RetireJapan

ToushiTime wrote: ↑Fri Jun 16, 2023 10:55 am Interesting to see in the NYU link that the value of 100 dollars invested in the S&P in 1928 grew to 143.81 in 1929 and was back at 146 in 1936...

This hits a very good point and one that I've raised in other threads. People tend to compare peak to trough drops for the worst crashes and lament how bad the figures seem. But they forgot about the gains they have accumulated after many years of investments made on the way up to that point.

Only people born with a silver spoon or inheriting one at the time of the peak, or new investors starting out and catching it very unlucky, are potentially in the scenario entering 100% at a peak and the following drop is a real loss of a substantial proportion of their net worth, and not just a paper one.

Then the people forget about dividends and grossly over-estimate the historical times for recovery from crashes.

So thinking about accumulated gains before the peak, and realistic real terms figures for the recovery times of equities after crashes, are good psychological aids to help carry people through crash periods in my view.

The vast majority of equity returns over the medium to long term come from reinvesting dividends. Any analysis looking at index returns without considering dividends is deeply flawed.

Deep Blue wrote: ↑Fri Jun 16, 2023 3:52 pm The vast majority of equity returns over the medium to long term come from reinvesting dividends. Any analysis looking at index returns without considering dividends is deeply flawed.

Yeap, the numbers are in the NYU spreadsheet ToushiTime linked to. Since 1927 the S&P500 has returned on average +7.7% yearly due to price growth, and +3.6% yearly from dividends. So anything not including dividends misses a big part of the picture.

In recent time dividends have been becoming less important though. E.g. since 1974 it's +9.2% yearly from price growth, and just +2.8% yearly from dividends. This is probably for US tax efficiency reasons, as paying dividends is less advantageous for investors than earning via price growth.

The answer is here in what I wrote:

Sorry, it was late at night and I missed that part of the article and your explanation.

Just the Lord knows what the future holds.

In recent time dividends have been becoming less important though. E.g. since 1974 it's +9.2% yearly from price growth, and just +2.8% yearly from dividends. This is probably for US tax efficiency reasons, as paying dividends is less advantageous for investors than earning via price growth.

I had another look at the final graph in the BusinessInsider article you linked. It shows that real total returns took longer to recover from 2000 than from 1929 and the performance over the similar timeframe was worse.

From 2000, dividend yields were indeed worse, at below 2% for most of the slump. During the Depression, they were above 5%, rising to nearly 10% at one point. And from 2000, there was no deflation, just moderate inflation.

So if we cannot be sure of reinvested dividends and/or deflation to keep the next crash shorter than a decade or so, wouldn’t it make sense to have bonds in our portfolio?

Edit: I guess you could argue that someone investing in 2000 (assuming the worst possible start) was almost back to were they were by 2008 on that graph, though if they held on beyond that, they would have been down some way by 2010.
More importantly, none of this assumes DCA which I rely on largely for my investment, so maybe I am being too pessimistic.
On the other hand, the reduced dividends and lack of deflation outside Japan in recent slumps could prolong the next crash.

ToushiTime wrote: ↑Sat Jun 17, 2023 4:04 am

In recent time dividends have been becoming less important though. E.g. since 1974 it's +9.2% yearly from price growth, and just +2.8% yearly from dividends. This is probably for US tax efficiency reasons, as paying dividends is less advantageous for investors than earning via price growth.

So if we cannot be sure that there will be sufficient reinvested dividends and/or deflation to keep the next crash, or at least the time needed to recover from it, shorter than a decade or so, wouldn’t it make sense to have bonds in our portfolio?

To confirm no misunderstanding, the numbers don't indicate that crash recovery times will be longer now just because dividend yields are lower than 100 years ago. Remember it's the balance of three factors that determines the "real times" to recovery: price growth of the bare index, dividend reinvestments, and inflation (e.g. CPI). And on average the drop of dividend yields over the last 50 years compared to the 50 before has been offset by the faster growth of the bare index in the last 50 years.


Whether you hold bonds in your portfolio and at what percentage depends entirely on circumstances. If you don't expect to need to use the money in your portfolio over the next decade, many would argue there is no need for bonds. Because simply, bonds are a lower expected return asset class.

If you will need the money, then yes drawing down from bonds if there is a crash instead of stocks is a way to insure against the adverse effects of the crash.

This topic is known as sequencing risk. And it's not just the % of bonds you choose for your portfolio that matters. It's whether you will have flexibility to lower the amount of money you withdraw from your portfolio in the worst case of a crash. And there's the psychological aspect as VG1 raised (again a highly individual specific factor).

The finance YouTuber James Shack, who I often share, has some great videos on sequencing risk and how to decide a bond %. Here's two good ones to start with:
https://www.youtube.com/watch?v=oyzR7tMmj9o
https://www.youtube.com/watch?v=Eac2jZcelQw
Hope you find useful info for you in there!

ToushiTime wrote: ↑Sat Jun 17, 2023 4:04 am Edit: I guess you could argue that someone investing in 2000 (assuming the worst possible start) was almost back to were they were by 2008 on that graph, though if they held on beyond that, they would have been down some way by 2010.
More importantly, none of this assumes DCA which I rely on largely for my investment, so maybe I am being too pessimistic.
On the other hand, the reduced dividends and lack of deflation outside Japan in recent slumps could prolong the next crash.

Hey no problem. Exactly if you are investing DCA at the moment as you are earning, and over a number of years I presume, then exactly, looking at previous crash peak to bottom minus figures is too pessimistic. The situation will actually be much rosier for you.

See my above post about the effect of reduced dividend yields and lack of deflation. It can't be said from the numbers we have that recent crashes have been getting worse compared to the past.

As for what happens in the next crash, again I'll skirt that question - no guarantees from me about anything in the future
But I'm always happy to help with the interpretation of what the past numbers say.

Thanks. I watched both those videos and bookmarked them.

To confirm no misunderstanding, the numbers don't indicate that crash recovery times will be longer now just because dividend yields are lower than 100 years ago. Remember it's the balance of three factors that determines the "real times" to recovery: price growth of the bare index, dividend reinvestments, and inflation (e.g. CPI). And on average the drop of dividend yields over the last 50 years compared to the 50 before has been offset by the faster growth of the bare index in the last 50 years.

Yes, I am aware it is a matter of dividends + inflation/deflation + nominal price growth.
If I am reading it correctly, the last graph in the BusinessInsider article shows that total real returns took the same time, if not longer, to recover from 2000 compared with from 1929
https://www.businessinsider.com/when-di ... 929-2010-7

And the in the Bad Sequence in the 1st YouTube video (3:52 mins in) https://www.youtube.com/watch?v=oyzR7tMmj9o, the couple that started investing in 2000, would not have caught back up until 2020.
I guess that cannot be helped, and you just have to adjust drawdowns using the method he described.
And again, this is not factoring DCA.

ToushiTime wrote: ↑Sat Jun 17, 2023 6:52 am

To confirm no misunderstanding, the numbers don't indicate that crash recovery times will be longer now just because dividend yields are lower than 100 years ago. Remember it's the balance of three factors that determines the "real times" to recovery: price growth of the bare index, dividend reinvestments, and inflation (e.g. CPI). And on average the drop of dividend yields over the last 50 years compared to the 50 before has been offset by the faster growth of the bare index in the last 50 years.

Yes, I am aware it is a matter of dividends + inflation/deflation + nominal price growth.
If I am reading it correctly, the last graph in the BusinessInsider article shows that total real returns took the same time, if not longer, to recover from 2000 compared with from 1929
https://www.businessinsider.com/when-di ... 929-2010-7

Fair point, the first "real terms" break even time is shorter for the 1929 crash, as it crosses over break even line first in 1936 - so in 7 years.

However what the BusinessInsider graph does not show, as the article was only written in 2010, is what happened after that. Here's that data:
The first real terms break even after the 2000 crash occurs in mid 2013. That's also the break free point at which the return never goes below the break even "loss" line again. So break free for the 2000 crash occurred 7 years earlier than the equivalent for the 1929 crash, which was 7/12/1949 (BusinessInsider graph 1).

Moreover despite the 1929 crash sneaking over the break even line in 1936, for the majority of the crash period the losses are deeper and more sustained for the 1929 crash than the 2000 crash - see the red line is below the blue line in the main on the last graph on BusinessInsider.

So TLDR it's not universal, but for most investors the 1929 crash would have felt worse than the 2000 crash.

ToushiTime wrote: ↑Sat Jun 17, 2023 6:52 am And the in the Bad Sequence in the 1st YouTube video (3:52 mins in) https://www.youtube.com/watch?v=oyzR7tMmj9o, the couple that started investing in 2000, would not have caught back up until 2020.
I guess that cannot be helped, and you just have to adjust drawdowns using the method he described.
And again, this is not factoring DCA.

The example couple in that video are retirees withdrawing 22,500 GBP from their pension pot yearly. So it's apples and oranges, and completely irrelevant for the question of whether recent crashes "are worse now" than previously was the case.

TBS wrote: ↑Sat Jun 17, 2023 1:15 am

Deep Blue wrote: ↑Fri Jun 16, 2023 3:52 pm The vast majority of equity returns over the medium to long term come from reinvesting dividends. Any analysis looking at index returns without considering dividends is deeply flawed.

Yeap, the numbers are in the NYU spreadsheet ToushiTime linked to. Since 1927 the S&P500 has returned on average +7.7% yearly due to price growth, and +3.6% yearly from dividends. So anything not including dividends misses a big part of the picture.

In recent time dividends have been becoming less important though. E.g. since 1974 it's +9.2% yearly from price growth, and just +2.8% yearly from dividends. This is probably for US tax efficiency reasons, as paying dividends is less advantageous for investors than earning via price growth.

Run the numbers for the difference compounding 9.2% for 50 years vs 12% for 50 years......

The first real terms break even after the 2000 crash occurs in mid 2013. That's also the break free point at which the return never goes below the break even "loss" line again. So break free for the 2000 crash occurred 7 years earlier than the equivalent for the 1929 crash, which was 7/12/1949 (BusinessInsider graph 1).
Moreover despite the 1929 crash sneaking over the break even line in 1936, for the majority of the crash period the losses are deeper and more sustained for the 1929 crash than the 2000 crash - see the red line is below the blue line in the main on the last graph on BusinessInsider.

Point taken regarding the break-even versus break-free point, and the deeper, more sustained losses during the Depression. Thank you.

The example couple in that video are retirees withdrawing 22,500 GBP from their pension pot yearly. So it's apples and oranges, and completely irrelevant for the question of whether recent crashes "are worse now" than previously was the case.

I'm having trouble understanding that video at 3:28mins here https://www.youtube.com/watch?v=oyzR7tMmj9o

In the bad sequence from 2000, they start of with:

450,000 pounds of savings (which continue to be reinvested, I think)
and draw down 22,500 pounds a year (he said 5% but then said 22,500 fixed each year, so I am guessing he didn't mean 5% of the remaining balance)

He said they would only have 196,150 pounds left by 2010

450,000 - (10 x 22,500 = 225,000) = 225,000 pounds = the amount left after draw-downs before factoring in the declines in investment value

So they "only" lost 28,850 pounds (225,000 - 196,150) due to the slump from 2000 until 2010?

He then said even if they did see the "phenomenal returns" from 2010 until 2020, they would never have been able to catch up, which seems a long time, but I guess the draw downs reduce the amount reinvested and therefore delay the recovery.

Presumably, this would have been even worse for a couple retiring in 1929 based on the deeper and longer troughs, and the time taken to reach the break-free point.

Am I understanding this correctly?