In the Historical Analysis tab, Under the Historic Asset Class, I'm not seeing anything for international equity returns. Does that mean I have to input my own numbers to include an international asset class to perform a Historical Analysis (given I have international stock index in my portfolio)?
@gbjordan222 Yes. You have to map all of your asset classes to some historical asset class. Pralana doesn't contain international equity returns in its database so you will need to provide that yourself.
Stuart
FWIW, I used the returns listed at this site to map International returns: https://novelinvestor.com/historical-returns/
It only starts at 1970, though. For the periods before that, I have it roughly correlated to US stocks, but with an offset for slightly lower returns. If someone has Int'l market returns from 1926-1969 that they'd like to share, I'd be grateful.
Edit: Actually, I forgot that user Chris Zuiker linked to Early Retirement Now's tool that includes monthly returns for World Excluding US markets that go back to the 1800's. That can be found in this post: https://pralanaretirementcalculator.com/community/postid/6700/
I took those numbers, and summed up the annual returns for 1928-2024 in a CSV list that can be imported into Pralana to map to that asset class:
1928,38.91%,1929,-5.15%,1930,-18.84%,1931,-37.02%,1932,20.65%,1933,55.50%,1934,-1.61%,1935,37.64%,1936,28.17%,1937,-42.96%,1938,39.92%,1939,2.67%,1940,-7.41%,1941,-20.90%,1942,11.44%,1943,21.26%,1944,16.06%,1945,29.98%,1946,-23.02%,1947,-2.90%,1948,4.21%,1949,19.19%,1950,21.73%,1951,17.02%,1952,16.74%,1953,-1.27%,1954,44.02%,1955,27.96%,1956,4.61%,1957,-13.67%,1958,34.98%,1959,10.07%,1960,-0.09%,1961,23.65%,1962,-8.57%,1963,19.44%,1964,14.15%,1965,10.16%,1966,-13.24%,1967,19.06%,1968,6.61%,1969,-13.71%,1970,-18.52%,1971,26.95%,1972,31.93%,1973,-17.00%,1974,-30.28%,1975,23.90%,1976,-0.45%,1977,9.88%,1978,21.15%,1979,-1.41%,1980,12.82%,1981,-9.45%,1982,-2.35%,1983,18.84%,1984,0.94%,1985,38.66%,1986,52.42%,1987,20.50%,1988,21.47%,1989,7.59%,1990,-27.69%,1991,10.29%,1992,-14.63%,1993,27.11%,1994,5.58%,1995,9.36%,1996,3.86%,1997,2.04%,1998,17.86%,1999,23.17%,2000,-16.49%,2001,-23.63%,2002,-17.50%,2003,33.08%,2004,16.14%,2005,11.16%,2006,21.38%,2007,8.63%,2008,-51.65%,2009,30.23%,2010,9.85%,2011,-13.72%,2012,15.27%,2013,18.81%,2014,-4.16%,2015,-2.27%,2016,1.94%,2017,20.32%,2018,-15.77%,2019,19.44%,2020,9.84%,2021,5.74%,2022,-18.73%,2023,15.22%,2024,2.76%,
FWIW, I used the returns listed at this site to map International returns: https://novelinvestor.com/historical-returns/
It only starts at 1970, though. For the periods before that, I have it roughly correlated to US stocks, but with an offset for slightly lower returns. If someone has Int'l market returns from 1926-1969 that they'd like to share, I'd be grateful.
Edit: Actually, I forgot that user Chris Zuiker linked to Early Retirement Now's tool that includes monthly returns for World Excluding US markets that go back to the 1800's. That can be found in this post: https://pralanaretirementcalculator.com/community/postid/6700/
I took those numbers, and summed up the annual returns for 1928-2024 in a CSV list that can be imported into Pralana to map to that asset class
Be careful. The ERN data you cited also starts at 1970 for international. Jeske uses the US stock market data for international prior to that.
It only starts at 1970, though. For the periods before that, I have it roughly correlated to US stocks, but with an offset for slightly lower returns. If someone has Int'l market returns from 1926-1969 that they'd like to share, I'd be grateful.
There is excel data from Òscar Jordà & Katharina Knoll & Dmitry Kuvshinov & Moritz Schularick & Alan M Taylor, 2019. "The Rate of Return on Everything, 1870–2015," The Quarterly Journal of Economics, vol. 134(3), pages 1225-1298 that you can download here. It's a humbling article, btw. (Skinny: Historically and globally, stocks have not done better than real estate, both about 7% nominal.) The article has had a lasting influence on me, with a reality check on how unique our own investment lifetimes have been in the big picture.
Jorda et al have nominal figures for equities (as well as bonds and real estate) from 1870 for these countries: Australia, Belgium, Germany, France, Italy. From 1900 on: Australia, Belgium, Switzerland, Germany, Denmark, Spain, Finland, France, UK, Italy, Japan, Netherlands, Norway, Portugal, Sweden, USA.
I translated all their returns into real dollars in my own excel file, but am not sure how to get you the data.
@jkandell I'm not sure if this is the right way to go about it, but...
If I download that dataset,
- pull out the eq_tr column for each country, from 1928-2020
- take the average of the returns for each year across all countries (excluding USA)
is that a legitimate result?
I see that those are nominal values, but Pralana uses nominal returns for the other asset classes in Historical Analysis.
Example, I calculate the 1928 average return to be 17.88%, and 1929 is -7.13%.
My results don't match ERN's values from 1970 onwards, though some years are close.
If that seems legit, I can post my CSV values for these results, like I did above.
@tcbarney I am not sure there is any one legitimate way to use this data. One way would be to average all the countries year by year, as you describe, to make an aggregate. (Do note: It won't be the same as the series used by ERN and others since theirs is based on EAFE, which includes Asia and is cap-weighted.) You could also choose only the largest 7 economies to average. You could also construct GDP-weighted results. (The GDP data is there too.). The authors do all of the above in their own analysis in the article, which is worth reading.
Note too that WW2 had an impact on many of these markets in big ways the USA did not, including liquidation. You might want to consider starting after the war, in order to have the same world economic regime. There are local situations in some countries, including hyperinflation, you might want to exclude, since they bias the average.
A problem with using nominal results is that Pralana will adjust for USA cpi, which I think would give an incorrect result without currency conversion. They do include a currency factor to usd, so you could in theory show how a US investor would have fared investing in that market. You'll see there is an inflation entry for each year for each country also, so you could calculate the real return from the perspective of someone in that country. That's what I elected to do for my own primary analyses, since I was interested in how stocks might appear to a local investor. They do include a currency factor to usd, so you could in theory show how a US investor would have fared investing in that market.
Given all the complications, I ended up using the data not for simulation entries but just to get a lay-of-the-land. I wanted to see just how badly or well things could go beyond our history. So I looked at percentile bands within and across countries, in different time periods, rather than averages, since there are so many outliers. And I was also interested in how equities compared with bonds (the equity premium) across countries and time. Hope that helps.
@jkandell Whew! Yes, that does help, and gives me a lot more to think about. I had considered the inflation question, but honestly, this kind of financial analysis - well, it's not completely over my head, but I'm treading water in the deep end.
For now, ExUS equities are only about 10% of my portfolio, so the impact on my overall picture isn't huge if I'm off by a bit.
I like using historical analysis more than Monte Carlo. The tricky part being, of course, that the world has changed quite a bit and you have to ask - Will markets ever react the same way as they've done in the past, even to similar, global, catastrophic events?
But, excellent food for (more) thought. Appreciate it!
If I focused on historical analysis, I think I would use a spliced index using ERN figures (or from Bogleheads Simba's spreadsheet) for ex-US 1970 on, combined with Pralana's US figures prior to that. This is on the presumption that "historical" analysis should stick to actual history, and US investors couldn't really invest outside the country in indexes prior to 1970. I'd personally use Pralana's data "as is" for intl through 1969 and ERN's converted to nominal using Pralana's inflation to generate entries after 1970.
Of course nothing stops you putting in the Jorda data 1928-2024 too, for fun. If so you might just stick with the largest 7 countries (rather than all 16) and convert to USD. (The conversion factors are one of the data columns.) With your 10% allocation it won't make much difference either way, to be honest; if it was me I'd just skip it entirely since it won't move the dial.
For what it's worth, though my own portfolio is split 50/50 Domestic and International, I still use a single combined estimate for "Stocks" in the monte carlo. My reasoning is simple: any small difference I suss out historically between international and US is drowned by the "error bands" of low confidence. Who knows if US will dominate the next forty years like it has. Having said that, having read Jorda et al and the more recent Pederson et al, my forward prediction for stocks is lower than the US historical average on the presumptions (a) US valuations are high internally and (b) "reversion to the geographic mean" will take away US exceptionalism historically in the 42 years of my plan. I've learned in modeling to always grossly simplify to capture key relationships, and for me the Stock v Bond v Cash ratio is the key, rather than modeling the subdivisions within each bucket, (even though my actual portfolio is sub-divided into subclasses for diversification). I even assume there is no value nor size bonus in modeling--all just "stocks"; but that's another story.
Using US returns for pre-1970 International is a bit strange. Within Pralana's historical time frame, WWII smashed not only Europe but Japan and China to rubble all at the same time. How would International investing look if you included that?
WWII wasn't unique, WWI smashed Europe just a generation before. It's not like such chaos is behind us, just look at what Russia is doing to Ukraine and realize that the rest of Europe is next on their menu if they succeed. China seems to be sharpening axes to come after Taiwan. N. Korea is getting help from the Russians on weapons and tactics so may be a bigger threat to S. Korea than we imagine, etc.
