Especially since we only have one 30 year period where people retired with a CAPE > 30 (beginning on black Tuesday). I don't know if that's a failure timeline, but the Dow was 4,800 in June of 1929 and didn't get back to that number until 1958.
http://www.macrotrends.net/1319/dow-jones-100-year-historical-chart
That's why I think I want to go from a 4% WR (25x) to 3.3-3.5 WR (28.5-30x). When I'm already at 25x spending, I don't think it will take very long to get to 28.5 - 30x. At that point, I'll likely be (1) saving 50% of my income (1x spending) and (2) potentially also earning another 1x spending through investment gains (if it's 4%). If I get a 10% return in year 1, then I'll get to 28.5x in one year.
It's probably too conservative, but I'd rather work a few extra years with a very high salary than be looking for more menial jobs during a downturn.
You have to remember though that there was significant deflation over parts of that time period, so the nominal return isn't nearly as bad as it looks. I believe the worst year to retire was sometime in the 60's or maybe early 70's because you had poor nominal returns combined with high inflation.
Still probably the third worst time frame to retire, after the time in the 60's or early 70's and then around 2000 (which may or may not be a failure time period, but is certainly off to a bad start)
[/quote]
I agree that the 1930 timeline isn't the worst ever and deflation "helped" those that retired just before the crash. The problem is that it's the only case where CAPE was 30 (and we're not > 33). "A test of one is a test of none", but I think it's reasonable to assume that sequence of returns risk is higher now than it was in 2010 when the CAPE was ~ 20).
My points were:
(1) the first 10 years after retirement are an important indicator for sequence of returns risk,
(2) CAPE is apparently a reasonable predictor of market performance in the 8-10 year timeline, and
(3) we only have one timeline where CAPE at retirement was ~ 30 (it's currently 33.5). That timeline wasn't particularly good (although I don't know if it's a timeline that failed in the Trinity study).
From this, I would not be super confident in the success rates attributed to WRs from Trinity Study type analyses if I were retiring today. There isn't enough data to assess the sequence of returns risk is now and saying it's the same as it was when CAPE was generally between 10 --> 25 might not be reasonable.*
I do agree with folks who argue that flexibility is important. But, working for 1-2 extra years means that it's harder for us to get wiped out if we're put in an inflexible situation (i.e. no one wants to hire you during a recession when you have no recent work experience OR large "unforeseen" medical expense).
* Note: Has anyone seen anything that looks at the correlation between SWRs and CAPE using a Trinity type analysis? My whole argument here is that they are related because I think CAPE should be correlated to SoR risk...but maybe it's not. It feels like this would be a cool idea for a paper / blog post. Hopefully, someone's written it already.
[/quote]
I'm an idiot. Completely missed the bolded statement.
And the blog earlyretirementnow.com does a pretty good dive into how higher CAPE valuations elevate sequence of return risk and lower safe withdrawal rates.