What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Topic Author
Kevin K
Posts: 360
Joined: Sun Aug 26, 2007 7:47 pm

What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by Kevin K »

Tyler at Portfolio Charts (who also frequently posts on these forums) has an impressive new post on this topic:

https://portfoliocharts.com/2024/04/01/ ... more-70357

It's a long and detailed post that I think will be of particular interest to folks who've been reading the Cederburg study (which is discussed along with the work of Kitces and Pfau). I appreciate the way that Tyler politely challenges conventional thinking, backed by a ton of data.
Wannaretireearly
Posts: 4986
Joined: Wed Mar 31, 2010 4:39 pm

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by Wannaretireearly »

Thanks for sharing this.
“At some point you are trading time you will never get back for money you will never spend.“ | “How do you want to spend the best remaining year of your life?“
User avatar
Lawrence of Suburbia
Posts: 697
Joined: Mon Aug 08, 2022 12:04 pm

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by Lawrence of Suburbia »

Did I read his comment correctly, about how great a big allocation to cash allows for great portfolio survivability vs. inflation? That was shocking. As is the idea about owning ex-U.S. stocks being superior in some scenarios.

My ADD probably has me screwed up; I couldn't follow a lot of that piece. Probably have it backwards.
74% VTHRX/8% DODWX/12% TIAA Traditional/6% SWVXX
DonIce
Posts: 1457
Joined: Thu Feb 21, 2019 5:44 pm

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by DonIce »

The biggest and most controversial (to bogleheads) conclusion seems to be the large (30%) allocation to gold that comes out in the top SWR portfolios. But I've generally seen the same playing around with historical data over varying time periods on portfolio visualizer... you often get the best results with gold heavy portfolios across all the available historical data in PV.

As for the large allocation to foreign stocks... got some questions about the methodology here. It seems like the best SWR portfolio being composed of a larger percentage of foreign stocks is done from the perspective of looking from within different countries. So for example if one looks at Italy and decides on a domestic vs foreign allocation, you'll want a large foreign allocation since that's where the historically higher performing US stocks were. Same for looking from inside Japan, from inside France, Spain, etc. For investors in countries besides the US who invest in foreign stocks according to market cap weight, US stocks make up >50% of that foreign stock allocation. And then when you average across all these perspectives, that tells you to invest in foreign stocks. But really all that it's telling you is that from the perspective of being in most countries, you would have wanted a large percentage allocation to US stocks and a smaller allocation to domestic stocks. So this data probably needs to be normalized based on something rather than just giving each country's perspective an equal weight. I think once you do a reasonable normalization here, the best result regardless of one's nationality would be roughly to invest in world stocks according to their global market cap weight.
User avatar
Random Musings
Posts: 6834
Joined: Thu Feb 22, 2007 3:24 pm
Location: Pennsylvania

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by Random Musings »

Very interesting study. As equity valuations will change from current levels and government fiscal policies will change as well, the next 50 year period could be a bit different which brings us back to "nobody knows nothing".

RM
I figure the odds be fifty-fifty I just might have something to say. FZ
nulka
Posts: 16
Joined: Mon Apr 01, 2024 3:12 am

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by nulka »

Great post. I've wanted to do this analysis myself.

I'd love to see a sensitivity analysis on how things like if, for the y-axis, instead of picking the worst case, you pick the 2nd/3rd/4th/5th worst case.
klaus14
Posts: 961
Joined: Sun Nov 25, 2018 6:43 pm

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by klaus14 »

his study places US in an equal footing to international for analyzing domestic vs foreign stocks allocation. Many would say history of Japan does not apply to a US investor.

But the best thing is he also shares the awesome tool where you can change the list of countries and assets you want to search over. The golden result doesn't change much even if you just loook at US.

I have only one qualm about this. 30 year SWR doesnt have many independent samples in the post 1970 period. A more interesting number to optimize would be his long term SWR concept he introduced recently. That one can work with 15 year samples.
My investment algorithm: https://www.bogleheads.org/forum/viewtopic.php?f=10&t=351899&p=6112869#p6112869
ScubaHogg
Posts: 3661
Joined: Sun Nov 06, 2011 2:02 pm

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by ScubaHogg »

Another home run from Tyler9000! Excellent article!

I was mildly surprised though that the creator of the Golden Butterfly portfolio didn’t include SCV in his analysis?
“Conventional Treasury rates are risk free only in the sense that they guarantee nominal principal. But their real rate of return is uncertain until after the fact.” -Risk Less and Prosper
seajay
Posts: 1700
Joined: Sat May 01, 2021 3:26 pm
Contact:

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by seajay »

Lawrence of Suburbia wrote: Mon Apr 01, 2024 8:53 pm Did I read his comment correctly, about how great a big allocation to cash allows for great portfolio survivability vs. inflation? That was shocking. As is the idea about owning ex-U.S. stocks being superior in some scenarios.
World for stocks IMO. As for inflation - depends upon your objectives. We run a generational portfolio, so base to future/distant spending (inflation) of house prices, stock prices and consumer prices, which a third each equal weightings historically was > just CPI inflation alone (1.5% more). With that objective thirds each in home/global stocks/gold might reasonably be expected to yield a low Ulcer Index. Three currencies, three assets, three income streams (imputed rent/stock dividends/SWR).
Topic Author
Kevin K
Posts: 360
Joined: Sun Aug 26, 2007 7:47 pm

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by Kevin K »

ScubaHogg wrote: Tue Apr 02, 2024 6:34 am Another home run from Tyler9000! Excellent article!

I was mildly surprised though that the creator of the Golden Butterfly portfolio didn’t include SCV in his analysis?
Agreed on both points: a really great article even by Tyler's high standards, and I was also surprised not to see SCV but perhaps it's an issue of insufficient data for SCV globally?

Interesting to me that the "bond" default is 10 year government, when (as McQ pointed out in his excellent thread on ITT's vs. BND) 5 year duration was the historical gold standard duration for ITT's.
User avatar
Tyler9000
Posts: 757
Joined: Fri Aug 21, 2015 11:57 am

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by Tyler9000 »

Kevin K wrote: Mon Apr 01, 2024 4:36 pm Tyler at Portfolio Charts (who also frequently posts on these forums) has an impressive new post on this topic:

https://portfoliocharts.com/2024/04/01/ ... more-70357

It's a long and detailed post that I think will be of particular interest to folks who've been reading the Cederburg study (which is discussed along with the work of Kitces and Pfau). I appreciate the way that Tyler politely challenges conventional thinking, backed by a ton of data.
Thank you for sharing this, Kevin K. Yeah, it's a bit long. But it's the culmination of a crazy amount of work so I guess I had a lot to say. :D I don't claim to have all the answers, but I do have a lot of data for discussion. I hope everyone sticks with it, learns something, and perhaps thinks about withdrawal rates in a new way.
User avatar
Tyler9000
Posts: 757
Joined: Fri Aug 21, 2015 11:57 am

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by Tyler9000 »

Lawrence of Suburbia wrote: Mon Apr 01, 2024 8:53 pm Did I read his comment correctly, about how great a big allocation to cash allows for great portfolio survivability vs. inflation? That was shocking.
You read it correctly. To be clear, though, when the article talks about cash it's referencing T-Bills that earn interest (for example, a treasury money market fund or an ETF like BIL), not a pile of money stuffed under a mattress. This article expands on that thought: https://portfoliocharts.com/2017/05/12/ ... -investor/
gt4715b
Posts: 541
Joined: Mon Jun 11, 2007 10:29 am

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by gt4715b »

I don't think it's meaningful to try to predict a domestic vs. foreign stock allocation. The safest bet is just to market weight, e.g., VT, IMO.
User avatar
Tyler9000
Posts: 757
Joined: Fri Aug 21, 2015 11:57 am

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by Tyler9000 »

DonIce wrote: Mon Apr 01, 2024 9:21 pm As for the large allocation to foreign stocks... got some questions about the methodology here. It seems like the best SWR portfolio being composed of a larger percentage of foreign stocks is done from the perspective of looking from within different countries. So for example if one looks at Italy and decides on a domestic vs foreign allocation, you'll want a large foreign allocation since that's where the historically higher performing US stocks were. Same for looking from inside Japan, from inside France, Spain, etc. For investors in countries besides the US who invest in foreign stocks according to market cap weight, US stocks make up >50% of that foreign stock allocation. And then when you average across all these perspectives, that tells you to invest in foreign stocks. But really all that it's telling you is that from the perspective of being in most countries, you would have wanted a large percentage allocation to US stocks and a smaller allocation to domestic stocks. So this data probably needs to be normalized based on something rather than just giving each country's perspective an equal weight. I think once you do a reasonable normalization here, the best result regardless of one's nationality would be roughly to invest in world stocks according to their global market cap weight.
This is a great insight.

From playing with the data more than just the very top results, I don't think it's as simple as the best portfolios favoring US stocks as much as possible. Surfing the cloud, I see lots of portfolios where at least some amount of home country bias is important even in countries outside of the US. Based on other research, I believe that this is due to the importance of tracking local purchasing power in retirement. Even investing in the same US assets, every global investor has a different experience based on local inflation and exchange rates.

That said, there is definitely some nuance to be aware of with the domestic/foreign paradigm of the approach. For example, in the #1 Ulcer Index portfolio in the article, the worst case for that top portfolio with no domestic allocation was in the US. It’s fair to point out that perhaps the US was the worst case for this portfolio simply because it was the only country without an allocation to US stocks. And it's also true that adding a domestic allocation helps the US outcomes (while hurting some of the others).

I've just amended that section to point this out, and also added a note in the takeaways about the looking at the worst cases for opportunities for improvement in specific countries. So thanks for the contribution.
User avatar
martincmartin
Posts: 904
Joined: Wed Jul 02, 2014 3:04 pm
Location: Boston, MA USA

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by martincmartin »

The problem with including gold, foreign and many other things, is that it restricts you to a small amount of recent data. This article says "30-year retirement scenarios starting in 1970." Starting in 1970 you miss ALL worst-case scenarios from Bengen/Trinity studies:

- Retire just before the bank panic of 1907
- Retire just before the stock market crash of 1929 and the Great Depression
- Retire in 1966 at the start of stagflation.

You do get a muted form of the last one by starting in 1970. Still, I wouldn't generalize from a single data point. Or more like, 80% of a single data point.
User avatar
Tyler9000
Posts: 757
Joined: Fri Aug 21, 2015 11:57 am

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by Tyler9000 »

ScubaHogg wrote: Tue Apr 02, 2024 6:34 am I was mildly surprised though that the creator of the Golden Butterfly portfolio didn’t include SCV in his analysis?
Kevin K wrote: Tue Apr 02, 2024 10:11 am I was also surprised not to see SCV but perhaps it's an issue of insufficient data for SCV globally?

Interesting to me that the "bond" default is 10 year government, when (as McQ pointed out in his excellent thread on ITT's vs. BND) 5 year duration was the historical gold standard duration for ITT's.
I would love to include SCV! But my main goal here was to explore identical portfolios from a global perspective, and because we only have good domestic SCV data for the United States I knew that would skew the results. Once you narrow in on an option, feel free to use the Withdrawal Rates tool to see how factors affect the numbers.

I agree that 5 year bonds are the better measure for traditional IT bonds, especially in the US. I went with 10Y mostly because of data availability. Practically speaking, they're somewhere between IT and LT in terms of behavior. And it's also reasonably close to the weighted average maturity of total bond market funds popular outside of the US.
Last edited by Tyler9000 on Tue Apr 02, 2024 10:57 am, edited 3 times in total.
User avatar
Tyler9000
Posts: 757
Joined: Fri Aug 21, 2015 11:57 am

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by Tyler9000 »

martincmartin wrote: Tue Apr 02, 2024 10:42 am The problem with including gold, foreign and many other things, is that it restricts you to a small amount of recent data. This article says "30-year retirement scenarios starting in 1970." Starting in 1970 you miss ALL worst-case scenarios from Bengen/Trinity studies:

- Retire just before the bank panic of 1907
- Retire just before the stock market crash of 1929 and the Great Depression
- Retire in 1966 at the start of stagflation.

You do get a muted form of the last one by starting in 1970. Still, I wouldn't generalize from a single data point. Or more like, 80% of a single data point.
All good points. Be sure to read the Methodology portion of the article and look for the "Trust, but Verify" section. I provide detailed data that directly addresses your concerns.
User avatar
martincmartin
Posts: 904
Joined: Wed Jul 02, 2014 3:04 pm
Location: Boston, MA USA

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by martincmartin »

Tyler9000 wrote: Tue Apr 02, 2024 10:50 am All good points. Be sure to read the Methodology portion of the article and look for the "Trust, but Verify" section. I provide detailed data that directly addresses your concerns.
Yes, thanks, and it is a kind of maddening problem. On the one hand, in some sense even 150 years of stock & bond data in the US isn't enough. Three bad periods isn't really enough either. So we have to make due with limited data. On the other hand, it seems foolish to discount things like international bonds, commodities and gold simply because we only have about 55 years of data. So I wrestle with all of this.

As a follow on, what does your analysis say about glide paths? Does increasing stocks over time in retirement help in other countries as well? McClung in Living Off Your Money seems to think so.
User avatar
Tyler9000
Posts: 757
Joined: Fri Aug 21, 2015 11:57 am

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by Tyler9000 »

martincmartin wrote: Tue Apr 02, 2024 11:00 am Yes, thanks, and it is a kind of maddening problem. On the one hand, in some sense even 150 years of stock & bond data in the US isn't enough. Three bad periods isn't really enough either. So we have to make due with limited data. On the other hand, it seems foolish to discount things like international bonds, commodities and gold simply because we only have about 55 years of data. So I wrestle with all of this.
I hear ya. I've wrestled with the same. It's a balancing act, and there's also something to be said for reading both very long studies (with more start year datapoints) and very wide studies (with more portfolio datapoints) to let each perspective inform your thought process. Most of my development is in the latter, as I think that's where I can contribute the most. But I also like the work of guys like Pfau who focus on deep histories.

Also, one of the core ideas of the Global Withdrawal Rates tool is the concept of scenarios. To increase scenarios within a country, most retirement studies look back as far as possible. But by looking around to many different countries even with a shorter timeframe, one can identify far more scenarios for study. And with way more economic diversity, too. For example, the original Bengen study that people look to for the 4% rule was based on just 51 scenarios in the US. My analysis here covers 400 scenarios in 10 different countries. So there's more than one way to look for worst cases.

martincmartin wrote: Tue Apr 02, 2024 11:00 amAs a follow on, what does your analysis say about glide paths? Does increasing stocks over time in retirement help in other countries as well? McClung in Living Off Your Money seems to think so.
The study is built on the assumption of a regularly rebalanced fixed allocation, and I don't think you can draw any conclusions about glide paths. But that's a great question, and one I'd like to look at in the future.
User avatar
firebirdparts
Posts: 4460
Joined: Thu Jun 13, 2019 4:21 pm
Location: Southern Appalachia

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by firebirdparts »

nulka wrote: Tue Apr 02, 2024 1:09 am Great post. I've wanted to do this analysis myself.

I'd love to see a sensitivity analysis on how things like if, for the y-axis, instead of picking the worst case, you pick the 2nd/3rd/4th/5th worst case.
I don't have a whole lot on this, but there is one study that I like to reference that shows just how many of the cases are similar. They showed the 'Safe", the 10th percentile, and then the number of failures at 4% and also at 5%. This is from Advisor Perspectives here linked below. There are format issues with the tables that you have to figure out on your own.
https://www.advisorperspectives.com/art ... awal-rates

Go to figure 1 at the end of this article. What you see here is just how often is 4% the number, and the answer is pretty often. Lows are long and peaks are short. "Domestic" here is USA.

I would post figure 1 here, but sadly imgur has gone the way of photobucket.
This time is the same
User avatar
Tyler9000
Posts: 757
Joined: Fri Aug 21, 2015 11:57 am

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by Tyler9000 »

nulka wrote: Tue Apr 02, 2024 1:09 am Great post. I've wanted to do this analysis myself.

I'd love to see a sensitivity analysis on how things like if, for the y-axis, instead of picking the worst case, you pick the 2nd/3rd/4th/5th worst case.
Thanks!

Tweaking the tool to do that for every portfolio simultaneously is an interesting idea. I'll look into it. In the meantime, you can do something close to that now for individual portfolios.

On the Global Withdrawal Rates calculator page, choose Portfolio as the search option and enter the asset allocation you're interested in. Look for the worst case scenario. Then go to the country settings and remove that country from consideration. Using that process, you can peel back the results as many layers as you like.

Note that this won't show the 2nd worst scenario if it happens to be in the same country. But it can at least help you sample a spectrum of outcomes.
GAAP
Posts: 2646
Joined: Fri Apr 08, 2016 12:41 pm

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by GAAP »

Gold in the USA since 1970? Outside of dentists and personal jewelry, ownership of gold was a felony in 1970, and didn't become legal until 1974.

How is this handled with this analysis?

Does anyone know when these countries adopted a true fiat currency? Gold is a great deflation hedge in a gold-standard world, but fiat currencies don't really have that problem.
“Adapt what is useful, reject what is useless, and add what is specifically your own.” ― Bruce Lee
McQ
Posts: 1458
Joined: Fri Jun 18, 2021 12:21 am
Location: California

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by McQ »

Tyler9000 wrote: Tue Apr 02, 2024 10:47 am
ScubaHogg wrote: Tue Apr 02, 2024 6:34 am I was mildly surprised though that the creator of the Golden Butterfly portfolio didn’t include SCV in his analysis?
Kevin K wrote: Tue Apr 02, 2024 10:11 am I was also surprised not to see SCV but perhaps it's an issue of insufficient data for SCV globally?

Interesting to me that the "bond" default is 10 year government, when (as McQ pointed out in his excellent thread on ITT's vs. BND) 5 year duration was the historical gold standard duration for ITT's.
I would love to include SCV! But my main goal here was to explore identical portfolios from a global perspective, and because we only have good domestic SCV data for the United States I knew that would skew the results. Once you narrow in on an option, feel free to use the Withdrawal Rates tool to see how factors affect the numbers.

I agree that 5 year bonds are the better measure for traditional IT bonds, especially in the US. I went with 10Y mostly because of data availability. Practically speaking, they're somewhere between IT and LT in terms of behavior. And it's also reasonably close to the weighted average maturity of total bond market funds popular outside of the US.
In Tyler’s defense, financial historians have moved to the 10 year maturity as the best metric for “bonds,” even to the extent of re-expressing long bond data as inferred 10- year returns. I have my qualms, but Tyler has to work with the data out there.
You can take the academic out of the classroom by retirement, but you can't ever take the classroom out of his tone, style, and manner of approach.
nulka
Posts: 16
Joined: Mon Apr 01, 2024 3:12 am

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by nulka »

firebirdparts wrote: Tue Apr 02, 2024 1:11 pm I don't have a whole lot on this, but there is one study that I like to reference that shows just how many of the cases are similar. They showed the 'Safe", the 10th percentile, and then the number of failures at 4% and also at 5%. This is from Advisor Perspectives here linked below. There are format issues with the tables that you have to figure out on your own.
https://www.advisorperspectives.com/art ... awal-rates

Go to figure 1 at the end of this article. What you see here is just how often is 4% the number, and the answer is pretty often. Lows are long and peaks are short. "Domestic" here is USA.

I would post figure 1 here, but sadly imgur has gone the way of photobucket.
Thanks. Figure 1 also demonstrates how much things can change if you pick a different period. In the first half, international diversification fared a bit worse and then in the second half it fared much better.
User avatar
Tyler9000
Posts: 757
Joined: Fri Aug 21, 2015 11:57 am

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by Tyler9000 »

GAAP wrote: Tue Apr 02, 2024 3:05 pm Gold in the USA since 1970? Outside of dentists and personal jewelry, ownership of gold was a felony in 1970, and didn't become legal until 1974.

How is this handled with this analysis?

Does anyone know when these countries adopted a true fiat currency? Gold is a great deflation hedge in a gold-standard world, but fiat currencies don't really have that problem.
Richard Nixon effectively ended the gold standard globally when he negated the Bretton Woods agreement by suspending the convertibility of the US dollar to gold in 1971. Gold is a global commodity and was freely traded outside of the US. My data tracks the gold returns using the price history recorded by sources like LBMA.
seajay
Posts: 1700
Joined: Sat May 01, 2021 3:26 pm
Contact:

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by seajay »

GAAP wrote: Tue Apr 02, 2024 3:05 pm Gold in the USA since 1970? Outside of dentists and personal jewelry, ownership of gold was a felony in 1970, and didn't become legal until 1974.

How is this handled with this analysis?

Does anyone know when these countries adopted a true fiat currency? Gold is a great deflation hedge in a gold-standard world, but fiat currencies don't really have that problem.
When gold (silver) was money the inclination would be to deposit surplus gold (silver/money) for both safe keeping and some interest (more gold/silver). If/when gold wasn't permitted one might have been inclined to opt for the next best thing, silver. From a British perspective I consider 'gold' (non-fiat commodity currency) returns to be British T-Bill yields pre WW1 (at which point the convertibility was ended until 1925), silver from the start of 1914 to the end of 1925, T-Bills again 1926 until 1931 - when again convertibility was ended, silver from 1932 onward, with a swap out of silver for gold at some point during the 1970's. Potentially 1979/1980 when silver soared, but rather than capture that 'ideal' timing I tend to go with 1975 as the last year silver was held, gold from the start of 1976. So a collective "gold" (non-fiat/commodity currency/money) history that goes back to the 1800's.

US Dollar year end silver prices (macrotrends). Another source is the US geological survey web site but IIRC that yields year average prices

Code: Select all

1915	0.51
1916	0.67
1917	0.84
1918	0.98
1919	1.13
1920	1.03
1921	0.63
1922	0.68
1923	0.65
1924	0.67
1925	0.69
1926	0.62
1927	0.57
1928	0.58
1929	0.53
1930	0.38
1931	0.29
1932	0.28
1933	0.35
1934	0.48
1935	0.64
1936	0.45
1937	0.45
1938	0.43
1939	0.39
1940	0.35
1941	0.35
1942	0.38
1943	0.45
1944	0.45
1945	0.52
1946	0.8
1947	0.72
1948	0.74
1949	0.72
1950	0.74
1951	0.89
1952	0.85
1953	0.85
1954	0.85
1955	0.89
1956	0.92
1957	0.92
1958	0.89
1959	0.92
1960	0.92
1961	0.93
1962	1.09
1963	1.29
1964	1.31
1965	1.3
1966	1.31
1967	2.25
1968	1.97
1969	1.79
1970	1.63
1971	1.37
1972	2.03
1973	3.26
1974	4.47
1975	4.18
1976	4.36
1977	4.76
1978	6.02
1979	32.2
1980	15.5
1981	8.15
1982	10.87
1983	8.91
1984	6.29
1985	5.8
1986	5.28
1987	6.7
1988	6.05
1989	5.22
1990	4.19
1991	3.86
1992	3.67
1993	5.12
1994	4.85
1995	5.14
1996	4.8
1997	6
1998	5.01
1999	5.33
2000	4.58
2001	4.52
2002	4.67
2003	5.97
2004	6.82
2005	8.83
2006	12.9
2007	14.76
2008	10.79
2009	16.99
2010	30.63
2011	28.18
2012	29.95
2013	19.5
2014	15.97
2015	13.8
2016	15.99
2017	17.13
2018	15.52
2019	17.9
2020	26.4
2021	23.35
2022	23.96
2023	23.79
As up to the transition from the gold backed British Pound being a common international trade settlement currency (some still use the British Sovereign gold coin as part of international trade settlements) and adoption of the US Dollar instead, Britain was a global hub for international trade/accounting/law/finance, the British FT All Share stock index might be considered as a form of global stock index fund. Today for instance and the FT All Share is still in excess of 70% foreign earnings.

For a family (generational) portfolio I assume a inflation rate of equal thirds FT All Share stock price only, British average house price increases, consumer price inflation i.e. the next gen might spend a lump sum inheritance on buying a home, some stocks, some spending. I ignore imputed rent benefit (historically averaged 4.5%), assume that to be a 1.33% SWR type benefit, and draw 2% actual SWR (of total combined home/stock/gold value) to make a combined 3.33% 30 year return of inflation adjusted capital measure (airing on the conservative side as likely imputed rent benefit would have lifted that a little higher). On that measure the Ulcer index is low (aligns reasonably closely with that house/stock/CPI prices inflation rate), whilst the historic tendency was to end with around the same inflation adjusted wealth at the start (0.66 low (1980's start years), 1.5 highs (1950's), median of around 1.0).

Yearly rebalancing a portfolio that in part includes your home value is impractical, however broadly the difference in rewards between rebalanced and non-rebalanced tends to wash. With rebalanced you more closely maintain target weightings such as thirds each stock/home/gold, with non rebalanced you end with a high weighting in whichever asset(s) performed the best, low weighting in the asset(s) that performed the worst. 50/50 stock/gold for instance might drift to end with 80/20, time averaged 65/35 in the best/worst asset. Partial rebalancing of non-rebalanced is also achieved by adding savings to the most-down, drawing spending from the most-up. Starting for instance with thirds each land/stock/gold and drawing from stock or gold each month according to whether stock or gold value was higher at the time, in order to pay off that months credit card bill. Sometimes you may be selling at highs, other times at lows, time averages out that volatility. Only if you are making a very large withdrawal does interim volatility become a risk (benefit) factor.

So from a family perspective we have 2% of total wealth SWR disposable 'income', have our rent paid (1.33% of total wealth imputed rent benefit), and a potential to pass on a inheritance of similar inflation adjusted wealth/assets to subsequent generations. 30 year SWR periods x 3.33% combined SWR return of inflation adjusted capital/value, and ending with 100% of inflation adjusted wealth remaining. Noteworthy is that house/stock/CPI price only combined inflation rates have exceeded just CPI inflation rates by around 1.5%/year (log linear regression line slope on a log scaled chart).

Asset diversity of land/stocks/commodity; Currency diversity of GB Pounds (land), global fiat (stocks), global non-fiat (gold); Income diversity of imputed, dividends, SWR. With two thirds of wealth in-hand (land and gold). Such a three-asset portfolio is better IMO than the likes of the three-fund portfolio.

A low Ulcer index relative to just CPI alone, is more for those with no heirs and that are looking to fully spend their wealth, where the inclination is towards not spending it all (avoid outliving their money poverty risk).

Dislike bonds/cash-deposits, lending to others who get to set the rates/terms and conditions, that generate regular taxable events (interest). More prefer 50/50 stock/gold as the alternative. Home + imputed is inclined to somewhat align with stock + dividends, so in effect a 67/33 equity/gold type asset allocation.

Some generational European families have utilized the likes of thirds land/art/gold. Me, I can't tell the difference between a priceless artwork and a kindergarten painting so prefer stocks instead.

In most countries you might be able to buy into a global stock index tracker fund, buy your own home, hide away some gold. With increased wealth the land (home) diversity might be extended to include homes around the world, perhaps to follow pleasant weather conditions, where the travel between costs might be more than offset by the savings in heating/cooling costs.
seajay
Posts: 1700
Joined: Sat May 01, 2021 3:26 pm
Contact:

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by seajay »

For wealth accumulation IMO the more appropriate choice is to use leveraged equity during younger/accumulation years, i.e. buy a home using a mortgage. Once that is owned outright add stocks. In later years add gold as a risk hedge. Whatever might drive equities (home and/or stock prices) to halve might equally see the price of gold double, 67/33 -> 33/67 has no capital loss and rebalancing would have you double up on the number of shares (home value) being held (relatively richer compared to those that held no gold). Tendency to lag both the all-stock up and down sides (less volatility). But physiologically a difficult asset, requires patience. 1980 to 1999 for instance and gold halved in nominal price terms (so even more in real terms). 50/50 with stocks however and you saw the accumulation of 6 times more ounces of gold in your safe. In the 2000's that flipped around, fewer ounces of gold, more stock shares. Many wont have the required patience however, might for instance buy into gold more recently as its price has gained a lot (buy high i.e. the gold/silver ratio is presently around 100), sell later as that 'mean-reverts' or as other assets become the recent high-gain attraction. Jumping around to profit-chase involves considerable overall cost.

Some do yield 'dividends' from their gold, via the likes of gold lending. In India for instance 7% type rates of return. Where banks are few and far money lenders lend up to 70% of spot gold value against gold deposits, charge a pro-rata 7% interest rate. More recently there are US based gold bond options that do the leg work, but I believe are relatively new and limited to 'accredited' investors only. In more developed gold markets spreads on physical can be low/zero. Communities of gold investors often have individuals who are happy to sell/buy at a common/consistent level such as 1% above spot, primarily you have to built up a reputation of trust/reliability. Even through different dealers you can often buy at one, walk to another and sell for around a 1% overall round trip cost. Equally you can often identify where such a round-trip cost might exceed 10%, that those that dislike gold might highlight. The field is broad, PAXG for instance, where exchanging/moving between gold and other 'currencies' can be a single mouse click, better suited to the likes of moving 'gold' across international borders (converting physical to electronic and electronic transfer to the target destination and then repurchase physical involves costs, but those costs may be small compared to trying to move physical gold internationally only to end up losing it).
User avatar
siamond
Posts: 6013
Joined: Mon May 28, 2012 5:50 am

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by siamond »

Just caught up with the latest from Tyler. I love the top-level idea of varying the country of origin while studying a 'big cloud' of portfolios. I ran a study along the same line of thinking a while ago (see this blog), but I didn't use powerful tools like the ones Tyler developed! This is very cool.

The part I am less enthused about is to center the analysis on two dimensions of risks (SWR defined as the worst possible result for 30 years cycles; Ulcer index). Agreed, one has to separate short-term emotional risk (embodied by the Ulcer Index) and long-term income / purchasing power risk (embodied by absolute worst SWR). This presents severe issues though:
a) this totally ignores any concept of potential reward (e.g. improved purchasing power if one's starting year was NOT the absolute worst one).
b) the absolute worst SWR is just one single data point in history (not only the per-country/AA SWR number, but the chart's Y axis is the minimum across countries). Optimizing on such single data point is rather unfortunate, I'm afraid.
c) the absolute worst SWR can be severely misleading. Think to Japan/1990. Yes, the drawdown and subsequent SWR were absolutely terrifying. BUT this ignores the fact that the Japan stock market absolutely skyrocketed before 1990. A 2% SWR applied to a portfolio which doubled in the previous years isn't that different (emotions aside!) from 4% SWR applied a few years earlier. Note that anybody studying long-term US history and the 1929 crisis has a similar issue.

I am afraid that any conclusion one would draw (e.g. choice of AA) would be strongly skewed by those issues. I took a very brief, quick and dirty, look at the Ranked #1 portfolio, and I'm afraid it displayed rather poor properties when you dive a bit further than those two risks metrics.

PS. in all fairness, come to think of it, my own blog analysis suffered from a somewhat similar methodology issue... Those things are tough nuts to crack. :shock:
seajay
Posts: 1700
Joined: Sat May 01, 2021 3:26 pm
Contact:

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by seajay »

Part of the reason why I see averaging in to be superior to lumping in

Buy and hold is no different to costless lumping in every day, rather than a 100% transition into retirement on a single day, half retiring 6 months before, and half 6 months later i.e. migration (time diversifying) into your choice of retirement asset allocation will avoid having lumped all-in at a single point in time that is potentially the worst time to have done so. Similar to two adjacent 30 year SWR measures where the worst case will be diluted by a adjacent case (uplifts the SWR)

Image

In the attached chart 1986 was the worst case for a 30 year 3% SWR, ended 30 years with 70.6% of the inflation adjusted start date portfolio value, adjoining either side of that were 93% and 133% earlier/later respective outcomes, such that the worst case of 1985/1986 averaged 82%.

Another factor is granularity of data, the above is calendar yearly based, within that such as daily granularity might be expected to present even worse case outcomes.

Yet another factor is the tendency to select from what data remains available, where prior failures have no data, the US over the last century has been more a right-tail good/great case market

Image

making forward assumptions based on a good/great case historic outcome is being ... optimistic.

Additionally factor in costs/taxes, and inflation rate adjustments (such as using the house, stock, CPI measure I outlined in a prior post, that exceed CPI inflation by 1.5%) and again not doing so is a optimistic side bias. In say the 1950's brokers costs were much higher, as might have market makers made a killing from postal trades (stock purchases/sales), perhaps 10% spreads, to the extend where many investors literally bought and long term held perhaps a diverse bunch of 8 or 10 major Blue Chip stocks, as trading was otherwise too expensive.

Of course those that make their money from investors, that often are highly paid and work out of offices in expensive locations, aren't inclined to present anything other than mathematical gross (biased) figures, as that makes their products/services look more appealing.
User avatar
Tyler9000
Posts: 757
Joined: Fri Aug 21, 2015 11:57 am

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by Tyler9000 »

klaus14 wrote: Tue Apr 02, 2024 6:01 am I have only one qualm about this. 30 year SWR doesnt have many independent samples in the post 1970 period. A more interesting number to optimize would be his long term SWR concept he introduced recently. That one can work with 15 year samples.
I originally planned to use the LTWR for this analysis, but chose to hold back for two reasons. First, I wanted to take a pass with the standard 30-year SWR numbers so that I could properly compare the numbers to other major retirement studies that use that metric. And second, the LTWR is way more computation-heavy for so many portfolios.

I do hope to look at that in the future, though.
User avatar
Tyler9000
Posts: 757
Joined: Fri Aug 21, 2015 11:57 am

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by Tyler9000 »

siamond wrote: Wed Apr 03, 2024 3:48 am a) this totally ignores any concept of potential reward (e.g. improved purchasing power if one's starting year was NOT the absolute worst one).
That's true. The distribution of outcomes in non-worst-case scenarios is also different for all portfolios. My best tool for studying that is the Retirement Spending chart, which tracks the portfolio values for all start dates and can even account for variable spending strategies. It can only do one portfolio at a time, but it's a great way to dig into this issue. Maybe I can expand this multi-portfolio research in that direction over time.

siamond wrote: Wed Apr 03, 2024 3:48 am b) the absolute worst SWR is just one single data point in history (not only the per-country/AA SWR number, but the chart's Y axis is the minimum across countries). Optimizing on such single data point is rather unfortunate, I'm afraid.
Read Bengen's original paper (PDF) on withdrawal rates, and his intro gives a nice narrative of why specifically studying the worst cases is important in retirement research. Basically, the alternative up to then was to just advise people to assume average returns and hope they weren't the unlucky ones with below-average outcomes. So I personally see worst-case datapoints as enlightening rather than unfortunate. But knowing your particularly strong knowledge of financial history, I also understand your hesitance to read too much into single scenarios.

siamond wrote: Wed Apr 03, 2024 3:48 am c) the absolute worst SWR can be severely misleading. Think to Japan/1990. Yes, the drawdown and subsequent SWR were absolutely terrifying. BUT this ignores the fact that the Japan stock market absolutely skyrocketed before 1990. A 2% SWR applied to a portfolio which doubled in the previous years isn't that different (emotions aside!) from 4% SWR applied a few years earlier. Note that anybody studying long-term US history and the 1929 crisis has a similar issue.
I hear ya. Just remember that a Japanese investor expecting much more than 4% to be safe based on personal experience and learning the hard way that 2% was the true SWR may wish later to have been offered a perspective on the breadth of SWR outcomes.

siamond wrote: Wed Apr 03, 2024 3:48 am I am afraid that any conclusion one would draw (e.g. choice of AA) would be strongly skewed by those issues. I took a very brief, quick and dirty, look at the Ranked #1 portfolio, and I'm afraid it displayed rather poor properties when you dive a bit further than those two risks metrics.
That's certainly possible. Any time you optimize for one metric, you open yourself up for less than ideal outcomes in others. With ANY portfolio recommendation (from my site or elsewhere), I recommend that people do exactly what you did and make sure it works for them from different perspectives. That's why I offer so many different types of charting tools. :D

And it's also why I really love your work collecting and studying historical data. I always appreciate your perspective.
User avatar
siamond
Posts: 6013
Joined: Mon May 28, 2012 5:50 am

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by siamond »

Tyler9000 wrote: Wed Apr 03, 2024 9:55 am
siamond wrote: Wed Apr 03, 2024 3:48 am b) the absolute worst SWR is just one single data point in history (not only the per-country/AA SWR number, but the chart's Y axis is the minimum across countries). Optimizing on such single data point is rather unfortunate, I'm afraid.
Read Bengen's original paper (PDF) on withdrawal rates, and his intro gives a nice narrative of why specifically studying the worst cases is important in retirement research. Basically, the alternative up to then was to just advise people to assume average returns and hope they weren't the unlucky ones with below-average outcomes. So I personally see worst-case datapoints as enlightening rather than unfortunate. But knowing your particularly strong knowledge of financial history, I also understand your hesitance to read too much into single scenarios.
siamond wrote: Wed Apr 03, 2024 3:48 am c) the absolute worst SWR can be severely misleading. Think to Japan/1990. Yes, the drawdown and subsequent SWR were absolutely terrifying. BUT this ignores the fact that the Japan stock market absolutely skyrocketed before 1990. A 2% SWR applied to a portfolio which doubled in the previous years isn't that different (emotions aside!) from 4% SWR applied a few years earlier. Note that anybody studying long-term US history and the 1929 crisis has a similar issue.
I hear ya. Just remember that a Japanese investor expecting much more than 4% to be safe based on personal experience and learning the hard way that 2% was the true SWR may wish later to have been offered a perspective on the breadth of SWR outcomes.
Looking at the worst case in one country (like Bengen did) is one thing and it's definitely a good thing to do. Looking at the worst of per-country worst cases is going one (BIG) step further. Then running an optimizer solely based on such worst-of-the-worst data point is going one (GIANT) step further. And when we have this utterly odd Japan/1990 data point in the middle of all of this, taken without context of the meteoric rise that preceded it, then... I'm sorry, Tyler, but we have a result which seems pretty much meaningless to me.

Not that I have a good suggestion on how to improve this, to be honest. In my original study, I believe I used low percentiles instead of the worst case(s), but this isn't quite satisfying either. I continue to really like the core idea of running AA numbers across countries and looking at the cloud of outcomes. I just really don't like to ignore the cloud at the end and solely focus on one terribly odd snowflake... Will mull it over!
User avatar
Tyler9000
Posts: 757
Joined: Fri Aug 21, 2015 11:57 am

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by Tyler9000 »

siamond wrote: Wed Apr 03, 2024 5:31 pm Not that I have a good suggestion on how to improve this, to be honest. In my original study, I believe I used low percentiles instead of the worst case(s), but this isn't quite satisfying either. I continue to really like the core idea of running AA numbers across countries and looking at the cloud of outcomes. I just really don't like to ignore the cloud at the end and solely focus on one terribly odd snowflake... Will mull it over!
Fair enough.

I expected some people to have this perspective, which is why I thought it was important to share a tool that lets them select as many or few countries as they like. But I look forward to any other ideas you have to improve the approach.
Alpha4
Posts: 180
Joined: Tue Apr 17, 2012 8:47 pm

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by Alpha4 »

Tyler9000 wrote: Tue Apr 02, 2024 10:12 am
Kevin K wrote: Mon Apr 01, 2024 4:36 pm Tyler at Portfolio Charts (who also frequently posts on these forums) has an impressive new post on this topic:

https://portfoliocharts.com/2024/04/01/ ... more-70357

It's a long and detailed post that I think will be of particular interest to folks who've been reading the Cederburg study (which is discussed along with the work of Kitces and Pfau). I appreciate the way that Tyler politely challenges conventional thinking, backed by a ton of data.
Thank you for sharing this, Kevin K. Yeah, it's a bit long. But it's the culmination of a crazy amount of work so I guess I had a lot to say. :D I don't claim to have all the answers, but I do have a lot of data for discussion. I hope everyone sticks with it, learns something, and perhaps thinks about withdrawal rates in a new way.
Tyler,

Wow. Pretty impressive.

Two quick questions/suggestions:

One, are these worst-case results monthly (i.e. starting in, say, Sept 1929 or November 1965) or are they only starting on Jan 1st of each year? If the latter, worst-case SWRs will likely be a tad lower than what this tool shows...right?

Two: I was looking over the comparisons with the Bengen, Pfau, and Cederburg data and your "non-disaster worst case" (i.e. the worst case SWR assuming one's country not being a battlefield and/or destroyed and defeated in WWI or WWII) for Germany, Italy, France, Spain, Netherlands, and the UK all were in the 70s, 90s, or 2000s. This may not be the case in actuality. Would it be possible to add data for these countries for the 1960s (and maybe the 1950s as well)? The reason I ask this is because the early to mid 1960s were decidedly not halycon and booming times for the majority of European stocks. German stocks (the DAX index) went from 351 in December 1960 to 207 at the end of 1966 (granted, this is price-only and not interluding dividends...but it also is nominal and doesn't include any loss of purchasing power due to inflation). French equities (in nominal before-inflation total return terms including reinvested dividends) performed such that 1 Franc of them at the start of 1961 was worth just over 50 Centimes at the start of 1968 (so a loss of just under half or purchasing power in nominal terms and more still in inflation adjusted terms). Italian stocks lost roughly half of their purchasing power from the start of 1961 to the end of 1971 (again, this is price-only and not total return...but it also doesn't include the corrosive effects of inflation...which averaged around 4.2% in Italy over these years). The Netherlands and the UK weren't quite as bad (the former saw about a 1/3 drop in stock market value from the start of 1961 to the end of 1966; recovered to a bit above where it was in 1961 by late 1969; fell again until 1971 but never to its 1966 lows; the latter had two barnburner years in 1958 and 1959--up over 50% each year--but then had three "meh" years in 1960-1962 (averaged barely 1% nominal total return those years); had a decent year in 1963 (up almost 20% in nominal total return terms), and then had two somewhat bad years and one slightly decent one in the 1964-1966 period such that by the end of 1966 the portfolio in total return terms was just a hair under what it was at the start of 1960. 1967 and 1968 were good years (although nowhere near 1958 and 1959), 1969 and 1970 stunk, 1971 was very good (a roughly 34% total return) and then from 1972-1974 UK stocks put in such poor returns (1974 especially was awful with UK shares losing half their value in real total return terms in one year) that for every one Pound the investor had at the start of 1960 it was only worth approximately £0.54 at the end of 1974. Of course, 1975-1989 was virtually one huge (relatively) rarely interrupted bull market for UK stocks but given that the first 14-15 years of performance weigh very heavily in determining one's SWR this was probably too late to do much good for the UK investor who started withdrawing in 1960 or 1961 (IIRC I remember reading somewhere that 1960 was the worst post WWII start date for a balanced all-UK (i.e. only UK shares and gilts; no foreign stocks, no foreign bonds, and no gold) portfolio and that it did fail the 4% test.

FWIW Switzerland (even though it wasn't included in your list) also would've had its worst post-1950 SWR starting in 1962; using the Pictet & Cie data set for Swiss stocks and bonds I could not make it go much beyond a 3.3 or 3.4% 30-year SWR without failing for that start year no matter whether I set up the portfolio as 1/99 or 99/1 or anything in between.

In short, for many/most European countries the start years between, say, 1960 and 1965 very well have been the worst post-WWII start years and likely would have had lower SWRs (nowhere near their WWII or 1914 SWRs but also with considerably worse outcomes than for the post-1969 dates PortfolioCharts shows as their worst SWR start dates) than the ones with start dates in the 1970-2000 period.
User avatar
Tyler9000
Posts: 757
Joined: Fri Aug 21, 2015 11:57 am

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by Tyler9000 »

Alpha4 wrote: Wed Apr 03, 2024 7:50 pm
Two quick questions/suggestions:

One, are these worst-case results monthly (i.e. starting in, say, Sept 1929 or November 1965) or are they only starting on Jan 1st of each year? If the latter, worst-case SWRs will likely be a tad lower than what this tool shows...right?
My data is annual. And when looking for worst cases, you're correct that the more granular the data the more opportunities you have to find more worst cases.

Alpha4 wrote: Wed Apr 03, 2024 7:50 pm Two: ... In short, for many/most European countries the start years between, say, 1960 and 1965 very well have been the worst post-WWII start years and likely would have had lower SWRs (nowhere near their WWII or 1914 SWRs but also with considerably worse outcomes than for the post-1969 dates PortfolioCharts shows as their worst SWR start dates) than the ones with start dates in the 1970-2000 period.
This is why I compared my numbers since 1970 against the Pfau data that went back to 1900 for a bunch of different countries. In the non-war cases, my numbers measurably overstate the SWRs calculated from the much longer timeframes by an average of 0.2%. So whatever number you expected it to be, just keep that calibration point in mind.

I'm not opposed to adding more years, though. I'm just careful about sources and also have to keep in mind the assets that are harder to find.
User avatar
Nestegg_User
Posts: 2114
Joined: Wed Aug 05, 2009 1:26 pm

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by Nestegg_User »

Tyler

I wondered why, in your cloud visualization for optimum non-gold, that you used the lowest Ulcer index point for US based (with resultant 2.85 swr) whereas for other countries you used the highest return but with lowest Ulcer Index. It seems you should have been more consistent. By my eyeball, had the same characteristics been used for US data the SWR would have been ~3.5%.

Overall, seems fairly consistent with the work by Estrada who has done a lot of work on international developed market SWRs. He shows a SWR of 3.5% covers most developed nations, outside of war years where they might be very low (obviously RUS is zero, post 1918 and JPN and GER are close). That's also why I didn't really consider Pfau's "An International Perspective on SWR...." paper since it used 1900-2008 (for retirees 1900-1979, to give 30 year numbers), since it includes such war, and hyperinflation (for GER), in its numbers. {I'd agree with others that post-1950, or perhaps post-1960, might give a better indication of non-war years for developed, although there would obviously still be somewhat of a penalty to those who suffered the conflict, both "winners" and "losers" from only the perspective of war outcome.

Personally, I use a 3.5% wr as max, having examined both Estrada, Pfau, and others. Further, while a lower equity level for optimum SWR was 30-35%, there wasn't that much gained above ~50% equities for many developed markets (even Pfau's paper showed a leveling off for US).
Alas, there isn't sufficient data for TIPs, as I would find that a more preferred inflation diversifier than commodities (where one needs to be concerned with contango and backwardation), as TIPs still would deliver some return in " normal" markets assuming not purchased at negative yields as we had for a period.

Thanks for your work
seajay
Posts: 1700
Joined: Sat May 01, 2021 3:26 pm
Contact:

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by seajay »

Nestegg_User wrote: Thu Apr 04, 2024 3:36 am Overall, seems fairly consistent with the work by Estrada who has done a lot of work on international developed market SWRs. He shows a SWR of 3.5% covers most developed nations, outside of war years where they might be very low (obviously RUS is zero, post 1918 and JPN and GER are close).
Does that reflect costs/taxes? Or perhaps reflective of 10% nominal, 35% average costs/taxes? In the UK 1968 for instance the Chancellor of the time (Roy Jenkins) imposed a retrospective tax that took the top rate to 130%. Around those years the Beatles were singing 'Taxman' ... 19 for you 1 for me ... in reflection of 95% taxation rates. Even Joe-Average was paying near 50%. Under such circumstances you need movable assets that ideally generate no regular (taxable) income flow footprint and plausible deniability against otherwise geopolitical confiscation. Reasonable taxes are fair, punitive taxation is confiscation.
StillGoing
Posts: 378
Joined: Mon Nov 04, 2019 3:43 am
Location: U.K.

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by StillGoing »

Alpha4 wrote: Wed Apr 03, 2024 7:50 pm
Tyler9000 wrote: Tue Apr 02, 2024 10:12 am
Kevin K wrote: Mon Apr 01, 2024 4:36 pm Tyler at Portfolio Charts (who also frequently posts on these forums) has an impressive new post on this topic:

https://portfoliocharts.com/2024/04/01/ ... more-70357

It's a long and detailed post that I think will be of particular interest to folks who've been reading the Cederburg study (which is discussed along with the work of Kitces and Pfau). I appreciate the way that Tyler politely challenges conventional thinking, backed by a ton of data.
Thank you for sharing this, Kevin K. Yeah, it's a bit long. But it's the culmination of a crazy amount of work so I guess I had a lot to say. :D I don't claim to have all the answers, but I do have a lot of data for discussion. I hope everyone sticks with it, learns something, and perhaps thinks about withdrawal rates in a new way.
Tyler,

Wow. Pretty impressive.

The Netherlands and the UK weren't quite as bad (the former saw about a 1/3 drop in stock market value from the start of 1961 to the end of 1966; recovered to a bit above where it was in 1961 by late 1969; fell again until 1971 but never to its 1966 lows; the latter had two barnburner years in 1958 and 1959--up over 50% each year--but then had three "meh" years in 1960-1962 (averaged barely 1% nominal total return those years); had a decent year in 1963 (up almost 20% in nominal total return terms), and then had two somewhat bad years and one slightly decent one in the 1964-1966 period such that by the end of 1966 the portfolio in total return terms was just a hair under what it was at the start of 1960. 1967 and 1968 were good years (although nowhere near 1958 and 1959), 1969 and 1970 stunk, 1971 was very good (a roughly 34% total return) and then from 1972-1974 UK stocks put in such poor returns (1974 especially was awful with UK shares losing half their value in real total return terms in one year) that for every one Pound the investor had at the start of 1960 it was only worth approximately £0.54 at the end of 1974. Of course, 1975-1989 was virtually one huge (relatively) rarely interrupted bull market for UK stocks but given that the first 14-15 years of performance weigh very heavily in determining one's SWR this was probably too late to do much good for the UK investor who started withdrawing in 1960 or 1961 (IIRC I remember reading somewhere that 1960 was the worst post WWII start date for a balanced all-UK (i.e. only UK shares and gilts; no foreign stocks, no foreign bonds, and no gold) portfolio and that it did fail the 4% test.
Agree with you about Tyler's work here...

So far, all the worst years for the UK occurred pre-WWII. The following graph uses a 50/50 domestic stock/bond portfolio with stock returns and inflation from macrohistory.net and bond returns for the 'all stocks' gilt sector* (UK gilts are referred to as government stocks - hence the name) from a new database of gilt returns described at https://papers.ssrn.com/sol3/papers.cfm ... id=4742450 and available to dowload). Note that the 'all stocks' sector was dominated by undated gilts before about 1920 and had a much longer duration that it does now through to the 1930s and 1940s).

Image

For the 50/50 allocation, post-WWII, the UK just about scraped through with a 4% SWR. Obviously, using a different stock allocation or maturity of gilt sector would have changed the details (e.g., using over 15 year gilts - i.e., closer in maturity to those in the macrohistory.net database, the SWR dropped below 4% in a few of the retirements starting in the 1960s), but the worst cases were still pre-WWIl (for 100% stocks, the retirement starting in 1969 came fairly close to setting a record low)

cheers
StillGoing

* sector here refers to a range of maturities, e.g., 0 to 10 years, and is therefore analogous to a fund.
User avatar
Tyler9000
Posts: 757
Joined: Fri Aug 21, 2015 11:57 am

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by Tyler9000 »

Nestegg_User wrote: Thu Apr 04, 2024 3:36 am I wondered why, in your cloud visualization for optimum non-gold, that you used the lowest Ulcer index point for US based (with resultant 2.85 swr) whereas for other countries you used the highest return but with lowest Ulcer Index. It seems you should have been more consistent. By my eyeball, had the same characteristics been used for US data the SWR would have been ~3.5%.
For the 3-pane animation for non-gold portfolios, each image looks at all countries. I just switched between the same 3 measures that I looked at earlier -- #1 ranked by SWR, UI, and Risk-Adjusted. The US just happened to have the worst case for the top-UI portfolio.
User avatar
Nestegg_User
Posts: 2114
Joined: Wed Aug 05, 2009 1:26 pm

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by Nestegg_User »

seajay wrote: Thu Apr 04, 2024 6:20 am
Nestegg_User wrote: Thu Apr 04, 2024 3:36 am Overall, seems fairly consistent with the work by Estrada who has done a lot of work on international developed market SWRs. He shows a SWR of 3.5% covers most developed nations, outside of war years where they might be very low (obviously RUS is zero, post 1918 and JPN and GER are close).
Does that reflect costs/taxes? Or perhaps reflective of 10% nominal, 35% average costs/taxes? In the UK 1968 for instance the Chancellor of the time (Roy Jenkins) imposed a retrospective tax that took the top rate to 130%. Around those years the Beatles were singing 'Taxman' ... 19 for you 1 for me ... in reflection of 95% taxation rates. Even Joe-Average was paying near 50%. Under such circumstances you need movable assets that ideally generate no regular (taxable) income flow footprint and plausible deniability against otherwise geopolitical confiscation. Reasonable taxes are fair, punitive taxation is confiscation.
In his "Maximum Withdrawal Rates: An Empirical and Global Perspective" (J. Retirement, 2018) he doesn't include taxes in any of the 21 countries, which is likely difficult to determine over the 1900-2014 period it surveys, nor does he include expenses of the instruments (again likely a difficult enterprise over the years and countries, and I'm not aware of others including them over that time period with that level of international depth).

In his Appendix A2 he gives the max withdrawal for P1, P5, P10, P90, P95, and P99 for twenty countries. In looking at the P5 for the countries the SWR for 50/50 was generally 3.4 -3.8 for the non-war-torn countries, with the exception of Switzerland at 3.1 and Ireland at 2.7. This is why I considered it fairly consistent with his work.

It would be difficult to determine the level of equity investment in the US over that time frame, much less the rest of the developed markets; how much "flight" occured in markets like GER, Austria, South Africa in various periods and how did it affect their markets. Indeed, how much of the markets could be realized by "ordinary investors" for such early periods I consider mostly academic as I remember both the high cost of odd-lot stock purchases (average investor probably couldn't have reasonably afforded whole lot purchases for many stocks) and the high bid-ask spread in the early to mid-70's, and I doubt earlier periods were better.
StillGoing
Posts: 378
Joined: Mon Nov 04, 2019 3:43 am
Location: U.K.

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by StillGoing »

Nestegg_User wrote: Thu Apr 04, 2024 9:52 am
seajay wrote: Thu Apr 04, 2024 6:20 am
Nestegg_User wrote: Thu Apr 04, 2024 3:36 am Overall, seems fairly consistent with the work by Estrada who has done a lot of work on international developed market SWRs. He shows a SWR of 3.5% covers most developed nations, outside of war years where they might be very low (obviously RUS is zero, post 1918 and JPN and GER are close).
Does that reflect costs/taxes? Or perhaps reflective of 10% nominal, 35% average costs/taxes? In the UK 1968 for instance the Chancellor of the time (Roy Jenkins) imposed a retrospective tax that took the top rate to 130%. Around those years the Beatles were singing 'Taxman' ... 19 for you 1 for me ... in reflection of 95% taxation rates. Even Joe-Average was paying near 50%. Under such circumstances you need movable assets that ideally generate no regular (taxable) income flow footprint and plausible deniability against otherwise geopolitical confiscation. Reasonable taxes are fair, punitive taxation is confiscation.
In his "Maximum Withdrawal Rates: An Empirical and Global Perspective" (J. Retirement, 2018) he doesn't include taxes in any of the 21 countries, which is likely difficult to determine over the 1900-2014 period it surveys, nor does he include expenses of the instruments (again likely a difficult enterprise over the years and countries, and I'm not aware of others including them over that time period with that level of international depth).
Not SWRs exactly but Armitage, Returns after personal tax on UK equity and gilts, 1919-98 (available at https://www.pure.ed.ac.uk/ws/portalfile ... ug_02_.pdf) did look at returns after taxes - a reduction of about 2 percentage points on the arithmetic means for equities and 1.5 percentage points for gilts. As the author says, "The main reason why returns have been measured before personal tax is probably that investors have different tax circumstances".

I also note that until 1996(?) gilt (government bond) yields in the UK were sometimes quoted both before and after tax (if my understanding is correct, they were taxed at source before that date). The books by Pember and Boyle listing all UK gilts from 1900 onwards (at https://www.escoe.ac.uk/research/histor ... scal-data/) has some examples of this.

cheers
StillGoing
User avatar
nisiprius
Advisory Board
Posts: 52446
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by nisiprius »

GAAP wrote: Tue Apr 02, 2024 3:05 pm Gold in the USA since 1970? Outside of dentists and personal jewelry, ownership of gold was a felony in 1970, and didn't become legal until 1974.

How is this handled with this analysis?

Does anyone know when these countries adopted a true fiat currency? Gold is a great deflation hedge in a gold-standard world, but fiat currencies don't really have that problem.
The problem with looking at gold is that, like commodities, 1) gold is extremely bursty--far burstier than stocks, which are bursty enough, and 2) the mean time between bursts isn't much shorter than the mean time between government policy changes that create discontinuous data regimes and fresh restarts. With gold, every time period is a special case.

The worst inflation in the United States since the beginning of the CPI in 1913 occurred around 1915 to 1920, when the United States was on the gold standard. I'm not even sure I know what it means to have an allocation to gold before 1971--all portfolios had an automatic, built-in gold allocation before 1971. (Some gold advocates have tried to say that only legitimate meaning for inflation is an increase in the price of gold, and therefore there was no inflation from 1913-1934 because the price of gold remained constant at $20.67 throughout this period).
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.
User avatar
aj76er
Posts: 1194
Joined: Tue Dec 01, 2015 10:34 pm
Location: Austin, TX

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by aj76er »

I read through the high-level takeaways and looked over the top portfolio. The first thing that comes to mind is that it loads up on traditional inflation fighters (e.g. gold and commodities), but in today’s world we have an opportunity to buy positive real yielding TIPS and IBonds. I understand that the data on these assets are limited, but you don’t need data to tell you that a duration matched TIPS ladder yielding 2% real is going to give much more consistent returns than gold and commodities.
"Buy-and-hold, long-term, all-market-index strategies, implemented at rock-bottom cost, are the surest of all routes to the accumulation of wealth" - John C. Bogle
randomguy
Posts: 11313
Joined: Wed Sep 17, 2014 9:00 am

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by randomguy »

DonIce wrote: Mon Apr 01, 2024 9:21 pm The biggest and most controversial (to bogleheads) conclusion seems to be the large (30%) allocation to gold that comes out in the top SWR portfolios. But I've generally seen the same playing around with historical data over varying time periods on portfolio visualizer... you often get the best results with gold heavy portfolios across all the available historical data in PV.

As for the large allocation to foreign stocks... got some questions about the methodology here. It seems like the best SWR portfolio being composed of a larger percentage of foreign stocks is done from the perspective of looking from within different countries. So for example if one looks at Italy and decides on a domestic vs foreign allocation, you'll want a large foreign allocation since that's where the historically higher performing US stocks were. Same for looking from inside Japan, from inside France, Spain, etc. For investors in countries besides the US who invest in foreign stocks according to market cap weight, US stocks make up >50% of that foreign stock allocation. And then when you average across all these perspectives, that tells you to invest in foreign stocks. But really all that it's telling you is that from the perspective of being in most countries, you would have wanted a large percentage allocation to US stocks and a smaller allocation to domestic stocks. So this data probably needs to be normalized based on something rather than just giving each country's perspective an equal weight. I think once you do a reasonable normalization here, the best result regardless of one's nationality would be roughly to invest in world stocks according to their global market cap weight.
All these studies basically come down to finding out what performed well in the 1930s and the 1966-1981 period. Gold sort of handles the latter. So do international stocks. The 1930 case is harder in that gold wasn't really traded in the US and international stocks are a bit harder to evaluate. The crash and recoverys were offset enough to gain some diversification bonuses but how much will depend on what data sets you are using

And yes to some extent this is telling everyone to diversify like mad. That 1989 japanese guy who was market cap weighted did OK. So did the 1917 russian investor. And it is probably even more important in fixed income to avoid things like the 1920s germany were your bonds became worthless. The hard part is that the odds are some other strategy will work better. That 1980 retire who sunk 25% of their money in gold had a long period of a huge amount of portfolio drag. The 2010 retiree has seen international do the same.
User avatar
watchnerd
Posts: 13800
Joined: Sat Mar 03, 2007 10:18 am
Location: Gig Harbor, WA, USA

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by watchnerd »

Lawrence of Suburbia wrote: Mon Apr 01, 2024 8:53 pm Did I read his comment correctly, about how great a big allocation to cash allows for great portfolio survivability vs. inflation? That was shocking. As is the idea about owning ex-U.S. stocks being superior in some scenarios.

My ADD probably has me screwed up; I couldn't follow a lot of that piece. Probably have it backwards.
Why is that shocking?

Cash held from 1972 to the present has, in aggregate, beaten inflation.

$10,000 in cash in 1972 is worth $13,483, *inflation adjusted*, today.

https://www.portfoliovisualizer.com/bac ... zhrurnRA8y
Global stocks, IG/HY bonds, gold & digital assets at market weights 75% / 19% / 6% || LMP: TIPS ladder
Topic Author
Kevin K
Posts: 360
Joined: Sun Aug 26, 2007 7:47 pm

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by Kevin K »

aj76er wrote: Thu Apr 04, 2024 2:39 pm I read through the high-level takeaways and looked over the top portfolio. The first thing that comes to mind is that it loads up on traditional inflation fighters (e.g. gold and commodities), but in today’s world we have an opportunity to buy positive real yielding TIPS and IBonds. I understand that the data on these assets are limited, but you don’t need data to tell you that a duration matched TIPS ladder yielding 2% real is going to give much more consistent returns than gold and commodities.
I think this is an important point, though Tyler himself has written eloquently about gold NOT being an inflation-fighting asset per se, but rather being valuable when interest rates are low and during market panics:

https://portfoliocharts.com/2020/08/21/ ... e-of-gold/


Tyler doesn't include TIPS because there's so little date (only since 1997) but I think there's an argument to be made that the gold data that ought to be used is for the paper version - and GLD (the first gold ETF) was introduced in 2004. Interestingly enough its sponsoring firm thinks the existence of GLD has had a major impact on the gold market altogether:

https://www.spdrgoldshares.com/media/GL ... f_Gold.pdf

So paper gold - the default choice for most investors - hasn't been around much longer than TIPS. And invaluable as a duration-matched TIPS ladder can be, a simple short-term TIPS ETF like VTIP has been a much better inflation hedge (and had far better risk-adjusted returns) than both longer-term TIPS funds and GLD since its inception.
abc132
Posts: 2501
Joined: Thu Oct 18, 2018 1:11 am

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by abc132 »

I think the value of a tool like this for me is to consider what to do if we have run-ups that are much higher than the norm. So in Japan I have just had my stock portfolio go up by a factor of 6 (say 2 million to 12 million) in 10 years (1980's) and I wonder what to do with my portfolio. Using this tool, 10% Japan stock, 30% non-Japan stock, 20% bond, 10% commodities, and 30% gold would be an appropriate portfolio after having my portfolio go up by a factor of 6 in 10 years.

It's tough for me to see how the worst global sequence is helpful under any other scenario. They usually follow the biggest run-ups and the SWR results are not typical for any other condition.

I can say if my goal is 3 million (say 100k/year spending) that I will have taken this 12 million and it will already have 3 million in a TIPS ladder and I will be unconcerned with meeting that 1.5% safe withdrawal rate. It looks like I could withdraw 1.5% of that extra 9 million so I am at around 2.5x what I desired in withdrawal (250k/year) with this 1.5% SWR. It's hard to justify fixating on that 1.5% SWR as this is still a great retirement sequence. I need not plan around this ultra-low SWR when I am at 2 million and in normal market conditions.
seajay
Posts: 1700
Joined: Sat May 01, 2021 3:26 pm
Contact:

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by seajay »

StillGoing wrote: Thu Apr 04, 2024 1:00 pm
Nestegg_User wrote: Thu Apr 04, 2024 9:52 am
seajay wrote: Thu Apr 04, 2024 6:20 am
Nestegg_User wrote: Thu Apr 04, 2024 3:36 am Overall, seems fairly consistent with the work by Estrada who has done a lot of work on international developed market SWRs. He shows a SWR of 3.5% covers most developed nations, outside of war years where they might be very low (obviously RUS is zero, post 1918 and JPN and GER are close).
Does that reflect costs/taxes? Or perhaps reflective of 10% nominal, 35% average costs/taxes? In the UK 1968 for instance the Chancellor of the time (Roy Jenkins) imposed a retrospective tax that took the top rate to 130%. Around those years the Beatles were singing 'Taxman' ... 19 for you 1 for me ... in reflection of 95% taxation rates. Even Joe-Average was paying near 50%. Under such circumstances you need movable assets that ideally generate no regular (taxable) income flow footprint and plausible deniability against otherwise geopolitical confiscation. Reasonable taxes are fair, punitive taxation is confiscation.
In his "Maximum Withdrawal Rates: An Empirical and Global Perspective" (J. Retirement, 2018) he doesn't include taxes in any of the 21 countries, which is likely difficult to determine over the 1900-2014 period it surveys, nor does he include expenses of the instruments (again likely a difficult enterprise over the years and countries, and I'm not aware of others including them over that time period with that level of international depth).
Not SWRs exactly but Armitage, Returns after personal tax on UK equity and gilts, 1919-98 (available at https://www.pure.ed.ac.uk/ws/portalfile ... ug_02_.pdf) did look at returns after taxes - a reduction of about 2 percentage points on the arithmetic means for equities and 1.5 percentage points for gilts. As the author says, "The main reason why returns have been measured before personal tax is probably that investors have different tax circumstances".

I also note that until 1996(?) gilt (government bond) yields in the UK were sometimes quoted both before and after tax (if my understanding is correct, they were taxed at source before that date). The books by Pember and Boyle listing all UK gilts from 1900 onwards (at https://www.escoe.ac.uk/research/histor ... scal-data/) has some examples of this.

cheers
StillGoing
Thanks for the net returns pointers StillGoing

From my own personal perspective, third in land (home value), 4% assumed imputed rent benefit (1.33% proportioned). Return of capital over 30 years (3.33% SWR) where 1.33% is via imputed rent, leaves 2% (3% proportioned to remainder two-thirds of liquid assets). Ideally ending 30 years with home + some residual inflation adjusted liquid asset wealth still remaining (for heirs).

Previously I had measured gold and bonds providing similar outcomes when combined with stocks. However confiscations/partial defaults do repeatedly occur across history and in a Fiat era the methodology of that is simplified ... via taxation, typically income (interest/dividends) taxation. 1960's UK for instance when the Beatles complained about 95% taxation rates (Taxman). Assets that generate no regular income flows, capital gains only (such as gold) are more defensive in that respect, as-is having some assets outside of the regular financial system, might be considered as being a better risk-reduction factor/element

With bond (Gilts) interest and stock dividends taxed at the typical (average) rate i.e. applying the 1919-1999 data source you highlighted ...

Image

I remain comfortable that thirds each land/stock/gold (or silver if gold is prohibited) remains a reasonable choice.

Rebalanced or not - is inclined to yield similar outcomes, with non rebalanced you just end up with a higher weighting in the asset(s) that performed the best. Spending from the ongoing best in part reduces the tendency to see asset weightings drift, and there's always the option to rebalance infrequently if/when weightings drift is deemed to have become too excessive.

Given easier availability of global stock nowadays, I consider a global stock index to be a reasonable choice to combine with a domestic home/land, and gold, where gold should be physical (in-hand), otherwise it is within the regular financial-system and more open to being fully or partially confiscated/taxed.

I've striven to avoid morality/legality issues here, but would point out that when fair taxation is increased to punitive levels, such as UK 1968 when top rate taxation was increased to 130%, some/many celebrity names opted to flight the country. Rolling Stones, David Bowie ...etc. i.e. for the comfortably rich, having multiple homes/land in different geopolitical regions is yet another form of risk-reduction/diversification. Countering that there's been a tendency towards having to report all your wealth/assets to state(s), which is perhaps indicative that the next round of punitive taxation (confiscation) is perhaps closer rather than more distant.
seajay
Posts: 1700
Joined: Sat May 01, 2021 3:26 pm
Contact:

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by seajay »

martincmartin wrote: Tue Apr 02, 2024 11:00 am
Tyler9000 wrote: Tue Apr 02, 2024 10:50 am All good points. Be sure to read the Methodology portion of the article and look for the "Trust, but Verify" section. I provide detailed data that directly addresses your concerns.
Yes, thanks, and it is a kind of maddening problem. On the one hand, in some sense even 150 years of stock & bond data in the US isn't enough. Three bad periods isn't really enough either. So we have to make due with limited data. On the other hand, it seems foolish to discount things like international bonds, commodities and gold simply because we only have about 55 years of data. So I wrestle with all of this.

As a follow on, what does your analysis say about glide paths? Does increasing stocks over time in retirement help in other countries as well? McClung in Living Off Your Money seems to think so.
In the absence of insufficient data, decisions may be made based on assumptions/general observation. Investment rewards arise out of three elements, price appreciation, income, volatility capture. Broadly each of those might compare equally, for if that were not the case then investors would concentrate into the sole more consistently productive choice. Options traders might focus upon volatility, income investors might seek out dividends/bond interest, price appreciation investors might focus on growth stocks. A reasonable choice is to diversify across all three elements.

Some assets have expectation of negating inflation, bonds/gold, other assets have expectations of seeing both the price of the asset rise with inflation, whilst also generating dividends (or imputed rent) along the way, perhaps where the dividends might broadly compare to cash deposits.

Stocks and/or land (home+imputed, worked land that yields harvests/dividends) are distinctly different to idle land, gold, bonds, might be categorized as being ‘equities’ or ‘productive assets’ that perhaps broadly yield 2i (2x inflation). Bonds/gold/idle-land are non-productive – might generally/broadly yield 1i.

Should you invest in productive (2i) assets solely? Well many assets are volatile, may see real prices double, only to later halve back down again. Or halve, and then double, providing the potential to trade that volatility productively (yield a form of ‘dividend’).

Foreign stocks incorporate elements of stock gain/loss plus/minus currency (FX) gain/loss. Sometimes the FX may attenuate volatility, other times amplify volatility.

For a British investor, a third of their wealth in a home (that yields imputed rent benefit), a third in US stocks (so stock +/- FX), a third in gold (that tends to be volatile) along with rebalancing, can yield overall total return rewards similar to that of stocks along. As can non-rebalanced. With non-rebalanced the tendency is to end with a higher weighting in the assets that yielded the highest total return reward over that period, low weighting in the asset(s) that performed poorly. That tends to overall compare to had you rebalanced the portfolio yearly. Multiple assets that are volatile and don’t correlate tends to also reduce overall portfolio volatility.

Three currencies, GB Pounds invested in land (home), US Dollars invested in stocks, gold commodity global non-fiat currency; Three assets, land, stocks, commodity. If over one period stocks halved in value so might gold double in value; Perhaps later seeing the opposite, gold halving, stocks doubling. Yielding three sources of income, imputed rent, dividends, volatility capture (or the benefits arising from ending with a high weighting in the most productive asset(s)).

From a geopolitical and/or counter-party risk perspective, holding (in-hand) two thirds in land and gold along with a third in stocks that have counter-party risk but where liquidation time is just a couple of days (T+2) has concentration risk diluted down to around a third. Concentration risk is a major risk factor, far better to be down a third than enduring a 100% loss due to over-concentration into a single asset/basket.

You can dig out historic data and backtest, or you can just make some assumptions such as above. Primarily however the fundamentals are the same, a combination of assets with different general characteristics/tendencies (high/low volatilities, correlations, sources of gains, geopolitical risk factors etc.) that in combination has the tendency to smooth down the overall portfolio volatility whilst maintaining overall real gains/rewards. Time diversification, accumulating over many years (so sometimes buying high, other times buying low) and the same for drawdown/retirement, drawing over many years, builds in yet another form of smoothing (volatility reduction).

The average investor might be investing for 70 years, 40 years accumulation, 30 years of drawdown. Much can happen over such lengths if time. Diversifying across multiple currencies, assets, geopolitical risks is less inclined to endure critical failure.

Even the ancient Talmud appreciated concentration risk millennia ago, when they advocated a third in merchandise (stocks), a third in-hand (gold), a third buried in the ground (land). That still holds and for instance a Japanese investor who owned a domestic home, held US$ invested in US stocks, and gold … rode comfortably though periods of high Japanese stock volatility since the 1970’s. As would a British investor over the last 120+ years despite the collapse of the British Empire, two world wars etc. Even for countries where perhaps their home was lost, having a additional two thirds in US$ invested in US stocks, and in-hand gold, have had them far better placed (relatively wealthier) than those who had fully concentrated into their domestic currency/markets alone.
User avatar
Lawrence of Suburbia
Posts: 697
Joined: Mon Aug 08, 2022 12:04 pm

Re: What Global Withdrawal Rates Teach Us About Ideal Retirement Porfolios

Post by Lawrence of Suburbia »

watchnerd wrote: Thu Apr 04, 2024 6:21 pm
Lawrence of Suburbia wrote: Mon Apr 01, 2024 8:53 pm Did I read his comment correctly, about how great a big allocation to cash allows for great portfolio survivability vs. inflation? That was shocking. As is the idea about owning ex-U.S. stocks being superior in some scenarios.

My ADD probably has me screwed up; I couldn't follow a lot of that piece. Probably have it backwards.
Why is that shocking?

Cash held from 1972 to the present has, in aggregate, beaten inflation.

$10,000 in cash in 1972 is worth $13,483, *inflation adjusted*, today.

https://www.portfoliovisualizer.com/bac ... zhrurnRA8y
I mistakenly assumed 'cash' meant like money in a checking account, rather than something like <three-month Treasury bills(?).

This does make me think, "Why TIPS ladders ...?"
74% VTHRX/8% DODWX/12% TIAA Traditional/6% SWVXX
Post Reply