I did not want to participate in this debate because I don't think it is a binary problem. But after reading a recent paper by AQR, I thought I would do an experiment using Perplexity AI.
I am attaching three documents.
Two are from AQR: 1. Speaks to how investors determine long term expected returns and 2. US vs Non-US equities. There will be more papers in this series.
The third document is from AI. I used Perplexity and uploaded the AQR paper on US vs Non-Us equites and ask it to find research that supported the AQR conclusions and the second question I asked it to find research that opposed the AQR conclusions.
My conclusion after reading the AI document is that an investor can find confirming evidence for whatever argument they want to make about the debate. However, I am in the diversify camp that AQR and many others that I respect support. I continue to hold international equities not because I think they will outperform the US but because I don't know which one will outperform. It is simply a risk mitigation approach and then I adjust my ROR assumptions such that they are conservative and allow for underperformance relative to history.
@pizzaman You did not rub me the wrong way but it seemed like this thread was just going back and forth on the subject and that is what I meant by binary. I knew that whatever opinion I had would not provide any additional information for or against the debate and that is why I have not participated in the thread up until now. Not sure what I posted warrants a time out though - my intentions were to provide information that I have not seen on this thread up to this point.
I was not trying to pick winners and losers or try to determine who is right or wrong. I just thought I would share the experiment that I conducted after reading the recent AQR article. I did not want to suffer from confirmation bias and it seemed like most of the podcasts and papers were advocating diversification into non-us equites. I have been experimenting with Perplexity for other research topics and thought I would try to better understand the other side by asking it to research studies that did not support AGR's and my views. I found it to be an interesting experiment and will probably use the same approach for other topics that don't seem to have a clear answer. I posted it on this thread thinking the additional information might be useful to someone that is still trying to make a decision on this topic. I also provided my view since additional research did open my eyes to other equally rational view but not enough to deviate from "my plan". I agree with you that we all have different perspectives and preferences and I respect yours so I ask that you also respect mine.
@golich428 My time out was not aimed at you, but the general direction the thread started going over the last few weeks.
I thought the Perplexity result was very interesting. I have not heard of it. Can you explain a little more about how it works?
@pizzaman Perplexity is an AI tool that I use that is somewhat unique because it uses different models that are out there and you as a user can ask it to use a specific model or all models. I was a Pro member for awhile and that gave me access to deep research (this is what they call it). I was pretty impressed with the number of sources that they it used and it did not give an answer immediately like Chat GPT, it would show you what it was doing (I could not keep up with the AI thought process) and then provided the answer to the question in a summary level but included all the references. I just stumbled onto it and found it interesting. I have used it to give me research on providers of MYGAs and other types of annuities to determine there corporate structure. I am concerned about private equity companies buying insurance providers and I can ask Perplexity to tell me what the corporate structure is and if it is private, mutual or public. I do follow up with it's results but so far it has bee pretty reliable.
Hope your retirement goes well even if we do not agree on the details. Perspective, Preferences, Personal, Preconditions, and Purpose are all attributes that make our plans unique! We should never try and allow our five P's dictate what others should do, we should only share our views in a respectful manner. However, dialog and sharing information is a good way to share our unique stories and maybe help others.
I have been thinking about starting a thread on the RISA approach by Wade Phau on how to determine your retirement style. I participated in the research and then used the tool they provide to inform my Retirement Style. It validated what I already was thinking but that is ok. I think we all have different styles and should recognize that others do not see it the same way. I am a safety first style and that is probably why I want to diversify into international and convert my pensions into monthly annuities rather than lump sums as some one with a total return style might do. I also don't like taking money out of my portfolio so getting a monthly check is a behavioral win.
Thanks to anyone who reads this and best of luck in your retirement.
@golich428 Thanks for your response and understanding.
I was in the very first group that got to use the RISA Retirement Style by Wade Phau a few years back. Didn't have to pay for it as we were the first test subjects. It was very interesting. Could be a good new thread.
Interesting take on when foreign markets start the year very good:
https://finance.yahoo.com/news/happens-u-stocks-lag-global-120013256.html
I am attaching three documents.
Two are from AQR: 1. Speaks to how investors determine long term expected returns and 2. US vs Non-US equities. There will be more papers in this series.
I think you only attached the third?
@golich428 @jkandell I read the first report on US vs Non-US-Equities. Can you explain what this statment means:
Using AQR’s capital market assumptions as of December 2024, we calculate a required growth edge of 2.2% p.a. for US equities to earn the same return as Non-US equities hedged to USD.
I read the footnote but still don't get it.
@pizzaman I will try to explain what they mean. AQR uses a building block approach for estimating forward looking returns using dividend yield, earnings growth rate and valuations. They estimate that US will return 4.2% and non US will be 6.1%. Based on their estimate at the end of 2024 keeping the dividend yields and valuations constant, the earnings growth rate for US relative to non US would need to be 2.2% per annum. They assumed it to be 0.3%. They do state that 1% might be reasonable but not 2.2%. The also point out that if relative valuations narrow towards more historical values that would also push non US returns higher relative to US. This also assumes no change in currencies relative to each other.
I still don't get it. How does the 2.2% per annum relate to the 4.2% and 6.1% returns?
@pizzaman For the US return to equal 6.1%, the earnings growth advantage would need to be 2.2% not the 0.3% they assumed in their calculation. There are nuances to the building block approach but you might get a clearer picture by reading part 1 of the series or some other paper that use a similar method. What we need to keep in mind is that all the CMA approaches have a range of probable outcomes and we should not focus on the point estimate. Not sure how I got this font?
Still not sure what to do with those numbers in terms of asset allocation of US vs developed foreign markets. No matter, the conclusion of the article isn't anything new, the US stock market is way over valued. Everybody, even me, acknowledges that. Well, so what. A correction is well over due says history. Of course the pundits have been saying that for over a decade. Will it happen sooner or later, you bet. But does that mean you should significantly change your asset allocation now? Not sure. How long will the US correction last? How big will it be? Don't know. What we do know is that the US market is still the largest by far. A million car freight train isn't just going to stop. Plus the US has what no other country or region has, the magnificent 7. High level technology is here to stay, especially AI. Nobody is even close. So my guess is that after the US correction, the US stock market will come back bigger and faster than anyone else, so if you are looking at the long term, stay the course. Even TSMC has unveiled a $100 billion investment, bringing their chip manufacturing to the US, drawing global attention and prompting concern in Taiwan. They make the highest tech chips that power AI. Now the supply chain for those chips is almost completely in the US. Good for our future tech dominance to continue.
@pizzaman one question most investors should ask themselves is “What % should I allocate to any asset class - international, REITS, gold, crypto, etc. The possible answer can be anywhere from zero to 100%. Each individual has the option to make their own decision. However, we should recognize that we all have biases that can interfere. That is one reason to get outside perspectives that may provide insights that we might be blind to. I tend to lean away from trying to predict the future and instead lean toward being prepared for the things that are uncertain. That approach is not a return maximization strategy but it may reduce being surprised.
@golich428 Great points!! So lets take it to the next level with a thought experiment. Take 12 asset classes (I think that's how many the online version of PRC can handle), US stocks, foreign stocks, bonds, REITs, gold, crypto, etc. Put each one into a black box. The black boxes are only labeled A,B,C... through L. We do not know which asset class is in each box. What we do know are the attributes on the contents of each box. ROR, average/mean return over all years of return, best year return, worst year return, how risky it is (all metrices you can think of, beta, gamma and any other Greek letter 😋), plus all other attributes you want to help you decide your asset allocation. Would you be willing to use such a system. By you I mean everybody on this thread. That would eliminate biases both personal and financial/blog-is sphere influence's. I think if we put our heads together we could build such a system. What do you think??
