Investing in a Bubble

For those who are interested in comparing today’s frothy market to that of the 1995-2000 period, I thought this article made some interesting points.

https://mailchi.mp/verdadcap/investing-in-a-bubble?e=b4dfbc82f2

Thanks for posting Yuval,

Besides trend following I also use uncorrelated assets (Gold, Global Bonds, Global Stocks, Commodities) and sector rotation. Hopefully that will help lessen the drawdown when the great reset happens.

Cheers,
MV

Thanks - interesting!

I’ve been using trend following for a few years as a hedge, but 2020 has reduced my confidence that it will continue to work in the future.

I think the Q3 commentary by Newfound Research (which started the ROMO trend following ETF) was worth reading:
https://blog.thinknewfound.com/2020/10/q3-2020-commentary/

A quote:

“As long-time proponents of trend following, this research suggests we were wrong, or at the very least, the circumstantial evidence supporting this narrative suggests it. And by wrong we do not mean, “the market path that occurred from 2010 to 2020 is but a single realization of a large number of possible realizations and we just got unlucky.” Rather, it suggests that the very distribution defining how market returns materialize has slowly, but fundamentally, changed over the last decade.”"

Thank you Yuval. Excellent article.

Personally, I think I am Warren Buffett and Ray Dalio worrying about this now. My guess is, like then, it will not happen for another 3 years. For sure, everyone will try to keep it at bay until after the NEXT election, including the FED.

That having been said, I do what Mark (mv388158) does above. I am pretty diversified. Still, I have increased by small-cap value positions recently.

I read recently that small-cap value does well with inflation.

So here is my question. I think we are in a bubble. But could it be different than 2000?

There is a lot of money-printing going on and we haven’t even begun to hear people (politicians and the news) tell us about Modern Monetary Theory. Explain how not understanding this has been our problem all along and how this time will be different.

Different than 2000, perhaps. A lot different? If so, how? Will the money-printing (or something else) make it different? What about that petro-dollar thing. Will it hold up? Multiple opinions on petro-dollar, Ethereum (Bitcoin), money-printing, and
dollar-devaluation here: https://www.lynalden.com

Anyway, I am buying (not selling) small-cap value. Using P123 to do that.

-Jim

I’m not sure trend following can improve alpha, but maybe it’s sacrificing alpha for downside fat tail insurance … which might be provide utility, depending on where you are in your investment career. I’m hesitant to say trend following as a hedge can’t work again because it didn’t work in March 2020, because March 2020 is such a unique correction born on by an external factor that seems like it might be temporary. But I think it’s important to follow a the trend of an index that reflects your model. If your model is based on small cap value, I’m not sure you get much benefit from trend following against the SP500 that is currently dominated by largecap growth. That said, I created a custom equal weight index of the top 150 ranked positions on one of my microcap rankings, and if I would have followed the close(0)/sma(200) trend it would have whipsawed me out of the smallcap bear market of September 2018 and wouldn’t have got me back in time for the rebound in January-February 2019. So I’m not sure … maybe I’m wrong … maybe all rebounds are v shaped now and too hard to capture with simple trend following.

I like how the article starting it’s very interesting to read. But for the conclusion not so sure, for this year, for example, probably lost a major market up-trend. I believe a combination of various factors to monitor the overall market is a better strategy than just trend-following.

Terry

I confess I’m very confused by the last few comments. I’ve always thought “trend following” meant technical analysis: identify a short-term trend in a security, try to catch it at its inception, and then sell when the trend looks like it’s turning. See https://en.wikipedia.org/wiki/Trend_following, for instance. But unless I’m mistaken, that’s not what “trend following” means for Newfound Research. Nor for InmanRoshi, who mentions trend following as a hedge, and says “it’s important to follow the trend of an index that reflects your model.” There must be some new meaning for the term that I’m missing . . . If someone could explain, I’d be grateful.

Yuval - Because trend following removes the left tail, it is considered a hedge. It is not free as the cost is the chop where false positives get you out a lower price only to chase back at a higher price, missing some of the recovery. One way this is addressed is to limit the decision on the last trading day of the month - so ignore trend violation unless closing price stays below a moving average at the end of the month. This causes some delay in getting out and back, but reduces false positives.

Value works when gdp and inflation is up on relative terms (like this quarter!)
Growth also works when gdp is down and inflation is up on relative terms.
And industry groups work totally different depending on gdp and inflation regime.

My take: make sure you have an industry momentum part in your ranking (my one is 25%), so groups that do poorly get a lower rank.

Also to prevent long term underperformance on one single factor like momentum, growth, value, quality combine them in your ranking system.
For example value might have had a bad stretch until october 2020, but if combined with momentum and other factors there was
almost no “bad time”.

Valuation of the market is the worst timing factor, it simply does not work, also bc its relative, expensive gets more expensive
in gdp up and inflation up, get both down, everything gets sold off independend from valuation.

Also be carefull with correlations of the past, I believe there is no hedge in the future besides cash,

I think the 20s are going to be great to invest. Gen Y is kicking in and its bigger then the boomer generation, just look at housing!!! The US will get on track again,
fiscal spending is coming to prop the middle class, probably we get higher inflation which would be great for value stocks and certain
industries.
Just look at the energy sector, tons of stocks have a PB below 1, if oil gets more expensive energy stocks are going to rally like from 2003

Until May 2021 I see no reason not to be long. The comps in inflation and gdp are easy to beat, q3 is different, that is going to be hard
to beat so gdp and inflation could be down y/y on a relative basis, then hell could break loose, so be carefull in the second
part of 2021.

Right know, ride the bull market.

Right. I haven’t actually implemented it with real money, but applying trend following would be merely a means of wealth preservation in a systematic way, not an attempt at a wealth maximization strategy. Right now I’m just a basic retail investor in the wealth accumulation phase of my life, so I’m just focused singularly focused on maximizing returns. At some point I’m (hopefully) in a wealth preservation point of my life, and I’m more focused on protecting against left tail risk. Even if the insurance it comes at a cost, I don’t want to have to make the money twice.

Unlike the late 1990s where all tech stocks experienced an astronomic rise, the current bubble is pretty much focused on the FANGs. Other sectors have severely lagged behind, and I agree with Andreas that the energy sector looks promising.
If we assume larger inflation going forward, a diversification into other asset classes such as precious metals and bitcoin will likely be advantageous.

Yuval,

Recently I was listening to a series of podcasts by an interesting guy named Mad Fientist (strange name, I know). And in one episode his guest mentioned he was doing something called Dual Momentum investing with his money. That guy sounded quite intelligent and knowledgeable, so I noted that. And then I read the book explaining the stuff - “Dual Momentum Investing”. So if you are interested and want to know more about this, I highly recommend that book, it is a short one: https://www.amazon.com/Dual-Momentum-Investing-Innovative-Strategy/dp/0071849440
Or you can just lookup his papers on SSRN (Gary Antonacci).

Among other things, the author explains in that book how this terminological confusion came about (with a bit of humor).
And he maintains a website where he updates results of such strategies to this day.
There are a lot out-of-sample data (the whole period is 1950-2020), so it looks quite convincing.
As a couple of primitive examples in PV, it removed big draw-downs quite effectively in the past:
https://tinyurl.com/yyam7l6y
https://tinyurl.com/y5tdywe6

I noticed that your stock picks, which you sell on SA via subscription, took a dip dive last spring. That stopped me from subscribing, because I agree with the opinion on the link you provided above - that we are approaching a storm (or maybe a “bound-ranged / sideways market” as Vitaliy Katsenelson puts it in his “Active Value Investing” book).
So I’m curious what you would think about using momentum to prevent large losses.
I read your posts on TypePad and, unless I’m missing something, I did not see you even considering it.
And I got a feeling that momentum might serve as a valuable addition to your stock-picking strategies.

BTW, some time ago I stumbled upon a Mechanical Investing forum on The Motley Fool.
And I remember one confession from an old-timer of that forum.
He said that many of the folks of their community got burned badly in the dot com crash (I believe they followed strategies similar to yours), and after that they had to learn in a hard way how to introduce momentum into their strategies.

So I would be very interesting in hearing your opinion on that. After reading your TypePad I’m impressed with your approach and logic.

So, trend-following only works if there is some “autocorrelation.” Correlation (or some relation) in prices from one month to the next is a must for this to work.

If there is no correlation it is just some variation on the gambler’s fallacy isn’t it? The Roulette table has hit red 9 times in a row so black is due or red is on a winning streak: let it ride.

Where in fact, the history has no impact whatsoever for things like Roulette wheels where there is no autocorrelation. The odds remain a little less than 50-50 no matter how excited you are with your money that just doubled 9 times.

This is also a little like the debate as to whether there really is a “hot hand” in basketball or whether we just remember the streaks: seeing a pattern that may not be there. Book have been written about this: The Hot Hand: The Mystery and Science of Streaks

Then if there is autocorrelation we tend to move freely from the “trend will continue” to “reversion-to-the-mean.” I personally believe there is a lot of BS out there. A lot of hand-waving. Go with whatever explanation works to overfit this data.

And everyone has access to good pricing data through Yahoo! (even people at Robinhood) so I think some of this is being arbitraged away—even if it exists.

But correlation is a sine qua non. Or translated: without which there is not. If there is no correlation then none of this has any rational basis whatsoever.

Is there a correlation? Maybe at the ETF level (image 1 for SPY). Maybe at the ticker level also (image 2). AMZN may be anecdotal but is was the first ticker I looked at and was not cherry-picked.

Both are monthly autocorrelations (with various lags shown).

Surely there must be much more to be said on this topic. But maybe. And I do use relative strength of ETFs myself—even while often waking in the middle of the night after some nightmare involving a Roulette Table with TSLA no it.

BTW, the pattern with SPY seems pretty common with ETFs (GLD being a notable exception). By that I mean a mild one month positive correlation followed by inverse correlation at 2 months and a stronger positive correlation at 3 months. Therefore it may not be a coincidence that I tend to use 3 month relative strength on ETFs (after backtesting).

Jim



Jim - autocorrelation and correlation are pretty broad terms. Can you be more specific about how you are calculating this?

Steve,

This from Portfolio Visualizer so they may not do it right

In Excel I line up the monthly returns of SPY next to SPY with the same date rows.

Then I delete the top cell in either one. Does not matter which (symmetry).

Then measure the correlation of the 2 rows this is a lag of one. Delete 2 cells that is a lag of 2.

Consistent with the way R does it for sure.

Jim

Hi Mike,
no momentum strategy can save you from drawdowns. There have been several threads in this forum on this subject, and I would say that most approaches ultimately fail to deliver - simply because they are modeled with the knowledge of the past.

Jim - what is that supposed to tell you?

Steve,

Really just another confirmation that looking at trends may work.

I think my avatar is actually an example of this autocorrelation in the form of a linear regression but I forget which ETF (3 month returns for the x-axis and the following month’s returns on the y-axis).

Suggests that past trends may predict the future at times or at least the data does not disprove it.

Jim

Mike –

Thanks for your comment. I’ve written about momentum here: https://blog.portfolio123.com/why-momentum-works/ Momentum is a very valuable factor and I use it in all my ranking systems. I’m especially enamored of industry momentum.

Trend-following always struck me as, in a way, the opposite of momentum. If you follow a momentum strategy, you want to buy stocks that are in a general upward trend considering the last six to eighteen months. But it’s best to buy them on a dip. That way you get a little extra oomph on your returns. When I measure momentum, I always ignore the price over the last four weeks. For example, I’ll look at the stock’s price four weeks ago compared to its 52-week high, or its price seven months ago. And then I’ll couple that with a factor that favors stocks whose price has fallen in the last two to four weeks—unless that coincides with an earnings report. I get more bang for my buck that way.

Trend-following, on the other hand—as I always understood it, that is—pays close attention to price swings over the last few weeks. You buy stocks that have been going up recently and sell them when they start to go down, using stop orders. I did that for a while in 2014 and 2015, using stochastic indicators and ATR-based stops, and lost tons of money in the process.

Maybe trend-following and momentum have come to mean the same thing now. Maybe the meanings of these terms have changed. Maybe that’s what’s causing my confusion.

My article explains why momentum works over six-month and year-long periods and why mean reversion works over two- to four-week periods. The trouble with trend-following, as I understand the term, is that it goes against mean reversion.

Last spring, as you point out, I lost 33% of my portfolio. I don’t know if trend-following would have limited my loss.

But considering that I ended up with a 105% gain for the year, I feel pretty certain it would have interfered with my overall return.

Yuval,

Thank you for your comments.
Indeed, my bad, I saw that article of yours last summer. I guess I tried to consume too much information too fast, so I forgot most of it as a result.

Yeap, then my links look a bit silly I guess, having your level of expertise and experience.

You are right, that dual momentum strategy yielded 2% in 2020 as I see on the website of the guy. So indeed it would have interfered with your 105% big time I guess.

As I see you are quite tolerant and respectful, unlike some commenters on p123 I observed during my brief stay on this forum. So I venture a couple more questions.

In my humble opinion, we all mostly got lucky last spring, when FED provided unthinkable liquidity, so the party got some more fuel to burn. Maybe I’m wrong here, but I like reading history and what I see does not look as a lasting situation.

So here is a question.
Maybe it is a one more of silly ones, but anyway I’m curious to hear your take on it.

What do you think would happen to your strategy if we had a re-run of 2000-2003 or 2008-2009 last spring?

Do you have any protection from the downside in your strategies?
Or do you just ride to the bottom with the equities market (whatever that bottom would be), hoping that following returns will cover the drawdowns?

I am quite risks-averse, but I think I could be fine with volatility - but only if it does not result in a permanent or a long-term loss of capital.

And, to justify your time spent answering, this is not a theoretical question. I’m considering investing my one year salary in your picks.

Again, I was about to do that last summer, but then I saw the drawdowns in your numbers and freaked out. OK, with hindsight, it cost me a lot, but that is my sleeping point at the moment anyway.

So the best alternative I see now is either cash or something low-volatile like Permanent / All Weather Portfolio. My line of reasoning is that I’m OK loosing 10-15% next 2-3 years due to inflation, but not 50% due to mean-reversal. And then, after CAPE is down to a reasonable levels again, I will be happy to switch to 60/40 or whatever.

I have recently invested one year salary with another service, so I’m really interested to know your downside protection (if any). What is your plan for the times when hell breaks loose?