A short study of the recent Japanese crisis

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
User avatar
Topic Author
siamond
Posts: 6003
Joined: Mon May 28, 2012 5:50 am

A short study of the recent Japanese crisis

Post by siamond »

I finally completed a small project I wanted to work on for a while...

Based on public data sources providing Japanese stock returns and bond interest rates, this article provides a quantitative analysis of the trajectory a Japanese passive investor could have experienced with his portfolio during the 1980-2016 time period. Japan suffered from what was probably the most severe stock crisis in modern history, starting by the end of 1989, taking nearly two decades to finally somewhat recover, and then being slammed again by the financial crisis of 2008. There is no equivalent in U.S. history of such long lasting crisis, and as such, it is a sobering example of what could happen.

Read more of this blog post on Financial Page. Feedback welcome. Many thanks to AlohaJoe and his invaluable help modeling bond fund returns.

-------- Follow-up article --------

This article is a follow-up to the "short study of the recent Japanese crisis" which was published on this blog a few days ago. The study was actively discussed on the Bogleheads forum. Various interesting questions were raised that weren't directly addressed by the original write-up, and triggered the author to do more research about domestic tilts (or lack thereof), variable withdrawals, and past performance of the stock market in Japan.

Read more of this second blog post on Financial Page.
Last edited by siamond on Wed Feb 08, 2017 11:31 pm, edited 1 time in total.
User avatar
nedsaid
Posts: 19249
Joined: Fri Nov 23, 2012 11:33 am

Re: A short study of the recent Japanese crisis

Post by nedsaid »

Siamond, you did great work. First, you showed that denominated in YEN the long bear market in Japanese stocks was not as bad as perceived. Secondly, you showed the value of International diversification. Thanks for posting.
A fool and his money are good for business.
livesoft
Posts: 85973
Joined: Thu Mar 01, 2007 7:00 pm

Re: A short study of the recent Japanese crisis

Post by livesoft »

Can I ask a question related to this please?

My impression from having many Japanese friends is that "a Japanese passive investor" has never invested that much in the stock market. They are simply less concerned about what happened to the Japanese stock market than Americans are concerned about the Japanese stock market. They have virtually lifetime employment and retirement plans backed by bonds, so stocks never figured much into their thoughts. They do have concern about real estate prices, but not stock prices.

Is that a good impression? Because that's the one I have.
Wiki This signature message sponsored by sscritic: Learn to fish.
User avatar
nedsaid
Posts: 19249
Joined: Fri Nov 23, 2012 11:33 am

Re: A short study of the recent Japanese crisis

Post by nedsaid »

livesoft wrote:Can I ask a question related to this please?

My impression from having many Japanese friends is that "a Japanese passive investor" has never invested that much in the stock market. They are simply less concerned about what happened to the Japanese stock market than Americans are concerned about the Japanese stock market. They have virtually lifetime employment and retirement plans backed by bonds, so stocks never figured much into their thoughts. They do have concern about real estate prices, but not stock prices.

Is that a good impression? Because that's the one I have.
From what I read and hear, this is mostly true in Europe as well.
A fool and his money are good for business.
stlutz
Posts: 5585
Joined: Fri Jan 02, 2009 12:08 am

Re: A short study of the recent Japanese crisis

Post by stlutz »

Wow, nice writeup Siamond! Might this be the first in a series on market crashes? :wink:

I think my takeaways based solely on this situation are:

a) It's more important to focus on being globally diversified than on selecting the correct stocks in one's home market.

b) If one does (a), the 4% rule looks pretty good.

Again, conclusions I draw based on this example only.
stlutz
Posts: 5585
Joined: Fri Jan 02, 2009 12:08 am

Re: A short study of the recent Japanese crisis

Post by stlutz »

My impression from having many Japanese friends is that "a Japanese passive investor" has never invested that much in the stock market.
Currency speculation is a big thing among Japanese women. Google "Mrs. Watanabe" to find a bunch of stories on this.
Dirghatamas
Posts: 581
Joined: Fri Jan 01, 2016 5:18 pm

Re: A short study of the recent Japanese crisis

Post by Dirghatamas »

Siamond this is excellent work and must have taken a lot of time. I do have a small nitpick. You write
"
One basically needed twice as many Yen to buy US dollars 30 years ago compared to now. Note that the strong depreciation of the Yen against the dollar occurred before the stock crisis of 1989 hit Japan"


I was confused because having lived through the late eighties/ early nineties my impression was opposite. The yen APPRECIATED like crazy ~2X vs. the dollar in a very short time during the mid eighties and that preceded the stock market going bust. Basically the yen went from ~240/dollar to ~120 Y/dollar. So it appreciated NOT depreciated. Just checking. Did I miss it or just a typo?

I remember that time because of my love of bicycling. Graduating in 1992, I had wanted a Miyata touring bicycle as probably the best long distance cycles made. Unfortunately, the yen appreciated so much just before my graduation, that "Made in Japan" steel bicycles had become unprofitable/ too expensive due to the yen and the market had shifted to Taiwan :( I still have a soft spot for Japanese steel bicycles from the eighties. The quality of construction was remarkable.
User avatar
Kevin M
Posts: 15750
Joined: Mon Jun 29, 2009 3:24 pm
Contact:

Re: A short study of the recent Japanese crisis

Post by Kevin M »

Very nice, siamond!

Kevin
If I make a calculation error, #Cruncher probably will let me know.
Clive
Posts: 1950
Joined: Sat Jun 13, 2009 5:49 am

Re: A short study of the recent Japanese crisis

Post by Clive »

Dirghatamas wrote:You write
"
One basically needed twice as many Yen to buy US dollars 30 years ago compared to now....
When the Japanese Yen was introduced in 1871, its value was more or less the same as the US dollar 1 Yen bought 1 US$ :)

There's historic Yen/US$ price data since 1949 here
stlutz
Posts: 5585
Joined: Fri Jan 02, 2009 12:08 am

Re: A short study of the recent Japanese crisis

Post by stlutz »

I remember that time because of my love of bicycling. Graduating in 1992, I had wanted a Miyata touring bicycle as probably the best long distance cycles made. Unfortunately, the yen appreciated so much just before my graduation, that "Made in Japan" steel bicycles had become unprofitable/ too expensive due to the yen and the market had shifted to Taiwan :( I still have a soft spot for Japanese steel bicycles from the eighties. The quality of construction was remarkable.
One of my big regrets in life was giving away my Miyata that I originally purchased in 1987. :(
User avatar
Topic Author
siamond
Posts: 6003
Joined: Mon May 28, 2012 5:50 am

Re: A short study of the recent Japanese crisis

Post by siamond »

Dirghatamas wrote:Siamond this is excellent work and must have taken a lot of time. I do have a small nitpick. You write
"
One basically needed twice as many Yen to buy US dollars 30 years ago compared to now. Note that the strong depreciation of the Yen against the dollar occurred before the stock crisis of 1989 hit Japan"


I was confused because having lived through the late eighties/ early nineties my impression was opposite. The yen APPRECIATED like crazy ~2X vs. the dollar in a very short time during the mid eighties and that preceded the stock market going bust. Basically the yen went from ~240/dollar to ~120 Y/dollar. So it appreciated NOT depreciated. Just checking. Did I miss it or just a typo?
Er... I rewrote this sentence after reading the Wikipedia article, and I found a way to get it backwards. You're absolutely right. I fixed the post. Thank you for pointing this out.

Here is the whole trajectory since 1971... A 3x change, that is quite dramatic.

Image
User avatar
Topic Author
siamond
Posts: 6003
Joined: Mon May 28, 2012 5:50 am

Re: A short study of the recent Japanese crisis

Post by siamond »

stlutz wrote:Wow, nice writeup Siamond! Might this be the first in a series on market crashes? :wink:

I think my takeaways based solely on this situation are:

a) It's more important to focus on being globally diversified than on selecting the correct stocks in one's home market.

b) If one does (a), the 4% rule looks pretty good.

Again, conclusions I draw based on this example only.
Yes, I have a roughly similar thesis, but it remains to be substantiated! I would love to have a copy of the full 'DMS dataset' (from the Triumph of the Optimists authors) to play with all sorts of International Returns, but this is way too expensive for my taste. I know Prof. Pfau has it though, and ran a lot of SWR math about it. So I shot him an e-mail earlier today asking if he'd be interested to extend his SWR studies along the same lines, comparing domestic-only portfolios to internationally diversified portfolios. We'll see if he bites!

In the mean time, I do plan to run the UK numbers through a variation of my Excel spreadsheet, as the full history of stocks and bonds (aka gilts) can be extracted from a Barclays study I know of. This will probably make for the next blog entry... :wink:
User avatar
Topic Author
siamond
Posts: 6003
Joined: Mon May 28, 2012 5:50 am

Re: A short study of the recent Japanese crisis

Post by siamond »

livesoft wrote:My impression from having many Japanese friends is that "a Japanese passive investor" has never invested that much in the stock market. They are simply less concerned about what happened to the Japanese stock market than Americans are concerned about the Japanese stock market. They have virtually lifetime employment and retirement plans backed by bonds, so stocks never figured much into their thoughts. They do have concern about real estate prices, but not stock prices.
I do not have any additional information about the matter, but I wouldn't be surprised. I also read that the prolonged crisis of the past few decades had an even worse predecessor after World War II, which would hardly be surprising given what happened to Japan by then. In such context, culturally, I don't quite see how trust in the stock market could have been established.

My primary point wasn't so much about Japanese investors though. I mostly wanted to play a 'what if' scenario, pondering what would be the effect on our investment lives if a crisis similar to this one would hit closer to home... I read various soundbites, but I wanted to establish a stronger quantitative basis for analysis.
Dirghatamas
Posts: 581
Joined: Fri Jan 01, 2016 5:18 pm

Re: A short study of the recent Japanese crisis

Post by Dirghatamas »

siamond wrote:So I shot him an e-mail earlier today asking if he'd be interested to extend his SWR studies along the same lines, comparing domestic-only portfolios to internationally diversified portfolios. We'll see if he bites!
This would be awesome. I have tried to model this in the past but simply don't have the underlying data. I have looked at Credit Suisse Yearbooks and they DO have the total returns for world stocks starting from 1900-2015 including countries that went to 0, including real returns converted to USD. They have pages for each country as well as total world stock as well as total world ex US. However, the volatility is difficult to figure out at the level of accuracy needed without the actual detailed source data.

My curiosity is not just academic. I have invested this way (global cap weighted) my whole investing life but done it intuitively rather than with precise data. It has been intuitive to me that if the expected returns of US stocks and say ex US are ~ similar, it is a no brainer to diversify so I always hold global cap weight.

However, I have never sat down and worked out the SWR as it has always felt like something for the far future..

One other request would be to compare a 100% global stock portfolio to not just 100% US stock portfolio but to say a 80/20 or 60/40 US stocks/bonds portfolio. The arguments for bonds has been volatility and draw downs for a stock heavy portfolio. Fair enough..but my intuition is that global ex US stocks are a better long term diversifier for US stocks, rather than US bonds (which place both stocks and bonds with geographic risk).

In the final analysis, bonds in any country come from the earnings of stocks (economic engine) so if stocks/earnings do poorly for decades, bonds in that country will too. Corporate bonds are directly linked to company earnings and so are Govt bonds (through taxation). Geographic diversification in higher expected return class (stocks) seems like a better long term bet.
User avatar
Topic Author
siamond
Posts: 6003
Joined: Mon May 28, 2012 5:50 am

Re: A short study of the recent Japanese crisis

Post by siamond »

Dirghatamas wrote:My curiosity is not just academic. [...]
Yes, we think alike, I've been asking myself exactly the same questions as the ones you articulated, and I am also an avid reader of the Credit Suisse Yearbooks (the DMS dataset is the underlying data for those publications). Global Financial Data (GFD) also has a huge treasure trove of fascinating data, I saw a very cool demo, but alas, not exactly cheap either.

About the question of International diversification vs. Bonds, this was exactly what I was getting at with the last test I described in my write-up. 2008 painfully illustrated that worldwide equities can sync up at the worst time though, and my guess is that this will happen again, and more often. I was actually taken aback to see the dramatic effect the financial crisis had on the Japanese returns, I thought it was more a US/Europe thing, I was very wrong. The world is flat nowadays, as they say.

I'll keep working on those themes, I have a few ideas on how to proceed, at least for 1970-onward. Before that, this is trickier, and well, the deep and prolonged crisis of bonds after World War II in many countries really has to be taken in account, and that is the part where I hope Prof. Pfau will step in.
AlohaJoe
Posts: 6609
Joined: Mon Nov 26, 2007 1:00 pm
Location: Saigon, Vietnam

Re: A short study of the recent Japanese crisis

Post by AlohaJoe »

It is also interesting to look at the traditional "safe withdrawal rate" metric.

Image

(I don't have data before 1975 and 1986 is the latest retirement that allows a full 30 years of data (i.e. 1986-2016). This shows allocations from 0% stocks to 100% stocks.)

With any reasonable allocation (i.e. not 100% equities) the SWR is above 5% even for a someone who retired shortly the stock market collapse.

There isn't enough data (yet) to see what a True 30-Year SWR looks like for a Japanese investor but if we're willing to cheat slightly we can get a picture.

Instead of calculating a 30-year SWR we can calculate a 25-year SWR. That means what you see below should seen as indicative rather than something that you can directly compare to other SWRs quoted in studies & papers. Remember that this means they have totally run out of money after 25 years.

Image

While things bottom out in 1990, it looks like most asset allocations with a dollop of bonds recovered reasonably well. Even a 60/40 allocation has a shot at ending up above 4% depending on how the next 5 years go.

However, imagine if you were a Japanese investor who saw the first chart -- the "safe" withdrawal rate seemed to be 5% or maybe even 6%. And then...whoops! Along comes the second chart and, nope, sorry, it is more like 3%.
AlohaJoe
Posts: 6609
Joined: Mon Nov 26, 2007 1:00 pm
Location: Saigon, Vietnam

Re: A short study of the recent Japanese crisis

Post by AlohaJoe »

siamond wrote:The world is flat nowadays, as they say.
This is true....ish. But I'm not sure what to make of it. The data is from a relatively old paper but...

Here's the correlations between global markets going back to 1860:

Image

Are correlations higher today? Or was the post-WW2 just abnormally low and due to recency bias that's what we consider "normal"?
User avatar
Topic Author
siamond
Posts: 6003
Joined: Mon May 28, 2012 5:50 am

Re: A short study of the recent Japanese crisis

Post by siamond »

Dirghatamas wrote:One other request would be to compare a 100% global stock portfolio to not just 100% US stock portfolio but to say a 80/20 or 60/40 US stocks/bonds portfolio.
Let's stay Japan-centric for this thread if that's ok with you. In the blog post, I did compare an 80/20 portfolio to a 60/40 portfolio, for somebody retiring at the end of 1988 and using a 4% SWR withdrawal. You'll find below the outcome of one more test, comparing the 80/20 portfolio to a 100/0 (no bonds) portfolio, while keeping the domestic tilt (20% Total-Japan, 10% Small Value). One might argue that the 100/0 portfolio worked ok, although the 2008 drop (and following years of uncertainty) would have been very hard to bear. Other starting years (except the extreme case of Dec-89) would have been fine.

Image

Now let's push the logic one step further as you suggested. Let's compare the 100% portfolio with the domestic tilt (AA#2) with a 100% World portfolio (AA#1), while keeping the returns in Yen, hence impacted by the trajectory of the exchange rate. The latter certainly looks much better, isn't it? Still, on both lines, the 2008 drop exceeds 60%, which would severely test anybody's nerves.

Image
User avatar
Topic Author
siamond
Posts: 6003
Joined: Mon May 28, 2012 5:50 am

Re: A short study of the recent Japanese crisis

Post by siamond »

AlohaJoe wrote:Here's the correlations between global markets going back to 1860
What is exactly displayed? Correlation between what and what? Could you also share the source? It is certainly a thought provoking graph.
User avatar
Topic Author
siamond
Posts: 6003
Joined: Mon May 28, 2012 5:50 am

Re: A short study of the recent Japanese crisis

Post by siamond »

AlohaJoe wrote:However, imagine if you were a Japanese investor who saw the first chart -- the "safe" withdrawal rate seemed to be 5% or maybe even 6%. And then...whoops! Along comes the second chart and, nope, sorry, it is more like 3%.
Indeed, great point. This demonstrates once again the foolishness of a constant-withdrawal method. The SWR metric itself makes relatively little sense if you ask me. I used a variant of VPW in my own tests in lieu of a 4% SWR, but stopped short of documenting it in the blog entry, to avoid confusing readers with a second-level topic.
AlohaJoe
Posts: 6609
Joined: Mon Nov 26, 2007 1:00 pm
Location: Saigon, Vietnam

Re: A short study of the recent Japanese crisis

Post by AlohaJoe »

siamond wrote:
AlohaJoe wrote:Here's the correlations between global markets going back to 1860
What is exactly displayed? Correlation between what and what?
"Average correlation of capital appreciation returns for all available markets"
Could you also share the source? It is certainly a thought provoking graph.
"Long-Term Global Market Correlations" (2001), NBER Working Paper 8612

NBER Working Paper W8178, "The Great Reversals: The Politics of Financial Development in the 20th Century" (despite its vague name) also gives an interesting picture of the past:
By most measures, countries were more financially developed in 1913 than in 1980 and only recently have they surpassed their 1913 levels.[...] What is especially interesting is that indicators of financial development fell in all countries after 1929, reaching their nadir around 1980. Since then, there has been a revival of financial markets.
JoinToday
Posts: 1110
Joined: Sat Mar 10, 2007 8:59 pm

Re: A short study of the recent Japanese crisis

Post by JoinToday »

Dirghatamas wrote:........
In the final analysis, bonds in any country come from the earnings of stocks (economic engine) so if stocks/earnings do poorly for decades, bonds in that country will too. Corporate bonds are directly linked to company earnings and so are Govt bonds (through taxation). Geographic diversification in higher expected return class (stocks) seems like a better long term bet.
Is this a good argument for international diversification of bonds?

I am currently 60% equity, 40% bonds; my bonds are 100% US bonds (Govt & investment grade corporate). I didn't find the need for international bonds to be very compelling, from arguments made on this forum and Vanguard's paper on international bonds. But if bonds are related to a company's & country's performance, maybe it makes sense to diversify bonds between US and international.
I wish I had learned about index funds 25 years ago
AlohaJoe
Posts: 6609
Joined: Mon Nov 26, 2007 1:00 pm
Location: Saigon, Vietnam

Re: A short study of the recent Japanese crisis

Post by AlohaJoe »

Dirghatamas wrote:In the final analysis, bonds in any country come from the earnings of stocks (economic engine) so if stocks/earnings do poorly for decades, bonds in that country will too. Corporate bonds are directly linked to company earnings and so are Govt bonds (through taxation).
If government bonds are directly linked to stock returns, why doesn't it show up in the correlations between them?

Image
User avatar
sperry8
Posts: 3065
Joined: Sat Mar 29, 2008 9:25 pm
Location: Miami FL

Re: A short study of the recent Japanese crisis

Post by sperry8 »

This is a great post.

My takeaway is that one must remember how in Yen, the data for a Japanese investor wouldn't be as bad... it is only for foreigners who invested returns have stayed so low. So when we worry about the US possibly achieving similar low returns - we must remember that we are seeing returns inclusive of the Yen - which makes it look worse to foreigners.
BH Contests: 23 #89 of 607 | 22 #512 of 674 | 21 #66 of 636 |20 #253/664 |19 #233/645 |18 #150/493 |17 #516/647 |16 #121/610 |15 #18/552 |14 #225/503 |13 #383/433 |12 #366/410 |11 #113/369 |10 #53/282
User avatar
sperry8
Posts: 3065
Joined: Sat Mar 29, 2008 9:25 pm
Location: Miami FL

Re: A short study of the recent Japanese crisis

Post by sperry8 »

siamond wrote:I finally completed a small project I wanted to work on for a while...

Based on public data sources providing Japanese stock returns and bond interest rates, this article provides a quantitative analysis of the trajectory a Japanese passive investor could have experienced with his portfolio during the 1980-2016 time period. Japan suffered from what was probably the most severe stock crisis in modern history, starting by the end of 1989, taking nearly two decades to finally somewhat recover, and then being slammed again by the financial crisis of 2008. There is no equivalent in U.S. history of such long lasting crisis, and as such, it is a sobering example of what could happen.

Read more of this blog post on Financial Page. Feedback welcome.

PS. many thanks to AlohaJoe and his invaluable help modeling bond fund returns.
Note: your "Growth Chart Japan" still shows mid-value on the graph keys even though you removed the line. Took a while before I realized you removed it on purpose. Perhaps remove it from the right hand side of the graph too.
BH Contests: 23 #89 of 607 | 22 #512 of 674 | 21 #66 of 636 |20 #253/664 |19 #233/645 |18 #150/493 |17 #516/647 |16 #121/610 |15 #18/552 |14 #225/503 |13 #383/433 |12 #366/410 |11 #113/369 |10 #53/282
User avatar
AtlasShrugged?
Posts: 699
Joined: Wed Jul 15, 2015 6:08 pm

Re: A short study of the recent Japanese crisis

Post by AtlasShrugged? »

siamond....Great article. Your closing comments are the ones that intrigued [scared?] me the most. You made mention that the Japanese market cap was actually greater than the US at one point . The statement you made was that there is no economy that is too big, or too strong to have a prolonged run of very low returns. Great cautionary note.

Two questions.

One, could the hypothetical Japanese retiree have made the portfolio last with annual rebalancing [the retiree from 1988, 89] to a specific ratio of equities/bonds? Was there an optimal rebalance and ratio for that hypothetical retiree?

Two, how much better could the hypothetical Japanese retiree [the case where the portolio failed over 30 years] done using VPW as opposed to SWR of 4%?

Apologies if the questions are not clear.

What is the chief lesson you would want a US investor to take away from your article?
“If you don't know, the thing to do is not to get scared, but to learn.”
Chadnudj
Posts: 1266
Joined: Tue Oct 29, 2013 11:22 am

Re: A short study of the recent Japanese crisis

Post by Chadnudj »

Loved this article. The big takeaways for me:

1. Even in the worst case scenario of retiring at the end of 1989, in Japan, fully invested domestically in Japan with a 60/40 stock/bond asset allocation, a 3.3% SWR works for a 30 year retirement period; add a bit of international diversification and/or the ability to use a variable withdrawal method, and you could have easily survived using a 4% or higher initial SWR. I'd urge those among us working extra years so that they can have a 2% or 2.5% SWR "number" to look at this article -- you're probably working too long to gain too little in security.

2. Related to 1, diversification is so very, very, very important.

3. Perhaps this is a general note for all of us considering SWR/when to retire, but thinking about Japan and what would happen if you retired in Japan at the end of 1989, is the lesson in all of this as simple as "be sure to not retire at an all-time market high, but if you do, make sure you have X% above your number as an additional buffer"? I mean, generally speaking with an upwards trending market, many of us/most of us will retire at or near all-time market highs....but in terms of "one more year" syndrome, maybe we should consider only retiring if our "number" is reached with stocks a certain percentage off/below all-time highs (no idea what that number is, though) or, alternatively, retire if our assets are a certain percentage (again, not sure what that percentage would be) above what they'd need to be if markets are at an all-time high and haven't adjusted downwards meaningfully..... food for thought, at least.
AlohaJoe
Posts: 6609
Joined: Mon Nov 26, 2007 1:00 pm
Location: Saigon, Vietnam

Re: A short study of the recent Japanese crisis

Post by AlohaJoe »

JCE66 wrote:One, could the hypothetical Japanese retiree have made the portfolio last with annual rebalancing [the retiree from 1988, 89] to a specific ratio of equities/bonds?
All of the graphs shown above include annual rebalancing. Rebalancing isn't magic. No amount of rebalancing would have improved things.
Was there an optimal rebalance and ratio for that hypothetical retiree?
Look at the charts I posted in this thread.
Two, how much better could the hypothetical Japanese retiree [the case where the portolio failed over 30 years] done using VPW as opposed to SWR of 4%?
Maybe siamond will post a fuller writeup of how VPW would fare in Japan but...VPW isn't a silver bullet. Your portfolio won't run out, you'll just be moving in with the grandchildren ;)

Image

This assumes 50% bonds, annual rebalancing, and retiring in 1990. You can see that you have to cut your spending by 30% the first year and will eventually need to cut it by 50%. Other than 2 brief spikes, you spend most of your retirement withdrawing 40% less than you did your first year.
AlohaJoe
Posts: 6609
Joined: Mon Nov 26, 2007 1:00 pm
Location: Saigon, Vietnam

Re: A short study of the recent Japanese crisis

Post by AlohaJoe »

Chadnudj wrote: 1. Even in the worst case scenario of retiring at the end of 1989, in Japan, fully invested domestically in Japan with a 60/40 stock/bond asset allocation, a 3.3% SWR works for a 30 year retirement period
It hasn't been 30 years since 1990 so....siamond's article isn't quite showing that 3.3% worked for 30 years. :)

If the next 4 years are flat (0% returns) then the 1990 retiree with a 60/40 allocation will end up with 3.1% SWR. A good/bad market could change that but not by more than about .1% in most cases.
User avatar
Topic Author
siamond
Posts: 6003
Joined: Mon May 28, 2012 5:50 am

Re: A short study of the recent Japanese crisis

Post by siamond »

JCE66 wrote:One, could the hypothetical Japanese retiree have made the portfolio last with annual rebalancing [the retiree from 1988, 89] to a specific ratio of equities/bonds? Was there an optimal rebalance and ratio for that hypothetical retiree?
All the charts are displayed with annual rebalancing. From my study of rebalancing algorithms a while ago (admittedly US-centric), I don't think rebalancing details would have made any significant difference.

A purely domestic investor would have been better off with 100% bonds, actually. But this is a finding in pure hindsight for a very specific starting point (near the height of the crisis). As a case in point, I would venture to guess that a 100% bonds allocation would be extremely unwise for a retiree starting now in Japan (given the negative interest rates!).
JCE66 wrote:Two, how much better could the hypothetical Japanese retiree [the case where the portfolio failed over 30 years] done using VPW as opposed to SWR of 4%?
Well, at least such retiree would not have gone straight in bankruptcy. But as AlohaJoe pointed out, this wouldn't have been pretty either. I gave it a quick run for a 60/40 domestic allocation, 30 years retirement cycle starting at the end of 1988 (ok, 2 years missing at the end!), adding a 10 years grace period (some folks live really long in Japan). Most of the VPW trajectory hovered around 3% of the initial portfolio (in real terms). VPW (and variations of it) is no silver bullet, just a way to automatically adjust the trajectory, and that is certainly invaluable.
JCE66 wrote:What is the chief lesson you would want a US investor to take away from your article?
Internationally diversify is by far the chief lesson in my humble opinion.

I would also add a more positive note, which is that if you do hedge your bets with your asset allocation, and if you are thoughtful about your withdrawals, don't agonize too much about the possibility of such prolonged crisis, you should be able to manage while maintaining a dignified life.
User avatar
Kevin M
Posts: 15750
Joined: Mon Jun 29, 2009 3:24 pm
Contact:

Re: A short study of the recent Japanese crisis

Post by Kevin M »

AlohaJoe wrote: Image

While things bottom out in 1990, it looks like most asset allocations with a dollop of bonds recovered reasonably well. Even a 60/40 allocation has a shot at ending up above 4% depending on how the next 5 years go.

However, imagine if you were a Japanese investor who saw the first chart -- the "safe" withdrawal rate seemed to be 5% or maybe even 6%. And then...whoops! Along comes the second chart and, nope, sorry, it is more like 3%.
If I'm understanding your chart, the SWR for 100% bonds never dropped below 6%. Why isn't anyone remarking on this?

Kevin
If I make a calculation error, #Cruncher probably will let me know.
User avatar
Topic Author
siamond
Posts: 6003
Joined: Mon May 28, 2012 5:50 am

Re: A short study of the recent Japanese crisis

Post by siamond »

Chadnudj wrote:3. Perhaps this is a general note for all of us considering SWR/when to retire, but thinking about Japan and what would happen if you retired in Japan at the end of 1989, is the lesson in all of this as simple as "be sure to not retire at an all-time market high, but if you do, make sure you have X% above your number as an additional buffer"? I mean, generally speaking with an upwards trending market, many of us/most of us will retire at or near all-time market highs....but in terms of "one more year" syndrome, maybe we should consider only retiring if our "number" is reached with stocks a certain percentage off/below all-time highs (no idea what that number is, though) or, alternatively, retire if our assets are a certain percentage (again, not sure what that percentage would be) above what they'd need to be if markets are at an all-time high and haven't adjusted downwards meaningfully..... food for thought, at least.
Yes, I agree with the spirit of your thought. I just didn't want to hone too much on Dec-89 as a starting point, because stocks were so insanely elevated in Japan by then (PE10 was close to 80, I believe) that I would have to hope that nobody in their right mind would have retired with a 4% expectation at this point.

Valuations and expected returns are tricky beasts, but I do see some value in computing 1/PE (for stocks) and combine it with bond yields (adjusted in real terms) in proportion to your AA to get a very rough view of expected returns over the coming decade or two. This should bring some perspective to retirement thinking at a market high point, and help with financial planning. Don't use that for market timing though, and don't forget that this is an extremely blunt statistical tool. Still, this might help.
User avatar
Topic Author
siamond
Posts: 6003
Joined: Mon May 28, 2012 5:50 am

Re: A short study of the recent Japanese crisis

Post by siamond »

Kevin M wrote:If I'm understanding your chart, the SWR for 100% bonds never dropped below 6%. Why isn't anyone remarking on this?
I just did - to some extent! See the post I wrote just before yours! :wink:

I will do a write-up about overloading on bonds during retirement one day, going through the consequences of doing so in the US, France, UK, etc after World-War II. I think this should cure anybody from ever thinking 100% bonds (or even 30/70).
User avatar
Kevin M
Posts: 15750
Joined: Mon Jun 29, 2009 3:24 pm
Contact:

Re: A short study of the recent Japanese crisis

Post by Kevin M »

siamond wrote:
Kevin M wrote:If I'm understanding your chart, the SWR for 100% bonds never dropped below 6%. Why isn't anyone remarking on this?
I just did - to some extent! See the post I wrote just before yours! :wink:
Perhaps, but the emphasis is very different. My point is not that bonds did better for someone retiring near the stock peak, but that the SWR was relatively high for 100% bonds for someone retiring at any time shown in the chart.
I will do a write-up about overloading on bonds during retirement one day, going through the consequences of doing so in the US, France, UK, etc after World-War II. I think this should cure anybody from ever thinking 100% bonds (or even 30/70).
Well we'll see. I will be a test case, since I'm at 30/70, and would be fine with 100% bonds as well. Once your withdrawal rate is low enough, you don't need to take stock risk, and you don't need to depend on the future resembling the past. Of course I'm assuming no total financial collapse in the US, which might be too optimistic.

Kevin
If I make a calculation error, #Cruncher probably will let me know.
User avatar
Topic Author
siamond
Posts: 6003
Joined: Mon May 28, 2012 5:50 am

Re: A short study of the recent Japanese crisis

Post by siamond »

Kevin M wrote:
siamond wrote:
Kevin M wrote:If I'm understanding your chart, the SWR for 100% bonds never dropped below 6%. Why isn't anyone remarking on this?
I just did - to some extent! See the post I wrote just before yours! :wink:
Perhaps, but the emphasis is very different. My point is not that bonds did better for someone retiring near the stock peak, but that the SWR was relatively high for 100% bonds for someone retiring at any time shown in the chart.
Understood, although the chart didn't include that many 30 years periods. Yes, certainly a fine time for bond holders in Japan. I would just point out that during the very exact same time periods, a 100% stock retiree in the US would have done very well with a 6% SWR, while a 100% bonds retiree would have had severe issues... :wink:

Ok, sorry, I don't mean to be argumentative, just to point out that rosy scenarios for a given AA aren't the most informative.

This being said, another lesson I got from this Japan analysis is the following. It is actually quite difficult to justify anything else than 100% stocks in the entire history of the US stock market, if one just looks coldly at numbers and compares asset allocations in presence of reasonable withdrawals. And it gets even harder when using a modicum of flexibility in the withdrawals. Please note that I (admittedly foolishly) removed any considerations about emotions and behavior from the previous statement. The Japan scenarios we're looking at here are NOT like that. Here, bonds DID save the day, and were pretty much necessary to do so. And this is an interesting fact per se. Again, speaking in favor of maximum diversification.
garlandwhizzer
Posts: 3562
Joined: Fri Aug 06, 2010 3:42 pm

Re: A short study of the recent Japanese crisis

Post by garlandwhizzer »

Kevin M wrote:
If I'm understanding your chart, the SWR for 100% bonds never dropped below 6%. Why isn't anyone remarking on this?
That graph looks compelling for 100% bonds over 25 years but I am almost certain those are nominal, not real, rates of return. During the 25 year periods the US suffered runaway inflation in the 70s and 80s with annual inflation rates peaking at 15%/yr and bond returns likewise. 15% per year means that a 6% nominal return is actually minus 9%/yr. in real inflation adjusted dollars which is the only kind you can use for living expenses. Many if not most investors today have been investing through a 3 decade plus period of declining inflation and haven't experienced the devastation of persistent high inflation which is bad for stocks and lethal for bonds in terms of real returns.

Graphs like this one are misleading and may give false assurance that 100% bonds provides a safe 4+% withdrawal rate for next 30 years of retirement. I believe, given todays low interest rates and the possibility of inflation on the horizon, the probability that a 100% bond portfolio will provide a real return of 4%+ going forward for the next 25 years is very close to zero. In retirement planning for safe withdrawal rates it's important to focus on real inflation adjusted returns, not nominal ones, especially over long time horizons. It's not a question of how much it costs to buy a loaf of bread today but how much it will cost in 25 years. Estimating future inflation is difficult/impossible but historically in the US I believe it has averaged a bit more than 3%. I don't expect the last decade of very low inflation to persist forever.


Garland Whizzer
AlohaJoe
Posts: 6609
Joined: Mon Nov 26, 2007 1:00 pm
Location: Saigon, Vietnam

Re: A short study of the recent Japanese crisis

Post by AlohaJoe »

garlandwhizzer wrote:
Kevin M wrote:
If I'm understanding your chart, the SWR for 100% bonds never dropped below 6%. Why isn't anyone remarking on this?
That graph looks compelling for 100% bonds over 25 years but I am almost certain those are nominal, not real, rates of return. During the 25 year periods the US suffered runaway inflation in the 70s and 80s with annual inflation rates peaking at 15%/yr and bond returns likewise.
The graph is real returns.

I'm not sure what US inflation has to do with anything since we're talking about Japan and Japanese investors.
User avatar
Kevin M
Posts: 15750
Joined: Mon Jun 29, 2009 3:24 pm
Contact:

Re: A short study of the recent Japanese crisis

Post by Kevin M »

garlandwhizzer wrote:
Kevin M wrote:
If I'm understanding your chart, the SWR for 100% bonds never dropped below 6%. Why isn't anyone remarking on this?
That graph looks compelling for 100% bonds over 25 years but I am almost certain those are nominal, not real, rates of return.<snip>
I think you may be confusing safe withdrawal rates (SWRs) with returns.
Graphs like this one are misleading and may give false assurance that 100% bonds provides a safe 4+% withdrawal rate for next 30 years of retirement. I believe, given todays low interest rates and the possibility of inflation on the horizon, the probability that a 100% bond portfolio will provide a real return of 4%+ going forward for the next 25 years is very close to zero.
Like here. A 4% SWR does not require a 4% real return. A 0% real return allows 4% real withdralwals for 25 years.

If my calculations are correct, you need 1.22% real to get to 30 years: =RATE(30,-4%,1,0). Since the yield on a 30-year TIPS is a little less than 1%, you can't reliably support a 30-year retirement at a 4% WR with just bonds, so we agree on at least this point.

Assuming you can earn about 0.5% real on a 30-year TIPS ladder, you will have to live with a 3.6% withdrawal rate with bonds in the current environment: =-PMT(0.5%,30,1,0).

My withdrawal rate is significantly less than 3.6%, which is why I would feel quite safe with 100% bonds, as long as I use TIPS or other fixed income that provides a decent inflation hedge. My withdrawal rate is low enough that even with 70% in safe fixed income I feel quite safe, and can think of the 30% allocation to stocks as my risk portfolio.

Kevin
If I make a calculation error, #Cruncher probably will let me know.
Clive
Posts: 1950
Joined: Sat Jun 13, 2009 5:49 am

Re: A short study of the recent Japanese crisis

Post by Clive »

"Safe" bonds - can look pretty scary

Image

More so when you consider that the UK has moved away from taxing real (after inflation) gains, towards taxing nominal gains.

To a large extent, post 1980's good bond gains are down to yields declining from exceptionally high levels down to more recent exceptionally low levels (yields up prices down, yields down prices up).

Much of Japan's big dip post 1990 was a consequence of big up's during the 1970's/80's. The same might be said for the US 1930's crash when prices halved, halved again, halved yet again (followed the "Roaring 20's" when prices doubled, doubled again and doubled yet again).

Japan's roaring 70's/80's saw it rise from very low global cap up to 50% levels, taking most from the US that dropped from 70% levels down to 25% levels. The subsequent post 1990 decline of Japan saw its global cap drop right off, whilst US recouped a large chunk of its former share back again. Most of the Japanese firms that endured the most pain were the large caps that had benefited the most from the rise. Other small/new firms faired much better. Indexes however tend to be dominated by the larger caps.
User avatar
Topic Author
siamond
Posts: 6003
Joined: Mon May 28, 2012 5:50 am

Re: A short study of the recent Japanese crisis

Post by siamond »

Clive wrote:Much of Japan's big dip post 1990 was a consequence of big up's during the 1970's/80's. The same might be said for the US 1930's crash when prices halved, halved again, halved yet again (followed the "Roaring 20's" when prices doubled, doubled again and doubled yet again).
Yes, I totally agree. SWR talk tends to really obscure (if not obfuscate) such considerations. The 'big dip' doesn't impress me very much considering the preceding bull market ('big up'). What is much more disturbing is the trajectory of the following decades AFTER the 'big dip'. That is a pattern we don't see in the history of the US stock market.
User avatar
Kevin M
Posts: 15750
Joined: Mon Jun 29, 2009 3:24 pm
Contact:

Re: A short study of the recent Japanese crisis

Post by Kevin M »

Clive wrote:"Safe" bonds - can look pretty scary <snip>
Nominal bonds really aren't safe in terms of meeting real liabilities, and the longer the term, the less safe they are. By contrast, long-term TIPS are quite safe in terms of meeting long-term real liabilities.

TIPS didn't exist during the big inflationary periods in the US--1940s and 1970s--don't know when inflation-linked bonds were introduced in the UK. TIPS exist now, so the mechanism exists to avoid the kinds of real drawdowns shown in the chart--at least drawdowns that matter. I don't care much about any drawdown in the real value of a 30-year TIPS, due to an increase in real yields, if the purpose of the 30-year TIPS is to pay for living expenses 30 years from now.

But yeah, nominal bonds sucked in the US in the 1940s and 1970s. The 1940s were actually worse for shorter-term bonds, since interest controls fixed nominal rates at low levels, so you couldn't benefit from short-term nominal rates following inflation higher like you could in the 1970s.

Kevin
If I make a calculation error, #Cruncher probably will let me know.
Dirghatamas
Posts: 581
Joined: Fri Jan 01, 2016 5:18 pm

Re: A short study of the recent Japanese crisis

Post by Dirghatamas »

JoinToday wrote:
Dirghatamas wrote:........
In the final analysis, bonds in any country come from the earnings of stocks (economic engine) so if stocks/earnings do poorly for decades, bonds in that country will too. Corporate bonds are directly linked to company earnings and so are Govt bonds (through taxation). Geographic diversification in higher expected return class (stocks) seems like a better long term bet.
Is this a good argument for international diversification of bonds?
JoinToday

I am the wrong guy on this one. I have never invested in bonds of any kind. The majority opinion on this board seems to be "take risk on equities and play it safe on bonds". Majority seems to vote to keep bonds in your own country..while on stocks there is much more of a vocal debate. For me, the first step is "why bonds" and only then the question of diversification. Bonds are different enough from stocks that the same arguments need not apply for diversification.
User avatar
Topic Author
siamond
Posts: 6003
Joined: Mon May 28, 2012 5:50 am

Re: A short study of the recent Japanese crisis

Post by siamond »

Kevin M wrote:TIPS didn't exist during the big inflationary periods in the US--1940s and 1970s--don't know when inflation-linked bonds were introduced in the UK.
Inflation-linked gilts were created in the UK in 1981. Which means that TIPS are basically a completely unproven instrument against severe inflationary periods (and worse, world wars when government priorities have very little to do with the protection of individual investors). I hope you're right and they will deliver in troubled times, but I'm not going to bet on that...

Anyhoo, back to Japan? :wink:
Dirghatamas
Posts: 581
Joined: Fri Jan 01, 2016 5:18 pm

Re: A short study of the recent Japanese crisis

Post by Dirghatamas »

AlohaJoe wrote:
Dirghatamas wrote:In the final analysis, bonds in any country come from the earnings of stocks (economic engine) so if stocks/earnings do poorly for decades, bonds in that country will too. Corporate bonds are directly linked to company earnings and so are Govt bonds (through taxation).
If government bonds are directly linked to stock returns, why doesn't it show up in the correlations between them?

Image
Right. AlohaJoe, I am aware of the well accepted view that Govt Bonds and stocks in a country show almost no correlation. However, my views (that I look forward to argue for and against as well as learn) look at causation and especially tail risks.

First, take revolutions and wars. Both the Bolshevik revolution as well as the Maoist one led to the same outcome. Bonds and stocks both went to 0. Similarly, both Germany and Japan bonds as well as stocks went to ~0 after the second world war. When there was hyperinflation e.g. Weimar Germany, bond holders lost their shirts.

Now, leaving aside serious wars an full scale revolutions, even the more modest ones that destroy an economy, usually have the same outcome. Consider Argentina whose stock market and economy conceivably looked poised to compete with US in 1900. Stock market went multiple times to 0 and so did bonds.

Leaving even these aside and going to the very mundane cases where there is no real catastrophe..lets look at causation. Over a short while, clearly bonds (both Govt and even corporate) have not much to do with the stock market. Over the long haul though, Govt bonds are valid/valuable mostly because the taxation capability of any Govt. Consider the US which has > 90% economic activity in the private sector. That engine (revenues) is what provides corporate taxes, payroll taxes and when paid through to employees and owners, income taxes. It is not possible for ANY Govt to maintain long term adequate REAL taxation, if the underlying economic engine falls apart. In that case (many countries have tried) the recipe is money printing causing large inflation. That can handle nominal bonds but then the real returns will be terrible. Countries like UK which had huge debts after the second world war and not great revival of economy (till recent decades) did indeed use this a lot. In the emerging market, there are cases after cases of this.

Going forward (not trying to make this post political), with aging population, slowing birth rates and huge Govt debt/GDP in basically every developed country from Japan to Germany to US..I don't see how decent REAL returns on bonds can exist without the economic engine also doing well. If the economic engine does well, stocks will do well and that's what you should own. If the economic engine tanks for the long haul, bonds don't have some unique method of getting paid..

What I said above is obviously true for corporate bonds in aggregate but also (in my view) about Munis/ Treauries/ any other Govt whatever.
AlohaJoe
Posts: 6609
Joined: Mon Nov 26, 2007 1:00 pm
Location: Saigon, Vietnam

Re: A short study of the recent Japanese crisis

Post by AlohaJoe »

siamond wrote:Anyhoo, back to Japan? :wink:
Back to Japan...

It is also interesting to think about how international diversification might have helped a US retiree. We know that the 1965-1975 period was the worst for retirement due to the terrible US market conditions. These were Japanese returns over that period:

Image

(Where two numbers are show, the first is for a Yen-investor, the second is for a USD-investor. When only one number was shown, there was a fixed exchange rate regime.)

Of course, hindsight is 20/20 but it isn't hard to imagine that having some of your portfolio in Japan when it was returning 30%, 40%, 50%, 60%, and 90% in single years would have helped whether American stagflation.

Japan is an interesting case where extending the historical record back is of limited use.

Image

It did so ridiculously poorly in the post-WW2 decade when the entire nation was rubble and starvation was widespread that it is hard to know what lessons to draw from that. Likewise, the rebound in the 1950s is so ridiculous that the lesson would be "invest in stocks on as much margin as you possibly can". I mean, 153% returns in a single year. A $100,000 portfolio turned into $400,000 portfolio in 24 months.
User avatar
Topic Author
siamond
Posts: 6003
Joined: Mon May 28, 2012 5:50 am

Re: A short study of the recent Japanese crisis

Post by siamond »

AlohaJoe wrote:
siamond wrote:
AlohaJoe wrote:Here's the correlations between global markets going back to 1860
What is exactly displayed? Correlation between what and what?
"Average correlation of capital appreciation returns for all available markets"
siamond wrote:Could you also share the source? It is certainly a thought provoking graph.
"Long-Term Global Market Correlations" (2001), NBER Working Paper 8612
Here is a direct link to the NBER article, which is quite an impressive piece of research. This is eye-opening, I had absolutely no idea that global correlations went up and down like that. Note that the figures are at the very end of the PDF file.

To answer my own question, the 'average correlation' appears to be computed by averaging all peer correlations in a matrix of all countries involved in the analysis for a given period of time. To take a trivial example, say that US, UK and France were involved. Then the correlations between US and UK, between US and France, and between UK and France, are computed and then averaged together.

The conclusion is fairly short and interesting, so let me quote it - for context, the paper was written at the end of 2001:
Long-term investing depends upon meaningful long-term inputs to the asset allocation decisions. One approach to developing such inputs is to collect data from historical time periods. In this paper, we collect information from 150 years of global equity market history in order to evaluate the stationarity of the equity correlation matrix through time. Our tests suggest that the structure of global correlations shifts considerably through time. It is currently near an historical high – approaching levels of correlation last experienced during the Great Depression. Unlike the 1930’s however, the late 1990’s were a period of prosperity for world markets. The time-series of average correlations show a pattern consistent with the “U” shaped hypothesis about the globalization at the two ends of the 20th century. Decomposing the pattern of correlation through time, however, we find that roughly half the benefits of diversification available today to the international investor are due to the increasing number of world markets and available to the investor, and half is due lower average correlation among the available markets. An analysis of the capital-weighted portfolio suggests that benefits are less than the equal-weighted strategy, but that the proportionate risk reduction by adding in emerging markets has actually been roughly the same over the past 25 years.
In other words, if you don't invest in emerging markets (say by solely using a Developed-only International fund), you're missing out on diversification. Which obviously makes sense, but it's interesting to view an EM investment as a risk reduction approach!
flyingaway
Posts: 3908
Joined: Fri Jan 17, 2014 9:19 am

Re: A short study of the recent Japanese crisis

Post by flyingaway »

livesoft wrote:Can I ask a question related to this please?

My impression from having many Japanese friends is that "a Japanese passive investor" has never invested that much in the stock market. They are simply less concerned about what happened to the Japanese stock market than Americans are concerned about the Japanese stock market. They have virtually lifetime employment and retirement plans backed by bonds, so stocks never figured much into their thoughts. They do have concern about real estate prices, but not stock prices.

Is that a good impression? Because that's the one I have.
Many people here use Japan's "lost two decades" as one of the bad things that might happen in our retirement planning (and market). However, I wonder if any of those people or "experts" who wrote about it had really lived in Japan for a long period of time.

Based on what I know from my Japanese friends and some reading about their life, most people in Japan are very happy with what happened in Japan in the past two decades. They have a much better quality of life. Maybe stock market is just one unimportant thing in their life.
garlandwhizzer
Posts: 3562
Joined: Fri Aug 06, 2010 3:42 pm

Re: A short study of the recent Japanese crisis

Post by garlandwhizzer »

Kevin M wrote in response to my post:

I think you may be confusing safe withdrawal rates (SWRs) with returns.
Sorry about the errors, and thanks for the correction. I didn't read the whole post, just zoomed to the graphs which caught my eye. My comments were based a trio of misperceptions: wrong country (US rather than Japan), using returns rather than safe withdrawal rates, and assuming the graphs which puzzled me had to be showing nominal rather than real returns.

Garland Whizzer
User avatar
Topic Author
siamond
Posts: 6003
Joined: Mon May 28, 2012 5:50 am

Re: A short study of the recent Japanese crisis

Post by siamond »

-------- Follow-up article --------

This article is a follow-up to the "short study of the recent Japanese crisis" which was published on this blog a few days ago. The study was actively discussed on the Bogleheads forum. Various interesting questions were raised that weren't directly addressed by the original write-up, and triggered the author to do more research about domestic tilts (or lack thereof), variable withdrawals, and past performance of the stock market in Japan.

Read more of this second blog post on Financial Page. Feedback welcome.
Northern Flicker
Posts: 15289
Joined: Fri Apr 10, 2015 12:29 am

Re: A short study of the recent Japanese crisis

Post by Northern Flicker »

Wade Pfau has studied SWR for US retirees with global diversification (50% US and 50% international):

https://www.advisorperspectives.com/art ... -rates.pdf

Figure 1 on page 11 is interesting. Int'l diversification increased the supported SWR in most years, but most of the years in which a 4% SWR failed were internationally diversified portfolios in the early part of the 20th century.

In a few years in the 1970's, the 4% SWR failed for US-only portfolios and internationally diversified portfolios succeeded.
Post Reply