Apr 20, 2017 7:45 amIn the end, you will need to make your own choice of SWR. Have you looked at portfoliocharts.com yet?
Sure! It should be applied to the market's All Time High, MSCI World or similar if you are an international investor.Tyler9000 wrote: Second, it's important to be clear on terms. Measuring your own investment account from its all time high is different than measuring the market from its all time high. SWRs are based on the assumption of regular ATHs in the market (aka "growth"). If you've never hit your savings goal, be very careful about using too much mental gymnastics to justify retiring anyway.
Re: Should we adjust the SWR based on last All Time High?
The SWR does have a point - 3% WR for life. Meaning 3% during market highs and 3% during market lows. You should never increase that number at any point for all the reasons pointed out above. Only the other side of the equation should change ($).