Indexing of Different Asset Classes
When it comes to stocks, index investing offers many advantages over other investment approaches. However, these advantages don’t always carry over to other asset classes. No investment style should be treated like a religion, indexing included. It pays to think through the reasons for using a given approach to investing.
Stocks
Low-cost broadly-diversified index investing in stocks offers a number of advantages over other investment approaches:
- Lower costs, including MERs, trading costs within funds, and capital gains taxes
- Less work for the investor
- Better diversification leading to lower volatility losses
Choosing actively-managed mutual funds or ETFs definitely has much higher costs. For investors who just pick some actively-managed funds and stick with them, the amount of work required can be low, but more often the investor stays on the lookout for better funds, which can be a lot of work for questionable benefit. Many actively-managed funds offer decent diversification. Ironically, the best diversification comes from closet index funds that charge high fees for doing little.
Investors who pick their own stocks to hold for the long-term, including dividend investors, do well on costs, but typically put in a lot of work and fail to diversify sufficiently. Those who trade stocks actively on their own tend to suffer from trading losses and poor diversification, and they put in a lot of effort for their poor results. Things get worse with options.
Despite the advantages of pure index investing in stocks, I make two exceptions. The first is that I own one ETF of U.S. small value stocks (Vanguard’s VBR) because of the history of small value stocks outperforming market averages. If this works out poorly for me, it will be because of slightly higher costs and slightly poorer diversification.
One might ask why I don’t make exceptions for other factors shown to have produced excess returns in the past. The reason is that I have little confidence that they will outperform in the future by enough to cover the higher costs of investing in them. Popularity tends to drive down future returns. The same may happen to small value stocks, but they seemed to me to offer enough promise to take the chance.
The second exception I make to pure index investing in stocks is that I tilt slowly toward bonds as the CAPE10 of the world’s stocks grows above 25. I think of this as easing up on stocks because they have risen substantially, and I have less need to take as much risk to meet my goals. It also reduces my portfolio’s risk at a time when the odds of a substantial stock market crash are elevated. But the fact that I think of this measure in terms of risk control doesn’t change the fact that I’m engaged in a modest amount of market timing.
At the CAPE10 peak in late 2021, my allocation to bonds was 7 percentage points higher than it would have been if the CAPE10 had been below 25. This might seem like a small change, but the shift of dollars from high-flying stocks to bonds got magnified when combined with my normal portfolio rebalancing.
Another thing I do as the CAPE10 of the world’s stocks exceeds 20 is to lower my future return expectations, but this doesn’t include any additional portfolio adjustments.
Bonds
It is easy to treat all bonds as a single asset class and invest in an index of all available bonds, perhaps limited to a particular country. However, I don’t see bonds this way. I see corporate bonds as a separate asset class from government bonds, because corporate bonds have the possibility of default. I prefer to invest slightly more in stocks than to chase yield in corporate bonds.
I don’t know if experts can see conditions when corporate bonds are a good bet based on their risk and the additional yield they offer. I just know that I can’t do this. I prefer my bonds to be safe and to leave the risk to my stock holdings.
I also see long-term government bonds as a different asset class from short-term government bonds (less than 5 years). Central banks are constantly manipulating the bond market through ramping up or down on their holdings of different durations of bonds. This manipulation makes me uneasy about holding risky long-term bonds.
Another reason I have for avoiding long-term bonds is inflation risk. Investment professionals are often taught that government bonds are risk-free if held to maturity. This is only true in nominal terms. My future financial obligations tend to grow with inflation. Long-term government bonds look very risky to me when I consider the uncertainty of inflation over decades. Inflation protected bonds deal with inflation risk, but this still leaves concerns about bond market manipulation by central banks.
Taking bond market manipulation together with inflation risk, I have no interest in long-term bonds. We even had a time in 2020 when long-term Canadian bonds offered so little yield that their return to maturity was certain to be dismal. Owning long-term bonds at that time was a bad idea, and I don’t like the odds any other time.
Once we eliminate corporate bonds and long-term government bonds, the idea of indexing doesn’t really apply. For a given duration, all government bonds in a particular country tend to all have the same yield. Owning an index of different durations of bonds from 0 to 5 years offers some diversification, but I tend not to think about this much. I buy a short-term bond ETF when it’s convenient, and just store cash in a high-interest savings account when that is convenient.
Overall, I’m not convinced that the solid thinking behind stock indexing carries over well to bond investing. There are those who carve up stocks into sub-classes they like and don’t like, just as I have done with bonds. However, my view of the resulting stock investing strategies, such as owning only some sub-classes or sector rotation, is that they are inferior to broad-based indexing of stocks. I don’t see broad-based indexing of bonds the same way.
Real estate
Owning Real-Estate Investment Trusts (REITs) is certainly less risky than owning a property or two. I’ve chosen to avoid additional real estate investments beyond the house I live in and whatever is held by the companies in my ETFs. So, I can’t say I know much about REITs.
However, I don’t think the advantages of indexing that exist with stocks automatically carry over to real estate without checking the details. Someone would have to examine REITs to see if they can play a positive role in investor portfolios. Do REITs have hidden costs? Are other market participants able to exploit REITs when trading properties? What are the costs of managing passively-owned properties? Perhaps someone has examined these questions in sufficient detail, but I haven’t seen the analysis.
Some of these problems apply to stock indexing as well. For example, there are circumstances where traders can exploit the way that index funds respond to changes in the list of stocks making up indexes. However, major index ETF providers have responded with changes to how they do business that reduce such costs to low levels. I don’t know if REITs are able to do the same.
Commodities, Hedge Funds, Venture Capital, Collectibles, Cryptocurrencies, other Alternatives
Any time I look at index-style investing in other asset classes, I find more questions than answers. What answers I do find often keep me away from these asset classes rather than draw me in. It would take a lot to convince me to make any of these asset classes part of my own portfolio.
Conclusion
I’ll never claim that how I invest is the best way. I seek reasonable long-term results with a reasonable amount of protection against extreme events, such as stocks market crashes, high inflation, or outright corruption. In some cases, I stay away from asset classes for well thought out reasons, and in other cases, I stay away out of my own ignorance. I’m good with that.
I am sceptical about CAPE. The theory seems great but does not seem to have much predictive power. Fixed allocation (70/30) is good enough for me. It does mean that when stocks go up then I buy FI.
ReplyDeleteIt's true that CAPE doesn't have strong predictive power, but it seems to have more predictive power than just about anything else. People who try to use it to guess when the next market crash is coming are hopelessly misguided. Others try to use it to predict average returns for the next decade, but that is quite weak as well.
DeleteBut weak is better than nothing. A high CAPE indicates that the odds of a severe market crash are higher than otherwise. People want this to mean that CAPE can predict market crashes; it can't. Using CAPE is like playing craps with loaded dice. Your odds of winning have changed, but both winning and losing are still possible. I assume that CAPE will decline to 20 by the time I die, so that the returns I expect over the decades will be slightly lower than normal. This causes me to be a little more conservative with how much I spend each month in retirement.
I also use high CAPE values to tip slightly more than usual to bonds. So far, the biggest tipping I've done is 7 percentage points, a smallish shift that is consistent with CAPE being a weak indicator. However, if CAPE were to get into territory never seen before for all the world's stocks, such as 50 or 60, my shift to bonds would be much larger.
Overall, my use of CAPE is fully automated in a way that I hope will behave reasonably if we have extreme events in stock markets of a kind never seen before. If CAPE stays in the range we have seen before, the effects on my portfolio and spending will be modest.
I am not even sure that the dice is loaded. For one thing, CAPE/adjusted cape and the like can only look backwards. The “real” value of the market is based on future profits which, is of course, unknown.
DeleteI prefer to look at the market as a random walk in the short term with an upward longterm trend.
That said, Japanese scenario would be a bit of a bummer if it were to replicate itself in the US. Your method would help to mitigate that.
If you look at a long-term plot of CAPE in a given month vs. the following 10-year average return, there is a negative correlation, but it isn't strong enough to be very predictive. There is just a gentle pressure to revert to the mean. It makes sense that this would be true; investors would be irrational if they were willing to paying ever high prices for a dollar of profit.
DeleteHowever, there are those who like to plot CAPE vs. 10-year average return over shorter periods and claim strong correlation. This is nonsense as I've explained before: https://www.michaeljamesonmoney.com/2021/06/how-to-lie-to-yourself-about-stock.html
As you say “ Now we see how little information there is for the CAPE above 30. But it gets worse. Those five points are from consecutive years during the year 2000 tech boom and bust, so they all overlap by six to nine years. Any strategy we develop based on high CAPE values is just guessing that the tech bust 20 years ago will repeat.”
DeleteA single stock market/economic event from history isn’t enough for any statistical analysis.
CAPE plots do show positive correlation with returns but they would if it was any other data mining method. CAPE is a relatively recent “invention”. What do we have since? Just over 2 decades worth of data? Hasn’t Shiller predicted abysmal returns ever since 2015?
Sorry, meant “negative correlation”.
DeleteI don't know what Shiller has said, but lots of others make predictions in his name. Perhaps he makes predictions too.
DeleteI'm satisfied that a negative correlation exists. However, it has little predictive power over the short term. Even over the long term it shifts probabilities rather than make definite predictions. Starting from a CAPE of 36, are we more likely to see CAPE double to 72 first, or get cut down to 18 first? I'm satisfied that CAPE is more likely to get to 18 first in the coming decades than it is to get to 72 first, but neither is certain.
Mordko
DeleteYes, he also likes making dire predictions.
DeleteLike here. Talking about housing and “all assets”. 2015.
https://web.archive.org/web/20151226020712/http://www.foxbusiness.com/business-leaders/2015/08/25/yales-robert-shiller-stock-market-turmoil-not-over-yet/
And he seems to think than CAPE is “investable”. Barclays uses his “CAPE” based products.
https://blogs.cfainstitute.org/investor/2019/01/07/robert-j-shiller-on-bubbles-reflexivity-and-narrative-economics/
I didn’t look in detail but seems the opposite of the momentum based “hot potato” strategy.
To answer your question re CAPE of 36… I have no idea. I am not even sure that the 1990s tech bubble was in fact a bubble. Yes, the markets dropped but with the hindsight, tech valuations from 1999 were far more accurate than from 2001 if we were to look at something like NASDAQ.
Shiller thought all assets were expensive in 2015. He didn't call for people to get out of the market (at least in that first article) as others do when they cite a high CAPE.
DeletePeople who invest based on CAPE do so in very different ways. Those who get out of the market when the CAPE is high and make predictions of impending doom are vastly overstating CAPE's predictive power. Neither example shows Shiller doing this. Barclay's CAPE-based indexes look like some peculiar mix of sector rotation and value investing. I'm not interested, but it's not crazy. It might do okay before costs, but I fear it might trail a regular index after costs.
Unfortunately, whether you have any idea if CAPE matters at all or not, you have to make your bet. To decide on a strict version of a random walk is to take a side with your money. I have taken the side that high prices become slightly more likely to moderate than they are to keep getting more expensive, and I've invested accordingly. Each one of us is free to bet whichever way we like.
I'm wondering why you say that owning REITs is less risky than owner 'a property or two'. I would think that owning actual real estate in the right places would be a pretty sure bet.
ReplyDeleteWe can't know if we've owned in the right places until after the fact. There are no sure bets when buying a single property. Some buildings are lemons, some tenants cause problems, and some areas fall out of favour. Diversifying investments reduces risk.
DeleteHi Michael, I also do not own real return bonds, but for different reasons. The current ~1.5% Canadian real return yield looks good (per Bank of Canada website), but the ETFs have MERs of 0.3% or more. If there was one with a 0.1% MER I would likely include it in my fixed income investments. Best Regards, Jim
ReplyDeleteI look at real return bonds every so often, but I'd rather own a few of them directly (which I'm not sure how to do) rather than owning one of the expensive ETFs.
DeleteI don’t think the market is very transparent for direct RR bonds. When I looked into it maybe 10 years ago on my TD trading site, I recall I needed to call a bond broker to get a quote. I did not do so. Not sure if things have changed. Cheers, Jim
ReplyDeleteI suspect it's similar at BMO Investorline, but I haven't called for a quote.
DeleteHave you found a reliable source for CAPE data online? At best I find this data hard to come by.
ReplyDeleteI get US CAPE data here: https://www.multpl.com/shiller-pe/table/by-month
DeleteFor the rest of the world, I've been using the export CSV option at https://indices.cib.barclays/IM/33/en/indices/static/historic-cape.app
Periodically, I record these CAPE values, and use them to produce a blended CAPE for each of my ETFs (based on the ETF's breakdown of country stocks). In between recordings, I estimate CAPE as scaling with the ETF's price. Having set all this up once in a spreadsheet, I just have to periodically copy the relevant line from the CSV into my spreadsheet, and capture the current ETF prices.
That's a cool approach (scaling on ETF price in between updates) and probably "accurate enough". Don't suppose you'd be willing to share your Sheet that does this?
DeleteThe relevant part of my spreadsheet is entangled with many other things. The Google sheets calls to get prices look like the following.
DeleteGoogleFinance("TSE:VCN","price")
GoogleFinance("VTI","price")
If you know the price p of an ETF on a given day and its CAPE c on the same day, then when the price changes to p' in the future, the new CAPE is roughly c*p'/p. This assumes that corporate earnings don't change (much).