A New and Improved Analysis of the Factor Weighted Three Fund Blend

[Benjamin Buffett]

In response to Morik's excellent points about momentum as a factor, the academic definitions, and the different correlations based on the time windows. I have decided to completely re-work my factor analysis on the non HML and SMB part of the Three Fund Factor Weighted portfolio. This portfolio uses only long tilts, or exposures. I consider any short strategy to be complex and suicidal its limited upside and unlimited potential down side. Short positions will not be considered for this portfolio.

The Factors Used for This Analysis

Market Risk, Value, Size, Carry, and Low Volatility.

For those who have never read about Carry and Low volatility:

Carry: The carry factor is the tendency for higher-yielding assets to provide higher returns than lower-yielding assets. This refers specifically to interest or dividend yields paid to an investor to hold the asset outside of price appreciation or depreciation.

In Swedroe's Book this is extracted via a Carry Trade by using long and short exposures. In his book he cites research that finds a Carry premium across nearly all asset classes, and that research paper finds this premium in equities. It is related to value, but is not always correlated. An example of the difference in carry exposure would be an asset like gold which pays no dividends or buybacks but depends entirely on price appreciation for its returns. However an index fund benefits from company share buybacks and dividends. In the long term the index fund should in perform better because of its carry exposure.

Low Volatility: Assets with a lower volatility can out-perform their more volatility counterparts, especially during certain market conditions.

Breaking Down the Three Fund Blend by Factor Exposures

To save words, here is a table:


VUG: Has greater exposure to market risk, less size and value stocks.
VBR: Has exposure to Size, Value, and some exposure to carry. (Value stocks tend to pay larger dividends on average, and sometimes perform more share buybacks)
USHY or VGLT: The High Yield Bond fund has about half the volatility of equities and provides most of its return in the form of interest payments, and thus, relative to the equities has greater exposure to low volatility and carry factors. Per input from Chocolate Bar, and rhusky, Long Term Treasuries may meet these needs better and with lower volatility in many time periods.

An Analysis of the Three Fund Blend's Components During Different Segments of the Market Cycle



How the Components Would have performed during the The Nikkei Stock Collapse and the Great Depression
The Great Depression were both crashes accompanied by deflation. That deflation provides an advantage to carry. Bonds typically performed well in both markets. The Great Depression would have seen substantially higher bond default rates, but the deflation made the remaining payments more potent. Dividends that were not cut were also more potent due to deflation. Small cap value took approximately five years to recover while the larger market took about 25 years to see their prices recover. In theory, if diversified, in this most adverse investing period, this portfolio would have had some benefits from recovering faster, and paying interest during a time of deflation.

An interesting article discussing the importance of Carry (dividend yields and interest) in the Great Depression's market collapse: https://www.nytimes.com/2009/04/26/your ... 6stra.html

There is more data about high yield bonds, long term treasuries, and small cap value during the Nikkei Stock Collapse, these preformed quite well in recovering faster. The high yield bonds paid interest during a period of deflation, and obviously recovered faster than the greater stock market, one estimate was around 5 years. Small cap value also only took about 5 years to recover.

This would mean two slices of the portfolio, in even recent history's most adverse market conditions, would be, after a short period of time, providing reasonable returns, and could sustain withdrawals without destroying portfolio value.

Implications for A Standard Cap Weighted Portfolio

As much as I hate to admit this, and I do, a standard Cap weighted three fund portfolio, with roughly 40% in bonds, would have done reasonably well, possibly even better during these time periods as well, in relation to having at least one slice that had appreciated in value. So long as your withdrawals only come from funds that are currently at a high valuation, or appreciating in value you are not destroying portfolio value by doing so.

Thanks to: Chocolate Bar, rhusky, Morik, and anyone else that hammered out the biggest flaws in my first factor based analysis.

[brightlightstonight]

High-yield bonds, during crashes, are notoriously volatile and have market risk (increased correlation to equities). E.g VWEHX had a 30% drawdown in 2008, a 26% drawdown from inception to 1990, a 20% drawdown in 2020... plus spikes in volatility.

[breakfastinbed]

USHY or VGLT: The High Yield Bond fund has about half the volatility of equities and provides most of its return in the form of interest payments, and thus, relative to the equities has greater exposure to low volatility and carry factors. Per input from Chocolate Bar, and rhusky, Long Term Treasuries may meet these needs better and with lower volatility in many time periods.

An Analysis of the Three Fund Blend's Components During Different Segments of the Market Cycle



How the Components Would have performed during the The Nikkei Stock Collapse and the Great Depression
The Great Depression were both crashes accompanied by deflation. That deflation provides an advantage to carry. Bonds typically performed well in both markets. The Great Depression would have seen substantially higher bond default rates, but the deflation made the remaining payments more potent. Dividends that were not cut were also more potent due to deflation. Small cap value took approximately five years to recover while the larger market took about 25 years to see their prices recover. In theory, if diversified, in this most adverse investing period, this portfolio would have had some benefits from recovering faster, and paying interest during a time of deflation.

I'm not sure this follows. USHY and other high-yield corporate bond funds are highly correlated with the total stock market in recent times (0.83). Backtesting 60/40 TSM + either VWEHX, TLT, or VBMFX favors TLT in terms of return, max drawdown, volatility, Sharpe, etc. This is partially because high-yield corporate bonds experienced a significantly greater drawdown in 2008 and 2020, similar to stocks (30% and 20% drawdowns, respectively) compared to treasuries (down 3% at the market's nadir in 2008 and again in 2020) and total bond (3.5% and 4.5%, respectively). While both times treasuries took longer to recover (and have yet to recover from 2020) in both instances you could have sold high on treasuries to buy stocks at a discount - since 2008 treasuries' two biggest years, 2008 and 2011, were also the total market's worst and fifth-worst year over that period.

Bonds have many purposes but one of the most useful is for non-correlated rebalancing. If you want them for that use, treasuries are the way to go. If you are looking for performance alone, you might want to look at junk bonds - since inception FALN, despite being more volatile, has outperformed VWEHX. Of course, you might ask yourself if treasuries + more equities would be even more preferable.

I do think an interesting argument could be had between high-yield corporate bonds vs total bond fund, but I think many Bogleheads hold things like treasuries or TIPS as part of their bonds. Also, you could make the argument that high-yield corporate bonds have significant credit risk - VWEHX is nearly 90% BB or below.

Overall it probably depends on what you want from your portfolio. There's no perfect answer.

[secondopinion]

USHY or VGLT: The High Yield Bond fund has about half the volatility of equities and provides most of its return in the form of interest payments, and thus, relative to the equities has greater exposure to low volatility and carry factors. Per input from Chocolate Bar, and rhusky, Long Term Treasuries may meet these needs better and with lower volatility in many time periods.

USHY is not even comparable to VGLT. I mean, that is almost like saying VUG and VBR are the same. To some, it is even worse.

[breakfastinbed]

USHY or VGLT: The High Yield Bond fund has about half the volatility of equities and provides most of its return in the form of interest payments, and thus, relative to the equities has greater exposure to low volatility and carry factors. Per input from Chocolate Bar, and rhusky, Long Term Treasuries may meet these needs better and with lower volatility in many time periods.

USHY is not even comparable to VGLT. I mean, that is almost like saying VUG and VBR are the same. To some, it is even worse.

Interestingly, USHY is more strongly correlated to VUG and to VBR than they are to one another, and is less correlated to VGLT (0.26) than VUG is to VGLT (0.3).

[secondopinion]

USHY is not even comparable to VGLT. I mean, that is almost like saying VUG and VBR are the same. To some, it is even worse.

Interestingly, USHY is more strongly correlated to VUG and to VBR than they are to one another, and is less correlated to VGLT (0.26) than VUG is to VGLT (0.3).

Right; they are far from comparable. If the OP really wants to explore portfolios constructed using opposing corners of a 9-square, they should use two bond funds like they do stocks.

[secondopinion]

How would large cap value, small cap growth, short-term treasuries, and long-term speculative-grade fixed-income relate to the analysis?

[Benjamin Buffett]

High-yield bonds, during crashes, are notoriously volatile and have market risk (increased correlation to equities). E.g VWEHX had a 30% drawdown in 2008, a 26% drawdown from inception to 1990, a 20% drawdown in 2020... plus spikes in volatility.

The reason I have chosen them is that they offer three unique characteristics:
1. They have a relatively low correlation to VUG, lower than VXUS does, and about half the volatility.
2. High Yield bonds provide this advantage while offering a decent return.
3. Their very high yield offers superior exposure to the carry factor premium.

For those investors who value stability over returns and potential yields, VGLT is a better choice. The choice is an individual one based on your need and tolerance for risk. For many investors who have not went through a bear market, I would think VGLT is more suitable. However, my portfolio is calibrated to meet a 7% after inflation risk/return. 7% was chosen because real expected returns are projected to be lower, so we really want at least a 5% real rate of returns, to make retirement attainable without a massive savings and investment rate. This choice was a deliberate one.

[Benjamin Buffett]

USHY or VGLT: The High Yield Bond fund has about half the volatility of equities and provides most of its return in the form of interest payments, and thus, relative to the equities has greater exposure to low volatility and carry factors. Per input from Chocolate Bar, and rhusky, Long Term Treasuries may meet these needs better and with lower volatility in many time periods.

USHY is not even comparable to VGLT. I mean, that is almost like saying VUG and VBR are the same. To some, it is even worse.

In a way it is, it depends on your perspective. I was looking at carry exposure, and risk exposure.

Both are boiling with risk:

VGLT has nearly 0 credit risk, but is very heavily saddled with term risk.

USHY has less term risk, but is very heavily saddled with credit risk.


They are both volatile and risky bond funds, though they are the near complete opposite of one another in how their risk is loaded. If bond funds had a bond style 9 square, they would indeed be on opposite corners. But both are heavily loaded with carry exposure, USHY is slightly superior on that exposure.

If your primary goal is carry and risk/return on the bond side, its really quite irrelevant, if you want that stream of interest payments, and don't much care about short term volatility. I think a lot of investors would be able to stomach VGLT much better. I prefer USHY because it meets the level of risk/return that I want for the bond side. Both bond funds are swimming with risk and volatility, just a different kind of risk.

[Benjamin Buffett]

Correlations



As we can see VUG's lowest correlation is obviously VGLT, its next lowest is JNK, its next lowest is VBR, and International is last on the non-correlation ranking.

Thus we see that VGLT is the best choice, with a lower yield/return, followed by JNK (more data for it than USHY) which has about half the volatility of VXUS and a lower correlation with VUG than VXUS has.

If you don't mind taking a hit on the expected returns, VGLT is superior, which is why I list it as an optional choice. All I really care is that it is a bond which pays a high interest yield, and it is swimming with risk and volatility that powers that yield. I want more risk, so I went USHY. If I wanted less risk I would have went VGLT. Both would serve largely the same purpose for providing risk and volatility.

[Chocolatebar]

For those investors who value stability over returns and potential yields, VGLT is a better choice.

I would highly recommend considering ultra-long treasuries (such as EDV, ZROZ, and GOVZ) for your portfolio. However, I suggest limiting their allocation to no more than 30%. I’m still a bit disappointed not to see any international exposure included

[Benjamin Buffett]

Interestingly, USHY is more strongly correlated to VUG and to VBR than they are to one another, and is less correlated to VGLT (0.26) than VUG is to VGLT (0.3).

Right; they are far from comparable. If the OP really wants to explore portfolios constructed using opposing corners of a 9-square, they should use two bond funds like they do stocks.

secondopinion you are quite right about their opposite natures, here is a simulation, quick and dirty no fine tuning:



The Three Fund Blend with VGLT does better in this time period, in others USHY had performed substantially better. The four fund has greater complexity, but still performs quite well.

[Benjamin Buffett]

How would large cap value, small cap growth, short-term treasuries, and long-term speculative-grade fixed-income relate to the analysis?

I purposely avoid small cap growth due to its higher quantity of lower quality companies, the junk problem hurts it worse than small cap value. Some academics think that the black hole of investing, small cap growth, has such low performance due to an accumulation of those companies. I am unsure about that explanation. You really have to work hard to cut the time windows in such a way that small cap growth does well, most time windows I looked it small cap growth struggled.

William Bernstein thought small cap growth stocks did badly because it contains "lottery stocks" his words, not mine.

In theory it should work out well, but I think in practice, empirically it does not. I was unable to find long term speculative grade fixed income as an option to test it in portfolio visualizer, but its possible I just missed it when scrolling through the options.

[secondopinion]

How would large cap value, small cap growth, short-term treasuries, and long-term speculative-grade fixed-income relate to the analysis?

I purposely avoid small cap growth due to its higher quantity of lower quality companies, the junk problem hurts it worse than small cap value. Some academics think that the black hole of investing, small cap growth, has such low performance due to an accumulation of those companies. I am unsure about that explanation. You really have to work hard to cut the time windows in such a way that small cap growth does well, most time windows I looked it small cap growth struggled.

William Bernstein thought small cap growth stocks did badly because it contains "lottery stocks" his words, not mine.

In theory it should work out well, but I think in practice, empirically it does not. I was unable to find long term speculative grade fixed income as an option to test it in portfolio visualizer, but its possible I just missed it when scrolling through the options.

One theory I have is that tax laws play a big role in the underperformance of junk. Let me explain.

Suppose we have an investment that is a coin flip in nine months. Heads, it doubles; tails, it is zero. My strategy is to sell on tails right after failure (getting short-term capital losses), and sell on heads after 12 months (getting long-term capital gains). If short-term capital gains is 40% and long-term capital gains is 25%, then a portfolio of such investments yields 7.5% in tax arbitrage. As long as enough short-term capital gains can be financially engineered from other sources, we profit from something that has no expected return.

Mutual funds and ETFs cannot take advantage of this; therefore, they underperform.

[breakfastinbed]

Correlations



As we can see VUG's lowest correlation is obviously VGLT, its next lowest is JNK, its next lowest is VBR, and International is last on the non-correlation ranking.

Thus we see that VGLT is the best choice, with a lower yield/return, followed by JNK (more data for it than USHY) which has about half the volatility of VXUS and a lower correlation with VUG than VXUS has.

If you don't mind taking a hit on the expected returns, VGLT is superior, which is why I list it as an optional choice. All I really care is that it is a bond which pays a high interest yield, and it is swimming with risk and volatility that powers that yield. I want more risk, so I went USHY. If I wanted less risk I would have went VGLT. Both would serve largely the same purpose for providing risk and volatility.

The issue is the difference between a correlation of 0.08 and of 0.75 is enormous, which explains your backtest results in a post below. Uncorrelated assets are huge for long-term returns, which is why all those risk parity portfolios with gold do so well in backtests. So even though treasuries have lower returns, their whole point is to support equities.

I don't know how to insert pictures (I'm dumb) but if you run regressions of 60% VTI combined with 40% of either total bond (VBMFX), JNK, intermediate treasuries (VFITX), or of 20% each JNK + VFITX, the highest return is 40% JNK by over 0.5% annually, but it also has the lowest risk-adjusted return. The counter-argument is "if JNK is riskier than treasuries, couldn't we just increase our equity portion with treasuries?" and it turns out that 70/15/15 of VTI/JNK/VFITX or just 75/25 of VTI/VFITX both ourperform it with lower volatility. So, if you're willing to "barbell" your bond exposure with a low-risk treasury + a higher-risk equity, that's the approach that worked best in backtests. If you DON'T want to do that, I do agree an argument could be made for high-yield corporates over investment-grade corporates.

You touch on this in your other post - a barbell of high-yield corporates + low-risk treasuries does outperform total bond. You split the risk.

I’m still a bit disappointed not to see any international exposure included

International actually has little role in an approach like this because international stocks (large cap, and to a lesser extent small cap) are highly correlated with their US equivalents. Their main use in a portfolio is for hypothetical risks - country/currency risk, US equity bubble protection, and withdraw stress test risk (what if something unexpected happens in retirement and you have to significantly increase your withdraw rate? Historically, anyway, including international ourperforms with withdraw rates over 7%). For me, these reasons (as well as valuations) are a sufficient enough argument to include it.

How would large cap value, small cap growth, short-term treasuries, and long-term speculative-grade fixed-income relate to the analysis?

Large cap value has a small advantage compared to large cap growth in terms of volatility, carry, valuations, etc, but at the cost of lower market risk exposure, to a sufficient extent that it theoretically is less helpful.

The argument for small cap growth is that it overweights market exposure compared to other assets, but theoretically small cap growth should have the lowest returns of any asset class, in part due to low profitability. Due to market exposure it's also more correlated with total market. The idea is that small cap growth are highly aggressive, quickly growing small companies with very high P/Es - they raise capital but aren't producing anything yet. This is why they're described as lottery tickets - they have very low current expected return but hypothetically high future return (if they actually succeed). From an academic perspective, historically removing small cap growth from a factor-tilted portfolio increases overall returns.

I actually don't know a long-term speculative-grade corporate fund, but the argument for these are similar to the argument for carrying both large growth (market risk) + small cap growth (size + value risk). Bogleheads carry lots of treasuries and TIPS and other things with interest risk, and avoid credit risk - hence the adage "take your risk on the equity side" - but historical data suggests BB/B-rated intermediate-duration corporate bonds actually have the best risk-adjusted, and even highest total return, of fixed income even after factoring in default rate. In this analogy, very high-grade corporates and treasuries are the total market, and high-yield corporates the small cap value - you "barbell" your risk for better overall return.

There's actually a lot of interesting research out there that credit risk is over-compensated, especially for municipal bonds. It's something I am looking into more myself and don't have a full grasp on yet, but outside of risk-conservative personal investment sites like Bogleheads it's fairly common advice to include speculative bonds in your fixed income.

[Benjamin Buffett]

I purposely avoid small cap growth due to its higher quantity of lower quality companies, the junk problem hurts it worse than small cap value. Some academics think that the black hole of investing, small cap growth, has such low performance due to an accumulation of those companies. I am unsure about that explanation. You really have to work hard to cut the time windows in such a way that small cap growth does well, most time windows I looked it small cap growth struggled.

William Bernstein thought small cap growth stocks did badly because it contains "lottery stocks" his words, not mine.

In theory it should work out well, but I think in practice, empirically it does not. I was unable to find long term speculative grade fixed income as an option to test it in portfolio visualizer, but its possible I just missed it when scrolling through the options.

One theory I have is that tax laws play a big role in the underperformance of junk. Let me explain.

Suppose we have an investment that is a coin flip in nine months. Heads, it doubles; tails, it is zero. My strategy is to sell on tails right after failure (getting short-term capital losses), and sell on heads after 12 months (getting long-term capital gains). If short-term capital gains is 40% and long-term capital gains is 25%, then a portfolio of such investments yields 7.5% in tax arbitrage. As long as enough short-term capital gains can be financially engineered from other sources, we profit from something that has no expected return.

Mutual funds and ETFs cannot take advantage of this; therefore, they underperform.

I think you are right in more way than one:

The turnover generates high taxes, and I think that those tiny little companies just can't take high inflation and tax loads. A large cap or mid cap growth company can survive a little taxes, government regulation, and inflation..... but I think any drag on a tiny junk company will just finish it off, and push it off the edge into bankruptcy.

So on the fund end taxes really hurt small junk, and on the company fundamentals, we see that taxes can really hurt small junk.



Not to bang on the carry factor drum again, but generally growth companies tend to, on average, return value to their shareholders in the form of price appreciation rather than dividend yields and buybacks. A tiny growth junk stock may be depending on its price appreciation to both access capital AND provide returns. This means the minute they hit a speed bump or pot hole (financial difficulty of some kind) that causes their stock to tank, it falls and can't get back up. Because growth is bought at a high PE PB multiple this means the losses are locked in, without dividends to help pull the price back up, and provide non-price related gains.

For example, if I held for example the logging MLP, I don't remember its ticker, but it has a price appreciation of about 0 and its volatile, its tiny, but its dividend yield is around 10%. It is risky, but as long as it pays its divided, I don't care if the market thinks that it is worth 10 cents a share, so long as it pays me the cash flow, because in my mind, on the carry end of the portfolio, I see it as a 10% or more coupon bond with a maturity set at infinity that just went on sale. So investors wanting the cash flow will pile back into it, and kind of pull it back up.

But even high dividend small junk tends to suffer because of taxes and interest rates. If I can get a 7% long term bond that is safe, and not on the edge of bankruptcy, and taxes are high, and maybe I can get that bond with tax favorable treatment, the attractiveness of the 10% MLP starts to fade away. I can get that carry from a safer source. However high interest rates usually happen when the economy is doing well and the Fed decided to raise interest rates a bit. When the market tanks instruments with carry tend to do a little better.

In short I completely agree: small junk companies are absolutely trashed and roasted by high tax rates, high interest rates, and high inflation rates.

[Benjamin Buffett]

For those investors who value stability over returns and potential yields, VGLT is a better choice.

I would highly recommend considering ultra-long treasuries (such as EDV, ZROZ, and GOVZ) for your portfolio. However, I suggest limiting their allocation to no more than 30%. I’m still a bit disappointed not to see any international exposure included

Excellent and wonder suggestion about the extra long duration treasuries. Also I have a heavy home nation bias, which may sound completely insane, but there it is lol. Here is the Four Fund Blend based on the above input. It only loses about 1% in performance, and splits the risk on the bond side, the result is a drawdown that is much lower:



Here is a comparison







As you can see that close to 1% difference in returns does start to really add up over nearly 50 years. But the draw down statics and sortino ratio improvements are undeniable. A 1% sacrifice in total returns for a massively improved portfolio stability isn't a bad trade.

Thank you breakfastinbed, you are quite right, as are you chocolatebar on the extended duration treasuries.

[Benjamin Buffett]

Splitting the risk by factor, risk type, and style and then taking advantage of the return differences does mitigate a lot of the weaknesses of a highly nation concentrated portfolio. I am probably going to alienate a lot of people by saying this: I think international offers similar returns to high yield bonds, with about twice the volatility, and it has roughly the same correlation. So you can kind of replace international in that way. I know there are still some hypothetical events that international would survive better, but with international you have other risks, in addition to 120 years of low performance relative to US stocks:
* Legal risk, your investments could be nationalized, or confiscated during a world war. (We were getting ready to sever Russia from SWIFT, and access for their financial instruments during our spat over Ukraine. In times of conflict the only thing you are guaranteed is that which you can take using your legal system. I can't get the US government to enforce a financial contract with someone in Russia or China.
* Accounting Risk: some nations have fuzzy accounting standards. The CCPs unrealistic GDP figures is an example of fuzzy accounting.

For the 200 years of low performance I cannot say it better than Nispirus in an earlier post:







That being said there are some really good arguments for having international diversification, they just do not resonate with me as well. This is why I began looking into factors, because PWL capital and AQR found they could provide similar levels of diversification as REITs and international, two things I wanted to replace with domestically sourced standard stock bond alternatives.

[Chocolatebar]

I know there are still some hypothetical events that international would survive better, but with international you have other risks, in addition to 120 years of low performance relative to US stocks

The claim of 120 years of underperformance is flawed. There have been numerous decades between 1900 and 2021 where ex-U.S. stocks outperformed the U.S. market. While backtesting prior to the 1960s is challenging, I believe the claim is heavily skewed by the recent period of U.S. outperformance. In reality, international stocks were likely a solid investment from 1900 to 2010, meaning the 14 years of U.S. outperformance are what primarily drive the perception of long-term underperformance.

If U.S. stocks were to collapse tomorrow, it wouldn’t be accurate to claim that the U.S. has underperformed for the past 124 years.

[breakfastinbed]

Splitting the risk by factor, risk type, and style and then taking advantage of the return differences does mitigate a lot of the weaknesses of a highly nation concentrated portfolio. I am probably going to alienate a lot of people by saying this: I think international offers similar returns to high yield bonds, with about twice the volatility, and it has roughly the same correlation. So you can kind of replace international in that way. I know there are still some hypothetical events that international would survive better, but with international you have other risks, in addition to 120 years of low performance relative to US stocks:
* Legal risk, your investments could be nationalized, or confiscated during a world war. (We were getting ready to sever Russia from SWIFT, and access for their financial instruments during our spat over Ukraine. In times of conflict the only thing you are guaranteed is that which you can take using your legal system. I can't get the US government to enforce a financial contract with someone in Russia or China.
* Accounting Risk: some nations have fuzzy accounting standards. The CCPs unrealistic GDP figures is an example of fuzzy accounting.

For the 200 years of low performance I cannot say it better than Nispirus in an earlier post:

This is an inaccurate view for a few reasons. International has performed US a number of times (over the past 55 years it's outperformed 25 times annually) and running a simple backtest and judging the best allocation based on CAGR is a poor way to analyze a portfolio. There are many reasons international stocks are recommended - you mention risks of holding international above, but:
-Inflation/currency risk for US dollar
-US stagnation/valuation risk (which is currently being predicted by pretty much every major financial institution, for what that's worth)
-Withdraw insurance - as discussed the "Beyond the Status Quo: A Critical Assessment of Lifecycle Investment Advice" thread, and elsewhere on this site, the major flaw of Boglehead retirement planning is that you pick a portfolio size that will support a set withdraw rate (X%), and then retire - say, 25x your annual expenses. This assumes a) you can actually reach that goal, i.e. you have sufficient control of both income and expenses, b) you get to choose when you retire (and don't have to retire early or get fired), and c) that your withdraw rate stays approximately the same. I work in a medical field full of bleak outcomes, and will tell you the unexpected costs of healthcare can be enormous - the reason that article recommends a high international allocation is because it survives massive left-tail risk (market crashes, massive withdraw rates, etc). Historically, anyway, international diversification reduces failure rate on non-idealized retirement portfolios, not on perfect, placid backtests.

That being said there are some really good arguments for having international diversification, they just do not resonate with me as well. This is why I began looking into factors, because PWL capital and AQR found they could provide similar levels of diversification as REITs and international, two things I wanted to replace with domestically sourced standard stock bond alternatives.

Be careful not to miss the total for a part - if you're going to say "PWL capital found [factor investing] could provide similar levels of diversification as REITs and international" as an argument against international, you need to answer why PWL overweighs international in their official sample portfolio.


Overall if your question is "how should I view international equities as part of an empirical, factor-focused portfolio?" the answer is as a defensive asset, i.e. something that might help in certain situations where both US large growth and small cap value are faltering. Is it the most "defensive" defensive asset? No, but when you talk about true defensive assets - short/foreign bonds, commodities, global REITs, gold, managed futures, etc - the benefit of international equities is that they are less deleterious to total portfolio return.

As before, it's really up to the individual investor. Personally I hold ~35% international (including international SCV) right now. Might end up reducing my long-term return, but I view it as insurance, and empirically it makes a lot more sense than a massive TIPS ladder or whatever else some people might recommend.

[Chocolatebar]

Overall if your question is "how should I view international equities as part of an empirical, factor-focused portfolio?" the answer is as a defensive asset, i.e. something that might help in certain situations where both US large growth and small cap value are faltering. Is it the most "defensive" defensive asset? No, but when you talk about true defensive assets - short/foreign bonds, commodities, global REITs, gold, managed futures, etc - the benefit of international equities is that they are less deleterious to total portfolio return.

This is absolutely spot on and, in my opinion, should be the primary justification for including international stocks. The significant valuation difference between VXUS and U.S. large caps makes this even more compelling.

[brightlightstonight]

Overall if your question is "how should I view international equities as part of an empirical, factor-focused portfolio?" the answer is as a defensive asset, i.e. something that might help in certain situations where both US large growth and small cap value are faltering. Is it the most "defensive" defensive asset? No, but when you talk about true defensive assets - short/foreign bonds, commodities, global REITs, gold, managed futures, etc - the benefit of international equities is that they are less deleterious to total portfolio return.

This is absolutely spot on and, in my opinion, should be the primary justification for including international stocks. The significant valuation difference between VXUS and U.S. large caps makes this even more compelling.

Yeah, +1000. I think far too often, conversation around international stocks (or anything beyond VTI-and-chill for equities) heads for "why do you think you can beat the market?" when a much more interesting and clarifying question is "what insurance do you want and how much are you willing to pay for it?". VXUS might outperform - I'd love it if it did - but that isn't why I hold it.

[Benjamin Buffett]

I ran multiple backtests and cut the years into different windows and found something interesting: In some time periods VT does pretty well when used instead of VUG. Also in more recent time periods, 10-15 years ago to now high yield bonds do better than VGLT. A few other time periods favor high yield bonds as well. There are limits to backtesting.

I suppose my Three Fund Blend Portfolio has become a Three Fund Decision Tree:

1. Do you want higher risk/return: Choose high yield bonds as the third fund. If you are slightly less aggressive and value lower volatility instead, VGLT is the best choice hands down. Do you want both? Can't Decide? Use both, 25% to each.

2. Do you believe in the arguments for international diversification? If so use VT instead of VUG. This does lower returns by about 200 basis points but that would be the premium you are paying for your insurance policy against severe market risks.

[Benjamin Buffett]

I ran multiple backtests and cut the years into different windows and found something interesting: In some time periods VT does pretty well when used instead of VUG. Also in more recent time periods, 10-15 years ago to now high yield bonds do better than VGLT. A few other time periods favor high yield bonds as well. There are limits to backtesting.

I suppose my Three Fund Blend Portfolio has become a Three Fund Decision Tree:

1. Do you want higher risk/return: Choose high yield bonds as the third fund. If you are slightly less aggressive and value lower volatility instead, VGLT is the best choice hands down. Do you want both? Can't Decide? Use both, 25% to each.

2. Do you believe in the arguments for international diversification? If so use VT instead of VUG. This does lower returns by about 200 basis points but that would be the premium you are paying for your insurance policy against severe market risks.

[Chocolatebar]

There are limits to backtesting.

It's important to remember that a well-constructed portfolio isn't just one that backtests well; it should also be capable of withstanding even the worst realistic scenarios that haven't occurred yet or been captured in historical data.

1. Do you want higher risk/return: Choose high yield bonds as the third fund. If you are slightly less aggressive and value lower volatility instead, VGLT is the best choice hands down. Do you want both? Can't Decide? Use both, 25% to each.

I don't see the value in holding both. It seems like an unnecessary addition that only adds complexity to the portfolio.

2. Do you believe in the arguments for international diversification? If so use VT instead of VUG. This does lower returns by about 200 basis points but that would be the premium you are paying for your insurance policy against severe market risks.

No one can say with certainty that it will lower your returns. Historically, it hasn’t, depending on the time period and how globally diversified you were. Either way, excluding it feels like performance chasing to me.

For what it's worth, my portfolio is:
S&P 500 (VOO)
U.S. Small-Cap Value (AVUV)
Total International (VXUS)
Ultra-long Treasuries (ZROZ)

It took me far too long to settle on something this simple. It sounds like you're considering a similar approach, so my "word to the wise" would be not to overthink it—sometimes simplicity is the best strategy.

[breakfastinbed]

There are limits to backtesting.

It's important to remember that a well-constructed portfolio isn't just one that backtests well; it should also be capable of withstanding even the worst realistic scenarios that haven't occurred yet or been captured in historical data.

Amen. I'm a big fan of risk parity style investing and part of the goal is to consider scenarios we haven't seen in the past. I've mentioned these before, but a few examples:
-High, sustained inflation - your fixed income implodes. Equities initially keep up, but eventually putter out
-Sustained deflation - your portfolio collapses as equities lose value
-Weakening dollar - your US holdings underperform. Interest rates and costs of imported goods rises, while US good plummet in price, along with profits
-Geopolitical turmoil - massive stock volatility as multiple companies across the globe tumble. Inflation rates soar, and fear drives down consumer spending
-Personal disaster - your child is hit by a car and suffers a traumatic brain injury. They survive, but will require 24 hour specialized care for the rest of their life. To afford it, you will need to draw down your portfolio at 8%/year, not the 3.5% you originally planned for.

None of these are reflected in backtests; this is why we hold a variety of uncorrelated assets that will behave differently in different environments. This is why people invest in portfolios like the Golden Butterfly, or the Weird Portfolio, or the ""Beyond the Status Quo" portfolio (>50% international), or what have you, because those portfolios survive stress tests that the standard Boglehead 60/40 portfolio does not.

It's for the above reasons, in part, that I think long-term treasuries - the darling of risk parity investing - are so valuable, compared to corporates: they don't perform like equities do, which is invaluable.


Since we're posting our portfolios, right now I'm doing
-Total stock market (FSKAX/VTI + VIIIX, may switch to VUG and/or AVUS once I start buying this again, not sure)
-AVUV, plus some DFSV for TLH purposes
-Total international (VTSNX)
-ADVD
-A small amount in mini sector strategy (FSELX + FSHOX, rebalanced with each other - an experiment that I may back out of)
-A small amount of speculative holdings

And I anticipate my retirement allocation to be the above, plus the following:
-Short-term municipal bond fund (something like FMNDX)
-Intermediate-term municipal bond fund (something like MUB or AVMU)
-Long-term treasuries
-Utilities fund (something like FUTY - like international, I view utilities as a semi-defensive asset, for different reasons)
-I-bonds (maybe)

Because I read too much risk parity stuff, I'm also considering preferred shares, CEFs, individual REITs (not REIT funds), speculative-grade corporate bonds, and of course gold (?!), but I'm not planning on retiring for 20+ years, so that'll probably change.

Ultimately I believe in factor investing and risk parity, which both say, in different ways, that concentration can out-perform broad diversification. But I also believe in diversification and risk-adjusted return, which is PWL Capital's answer to "why don't you tilt more?" How to blend the two into the perfect portfolio is anyone's guess, but that's why we're all here on this forum.

[Benjamin Buffett]

So far my portfolio is equally weighted between:
VUG (market risk)
VBR (value and small size, and some carry)
USHY (high yield bonds, for carry and low volatility relative to pure stocks)

Original targeted risk and return was that of an 80/20 stock bond (VTI/BND) portfolio with some factor based diversification. That has been accomplished.

My 401 plan is awful, no vanguard funds, few choices, very high expense ratios, my best choices were limited, so that portfolio is:

S&P500 index fund
Russell 2000 Smid Cap Index Fund
Target Date Fund (when I get closer to retirement it will fill with bonds, giving me about 20% in bonds)

Everything else had an expense ratio of about 0.5 or 1%. Horrible plan offerings.

I may be adding in the future:

VGLT or EDV or some other very long term Treasury fund (25% to each part, equally weighted)

VXUS or the dimensional international small value, if it gets a lower expense ratio, I do not know, have not decided which one, or if, I am to add it yet. I have very mixed feelings on international so I might just slide it in at 10% and break my equal weighting criteria. It will make rebalancing more annoying especially in that 401 its fee structure makes doing anything annoying, its user interface and website is awful. I would rather contribute to vanguard brokerage, but my employer uses nationwide to run their plan, could be worse, could be better.

They match 3.5% and I am contributing 15%. This limits what I can contribute to the superior vanguard brokerage

[breakfastinbed]

Running an efficient portfolio across multiple accounts is pretty challenging. For example, I do:

Taxable brokerage: Total US stock market + AVUV
Roth IRA: AVDV
HSA: right now DFSV, will eventually switch to AVDV
403B: some total US stock market, total international

I also have a brokeragelink account through my 403B that lets me pick any mutual funds (not EFTs) offered on Fidelity's website, which interestingly includes a number of Avantis funds, albeit with transaction fees. I haven't really started using this yet.

I wouldn't get too worked up over expense ratios if you think the fund will out-perform. Taxes, however, are worth worrying about. I personally wouldn't hold international SCV in my taxable account, solely because of my marginal tax rate.


Target date fund or just S&P500 seem like reasonable options for your 401K. 401Ks are perfect for international if you have that option.

My one comment would be that high-yield corporate bonds are among the least tax-efficient assets possible. You'll need to do the math to determine if holding that in a taxable account is worth the squeeze.

I plan to hold a lot of municipal bonds, in my taxable account, for this very reason.

[Benjamin Buffett]

Running an efficient portfolio across multiple accounts is pretty challenging. For example, I do:

Taxable brokerage: Total US stock market + AVUV
Roth IRA: AVDV
HSA: right now DFSV, will eventually switch to AVDV
403B: some total US stock market, total international

I also have a brokeragelink account through my 403B that lets me pick any mutual funds (not EFTs) offered on Fidelity's website, which interestingly includes a number of Avantis funds, albeit with transaction fees. I haven't really started using this yet.

I wouldn't get too worked up over expense ratios if you think the fund will out-perform. Taxes, however, are worth worrying about. I personally wouldn't hold international SCV in my taxable account, solely because of my marginal tax rate.


Target date fund or just S&P500 seem like reasonable options for your 401K. 401Ks are perfect for international if you have that option.

My one comment would be that high-yield corporate bonds are among the least tax-efficient assets possible. You'll need to do the math to determine if holding that in a taxable account is worth the squeeze.

I plan to hold a lot of municipal bonds, in my taxable account, for this very reason.

I prefer as low as an ER as I can reasonably get. Recently I found that SPHY is quite a bit better than USHY, as it is more passive, and has better tracking of the high yield aggregate index. Its improvement is not large enough I would consider a large switch in taxable, but in tax advantaged there is little reason not to get the cheaper fund.

Most of my accounts are being held in non-taxable, or in the process of being migrated there.

Here is the table of possible modifications to the Three Fund Blend:

[Benjamin Buffett]

I did some more research and stumbled onto a new risk ratio, and starting playing with it. Well new to me. I also have been looking for global factor funds that are passive and low expense ratio, the offerings are slim for the publicly accessible market it seems. I will continue to research factor weighted asset allocations.




Optimized Portfolio makes an exceptional argument for using the Calmar Ratio instead of the Sortino Ratio:

"…I maintain that a good definition for risk is permanent loss of capital. That risk is obviously most apparent and most impactful during drawdowns, or crashes. As such, it is a superior and much more “pure” measurement of risk than volatility. Consequently, I would submit that Calmar is superior to Sortino, and, as mentioned, Sortino is superior to Sharpe.” –Optimized Portfolio Website

Lets look at the Calmar Ratios for the Three Fund Blend, The Boglehead's Three Fund Portfolio, and the Three Fund Blend with Long Term Treasuries instead of High Yield Bonds.

Calmar Ratios at a glance:

Three Fund Blend: 0.38

Traditional Boglehead's Three Fund Portfolio: 0.28

Three Fund Blend with Long Term Treasuries: 0.13

Looking at Sortino Ratios gives us different results:

Three Fund Blend with Long-Term Treasuries: 1.00

Three Fund Blend: 0.88

Traditional Boglehead's Three Fund Portfolio: 0.75

I am surprised how well it scores on the Calmar ratio, and how poorly it performs on the Sortio Ratio. It is also interesting how well the Traditional Boglehead's Portfolio performs on the Calmar Ratio, edging out the Three Fund Blend with Long Term Treasuries by a significant margin.

I suspect that better Sortino Performance is due to its international exposure, but I could be wrong. It seems that the fast recovery times of Small Cap Value and High Yield Bonds is the reason behind the higher Calmar ratio.

From Rick Ferri's Book:




Factor based diversification does very well during severe financial crashes, including an all Japanese portfolio during the Nikkei Collapse. Small Cap Value recovered more quickly, much more quickly, and other research has shown that high yield bond funds in each nation seem to recover much faster than equities. However, for the International proponents, international would have had similar diversification benefits, loss reduction, and performance in each of these scenarios.

Useful Take Away for Style Based Indexers:

I read one study that expects international correlation to increase, and US performance to fall to about 6%, and international to rise to about 6% over the decades. However, other studies expect that the performance gap, which is significant, will close, but not necessarily the correlation. I suspect that returns in the US and international will eventually approach the same approximate level of return, and as time passes the most logical fund choices might be something like:

The best I could find so far are:

Global Large Cap Growth CGGO but it has a expense ratio of 0.47 which makes the next best choice VT though it is not as pure of a concentration
Global Small Cap Value No Single Fund seems to replicate this adequately, best seems to be a combination of VBR and DISV. If a single fund version of this were ever released that would be ideal, providing it has a low ER.
Global High Yield Bonds the closest I could find so far is PGHY

If a low expense ratio quality fund of each type above is offered, it might easily satisfy both international and factor based investing objectives with only three funds, in a simple and accessible way. I will continue to search until I can find a three fund solution that solves both international and factor based objectives simultaneously, in addition to having low expense ratios.

From CSI PWL video:
https://youtu.be/jKWbW7Wgm0w?t=618

[Benjamin Buffett]

Running an efficient portfolio across multiple accounts is pretty challenging. For example, I do:

Taxable brokerage: Total US stock market + AVUV
Roth IRA: AVDV
HSA: right now DFSV, will eventually switch to AVDV
403B: some total US stock market, total international

I love the efficient segmentation, it seems to be very simple and efficient.

My own portfolio is the same across all accounts:

Vanguard Taxable Three Fund Blend
M1 Roth and Taxable Three Fund Blend
457(b) Three Fund Blend
HSA: cash

My plan thus far has been to gradually move the glut of savings from earlier in my job from taxable to non-taxable in increments. Some of the taxable I spend instead of my paycheck which goes to max out my tax advantaged. It is a two step way to absorb it into a tax protected, and liability protected tax advantaged account.

[JohnW2]

I did some more research and stumbled onto a new risk ratio, and starting playing with it. Well new to me. I also have been looking for global factor funds that are passive and low expense ratio, the offerings are slim for the publicly accessible market it seems. I will continue to research factor weighted asset allocations.




Optimized Portfolio makes an exceptional argument for using the Calmar Ratio instead of the Sortino Ratio:

"…I maintain that a good definition for risk is permanent loss of capital. That risk is obviously most apparent and most impactful during drawdowns, or crashes. As such, it is a superior and much more “pure” measurement of risk than volatility. Consequently, I would submit that Calmar is superior to Sortino, and, as mentioned, Sortino is superior to Sharpe.” –Optimized Portfolio Website

Lets look at the Calmar Ratios for the Three Fund Blend, The Boglehead's Three Fund Portfolio, and the Three Fund Blend with Long Term Treasuries instead of High Yield Bonds.

Calmar Ratios at a glance:

Three Fund Blend: 0.38

Traditional Boglehead's Three Fund Portfolio: 0.28

Three Fund Blend with Long Term Treasuries: 0.13

Looking at Sortino Ratios gives us different results:

Three Fund Blend with Long-Term Treasuries: 1.00

Three Fund Blend: 0.88

Traditional Boglehead's Three Fund Portfolio: 0.75

I am surprised how well it scores on the Calmar ratio, and how poorly it performs on the Sortio Ratio. It is also interesting how well the Traditional Boglehead's Portfolio performs on the Calmar Ratio, edging out the Three Fund Blend with Long Term Treasuries by a significant margin.

I suspect that better Sortino Performance is due to its international exposure, but I could be wrong. It seems that the fast recovery times of Small Cap Value and High Yield Bonds is the reason behind the higher Calmar ratio.

From Rick Ferri's Book:




Factor based diversification does very well during severe financial crashes, including an all Japanese portfolio during the Nikkei Collapse. Small Cap Value recovered more quickly, much more quickly, and other research has shown that high yield bond funds in each nation seem to recover much faster than equities. However, for the International proponents, international would have had similar diversification benefits, loss reduction, and performance in each of these scenarios.

Useful Take Away for Style Based Indexers:

I read one study that expects international correlation to increase, and US performance to fall to about 6%, and international to rise to about 6% over the decades. However, other studies expect that the performance gap, which is significant, will close, but not necessarily the correlation. I suspect that returns in the US and international will eventually approach the same approximate level of return, and as time passes the most logical fund choices might be something like:

The best I could find so far are:

Global Large Cap Growth CGGO but it has a expense ratio of 0.47 which makes the next best choice VT though it is not as pure of a concentration
Global Small Cap Value No Single Fund seems to replicate this adequately, best seems to be a combination of VBR and DISV. If a single fund version of this were ever released that would be ideal, providing it has a low ER.
Global High Yield Bonds the closest I could find so far is PGHY

If a low expense ratio quality fund of each type above is offered, it might easily satisfy both international and factor based investing objectives with only three funds, in a simple and accessible way. I will continue to search until I can find a three fund solution that solves both international and factor based objectives simultaneously, in addition to having low expense ratios.

From CSI PWL video:
https://youtu.be/jKWbW7Wgm0w?t=618

Thanks for the shout-out!

I'd just add, as I noted in that blog post, that we probably shouldn't obsess too much over these ratios. We can easily construct portfolios where Calmar, Sortino, and Sharpe look fantastic but where the expected return is simply too low to even achieve the financial objective. Bill Sharpe himself never intended for his metric to be used how it is used today.

Re: global factor funds, have you looked at Avantis? AVGE, AVGV, and AVMA come to mind.

[Benjamin Buffett]

I did some more research and stumbled onto a new risk ratio, and starting playing with it. Well new to me. I also have been looking for global factor funds that are passive and low expense ratio, the offerings are slim for the publicly accessible market it seems. I will continue to research factor weighted asset allocations.




Optimized Portfolio makes an exceptional argument for using the Calmar Ratio instead of the Sortino Ratio:



Lets look at the Calmar Ratios for the Three Fund Blend, The Boglehead's Three Fund Portfolio, and the Three Fund Blend with Long Term Treasuries instead of High Yield Bonds.



Three Fund Blend: 0.38

Traditional Boglehead's Three Fund Portfolio: 0.28

Three Fund Blend with Long Term Treasuries: 0.13

Looking at Sortino Ratios gives us different results:

Three Fund Blend with Long-Term Treasuries: 1.00

Three Fund Blend: 0.88

Traditional Boglehead's Three Fund Portfolio: 0.75

I am surprised how well it scores on the Calmar ratio, and how poorly it performs on the Sortio Ratio. It is also interesting how well the Traditional Boglehead's Portfolio performs on the Calmar Ratio, edging out the Three Fund Blend with Long Term Treasuries by a significant margin.

I suspect that better Sortino Performance is due to its international exposure, but I could be wrong. It seems that the fast recovery times of Small Cap Value and High Yield Bonds is the reason behind the higher Calmar ratio.

From Rick Ferri's Book:




Factor based diversification does very well during severe financial crashes, including an all Japanese portfolio during the Nikkei Collapse. Small Cap Value recovered more quickly, much more quickly, and other research has shown that high yield bond funds in each nation seem to recover much faster than equities. However, for the International proponents, international would have had similar diversification benefits, loss reduction, and performance in each of these scenarios.

Useful Take Away for Style Based Indexers:

I read one study that expects international correlation to increase, and US performance to fall to about 6%, and international to rise to about 6% over the decades. However, other studies expect that the performance gap, which is significant, will close, but not necessarily the correlation. I suspect that returns in the US and international will eventually approach the same approximate level of return, and as time passes the most logical fund choices might be something like:

The best I could find so far are:

Global Large Cap Growth CGGO but it has a expense ratio of 0.47 which makes the next best choice VT though it is not as pure of a concentration
Global Small Cap Value No Single Fund seems to replicate this adequately, best seems to be a combination of VBR and DISV. If a single fund version of this were ever released that would be ideal, providing it has a low ER.
Global High Yield Bonds the closest I could find so far is PGHY

If a low expense ratio quality fund of each type above is offered, it might easily satisfy both international and factor based investing objectives with only three funds, in a simple and accessible way. I will continue to search until I can find a three fund solution that solves both international and factor based objectives simultaneously, in addition to having low expense ratios.

From CSI PWL video:
https://youtu.be/jKWbW7Wgm0w?t=618

Thanks for the shout-out!

I'd just add, as I noted in that blog post, that we probably shouldn't obsess too much over these ratios. We can easily construct portfolios where Calmar, Sortino, and Sharpe look fantastic but where the expected return is simply too low to even achieve the financial objective. Bill Sharpe himself never intended for his metric to be used how it is used today.

Re: global factor funds, have you looked at Avantis? AVGE, AVGV, and AVMA come to mind.

That is a fair argument regarding Sortino and related ratios, and their intended use. I love your website by the way. I also am one of your Patreon supporters.

20% Liquidity Objective

So what I was Targeting was the equivalent of 80% equity and 20% 5 year treasuries, three funds in equal proportions with an after inflation return that is around 7%. Interestingly enough the behavior of high yield fulfills the same riskiness of only having 20% in 5 year treasuries. However, because only about 12-25 percent of its risk is equity related, it behaves differently month to month. Andrew Smithers said you need at least 20% in stable value to prevent stock erosion during bear markets.

This portfolio meets that objective of holding 20% liquidity.

Barbelled Stock Holding to Reduce Market Risk

The second objective was to barbell the stock holdings to benefit from their different sources of volatility, via rebalancing, or selective selling during market recessions. Small Cap Value in multiple cases has recovered in about 5 years.....while the greater market stayed flat for decades:
* Japan Nikkei stayed flat but their small cap value rebounded rapidly, about 5 years
* FTSE went flat, but their small cap value performed pretty well after about five years
* In the Great depression the equivalent small cap value stocks saw a recovery in about five years.


In short I expect when the S&P500 and VTI are absolutely rising to the sky, SCV like AVUV will under-perform, but when the market crashes, and hard, or when inflation is very very high, small cap value will behave differently.

I have been thinking about the behavior of value stocks for some time, and I have a hunch.
The "premium" seems to have disappeared, but to be honest I have a suspicion, and that suspicion undermines the entire concept of "factor risk premia":

1. Value stocks tend to benefit from low PB ratios, and are priced based on concrete numbers rather than expected future returns. Yield support also is part of what makes them behave differently. I don't think your risk is due to volatility so much as that they underperform for long periods of time, until the greater market takes a dump.

2. Growth stocks, the opposite of value stocks, are bought with high PB ratios, do not benefit from yield support as much, and are assigned valuation based on the market's expected profitability. Whenever structural changes in the economy happen, or whenever these promised advancements in product or profitability do not happen they tank hard.

I think the long term average was because the larger data set contained high inflation and crash recovery periods. When we see a prolonged bull market in the S&P500 without the stagflation of the 70s, we see the "factor premium" disappear. I don't think it disappeared, but instead that under-performance is the premium we pay for that insurance against a prolonged recession of the greater market. The greater market was horrible for retires during the 70s, but during that time SCV did really well, likely because of inflation. Small Cap Value companies were seeing their leverage inflated away very rapidly. When we cut off those prolonged periods of market adversity the value premium appears weak.

I hate to say the above because it weakens my position on Value Factor investing, in as much as that it seems Value stocks do not outperform due to being more risky. I think the research about factor under and over performance holds, but the recent decades of data make me view value differently, not unlike how people here view their international allocation, as insurance against an S&P500 long period of under-performance. I think the Value premium exists, but it only shows itself when the greater market falls and stays down awhile.

Both splitting your equity allocation between factor exposures, or domestic and international, addresses this risk of prolonged local market recessions.

Those that want small cap value and international to out-preform may want to be careful what they are wishing for:
Do you really want to see a time when the greater market falls and goes flat for awhile? Because I think that is what it takes for Small Cap Value or International Stocks to be great again. The US has to take a hit, and prolonged under-performance, for either of these hedges to really shine.

Some Notes About Factor Exposures Across National Borders

There is a well noted value and growth cycle in each economy, they move opposite of one another, generally speaking. However value in one nation might be opposite of value in another nation's economy and business cycle. Value in the US would largely have lined up with Growth in the EU during some time segments. These effects cancel one another out, and as such I question the benefit of capturing value across national economies with the same fund. When I simulate it and look at value and growth movements across borders you lose the counter-cyclical advantage of small value vs large growth in the same economy.

I think there is a way to combine factor exposure and global investing, but I expect it would have be a lot more nuanced than what many products currently do.


All of the above could be wrong, but these are my hunches thus far based on obsessively pouring through research, papers, and different back-tests. I got the idea that I was looking at Value wrong from reading a funny post by William Bernstein who was discussing value, and its riskiness, or lack there of compared to growth companies. So I investigated it, and even though it partially undermines some of my points, I have to lean towards William Bernstein's view of Value as less risky than growth. Note it still does not change my view that value and growth are negatively correlated due to their opposite positioning on the value growth cycle.

http://www.efficientfrontier.com/ef/999/risk.htm

But perhaps the most persuasive argument against the risk premium story for value stocks comes from the current internet stock mania. These [growth] stocks trade at astronomical multiples of book value, and most will disappear without a trace of earnings. Yet the risk premium theory predicts that the low expected returns of these stocks (and boy, have they ever got that one right) is due to their safety. Sure.

In short my creeping suspicion is that small cap value outperforms during certain times, but it has nothing to do with inherent hidden riskiness, but other things that dictate that behavior, rather greater exposure to leverage (inflation helps small cap value), and valuations based on their actual returns rather than possible future growth. No one buys a value company expecting it to invent the next Google Browser or to market the next I Phone.

Anyway if you make Optimized Portfolio, consider putting more stuff on your Patreon, love your website, and excellent points regarding the misuse of risk ratios blindly.

[Benjamin Buffett]

Re: global factor funds, have you looked at Avantis? AVGE, AVGV, and AVMA come to mind.

Excellent Suggestions

AVMA looks like a nice one-fund solution for someone looking for conservative risk/return.

AVGE and AVGV could indeed be a way to capture both the factor exposures and national/international. They do not seem to suffer the same value growth cycle overlap issue I was mentioning above. This is probably because they are actively managed or have some way to rationally avoid that overlap. Avantis never ceases to impress me.