The Worst Years To Retire: The Surprise In The Data

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
BernardShakey
Posts: 1147
Joined: Tue Jun 25, 2019 10:52 pm

Re: The Worst Years To Retire: The Surprise In The Data

Post by BernardShakey »

Normchad wrote: Mon Oct 25, 2021 6:10 pm Oddly, a lot of folks here talk like they are convinced the worst tear to retire is *right now*. With all the hand wringing over 2% or lower SWRs…..
Exactly. Hoping 2022 doesn't become the new 1966. Would be just my luck. :shock:
An important key to investing is having a well-calibrated sense of your future regret.
nigel_ht
Posts: 4742
Joined: Tue Jan 01, 2019 9:14 am

Re: The Worst Years To Retire: The Surprise In The Data

Post by nigel_ht »

BernardShakey wrote: Tue Oct 26, 2021 10:05 pm
Normchad wrote: Mon Oct 25, 2021 6:10 pm Oddly, a lot of folks here talk like they are convinced the worst tear to retire is *right now*. With all the hand wringing over 2% or lower SWRs…..
Exactly. Hoping 2022 doesn't become the new 1966. Would be just my luck. :shock:
We’ve been assured it’s transitory…
alfaspider
Posts: 4816
Joined: Wed Sep 09, 2015 4:44 pm

Re: The Worst Years To Retire: The Surprise In The Data

Post by alfaspider »

Valuethinker wrote: Tue Oct 26, 2021 12:39 pm
alfaspider wrote: Tue Oct 26, 2021 10:52 am
BernardShakey wrote: Mon Oct 25, 2021 9:50 pm
Nestegg_User wrote: Mon Oct 25, 2021 9:21 pm I'm gonna still say that the ~1905 retiree was a decent second place....the 1907 recession was just ahead...massive inflation in the ww1/twenties years ....and yeah, that thing called the income tax arrived shortly after they retired...and of course they ended with the great depression
(of course they first had to get to a point to be able to retire... and if we look at the economic history of the prior ~50 years...it would have been "challenging")
Agree, except a 1905 retiree would very likely have been dead by the 1929 crash given life expectancy at that time.
Life expectancy changes have largely come through improvements in infant mortality. You were more likely to keel over unexpectedly from a heart attack in old age, but it wasn't unusual to make it to 80 even in that era.
That's what I had understood. But then I watched this series (PBS/ BBC):

https://www.pbs.org/about/about-pbs/blo ... ic-health/

the effect of antibiotics cannot be understated. Surgery was a very high risk activity before the invention of modern antibiotics. Plus a whole set of dread diseases (like smallpox) that were floating around.
However, I would say that "retirement" as we know it is a relatively recent invention. For most of human history, you worked until you were unable to do so. Not-working at all for long periods of time was generally a luxury of the aristocracy, although they often participated in activities that today might be considered "work" (i.e. office-type administrative tasks). In 1905, "retiring" just because you could was still not an expectation. They were still decades away from the creation of social security, which helped cement the idea amongst the general populace that one would have a long-term period of idleness not mandated by infirmity.
I think when the retirement pension was brought in by the Liberal government in the UK (1904 from memory) it was set at age 65 for men, and the average retired life expectancy at age 65 was only a couple of years longer than that.
If you look at this, men have really only gotten an extra 3 years on average since 1940.

https://www.ssa.gov/history/lifeexpect.html
McQ
Posts: 1425
Joined: Fri Jun 18, 2021 12:21 am
Location: California

Re: The Worst Years To Retire: The Surprise In The Data

Post by McQ »

nisiprius wrote: Mon Oct 25, 2021 6:24 pm If I'd been asked to guess without seeing the answer first, I'm pretty sure I'd have said "1966-1982."

I don't think it's a big surprise to anyone who's looked at a chart of stock market real returns. The famous and much-reproduced chart from Siegel's Stocks for the Long Run, for example. If I color in the places where the stock market returns are below Siegel's trend line--which might be characterized where it was performing below expectation--we can see that the magnitude of shortfall in 1966-82 was certainly comparable to 1929-1944, and that it was more sustained and unrelieved, with no letup and no opportunity to recoup shortfalls and rebuild a portfolio shrinking from constant withdrawals.

Image

1832-1850 looks worse than either. I don't think Simba's data goes back that far, though?
Hello all: I recompiled the 19th century history of US stock and bond returns and found lower stock returns before 1871 than Siegel, and higher bond returns after 1857. The paper is here: https://papers.ssrn.com/sol3/papers.cfm ... id=3805927

So I have data for the 1830s, 1840s, and 1850s.

Below are two tables that may be relevant to today’s discussion.

Image

This first table shows the worst doldrums in the US stock market, defined as a return <1% real annualized over a period of a decade or more. The stretch following 1966 is prominent among them but hardly unique.

Image

This second table shows the worst 20-, 30- and 50-year returns in the US stock market since 1793. Once again, early 1982 appears as a major bottom, but only on the relatively brief 20-year time scale.

Much worse drawdowns have occurred internationally, and not just in war-torn nations, but that is a topic for another day (see Table 5 in the paper).

None of these raw returns address the situation of a retiree who must make annual withdrawals and is thus subject to sequence of returns risk. A project under way does identify the mid-1960s as one of the very worst times to retire with a balanced portfolio. Around 1910 was another bad one. Conversely, as with 1929, none of the 19th century "retirement" dates were so bad ... because of deflation. The Bengen test goes down with deflation--fewer and fewer dollars have to be withdrawn, even as the real return on the bond portion of the portfolio soars.

In short: I have good 19th century data, but I did not find a worse date to retire than the mid-1960s in the US, and the culprit, as mentioned by other posters, is inflation.
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.
User avatar
Topic Author
martincmartin
Posts: 900
Joined: Wed Jul 02, 2014 3:04 pm
Location: Boston, MA USA

Re: The Worst Years To Retire: The Surprise In The Data

Post by martincmartin »

Thanks, this is very useful! I've come to realize, as have others, that analysis like the Trinity study ends up focusing on 1966 - 1982, because it's so bad. So when evaluating decisions, it might make sense to explicitly identify other bad stretches and back test separately with them. For example, I'm currently wondering about bond duration, my impression is "long term bad, short term good, intermediate term just ok." That's certainly true in 1966 - 1982, and therefore pops out of Trinity style analysis. So what other periods should I use?

cFIREsim shows us. Using a "stock heavy" fixed AA, i.e. 67/33, 30 years 4% SWR, we get:

1906 - 1935

1937 - 1966

1966 - 1995

1990 - 2019

For 20 years:

1906 - 1925

1937 - 1956

1966 - 1985

2000 - 2019
Last edited by martincmartin on Thu Oct 28, 2021 6:05 am, edited 1 time in total.
UpperNwGuy
Posts: 9479
Joined: Sun Oct 08, 2017 7:16 pm

Re: The Worst Years To Retire: The Surprise In The Data

Post by UpperNwGuy »

SnowBog wrote: Tue Oct 26, 2021 6:26 pm In theory, I'd agree with that...

But in this case, we have to account for politics which don't always operate on logic.
Let's keep politics out of it, please.
seajay
Posts: 1656
Joined: Sat May 01, 2021 3:26 pm
Contact:

Re: The Worst Years To Retire: The Surprise In The Data

Post by seajay »

A historic very bad case might have seen 4% SWR all spent after 20 years, seen a 65 year old through to their mid 80's, with a modest risk of not even living that long. With a plan B of a home value to fund downsizing or selling to fund care home costs, perhaps further supplemented with some pension income, even the worst historic cases might not have been so bad. In the average case the outcome was pretty good, in many cases still seeing the same or more wealth as the inflation adjusted start date wealth, in some cases considerably more, multiples of inflation adjusted start date wealth.

The indications I'm seeing are that it didn't really matter that much as to what asset allocation you held, T-Bills alone were the more assured to leave the least, all stock was inclined to maximise the legacy. Worst cases for either/both tended to align - as others have indicated driven by high inflation.

The greater risks are physical issues. Not living as long as hoped, or physical/mental issues that that have life being unbearable or that involve high levels of care costs.
User avatar
willthrill81
Posts: 32250
Joined: Thu Jan 26, 2017 2:17 pm
Location: USA
Contact:

Re: The Worst Years To Retire: The Surprise In The Data

Post by willthrill81 »

seajay wrote: Thu Oct 28, 2021 6:22 am The indications I'm seeing are that it didn't really matter that much as to what asset allocation you held, T-Bills alone were the more assured to leave the least, all stock was inclined to maximise the legacy.
If all you're talking about is varying the allocation between total stock market and bonds, then yes, tweaking the AA didn't usually change the SWR very much. But if we expand AA to include things like factors and gold (from 1975 and onward), then it absolutely made a big difference in the SWR.
The Sensible Steward
jharkin
Posts: 3469
Joined: Mon Mar 28, 2016 7:14 am
Location: Boston suburbs

Re: The Worst Years To Retire: The Surprise In The Data

Post by jharkin »

seajay wrote: Mon Oct 25, 2021 6:02 pm Therein lies often overlooked risk. Back then regular/basic rate taxpayers were seeing 38% tax rates in the UK, whilst richer individuals were paying 95% rates (or even more). The Stones self exiled from the UK into France due to such, whilst the Beatles sang 'Taxman' ... 19 for you, 1 for me ... in reflection of 95% tax rates. In 1968 there was even a retrospective taxation applied that increased the top rate tax to 130%. A Labour (Democrats equivalent) UK party with irresponsible tax/spend policies. Such governments might suddenly rise at any time and reek havoc.

Old Money, generational wealth favors 'a third, a third, a third' ... land, art, gold. Assets that can be just left to ride through 'bad times' without generating regular taxable income streams such as bond interest/stock dividends.
“Wreaking havoc” is a matter of perspective - if you are already wealthy and privileged of course you believe you are being penalized.

An alternative view is that 50-90% top marginal rates where an excellent deterrent against extreme income inequality and a major net benefit to the other 99.5% of us. The tax rates encouraged companies to reinvest in business growth and fair wages for the rank and file employees, rather than hoarding cash on the books and paying a handful of executive 500x what the average employee makes.

Those well paid rank and file employees turn round and spend a far greater proportion of income than the CEO class, in turn helping to drive overall economic growth and prosperity.
olyveoil
Posts: 243
Joined: Thu Feb 04, 2021 7:47 am

Re: The Worst Years To Retire: The Surprise In The Data

Post by olyveoil »

Not really seeing how this is a surprise in the data.
Firecalc points this out clearly.
seajay
Posts: 1656
Joined: Sat May 01, 2021 3:26 pm
Contact:

Re: The Worst Years To Retire: The Surprise In The Data

Post by seajay »

willthrill81 wrote: Thu Oct 28, 2021 10:23 am
seajay wrote: Thu Oct 28, 2021 6:22 am The indications I'm seeing are that it didn't really matter that much as to what asset allocation you held, T-Bills alone were the more assured to leave the least, all stock was inclined to maximise the legacy.
If all you're talking about is varying the allocation between total stock market and bonds, then yes, tweaking the AA didn't usually change the SWR very much. But if we expand AA to include things like factors and gold (from 1975 and onward), then it absolutely made a big difference in the SWR.
Yes from 1975. More broadly however for a constant/fixed AA I'm seeing indications that what worked better for one than the other over some periods, swung the other way around over other periods. Broadly averaging similar overall outcomes but where stock heavy tended to do better overall i.e. in good cases yielded multiple more times the inflation adjusted start date amount. Similar worst case outcomes, better good case outcomes, higher average overall.

To improve the worst case outcomes relative valuation adjustments can achieve that. Such as 1980 when Dow/Gold was down at 1.0 levels not weighting too much (if any) into gold. Or 1999 when Dow/gold was up at 40 reducing stock weighting ... or whatever relative valuation measure/method.

(UK data) 1912 start date and 4% SWR either for all-stock or 50/50 stock/tbills saw that last 20 years before all-spent. 2000 start year and 50/33/17 stock/gold/tbills did better than all-stock, both still supporting 4% SWR but where all-stock was recently down to 15% of the inflation adjusted start date capital remaining versus 50/33/17 having 80%. Over other periods that swung the other way around

Image

The indications I'm seeing are that 4% SWR more often is fine, but could only last 20 years so have a plan B to fall back upon. In some cases it didn't matter what AA was employed and that 20 year failure point could equally occur if you were all stock or in a blend of stock/gold/bonds. To address the risks 1) relative valuation based asset allocation could help reduce the risk, not including gold if the Dow/gold was at 1.0 levels as per 1980, or less stock in 1999 when Dow/Gold was up at 40. In the mid/late 1960's Dow/Gold was hitting new highs (see below). And 2) for a plan B, well 20 years for a 65 year old has a modest probability of not even living all of those years, or if you do having capital in a home value that might be downsized or sold to fund care home costs could be enough to see you through from age 85+.

Dow/Gold ratio ...

Code: Select all

1950	4.94
1951	5.85
1952	6.73
1953	7.54
1954	7.91
1955	11.47
1956	13.89
1957	14.19
1958	12.36
1959	16.56
1960	19.27
1961	16.87
1962	20.60
1963	18.45
1964	21.64
1965	24.73
1966	27.30
1967	22.19
1968	25.50
1969	22.50
1970	22.73
1971	22.45
1972	20.46
1973	15.77
1974	7.58
1975	3.29
1976	6.08
1977	7.47
1978	5.02
1979	3.59
1980	1.60
1981	1.64
With the baseline covered, the average or good case outcomes tend to be good enough outcomes - where wealth might be maintained or expanded multiples, more so for stock-heavy AA.
Blue456
Posts: 2152
Joined: Tue Jun 04, 2019 5:46 am

Re: The Worst Years To Retire: The Surprise In The Data

Post by Blue456 »

nigel_ht wrote: Tue Oct 26, 2021 2:36 pm
Blue456 wrote: Tue Oct 26, 2021 11:54 am
nigel_ht wrote: Tue Oct 26, 2021 8:00 am
Blue456 wrote: Tue Oct 26, 2021 5:14 am The above assumptions are that investors remains with US only acclamation. A more diversified international investor would do better.
Today, with a correlation of 0.8, I don’t think that VT would fare that much better than VTI in a 1929/Great Depression scenario…
Todays correlation doesn’t mean that future one will be the same.
True but it’s not a given that international large cap is actually useful diversification.
It seems like it was, at least between dates 1969 and 2000. For the 32 year period the average return of EAFE was 11.1% and SP500 12.17% as least according to Dr. Bernstein, The Four Pillars of Investing, Chapter 12. Given the US outperformance during 1990s, there had to be a period between 1969 and 1989 during which EAFE did much better than US.
nigel_ht
Posts: 4742
Joined: Tue Jan 01, 2019 9:14 am

Re: The Worst Years To Retire: The Surprise In The Data

Post by nigel_ht »

Blue456 wrote: Thu Oct 28, 2021 6:37 pm
nigel_ht wrote: Tue Oct 26, 2021 2:36 pm
Blue456 wrote: Tue Oct 26, 2021 11:54 am
nigel_ht wrote: Tue Oct 26, 2021 8:00 am
Blue456 wrote: Tue Oct 26, 2021 5:14 am The above assumptions are that investors remains with US only acclamation. A more diversified international investor would do better.
Today, with a correlation of 0.8, I don’t think that VT would fare that much better than VTI in a 1929/Great Depression scenario…
Todays correlation doesn’t mean that future one will be the same.
True but it’s not a given that international large cap is actually useful diversification.
It seems like it was, at least between dates 1969 and 2000. For the 32 year period the average return of EAFE was 11.1% and SP500 12.17% as least according to Dr. Bernstein, The Four Pillars of Investing, Chapter 12. Given the US outperformance during 1990s, there had to be a period between 1969 and 1989 during which EAFE did much better than US.
Outperformance and correlation/diversification are two different things. If international “only” drops 75% vs 80% for the US it outperforms but remains highly correlated to movements of the US market.

The general objective of diversification is to have a set of uncorrelated (or negatively correlated) assets…
Blue456
Posts: 2152
Joined: Tue Jun 04, 2019 5:46 am

Re: The Worst Years To Retire: The Surprise In The Data

Post by Blue456 »

nigel_ht wrote: Thu Oct 28, 2021 7:17 pm
Blue456 wrote: Thu Oct 28, 2021 6:37 pm
nigel_ht wrote: Tue Oct 26, 2021 2:36 pm
Blue456 wrote: Tue Oct 26, 2021 11:54 am
nigel_ht wrote: Tue Oct 26, 2021 8:00 am

Today, with a correlation of 0.8, I don’t think that VT would fare that much better than VTI in a 1929/Great Depression scenario…
Todays correlation doesn’t mean that future one will be the same.
True but it’s not a given that international large cap is actually useful diversification.
It seems like it was, at least between dates 1969 and 2000. For the 32 year period the average return of EAFE was 11.1% and SP500 12.17% as least according to Dr. Bernstein, The Four Pillars of Investing, Chapter 12. Given the US outperformance during 1990s, there had to be a period between 1969 and 1989 during which EAFE did much better than US.
Outperformance and correlation/diversification are two different things. If international “only” drops 75% vs 80% for the US it outperforms but remains highly correlated to movements of the US market.

The general objective of diversification is to have a set of uncorrelated (or negatively correlated) assets…
All you need is outperformance of international when US underperforms and vice versa to have a more stable return. Seems like international has done that during that time period.
User avatar
willthrill81
Posts: 32250
Joined: Thu Jan 26, 2017 2:17 pm
Location: USA
Contact:

Re: The Worst Years To Retire: The Surprise In The Data

Post by willthrill81 »

Blue456 wrote: Thu Oct 28, 2021 7:30 pm
nigel_ht wrote: Thu Oct 28, 2021 7:17 pm
Blue456 wrote: Thu Oct 28, 2021 6:37 pm
nigel_ht wrote: Tue Oct 26, 2021 2:36 pm
Blue456 wrote: Tue Oct 26, 2021 11:54 am
Todays correlation doesn’t mean that future one will be the same.
True but it’s not a given that international large cap is actually useful diversification.
It seems like it was, at least between dates 1969 and 2000. For the 32 year period the average return of EAFE was 11.1% and SP500 12.17% as least according to Dr. Bernstein, The Four Pillars of Investing, Chapter 12. Given the US outperformance during 1990s, there had to be a period between 1969 and 1989 during which EAFE did much better than US.
Outperformance and correlation/diversification are two different things. If international “only” drops 75% vs 80% for the US it outperforms but remains highly correlated to movements of the US market.

The general objective of diversification is to have a set of uncorrelated (or negatively correlated) assets…
All you need is outperformance of international when US underperforms and vice versa to have a more stable return. Seems like international has done that during that time period.
As discussed in this thread, since 1970, there was a little improvement in the 30 year SWR with global stock exposure rather than strictly U.S. (i.e., 4.4% vs. 4.6%), though this came at the expense of significantly lower maximum withdrawal rates since 1987.
The Sensible Steward
User avatar
abuss368
Posts: 27850
Joined: Mon Aug 03, 2009 2:33 pm
Location: Where the water is warm, the drinks are cold, and I don't know the names of the players!
Contact:

Re: The Worst Years To Retire: The Surprise In The Data

Post by abuss368 »

1966! Incredible but I think 1929 will forever come to mind for most folks. I recall listening and learning from my Grandparents how bad the Great Depression was. I had relatives come straight in from Italy through Ellis Island straight to Little Italy and Long Island!

Think Godfather II back then to see how it was.

Different times.

Tony
John C. Bogle: “Simplicity is the master key to financial success."
NiceUnparticularMan
Posts: 6103
Joined: Sat Mar 11, 2017 6:51 am

Re: The Worst Years To Retire: The Surprise In The Data

Post by NiceUnparticularMan »

Somewhat philosophical, but I do think the unexpected (in a bad way) is what ends up worst for long-term investors, and generally the worst possible combination is going to be unexpectedly high inflation combined with unexpectedly disappointing macroeconomic performance. Stagflation, at least in the loose sense of the term.

And here is a very uncomforting thought: valuations of both US stocks and US bonds were far lower in 1966 than they are today. What this means is you really wouldn't need a stagflation scenario nearly as bad as what happened then in order for real returns, safe withdrawal rates, and so on to be as bad or indeed worse (at least on US-dominated portfolios).

Sleep well!
troysapp
Posts: 62
Joined: Fri Mar 25, 2011 6:08 pm

Re: The Worst Years To Retire: The Surprise In The Data

Post by troysapp »

Hi All,
Below is a "SAFEMAX" chart I maintain. Though this chart is for hypothetical 30-year retirees, I also have a version for shorter time horizons if there's interest. At any rate, one major takeaway for me is that the hypo 50/50 portfolio beat the otherwise all Treasury portfolio over all scenarios while all 3 hypo portfolio temporarily converged roughly in the mid-60s through mid-70s.

Image
Post Reply