Study: The Rate of Return on Everything, 1870-2015

For investors outside the US. Personal investments, personal finance, investing news and theory.
Sister forums: Canada, Spain (en español)
---------------
KyleAAA
Posts: 7134
Joined: Wed Jul 01, 2009 5:35 pm
Contact:

Re: Study: The Rate of Return on Everything, 1870-2015

Post by KyleAAA » Mon Jan 15, 2018 10:46 am

WanderingDoc wrote:
Fri Jan 12, 2018 5:04 pm
My preference to real estate are two-fold: cash flow and tax efficiency.
The point is that direct real estate isn't special in any way along those lines.
WanderingDoc wrote:
Fri Jan 12, 2018 5:04 pm
Investing in a company with real estate experience (ie. REIT) gives you a measly 5-7% return. I won't even consider a deal if it doesn't produce a minimum 20% IRR. So no, I definitely don't want to 'leave it to the experienced ones' :wink: :beer
If you care ONLY about returns and not risk then sure, leverage up your limited amount of capital and hope for the best. Most rational investors prefer more balance.

WanderingDoc
Posts: 1341
Joined: Sat Aug 05, 2017 8:21 pm

Re: Study: The Rate of Return on Everything, 1870-2015

Post by WanderingDoc » Mon Jan 15, 2018 1:15 pm

KyleAAA wrote:
Mon Jan 15, 2018 10:46 am
WanderingDoc wrote:
Fri Jan 12, 2018 5:04 pm
My preference to real estate are two-fold: cash flow and tax efficiency.
The point is that direct real estate isn't special in any way along those lines.
WanderingDoc wrote:
Fri Jan 12, 2018 5:04 pm
Investing in a company with real estate experience (ie. REIT) gives you a measly 5-7% return. I won't even consider a deal if it doesn't produce a minimum 20% IRR. So no, I definitely don't want to 'leave it to the experienced ones' :wink: :beer
If you care ONLY about returns and not risk then sure, leverage up your limited amount of capital and hope for the best. Most rational investors prefer more balance.
Actually, direct real estate IS special in that regard. Why don't you ask Mr. Trump why he pays so little taxes? Hint: its not his mutual fund portfolio ;)

My real estate portfolio currently has very little debt. About 40-50% LTV in aggregate, which is quite low.
I'm not looking to get rich quick (stocks), I'm not looking to get rich slow (indexing), I'm looking to get rich, for sure (real estate) | Don't wait to buy real estate. Buy real estate.. and wait.

KyleAAA
Posts: 7134
Joined: Wed Jul 01, 2009 5:35 pm
Contact:

Re: Study: The Rate of Return on Everything, 1870-2015

Post by KyleAAA » Mon Jan 15, 2018 2:03 pm

WanderingDoc wrote:
Mon Jan 15, 2018 1:15 pm

Actually, direct real estate IS special in that regard. Why don't you ask Mr. Trump why he pays so little taxes? Hint: its not his mutual fund portfolio ;)
What, specifically, about real estate is special in that regard that companies in other industries can't and don't take advantage of? Please be specific about what exactly about real estate is special in that regard.

OkieIndexer
Posts: 429
Joined: Sun Aug 23, 2009 1:10 pm

Re: Study: The Rate of Return on Everything, 1870-2015

Post by OkieIndexer » Mon Jan 15, 2018 2:20 pm

Physical real estate or land that you own, especially if it has a pretty reliable income stream like rental property, is a true alternative investment. As the paper showed, rental property can even have higher returns than the stock market in some scenarios.

My Depression-era grandmother (born in 1914) bought a bunch of double-digit % long-term (probably 5 year) CDs in the early 80s, and that combined with her rental property was all she needed to fund retirement (plus Social Security/Medicare, no pension). She did great, and she would have done great even if the high inflation/bad stock market 70s had continued into the 80s and 90s.
"In bull markets, people say 'The more risk I take, the greater my return.' But when people aren't afraid of risk, they'll accept risk without being compensated." -Howard Marks, Oaktree Capital

WanderingDoc
Posts: 1341
Joined: Sat Aug 05, 2017 8:21 pm

Re: Study: The Rate of Return on Everything, 1870-2015

Post by WanderingDoc » Mon Jan 15, 2018 3:36 pm

KyleAAA wrote:
Mon Jan 15, 2018 2:03 pm
WanderingDoc wrote:
Mon Jan 15, 2018 1:15 pm

Actually, direct real estate IS special in that regard. Why don't you ask Mr. Trump why he pays so little taxes? Hint: its not his mutual fund portfolio ;)
What, specifically, about real estate is special in that regard that companies in other industries can't and don't take advantage of? Please be specific about what exactly about real estate is special in that regard.
You are not a company or industry. Neither am I. We are small time investors. And we were comparing tax advantages of equities vs. real estate, before you conveniently trying to deter the topic. A small time investor can benefit from killer tax strategies and pay very little taxes, WITHOUT being a big industry or corporation. Thats the point. This just isn't available to mutual fund investors.
I'm not looking to get rich quick (stocks), I'm not looking to get rich slow (indexing), I'm looking to get rich, for sure (real estate) | Don't wait to buy real estate. Buy real estate.. and wait.

KyleAAA
Posts: 7134
Joined: Wed Jul 01, 2009 5:35 pm
Contact:

Re: Study: The Rate of Return on Everything, 1870-2015

Post by KyleAAA » Tue Jan 16, 2018 11:00 am

WanderingDoc wrote:
Mon Jan 15, 2018 3:36 pm
You are not a company or industry. Neither am I. We are small time investors. And we were comparing tax advantages of equities vs. real estate, before you conveniently trying to deter the topic. A small time investor can benefit from killer tax strategies and pay very little taxes, WITHOUT being a big industry or corporation. Thats the point. This just isn't available to mutual fund investors.
The point you're missing is that the corporations I'm investing in are doing the same things you are doing with real estate. The benefit to me in both cases, as measured by the total amount of after-tax wealth adjusted for risk, are more or less the same averaged over the population. You are claiming that because the tax strategies happen on YOUR income statement instead of the company's income statement that there is some sort of advantage there. I'm saying I don't buy your claim: it's an accounting fiction. Indeed, the fact that terminal after-tax wealth is comparable in both cases is strong evidence to the contrary. In reality it's quite easy to pay very little taxes with equities: I've been doing it for years along with many others on this forum.
Last edited by KyleAAA on Tue Jan 23, 2018 1:04 pm, edited 1 time in total.

mkg
Posts: 1
Joined: Tue Jan 23, 2018 12:34 am

Re: Study: The Rate of Return on Everything, 1870-2015

Post by mkg » Tue Jan 23, 2018 1:30 am

Summary of explainers for why they obtained the result "housing returns > stock returns since 1870"

These factors might explain why the study observed excess returns for housing relative to stocks, while in fact the returns are not excess, rather they just compensate investors for hidden costs. These are all of the possible hidden costs of housing (that stocks don't have). Everything below is for housing relative to stocks. The word "house" is shorthand for an real estate property.


1. Illiquidity premium - housing returns more because it's harder to buy/sell
1a. Increased bid/ask spreads - lower volumes of housing transactions vs. stock transactions
1b. Increased other transaction costs - closing costs are typically high
1c. Less favorable tax treatment requires higher nominal returns

2. Housing index is unrealizable. This means that just like some stock backtests that show profits that couldn't possibly be achieved, this study is tracking housing prices through time in a way that couldn't possibly be captured by investors.
2a. Legally impossible to buy some housing - hard to/can't buy housing in a foreign country, particularly emerging markets
2b. Irrationally long holding periods - prices in the housing index have to be imputed from the inventory that's actually sold. However, if people refuse to sell their house under any circumstance (and obtain the legal ability to continue holding house irrationally long) then observing a single house sell in a neighborhood isn't actually a good measure for an avg sale price, because in order to obtain inventory in that neighborhood, you would have to offer people way above "market" prices because they are irrationally attached to the house.
2c. An index assumes fractional ownership of many houses is possible, but in reality it's not. Even modern REITs exclude owner-occupied mandated housing from their holdings and therefore couldn't accurate reproduce this index over long periods of time.
2d. If power law of returns holds in housing, then a small percent of housing would generate the vast majority of returns. However, due to the socially subsidized availability of debt financing (mortgages) the market will not force owners to sell to realize returns. This makes it impossible for a momentum type strategy to be executed because very little inventory will be available in the most successful markets.

3. Measurement errors in the price of a house - the index must account for changes in the price of a house due to improvements/deprovements. Assumptions about how to account for these underlying changes may be flawed.
3a. Houses have gotten bigger over time in a way that is hard to measure - people always have an incentive to report smaller than actual sqft to the government to avoid taxes, so maybe early measurement data is inaccurate. Or maybe entirely missing. How many farmers in 1928 could correctly measure heated sqft of their farm house?
3b. Owners of rental properties are incentivized to overstate the cost of improvements to housing to get larger tax depreciation write offs - this could inflate the price index, but due to point 2b, true market prices for the improvements are never established correctly.
3c. After a house becomes negative value (because the structure has rotted or some other damage), if the house is rebuilt without being sold in between, it's slowly decreasing value over time will never show up on the index even though the owner has incurred a significant expense while owning the asset. If this is common, then the index systematically ignores reconstruction costs making the price gains untrue.

4. Measurement errors in the rent collected on a house that DECREASE returns. (see below for those that increase returns).
4a. Collecting rent requires providing property management and rent collection services.
4b. This paper directly contradicts one of it's citations. It relies on Knoll, Katharina (2016) which is another study of returns on housing which finds that 80% of the increase is explained by land value increases (which can't possibly be direct collection of rent). However this paper argues 50% of the returns on housing come from rent.
4c. Rental data is not directly tracked by anyone until the advent of national realtors and is therefore unlikely to be representative of all housing. The author of Knoll, Katharina (2016) has the following to say about their original rent index:
"Where do the rent data come from? The construction of the dataset was in large part an investigative and assembly operation"... " I consulted a broad range of sources including publications of national statistical offices and central banks, publications of the International Labor Organization, economic and financial history books and journal articles. For most countries, I relied on the rent components of the cost of living or consumer price index as constructed by national statistical offices and combined them with information from other sources to create long-run series reaching back to the late 19th century"... " Constructing long-run rent series required pragmatic choices between the ideal and the available data."... " I chose indices that control for quality changes where available, that best concurred with the respective house price series, and opted for within-country consistency as well as historical plausibility." There are two major issues with this approach. 1) Reliance on an aggregation of already aggregate figures from governments in early 1900s (large error bars) and 2) filtering for historical plausibility means replacing outliers with more normal seeming data. However, in all investments, outliers are common (fat tails) and tend to represent the majority of returns (power law). So taking a data sanitization approach *probably* overstates the return and *certainly* understates the volatility.

Counter argument - there are some errors in the paper that could actually increase returns even beyond what they measure:

1. Measurement errors in the rent collected on a house that INCREASE returns - the index must account for rent collected but there could be errors in the the measurement of rent.
1a. Owners have a strong incentive to hide all or part of rental income [THIS FACTOR ACTUALLY INCREASES RETURNS ON HOUSING]. Obvious for tax reasons.
1b. Rent collected from family may be paid in unmeasurable ways. Unreported gift income, love between spouses, etc.


-=-=-=

Maybe there's something I'm also missing about leverage. Something like in 1870 the developed world was deep into the industrial revolution and no one owned their land (the factory owners and lords owned all the land). Since then, society adopted a slow but steady policy of increasing ownership through subsidized loans. This steadily increasing demand created true excess returns. However now most people have a house and post-2007 no one wants to subsidize home ownership so the returns available now are lower than they were in the past.... this is mostly made up nonsense but maybe someone can correct me.

User avatar
siamond
Posts: 4773
Joined: Mon May 28, 2012 5:50 am

Re: Study: The Rate of Return on Everything, 1870-2015

Post by siamond » Fri Feb 02, 2018 11:39 am

siamond wrote:
Mon Sep 11, 2017 11:49 am
I finally got an answer from the folks managing the database. Not terribly helpful, but keeping the door open...
The dataset on the website http://www.macrohistory.net/data/ currently does not contain the dividend and housing rent series. These new series will only be incorporated into the JST dataset that is available online at a later stage.
I asked again mid-December, didn't get an answer, I had given up... Out of the blue, I just received an answer this morning (more than 2 months later) stating the following:
Thank you for your interest in the paper and the data. We will make the data from "Rate of Return on Everything" available online once the paper - and the data - have been through the peer review process, and the study is published in a journal. Hopefully this will be months rather than years, but we cannot commit to a specific date because it depends on the review and publication process.
So I guess there is still hope to get the missing dividends series, one day or another. I marked my calendar, to ask from time to time.

csp256
Posts: 4
Joined: Sun Mar 12, 2017 3:14 pm

Re: Study: The Rate of Return on Everything, 1870-2015

Post by csp256 » Mon Jul 23, 2018 11:13 am

I wrote a script to scrape the results from this paper, because I wanted to play around with the data. Here is the Rent-Price Ratio for the US, in percent, from 1890 to 2015 (oldest first), scraped from the green line in Figure A.20 on page A92:

Code: Select all

7.023696682464455, 7.649289099526066, 7.592417061611374, 7.194312796208531, 6.028436018957346, 6.274881516587678, 7.165876777251185, 6.682464454976303, 6.18957345971564, 6.303317535545023, 5.355450236966824, 6.350710900473934, 5.516587677725118, 5.440758293838862, 5.402843601895735, 6.123222748815166, 5.213270142180095, 7.071090047393365, 5.232227488151659, 5.042654028436019, 4.701421800947867, 4.654028436018957, 4.4170616113744074, 4.492890995260663, 4.303317535545023, 4.691943127962086, 4.303317535545023, 4.180094786729858, 3.962085308056872, 3.8862559241706163, 4.14218009478673, 4.834123222748815, 4.881516587677726, 4.900473933649289, 5.042654028436019, 4.796208530805687, 4.909952606635071, 5.004739336492891, 4.796208530805687, 4.748815165876778, 5.298578199052133, 6.018957345971564, 6.606635071090047, 5.85781990521327, 5.7725118483412325, 5.222748815165877, 5.137440758293839, 5.251184834123222, 5.886255924170616, 5.781990521327014, 6.293838862559242, 7.1469194312796205, 6.947867298578199, 6.350710900473934, 5.469194312796208, 4.900473933649289, 4.056872037914692, 3.6966824644549763, 3.924170616113744, 4.180094786729858, 4.2274881516587675, 4.322274881516588, 4.398104265402844, 4.293838862559242, 4.360189573459715, 4.4454976303317535, 4.606635071090047, 4.597156398104265, 4.654028436018957, 4.777251184834123, 4.8530805687203795, 4.947867298578199, 4.976303317535545, 5.014218009478673, 5.014218009478673, 5.023696682464455, 5.071090047393365, 5.156398104265403, 5.213270142180095, 5.156398104265403, 5.061611374407583, 5.061611374407583, 5.071090047393365, 5.308056872037914, 5.308056872037914, 5.1469194312796205, 5.1469194312796205, 5.004739336492891, 4.786729857819905, 4.6350710900473935, 4.777251184834123, 5.023696682464455, 5.260663507109005, 5.374407582938389, 5.5260663507109005, 5.62085308056872, 5.5829383886255926, 5.440758293838862, 5.345971563981043, 5.289099526066351, 5.3364928909952605, 5.412322274881516, 5.38388625592417, 5.393364928909953, 5.38388625592417, 5.3649289099526065, 5.327014218009479, 5.308056872037914, 5.222748815165877, 5.127962085308057, 5.014218009478673, 4.872037914691943, 4.758293838862559, 4.61611374407583, 4.331753554502369, 4.009478672985782, 3.876777251184834, 3.990521327014218, 4.341232227488152, 4.691943127962086, 4.890995260663507, 5.165876777251185, 5.270142180094787, 5.023696682464455, 4.909952606635071, 4.786729857819905
Obviously these are only approximate and I make no guarantee of any type as to their validity. Let me know if you would like any of the other plots pulled out into numerical form.

User avatar
siamond
Posts: 4773
Joined: Mon May 28, 2012 5:50 am

Re: Study: The Rate of Return on Everything, 1870-2015

Post by siamond » Tue Jan 29, 2019 1:23 pm

I checked again the status of the new paper (supposedly including dividends data), here is the answer I received:
We are progressing through peer review; looks like the final version of the data and the paper will be up online within the next couple of months.
Geez, those peer reviews seem like an excruciating process! :shock:

s8r
Posts: 116
Joined: Thu May 10, 2018 1:50 pm
Location: Northern Europe

Study: The Rate of Return on Everything

Post by s8r » Tue Apr 02, 2019 12:42 am

[Thread merged into here, see below. --admin LadyGeek]

Interesting study by Harvard University on global returns of different asset classes during 1870-2015.
https://economics.harvard.edu/files/eco ... s28533.pdf

Things that caught my attention:
  • Housing has provided similar long-term returns as equities.
  • Performance of US equities does not stand out as much as expected.
  • Go Finland. :happy
Note that "period coverage differs across countries". I assume this applies only to the "full sample" results and not the "post 1950" or "post 1980" results.

Real arithmetic mean return in USD, full sample, ranked by equity return:

Code: Select all

	Equity	Housing	Bills	Bonds
FIN	12.08	12.00	1.93	6.48
GER	8.75	9.69	4.33	5.81
ITA	8.73	6.37	2.85	4.81
SWE	8.63	8.87	2.08	3.64
AUS	8.53	7.28	1.77	2.59
USA	8.46	6.10	1.52	2.33
JPN	8.00	8.74	2.37	4.15
DEN	7.94	8.93	3.63	4.12
NLD	7.91	8.67	1.87	2.93
GBR	7.66	6.25	1.89	2.66
CHE	7.62	7.14	2.04	3.62
ESP	7.47	6.38	0.93	2.36
BEL	7.36	8.92	0.96	3.27
NOR	6.72	8.88	1.64	3.04
PRT	6.09	7.10	0.23	2.11
FRA	5.16	8.87	1.05	3.10
Real arithmetic mean return in local currency, full sample, ranked by equity return:

Code: Select all

	Equity	Housing
FIN	10.03	9.58
USA	8.46	6.10
SWE	8.02	8.30
AUS	7.79	6.37
DEN	7.49	8.22
ITA	7.25	4.77
GER	7.11	7.82
NLD	6.96	7.28
GBR	6.83	5.44
CHE	6.51	5.63
BEL	6.23	7.89
JPN	6.00	6.54
ESP	5.83	5.21
NOR	5.67	8.03
PRT	4.51	6.31
FRA	3.21	6.39
Real arithmetic mean return in local currency, post 1950, ranked by equity return:

Code: Select all

	Equity	Housing
FIN	12.89	11.18
SWE	11.37	8.94
DEN	9.73	7.04
BEL	9.65	8.14
NLD	9.19	8.53
GBR	9.10	6.57
USA	8.89	5.76
CHE	8.37	5.64
ESP	7.75	5.83
AUS	7.53	8.29
GER	7.53	5.30
NOR	7.33	9.10
JPN	6.21	6.74
ITA	6.09	5.55
FRA	6.01	9.68
PRT	4.84	6.01
Real arithmetic mean return in local currency, post 1980, ranked by equity return:

Code: Select all

	Equity	Housing
FIN	16.32	9.47
SWE	15.87	9.00
DEN	13.30	5.14
NOR	12.22	9.82
ESP	11.96	4.62
NLD	11.51	6.41
BEL	11.49	7.20
GER	10.07	4.13
FRA	9.61	5.78
ITA	9.45	4.57
USA	9.31	5.86
CHE	9.29	6.19
GBR	9.11	6.81
AUS	8.70	7.16
PRT	8.60	7.15
JPN	5.62	3.58
EDIT:
These are in fact arithmetic means. Here are geometric means.

Real geometric mean return in local currency, full sample, ranked by equity return:

Code: Select all

	Equity	Housing
USA	6.66	5.80
AUS	6.37	5.90
NLD	6.27	7.16
DEN	6.23	7.95
SWE	6.08	7.92
FIN	5.78	8.52
GER	5.08	7.36
GBR	5.02	5.05
JPN	4.34	6.22
CHE	4.26	5.03
ESP	3.81	4.56
NOR	3.79	7.68
ITA	3.66	4.39
BEL	3.57	6.79
FRA	0.90	5.90
PRT	-0.67	4.79
Last edited by s8r on Tue Apr 02, 2019 4:29 am, edited 1 time in total.

User avatar
steve roy
Posts: 1642
Joined: Thu May 13, 2010 5:16 pm

Re: Study: The Rate of Return on Everything

Post by steve roy » Tue Apr 02, 2019 12:44 am

The Finnish have the secret sauce.

techcrium
Posts: 142
Joined: Sun Mar 24, 2013 12:09 am

Re: Study: The Rate of Return on Everything

Post by techcrium » Tue Apr 02, 2019 10:12 am

How is housing evaluated though?

Is it by buying a 1900s house and holding it for 100 years?? Or are they comparing by average new house built?

s8r
Posts: 116
Joined: Thu May 10, 2018 1:50 pm
Location: Northern Europe

Re: Study: The Rate of Return on Everything

Post by s8r » Tue Apr 02, 2019 11:16 am

techcrium wrote:
Tue Apr 02, 2019 10:12 am
How is housing evaluated though?

Is it by buying a 1900s house and holding it for 100 years?? Or are they comparing by average new house built?
Haven't looked at the details of the methodology, but I guess the margin of error here is high for housing returns. Note also that these housing returns include also rental return.

ohai
Posts: 679
Joined: Wed Dec 27, 2017 2:10 pm

Re: Study: The Rate of Return on Everything

Post by ohai » Tue Apr 02, 2019 11:28 am

If you hold a house from 1950 until now, that house will have to be renovated or rebuilt several times. The price is not continuous. So, it's hard to measure unit returns over time. Maybe a better way would be to track real estate investment company stock prices over the same time.

User avatar
siamond
Posts: 4773
Joined: Mon May 28, 2012 5:50 am

Re: Study: The Rate of Return on Everything

Post by siamond » Tue Apr 02, 2019 1:25 pm

This new thread should probably be merged with this older one, which discussed the same study.

s8r
Posts: 116
Joined: Thu May 10, 2018 1:50 pm
Location: Northern Europe

Re: Study: The Rate of Return on Everything

Post by s8r » Tue Apr 02, 2019 1:38 pm

siamond wrote:
Tue Apr 02, 2019 1:25 pm
This new thread should probably be merged with this older one, which discussed the same study.
Funny, I thought this was a new study since it says "March 2019" on the title page of the paper. :D

User avatar
siamond
Posts: 4773
Joined: Mon May 28, 2012 5:50 am

Re: Study: The Rate of Return on Everything

Post by siamond » Tue Apr 02, 2019 3:34 pm

s8r wrote:
Tue Apr 02, 2019 1:38 pm
siamond wrote:
Tue Apr 02, 2019 1:25 pm
This new thread should probably be merged with this older one, which discussed the same study.
Funny, I thought this was a new study since it says "March 2019" on the title page of the paper. :D
As you can see on the other thread, I've been actually waiting for their update for quite a while!

User avatar
dodecahedron
Posts: 4356
Joined: Tue Nov 12, 2013 12:28 pm

Re: Study: The Rate of Return on Everything

Post by dodecahedron » Tue Apr 02, 2019 3:38 pm

steve roy wrote:
Tue Apr 02, 2019 12:44 am
The Finnish have the secret sauce.
The discrepancy between the ranking of Finland´s arithmetic averages vs. their geometric averages tells us that Finland has had a much higher volatility roller coast ride than the other high ranked countries.

User avatar
steve roy
Posts: 1642
Joined: Thu May 13, 2010 5:16 pm

Re: Study: The Rate of Return on Everything

Post by steve roy » Wed Apr 03, 2019 1:30 am

dodecahedron wrote:
Tue Apr 02, 2019 3:38 pm
steve roy wrote:
Tue Apr 02, 2019 12:44 am
The Finnish have the secret sauce.
The discrepancy between the ranking of Finland´s arithmetic averages vs. their geometric averages tells us that Finland has had a much higher volatility roller coast ride than the other high ranked countries.
"Finnish"? Good golly. How about "The Finns"?

msk
Posts: 1190
Joined: Mon Aug 15, 2016 10:40 am

Re: Study: The Rate of Return on Everything

Post by msk » Wed Apr 03, 2019 2:02 am

OK, most BHs have already gone through a significant fraction of their accumulation phase. Nevertheless my advice to youngsters has always been:
Invest in your home as soon as you have stability and pay it off asap. This way you have a toehold on the RE market and you have an excellent tenant, yourself. Beyond that, invest in 100% stock index funds or whatever strikes your sober fancy. Your paid up home serves as a moderator, bond equivalent? Unfortunately people get tempted to buy homes that are over-priced compared to their incomes and then they carry mortgages their entire working lives having been tempted by low interest rates. Rules of thumb:

Save and invest 30% of after tax income (paying off principal on a home mortgage counts as investing, not the interest)
Never purchase a home worth more than 3x annual income (2.5x combined income)
Never acquire a car(s) worth more than 6 months income

Hobby purchases can skew some of those "rules" but do recognize that you are indulging a hobby, e.g. too expensive a home or a Porsche...
No need for detailed interest payment calculations in any of the above; just leads to endless indecision. I did very well even in the high teens interest environment of the early 1980s, and by the mid 1990s I already had 30 rental units fully paid up. Sold the lot in the 2000s and made a whole :moneybag in stocks. Start investing early. Be very careful about speculating but stay invested. We all pick up on different opportunities at different times. Be awake enough to recognize those that drop into your lap. They never repeat... Does not matter much what you are invested in. As long as you stay invested! For a couple of years I was the Head of Economics and Planning for a subsidiary of a Fortune 10 multinational. It was a real eye-opener to find that we had billions available for investment annually, but there were always far, far more investment opportunities than we could possibly fund. Last year I was in that role I set the cut-off at an Earning Power of 14%. Any project forecasted to do below that was simply dropped... Of course some do much better, some much worse. Yes. You cannot forecast the future.

s8r
Posts: 116
Joined: Thu May 10, 2018 1:50 pm
Location: Northern Europe

Re: Study: The Rate of Return on Everything

Post by s8r » Wed Apr 03, 2019 2:40 am

msk wrote:
Wed Apr 03, 2019 2:02 am
Nevertheless my advice to youngsters has always been: Invest in your home as soon as you have stability and pay it off asap.
Depends.

I'm a young highly-educated person who prefers rural areas to cities or suburbs. I'm in no hurry to pay off my mortgage. The house value is not that high and there is a reasonable likelihood it will depreciate (better to put my money elsewhere). I'm able to invest 50% of my after tax income (40% if i payed the mortgage at a "fast" rate).

Mortgage interest rates in my country are also very low: I think the average is around 0.8% with the lowest rates below 0.3%. These are adjustable rates. Fixed rates are probably 1.5-2.5%, but they are not very popular.
Last edited by s8r on Wed Apr 03, 2019 2:52 am, edited 1 time in total.

Valuethinker
Posts: 38113
Joined: Fri May 11, 2007 11:07 am

Re: Study: The Rate of Return on Everything

Post by Valuethinker » Wed Apr 03, 2019 2:49 am

msk wrote:
Wed Apr 03, 2019 2:02 am
OK, most BHs have already gone through a significant fraction of their accumulation phase. Nevertheless my advice to youngsters has always been:
Invest in your home as soon as you have stability and pay it off asap. This way you have a toehold on the RE market and you have an excellent tenant, yourself. Beyond that, invest in 100% stock index funds or whatever strikes your sober fancy. Your paid up home serves as a moderator, bond equivalent? Unfortunately people get tempted to buy homes that are over-priced compared to their incomes and then they carry mortgages their entire working lives having been tempted by low interest rates. Rules of thumb:

Save and invest 30% of after tax income (paying off principal on a home mortgage counts as investing, not the interest)
Never purchase a home worth more than 3x annual income (2.5x combined income)
That's not possible in many metropolitan areas. You live in (I imagine) SF area, NYC area, Washington DC area, Seattle area, Toronto/ Vancouver, London, Portland ((?), Amsterdam - it essentially means "rent for life".

4x income is allowed these days, first property I could see it being higher, particularly the implicit leverage (you may well borrow your downpayment from your parents).
Never acquire a car(s) worth more than 6 months income
That's probably a good rule, but for a first car that goes with a first job (most North American cities you cannot live without a car) that's probably not the case. A used car would work, but for a single woman, I generally would not suggest buying a used car unless absolutely necessary - the problem of being broken down somewhere you should not be looms large.

A better rule might be "plan to drive your cars for 10 years or 150,000 miles (250k km)". That makes one more cautious about car purchase. I am thinking of something like "divide the cost of the car by 10 (for depreciation) and double it (insurance, repairs, parking, a moderate amount of gasoline)" for a feel for the average annual cost of owning a car.

Or, if one leases (which most people seem to do, these days) then to treat the full leasing cost as a rental "expense" i.e. don't fool yourself that you are building up equity.
Hobby purchases can skew some of those "rules" but do recognize that you are indulging a hobby, e.g. too expensive a home or a Porsche...
No need for detailed interest payment calculations in any of the above; just leads to endless indecision. I did very well even in the high teens interest environment of the early 1980s, and by the mid 1990s I already had 30 rental units fully paid up. Sold the lot in the 2000s and made a whole :moneybag in stocks. Start investing early. Be very careful about speculating but stay invested. We all pick up on different opportunities at different times. Be awake enough to recognize those that drop into your lap. They never repeat... Does not matter much what you are invested in. As long as you stay invested! For a couple of years I was the Head of Economics and Planning for a subsidiary of a Fortune 10 multinational. It was a real eye-opener to find that we had billions available for investment annually, but there were always far, far more investment opportunities than we could possibly fund. Last year I was in that role I set the cut-off at an Earning Power of 14%. Any project forecasted to do below that was simply dropped... Of course some do much better, some much worse. Yes. You cannot forecast the future.
That tells you that companies are unrealistic about their actual returns in practice. The average corporate return on projects now is probably something below 8 per cent.

Yes there were times such as Asset Management businesses in the 1990s, software companies (generally), pharmaceutical companies, where returns were sustained well above normal corporate returns. But even the good times come to an end - look at pharma now compared to pharma then - financial engineering and cost cuts are now what drives earnings growth (FE via acquisitions, often).
Last edited by Valuethinker on Wed Apr 03, 2019 3:20 am, edited 2 times in total.

Valuethinker
Posts: 38113
Joined: Fri May 11, 2007 11:07 am

Re: Study: The Rate of Return on Everything

Post by Valuethinker » Wed Apr 03, 2019 2:57 am

ohai wrote:
Tue Apr 02, 2019 11:28 am
If you hold a house from 1950 until now, that house will have to be renovated or rebuilt several times. The price is not continuous. So, it's hard to measure unit returns over time. Maybe a better way would be to track real estate investment company stock prices over the same time.
Someone came up with an index of office buildings in New York (or Boston?) from the late 1800s until now. The structures still exist but there have been comprehensive renovations every 30-40 years.

What was striking is how low the actual appreciation was. In cities that have done very well in the 20th Century - this wasn't Cleveland or Buffalo.

What seems to have happened is that better transportation and communication technology have caused cities to spread out, so the locational advantages have diminished - even in Manhattan. Another thought that occurs to me is we build taller now - although for office buildings that's largely self defeating*.

All the data I have seen on houses (Neil Monnery "Safe as Houses: 8 centuries of housing prices") suggests that housing is, surprisingly, an indifferent bet - real appreciation is small, or not at all. You do get bubbles and busts though, and these can last decades.


* you need bigger banks of high speed elevators as you grow taller, thus smaller floorplate and hence less rentable space. I think 40 storeys is about the optimum. Above that it's brag factor as much as anything else -- construction of very tall office buildings correlates nicely with stock market peaks. The World Trade Center had an unusual structure and made compromises re location of stairwells which in retrospect were a mistake - but to preserve floorplate size.

For residential buildings, which have many fewer elevators, this is less of a problem - no reason not to go to 100 storeys except for local civil aviation rules.

User avatar
siamond
Posts: 4773
Joined: Mon May 28, 2012 5:50 am

Re: Study: The Rate of Return on Everything, 1870-2015

Post by siamond » Wed Apr 03, 2019 11:08 am

The updated paper was finally published, as discussed on this new thread:
viewtopic.php?f=22&t=277535

Unfortunately, following the link to the underlying database, the dividends are still missing. I'll wait a few days and ping the authors...

User avatar
LadyGeek
Site Admin
Posts: 54308
Joined: Sat Dec 20, 2008 5:34 pm
Location: Philadelphia
Contact:

Re: Study: The Rate of Return on Everything

Post by LadyGeek » Wed Apr 03, 2019 3:46 pm

siamond wrote:
Tue Apr 02, 2019 1:25 pm
This new thread should probably be merged with this older one, which discussed the same study.
I have merged SimpleGift's thread into the later discussion by s8r, which is in the Non-US Investing forum. Merged threads can only have one location. Since the focus is on everything, I chose the Non-US Investing forum.
Wiki To some, the glass is half full. To others, the glass is half empty. To an engineer, it's twice the size it needs to be.

User avatar
Tyler9000
Posts: 423
Joined: Fri Aug 21, 2015 11:57 am

Re: Study: The Rate of Return on Everything, 1870-2015

Post by Tyler9000 » Wed Apr 03, 2019 4:44 pm

siamond wrote:
Wed Apr 03, 2019 11:08 am
The updated paper was finally published, as discussed on this new thread:
viewtopic.php?f=22&t=277535

Unfortunately, following the link to the underlying database, the dividends are still missing. I'll wait a few days and ping the authors...
Thanks for keeping track, Siamond. This could be a really cool source if they get their act together on the dividends.

protagonist
Posts: 5778
Joined: Sun Dec 26, 2010 12:47 pm

Re: Study: The Rate of Return on Everything

Post by protagonist » Thu Apr 04, 2019 9:26 am

Doesn't anybody factor in the hassle factor? Is that not worth something? Do people here not consider that their time (and anxiety level) is worth money?

Maybe housing and equities provide similar returns. But I bought my index funds about two decades ago- I forget if it took a phone call or a couple of minutes on a computer- I have rarely thought about it since. Fidelity does all my accounting for me.

I have also owned one simple property in FL that I have rented over the last decade. The cash flow was good. I have had to deal with repairs, good renters, renters from hell, advertising, paying bills, lawyers, travel, idiots in the condo board, collecting rent, paperwork, etc. I finally sold it last year- more hassles with finding realtors, paperwork, lawyers, inspections, negotiations, etc. It was in FL so it lost about 2/3 of its value during the crash (fortunately I inherited it so it was not my money). It gained about 1/2 of that back by the time I sold it, but then there were closing costs.... And my taxes for 2018 were more complicated than ever.

Unloading it was like having an albatross lifted from my neck.

There are so many things in life more fun than managing real estate, and life is short. Sure, you can hire properrty managers, but you still have to deal with stuff. I get a statement from Fidelity once a month, glance at how my index fund is doing, file it away, and get back to dinner with my loved ones or playing my saxophone. Why anybody would voluntarily become a landlord , even if the return was a little superior to holding stock (and I am not convinced by the arguments above), is beyond me.

hoops777
Posts: 2573
Joined: Sun Apr 10, 2011 12:23 pm

Re: Study: The Rate of Return on Everything

Post by hoops777 » Thu Apr 04, 2019 9:33 am

protagonist wrote:
Thu Apr 04, 2019 9:26 am
Doesn't anybody factor in the hassle factor? Is that not worth something? Do people here not consider that their time (and anxiety level) is worth money?

Maybe housing and equities provide similar returns. But I bought my index funds about two decades ago- I forget if it took a phone call or a couple of minutes on a computer- I have rarely thought about it since. Fidelity does all my accounting for me.

I have also owned one simple property in FL that I have rented over the last decade. The cash flow was good. I have had to deal with repairs, good renters, renters from hell, advertising, paying bills, lawyers, travel, idiots in the condo board, collecting rent, paperwork, etc. I finally sold it last year- more hassles with finding realtors, paperwork, lawyers, inspections, negotiations, etc. It was in FL so it lost about 2/3 of its value during the crash (fortunately I inherited it so it was not my money). It gained about 1/2 of that back by the time I sold it, but then there were closing costs.... And my taxes for 2018 were more complicated than ever.

Unloading it was like having an albatross lifted from my neck.

There are so many things in life more fun than managing real estate, and life is short. Sure, you can hire properrty managers, but you still have to deal with stuff. I get a statement from Fidelity once a month, glance at how my index fund is doing, file it away, and get back to dinner with my loved ones or playing my saxophone. Why anybody would voluntarily become a landlord , even if the return was a little superior to holding stock (and I am not convinced by the arguments above), is beyond me.
+1
K.I.S.S........so easy to say so difficult to do.

User avatar
siamond
Posts: 4773
Joined: Mon May 28, 2012 5:50 am

Re: Study: The Rate of Return on Everything, 1870-2015

Post by siamond » Wed Apr 10, 2019 10:23 am

siamond wrote:
Wed Apr 03, 2019 11:08 am
The updated paper was finally published [...]

Unfortunately, following the link to the underlying database, the dividends are still missing. I'll wait a few days and ping the authors...
I just pinged one of the authors, who was kind enough to answer right away. The new data is available in Stata format (which I never played with), there are working on updating the Macrohistory public database (which provides an Excel pointer), it might take a month or so before it is made publicly available. Bottomline is... this is coming! Good things come to those who wait... :wink:

PS. the DMS/Morningstar folks must be a little miffed... They won't get much sympathy from me though... :twisted:

international001
Posts: 854
Joined: Thu Feb 15, 2018 7:31 pm

Re: Study: The Rate of Return on Everything, 1870-2015

Post by international001 » Mon May 27, 2019 5:53 pm

Very interesting paper. Trying to make sure I understand
I'm looking at https://www.frbsf.org/economic-research ... 017-25.pdf, appendix K, figure A.20

This would mean that in US the rental return of housing is about 5%. This doesn't count housing appreciation (but as discuss before, it's usually less than 1% after discounting inflation, it would be ~0% if we believe S&P/Case–Shiller ). It already discounts about the 2% maintenance costs. But it doesn't discount taxes, like property taxes (I wonder why not, since they are the same for everybody). It also doesn't discount income taxes (what it's important if you are renting the place to somebody else instead of just living there yourself).

So if I have ~1% on property taxes, let's say my total real return today is ~4%. If stocks are comparable, are those the returns am I expecting to get for stocks?

Did I get it right? Do I have any leaps somewhere?

TomCat96
Posts: 657
Joined: Sun Oct 18, 2015 12:18 pm

Re: Study: The Rate of Return on Everything, 1870-2015

Post by TomCat96 » Mon May 27, 2019 7:24 pm

whodidntante wrote:
Wed Aug 16, 2017 2:51 pm
Residential real estate investing is a job. I've already got one. Also it has costs and risks that if excluded will cause you to overestimate the returns and underestimate the risks of the investment. Taxes, loan costs and leverage risks, maintenance costs, insurance, legal costs (eviction, torts, etc.), transaction costs, vacancies, and black swans. I'm not saying it's a "bad investment" but you cannot compare it to lazily buying an index fund.
I agree.
You cannot say in a vacuum that Asset A performs better than Asset B, but Asset A gets to include additional personal time expended.
If anything, a fairer comparison would be the additional value of labor added to the stock market portfolio.

User avatar
siamond
Posts: 4773
Joined: Mon May 28, 2012 5:50 am

Re: Study: The Rate of Return on Everything, 1870-2015

Post by siamond » Mon Jun 10, 2019 5:21 pm

siamond wrote:
Wed Apr 10, 2019 10:23 am
siamond wrote:
Wed Apr 03, 2019 11:08 am
The updated paper was finally published [...]

Unfortunately, following the link to the underlying database, the dividends are still missing. I'll wait a few days and ping the authors...
I just pinged one of the authors, who was kind enough to answer right away. The new data is available in Stata format (which I never played with), there are working on updating the Macrohistory public database (which provides an Excel pointer), it might take a month or so before it is made publicly available. Bottomline is... this is coming! Good things come to those who wait... :wink:

PS. the DMS/Morningstar folks must be a little miffed... They won't get much sympathy from me though... :twisted:
That's it! The full data set is finally available in Excel format... Including price, dividends, total return data series for stocks and govt bonds for the various countries. Also bills, short-term and long-term bonds interest rates. And all the real estate data, inflation, GDP and more.

http://www.macrohistory.net/data/

Everything is free to share, redistribute, remix and build upon as long we provide the following citation:

JSTdatasetR4 (Release 4, May 2019)       
                                                                          
 To comply with the attribution requirement in the license, whenever it   
 is used the dataset must be cited as follows: 'Òscar Jordà, Moritz       
 Schularick, and Alan M. Taylor. 2017. Macrofinancial History and the New 
 Business Cycle Facts. In NBER Macroeconomics Annual 2016, volume 31,     
 edited by Martin Eichenbaum and Jonathan A. Parker. Chicago: University  
 of Chicago Press.' Those using any data pertaining to rates of return    
 should also cite: 'Òscar Jordà, Katharina Knoll, Dmitry Kuvshinov,       
 Moritz Schularick and Alan M. Taylor. 2019. The Rate of Return on        
 Everything, 1870–2015. Quarterly Journal of Economics. Forthcoming.'     

User avatar
Tyler9000
Posts: 423
Joined: Fri Aug 21, 2015 11:57 am

Re: Study: The Rate of Return on Everything, 1870-2015

Post by Tyler9000 » Mon Jun 10, 2019 7:12 pm

siamond wrote:
Mon Jun 10, 2019 5:21 pm
That's it! The full data set is finally available in Excel format... Including price, dividends, total return data series for stocks and govt bonds for the various countries. Also bills, short-term and long-term bonds interest rates. And all the real estate data, inflation, GDP and more.

http://www.macrohistory.net/data/

Everything is free to share, redistribute, remix and build upon
This is amazing! It will take a little while to unravel the many layers of sources to evaluate certain calculation methodologies before I can confidently endorse every number, but on the surface this appears to be an absolute treasure trove of free global data. Just in case they're reading this, huge kudos to the authors for doing such great work and sharing it with the world.

User avatar
LadyGeek
Site Admin
Posts: 54308
Joined: Sat Dec 20, 2008 5:34 pm
Location: Philadelphia
Contact:

Re: Study: The Rate of Return on Everything, 1870-2015

Post by LadyGeek » Mon Jun 10, 2019 8:59 pm

I should remind those using the Excel download that this is a perfect application for pivot tables. You can slice and dice the data in any way imaginable, then create pivot table charts to visualize the analysis.

Pivot tables are also working in LibreOffice Calc. (Just testing the format to see if it works - yes. I'll use MS Excel to actually look at the data.)
Wiki To some, the glass is half full. To others, the glass is half empty. To an engineer, it's twice the size it needs to be.

User avatar
siamond
Posts: 4773
Joined: Mon May 28, 2012 5:50 am

Re: Study: The Rate of Return on Everything, 1870-2015

Post by siamond » Wed Jun 12, 2019 10:54 am

Yup, using a pivot table to unpack the data is an excellent suggestion, Ladygeek.

I started to look at the Total-Return data (numbers as well as data sources). US first.

Image

Looking at the data sources, they used Shiller to compute the US equities total return (good choice for early years, unfortunate choice for more recent years imho, given Prof. Shiller's unfortunate habit of averaging the price over January). But I am afraid their TR annual math isn't quite right. Which is easy to do, as Shiller makes several undocumented assumptions which took me a while to unravel a few years ago. This being said, messing up such math is usually shifting one month of data onto the next or previous year, it makes year-to-year comparisons quite off while the CAGR over decades stays reasonably ok. So I'd give this data a half hearted 'pass'.

About bonds, they made some weird choices, splicing a 10yrs series with an LT series. I checked their LT series against the Barclays index I typically use, and although this isn't a perfect match, it isn't far off. Also note that the Damodaran math is about individual bonds while the Barclays index is akin to what a bond fund would do, which isn't the same thing.

Overall, I am afraid the various researchers in the various countries used their own assumptions about bond durations AND about the way to derive TR numbers from interest rates. I think we'd be better off ignoring the bond TR series, and redo the math in a more consistent manner based on the interest rates data series provided by the data set, e.g. using longinvest's bond simulator or Tyler's simpler simulator. Then splice with known indices for more recent years.

About equities, I also did a quick comparison between the Total-Return dataset for Spain, Germany, Italy, France and the corresponding MSCI data (GR series, expressed in local currencies) for 1970+. This is a mixed bag. Germany and Italy have a pretty good match, France less so, and Spain is completely off. I didn't try to figure out why.

Overall, the general sense I have about equities is that we should rely on more established sources for recent data (e.g. MSCI for 1970+) while the MacroHistory dataset probably represents the best we have for earlier years, and even if a good level of skepticism might be warranted, well, it is much better than nothing, and I am personally very grateful that this group of researchers went through the extraordinary effort of gathering such data and making it truly public. This is a real breakthrough moment imho.

User avatar
LadyGeek
Site Admin
Posts: 54308
Joined: Sat Dec 20, 2008 5:34 pm
Location: Philadelphia
Contact:

Re: Study: The Rate of Return on Everything, 1870-2015

Post by LadyGeek » Wed Jun 12, 2019 3:10 pm

siamond wrote:
Wed Jun 12, 2019 10:54 am
About bonds, they made some weird choices, splicing a 10yrs series with an LT series. I checked their LT series against the Barclays index I typically use, and although this isn't a perfect match, it isn't far off. Also note that the Damodaran math is about individual bonds while the Barclays index is akin to what a bond fund would do, which isn't the same thing.
The wiki has some background info: Aswath Damodaran - take a look at his website.
Wiki To some, the glass is half full. To others, the glass is half empty. To an engineer, it's twice the size it needs to be.

User avatar
siamond
Posts: 4773
Joined: Mon May 28, 2012 5:50 am

Re: Study: The Rate of Return on Everything, 1870-2015

Post by siamond » Fri Jun 14, 2019 5:06 pm

After swapping a couple of e-mails with one of the researchers involved in this project, I decided to open a separate thread to discuss the data sources (to ask clarifying questions and provide feedback/suggestions) and try to foster interactions with the authors.

The intent (data sources and resulting numbers) is different enough from the OP's thread, which is more centered on discussing the resulting findings of the study, that I think it deserves its own thread.

User avatar
siamond
Posts: 4773
Joined: Mon May 28, 2012 5:50 am

Re: Study: The Rate of Return on Everything, 1870-2015

Post by siamond » Fri Jun 14, 2019 7:14 pm

I assembled a custom version of the Simba backtesting spreadsheet, where I included the equity and bond total returns from the MacroHistory database, for the 16 countries which have such data series. You can download the custom spreadsheet here. It works best with Microsoft Excel, but it should work reasonably ok with LibreOffice.

The returns are all real returns, expressed in local currency and adjusted for local inflation. Therefore they are comparable, but cannot be mixed together, the only portfolios making sense would mix equities and bonds from the same country. No expense ratio were applied, so those are essentially index returns, not funds returns. Note that the MacroHistory dataset stops in 2015, and I didn't try to splice the dataset with anything else.

In the Analyze_Portfolio and Compare_Portfolios tabs, you can customize the first line and use drop-down menus to select the country of interest. In the cell below (line 2), you can select the % of bonds in the portfolio.

In the Lazy_Portfolios tab, I included all 16 countries and cell D2 allows to select the % of bonds. I also duplicated and slightly adjusted a couple of charts, notably the one comparing risk metrics, for a more friendly display.

Everything else works as usual, so those of you familiar with the Simba spreadsheet should easily find their way to focus on one aspect or another. Customizing the period of time being studied (lines 85 and 86) should notably be of interest.

User avatar
siamond
Posts: 4773
Joined: Mon May 28, 2012 5:50 am

Re: Study: The Rate of Return on Everything, 1870-2015

Post by siamond » Fri Jun 14, 2019 7:56 pm

For those of you not willing to play with a spreadsheet yourself, I assembled a few graphs, focusing on a 60/40 (all domestic) asset allocation. I focused on the 1925+ time period, because before that we have various data holes and some countries had a few years of hyper-inflation which made the numbers really weird. Also, even in the US, it is pretty hard to trust the numbers before the Cowles commission was created, etc.

First off, the annualized returns (real CAGR). Instead of simply looking at the CAGR over this long time period, I looked at the median value of 30-years cycles. Again, this is a median, by no mean the worse or best cases. Point being to get a rough idea about average returns during a typical accumulation or a retirement period. Most countries are between 4% and 5% (again, inflation-adjusted). But some European countries fared incredibly poorly, essentially due to WW-II and to the oil crisis (plus some local considerations). Surprisingly, Japan did ok, but this is really due to averaging, Japan's 30-years cycles are a mix of stellar performance and total disasters.

Image

Now let's look at maximum drawdowns. We have some incredibly bad situations. Note that bonds did NOT save the day, they actually largely contributed to the issues by totally faltering against inflation at times. This makes the US drawdowns look pretty mild in comparison.

Image

And finally, let's look at Safe Withdrawal Rates (defined as 95% success -no bankruptcy- when withdrawing X% of the initial portfolio every year, inflation-adjusted). If anybody still believes that a 4% SWR is some sort of constant of the universe, I don't know, think again...

Image

Bottomline:
1) when (censored) hits the fan, it can REALLY make a giant mess.
2) any perception we may have developed by studying the US history really needs to be put in perspective, there is an undeniable 'success bias' at work here.
3) I don't know about you, but personally, I would NEVER put all my eggs in one basket, however nice it looked in the past.

PS. Later on, I'll post the same graphs for 1950+, hence excluding WW-II. To a large extent, one could say that we just cannot plan for such cataclysmic events, all we can plan for is a somewhat steady time (while praying for it to last).

PS2. Yes, this dataset is more pessimistic than the DMS dataset (notably for France and a couple of other countries). See explanations in their write-up.

AlohaJoe
Posts: 4511
Joined: Mon Nov 26, 2007 2:00 pm
Location: Saigon, Vietnam

Re: Study: The Rate of Return on Everything, 1870-2015

Post by AlohaJoe » Fri Jun 14, 2019 8:16 pm

siamond wrote:
Fri Jun 14, 2019 7:56 pm
3) I don't know about you, but personally, I would NEVER put all my eggs in one basket, however nice it looked in the past.
I dunno...Sweden is looking might fine! :D

User avatar
aj76er
Posts: 676
Joined: Tue Dec 01, 2015 11:34 pm
Location: Portland, OR

Re: Study: The Rate of Return on Everything, 1870-2015

Post by aj76er » Sun Jun 16, 2019 1:55 pm

siamond wrote:
Fri Jun 14, 2019 7:56 pm
For those of you not willing to play with a spreadsheet yourself, I assembled a few graphs, focusing on a 60/40 (all domestic) asset allocation. I focused on the 1925+ time period, because before that we have various data holes and some countries had a few years of hyper-inflation which made the numbers really weird. Also, even in the US, it is pretty hard to trust the numbers before the Cowles commission was created, etc.

First off, the annualized returns (real CAGR). Instead of simply looking at the CAGR over this long time period, I looked at the median value of 30-years cycles. Again, this is a median, by no mean the worse or best cases. Point being to get a rough idea about average returns during a typical accumulation or a retirement period. Most countries are between 4% and 5% (again, inflation-adjusted). But some European countries fared incredibly poorly, essentially due to WW-II and to the oil crisis (plus some local considerations). Surprisingly, Japan did ok, but this is really due to averaging, Japan's 30-years cycles are a mix of stellar performance and total disasters.

Image

Now let's look at maximum drawdowns. We have some incredibly bad situations. Note that bonds did NOT save the day, they actually largely contributed to the issues by totally faltering against inflation at times. This makes the US drawdowns look pretty mild in comparison.

Image

And finally, let's look at Safe Withdrawal Rates (defined as 95% success -no bankruptcy- when withdrawing X% of the initial portfolio every year, inflation-adjusted). If anybody still believes that a 4% SWR is some sort of constant of the universe, I don't know, think again...

Image

Bottomline:
1) when (censored) hits the fan, it can REALLY make a giant mess.
2) any perception we may have developed by studying the US history really needs to be put in perspective, there is an undeniable 'success bias' at work here.
3) I don't know about you, but personally, I would NEVER put all my eggs in one basket, however nice it looked in the past.

PS. Later on, I'll post the same graphs for 1950+, hence excluding WW-II. To a large extent, one could say that we just cannot plan for such cataclysmic events, all we can plan for is a somewhat steady time (while praying for it to last).

PS2. Yes, this dataset is more pessimistic than the DMS dataset (notably for France and a couple of other countries). See explanations in their write-up.
Nice analysis. Thanks for the write-up!

It would be interesting to see how post 1970 data compares to your findings here: https://finpage.blog/2017/03/25/investi ... rld-part-2

I wonder if going back as far as 1950+ could be a bit misleading as many European countries were rebuilding from the aftermath of WWII between 1950 and 1970. I haven't studied the data, but I'm thinking that a lot of the massive drawdown numbers for 1929+ data are due to WWI and WWII ?
"Buy-and-hold, long-term, all-market-index strategies, implemented at rock-bottom cost, are the surest of all routes to the accumulation of wealth" - John C. Bogle

User avatar
siamond
Posts: 4773
Joined: Mon May 28, 2012 5:50 am

Re: Study: The Rate of Return on Everything, 1870-2015

Post by siamond » Mon Jun 17, 2019 3:17 pm

aj76er wrote:
Sun Jun 16, 2019 1:55 pm
It would be interesting to see how post 1970 data compares to your findings here: https://finpage.blog/2017/03/25/investi ... rld-part-2
Yes, I have such goal in mind, of course. Trouble is we don't have a proper total-world return series, which would be needed to confirm (or disprove!) the main conclusion of the blog article. The authors of the 'Rate of Everything' study assembled a world series based on GDP weights, but this doesn't seem quite right as GDP weights and market weights are (and have been) squarely different. So I am a little hesitant. Also I'd like to interact with the authors a little more before getting in a lot of graph building and analysis, as I have some doubts about some of their data series...
aj76er wrote:
Sun Jun 16, 2019 1:55 pm
I wonder if going back as far as 1950+ could be a bit misleading as many European countries were rebuilding from the aftermath of WWII between 1950 and 1970. I haven't studied the data, but I'm thinking that a lot of the massive drawdown numbers for 1929+ data are due to WWI and WWII ?
Well, we can always find a good reason for which a given time period isn't quite right... The graphs I assembled were based on aggregating results for all cycles of 30 years included in the time period, which should definitely reduce the impact of the starting point.

User avatar
siamond
Posts: 4773
Joined: Mon May 28, 2012 5:50 am

Re: Study: The Rate of Return on Everything, 1870-2015

Post by siamond » Mon Jun 17, 2019 7:07 pm

Here are the same graphs, restricted to the 1950-2015 time period, hence a time of relative peace and stability. 60/40 asset allocation, all domestic.

First off, the annualized returns (real CAGR). Instead of simply looking at the CAGR over this long time period, I looked at the median value of 30-years cycles. Again, this is a median, by no mean the worse or best cases. Some countries kept displaying dismal numbers. Note that Portugal suffered from a double-whammy in the 70s, not only the oil crisis, but also the repercussions of the Carnation Revolution.

Image

Then the drawdowns. Even without WW-II, the picture remains quite dramatic for some countries.

Image

And finally the Safe Withdrawal Rate (95% success).

Image

User avatar
siamond
Posts: 4773
Joined: Mon May 28, 2012 5:50 am

Re: Study: The Rate of Return on Everything, 1870-2015

Post by siamond » Mon Jun 17, 2019 7:19 pm

Now, it seems a tad surprising to see Japan leading the pack for the SWR graph in the previous post. But we're speaking of 30-years cycles, and a data set that stops in 2015. Which means we don't have a cycle that starts right before the asset bubble burst in Japan.

To put things in perspective, I compared all max withdrawal rates for cycles of 25 years, and I also included WW-II and the Great Depression for good measure - see below. Now we can see the really bad times for Japan, although somehow the asset bubble wasn't that much worse than starting in the mid-60s in the US, at least for a 60/40 domestic allocation.

Image

alex123711
Posts: 87
Joined: Sun May 20, 2018 5:01 am

Re: Study: The Rate of Return on Everything, 1870-2015

Post by alex123711 » Tue Jun 18, 2019 7:07 pm

siamond wrote:
Fri Jun 14, 2019 7:56 pm

2) any perception we may have developed by studying the US history really needs to be put in perspective, there is an undeniable 'success bias' at work here.

3) I don't know about you, but personally, I would NEVER put all my eggs in one basket, however nice it looked in the past.
. [/i]
This is what I have been questioning for a while..

Valuethinker
Posts: 38113
Joined: Fri May 11, 2007 11:07 am

Re: Study: The Rate of Return on Everything, 1870-2015

Post by Valuethinker » Wed Jun 19, 2019 12:10 pm

aj76er wrote:
Sun Jun 16, 2019 1:55 pm


It would be interesting to see how post 1970 data compares to your findings here: https://finpage.blog/2017/03/25/investi ... rld-part-2

I wonder if going back as far as 1950+ could be a bit misleading as many European countries were rebuilding from the aftermath of WWII between 1950 and 1970. I haven't studied the data, but I'm thinking that a lot of the massive drawdown numbers for 1929+ data are due to WWI and WWII ?
Yes. But European GDP was back to 1939 levels by, from memory, 1954.

The French call it something like "Les trentes glorieuses" - the 30 glorious years up to 1972. Germans call it the wirtschaftwunder (work wonder years, I think), Italians "Il Surpasso" ("the Surpassing" when their GDP exceeded that of the UK, even not counting the black economy).

(all renditions of foreign expressions approximate)

Once Germany had a hard currency again (Adenauer - 1948?) and the Marshall Plan kicked in the German economy exploded into recovery (General Marshall was almost hounded from office; he shall go down in history as one of the greatest Americans of the 20th century, the architect of American victory first in war and then in peace -- if FDR was the greatest American of the 20th century (Dr. Jonas Salk/ or Sabin, might be my second choice) then Marshall has a good claim to be the second one).

European stock markets (not UK) well outperformed US ones in the 1970s, also - I think that was largely a function of currency (American inflation was much higher than German or Swiss during that time period).

kfitz1313
Posts: 63
Joined: Sun May 19, 2019 3:20 am

Re: Study: The Rate of Return on Everything, 1870-2015

Post by kfitz1313 » Sat Jun 22, 2019 12:31 pm

siamond wrote:
Fri Jun 14, 2019 7:56 pm
Now let's look at maximum drawdowns. We have some incredibly bad situations. Note that bonds did NOT save the day, they actually largely contributed to the issues by totally faltering against inflation at times. This makes the US drawdowns look pretty mild in comparison.

Bottomline:
1) when (censored) hits the fan, it can REALLY make a giant mess.
2) any perception we may have developed by studying the US history really needs to be put in perspective, there is an undeniable 'success bias' at work here.
3) I don't know about you, but personally, I would NEVER put all my eggs in one basket, however nice it looked in the past.
Any practical takeaway advice from your above statements, especially with regard to the first about bonds not helping? Thanks a lot for the fascinating information and analysis.

Are you able to find the max drawdown for a global basket of equities instead of country by country? I've been trying to figure out what the worst drawdown on something like a total world equity index would have been.

Hulu
Posts: 177
Joined: Wed Oct 22, 2014 7:55 pm

Re: Study: The Rate of Return on Everything, 1870-2015

Post by Hulu » Sat Jun 22, 2019 1:45 pm

I think people miss the time value and leverage explosiveness of real estate. A 22 year old in some place like Columbus could conceivably buy a $400k four-family for $15k. And that property could easily return an average of 20% between appreciation, income and reinvested profits. And she could rinse and repeat as much as she likes. And then turn to stocks once family or career ramp up. But probably won’t because 20% returns are hard to turn down.

User avatar
siamond
Posts: 4773
Joined: Mon May 28, 2012 5:50 am

Re: Study: The Rate of Return on Everything, 1870-2015

Post by siamond » Sat Jun 22, 2019 2:53 pm

kfitz1313 wrote:
Sat Jun 22, 2019 12:31 pm
Any practical takeaway advice from your above statements, especially with regard to the first about bonds not helping? Thanks a lot for the fascinating information and analysis.
Well, when I said that 'bonds didn't help', I was referring to some of the truly abysmal cases when both stocks AND bonds totally cratered in real terms. In other cases, e.g. Japan and the asset bubble crisis of the 80s, bonds did help a lot, and this explains why the 1950+ SWR stats for Japan are actually similar to the US, although Japan faced one of the worst possible (depth and length) equity crises ever known. Also, it wasn't directly visible in the few graphs I posted, but in some countries, the premium of stocks against bonds was nowhere near as striking as in the US history.

Personally, my practical takeaway is DON'T PUT ALL YOUR EGGS IN ONE BASKET, MAXIMALLY DIVERSIFY AND BE ADAPTIVE. Don't believe that bonds are 'safe', sometimes they were, sometimes they were the biggest risk. Don't believe that equities will always provide a great premium, they often did, but sometimes the premium was rather underwhelming. Don't believe that domestic is better than international when one country has a great track record, things change and (self-censored) can hit the fan real bad. Don't use a flat inflation-adjusted spending budget for your retirement, use an adaptive variable method. Etc.
kfitz1313 wrote:
Sat Jun 22, 2019 12:31 pm
Are you able to find the max drawdown for a global basket of equities instead of country by country? I've been trying to figure out what the worst drawdown on something like a total world equity index would have been.
I am stuck on the 'global basket' theme, which I'd love to investigate further (as I did for 1970+ in the blog post I mentioned earlier). In the Rate of Return on Everything study, they assembled a global basket data series, but they simply weighted the countries real returns by the countries' GDP. This doesn't make much sense to me, this is mixing currencies, this is mixing inflation rates, and GDP weighting is totally different from market cap weighting, which is how real-life world funds operate. Case in point, the US GDP is currently roughly one quarter of the world, while its market cap is roughly half of the world. The currency/inflation issue isn't hard to solve, we have the data to do it right. But... I don't have historical market weights, so I am stuck. Ideas welcome...

EDIT: oh wait. The Triumphs of the Optimists folks did assemble a historical chart of market weights. Maybe we could reverse-engineer this chart, just to get rough orders of magnitude.

Post Reply