Crisis Investing

Research
21

progscon, agree, that is the trade + very dovish fed and CBs around the world + coronav. (Wall of worry), before coronav. we had a good swing
up in world economy. So quite complex macro situation…, nothing to do then trading your system…

22

Georg, thank your for your great work, will try out UnivAvg(“True”,“(EMA(3,0,##corpB) - EMA(3,0,##UST10YR))”) !!!

23

UnivAvg(“True”,“(EMA(3,0,##corpB) - EMA(3,0,##UST10YR))”)
and universe All Fundamentals.

24

I think you’re missing the point of the article, which is that during crises, small-cap value works BETTER than during bull runs. The lesson is to NOT use market timing since that will make you get out of the market precisely when you should be in it.

25

Yuval, I think you are missing my point.
If you read the VERDAD article it is all about market timing, i.e. when is the best period to be in small caps, like ETF VOE.

VERDAD’s timing rules for “financial accelerator” events are:

  1. Period starts in the first month when the High Yield Spread (HYS) >6.5% AND when for the preceding 24 months HYS < 6.5%.
  2. Start investing 3 months after HYS > 6.5%.

Rules 1 and 2 read like a recipe book, and are market timing rules, what else? No different to a moving average crossover strategy for S&P500, just more complicated and established by backtesting since 1970. They examined eight high-yield events since 1970.

Also there are no exit rules. And that’s a problem because this strategy failed in 2008, returning -43%. Why not follow a recession indicator strategy like my MAC-REC system to get out of equities and into Treasuries when times get tough, otherwise trade small-caps.
https://www.advisorperspectives.com/articles/2020/02/10/market-timing-with-the-s-p-500-golden-cross-and-a-recession-indicator?bt_ee=WcSnIP2aQCaZ97oeveT%2BSObziicKrgXVjhIahNsIfMs%3D&bt_ts=1581419214491

26

the question is: Time the market or not time the market with a very difersified ranking system.
I go both ways, I try to be out of deasaster times like 2008 but otherwise ride it out.

But I learned a ton on this project.

I tried for two years to get a better market timing and I have to say I only learned the following:

  • improving marketing timing by anticipating DDs of 10-20% → could not do it (2018 is a good example, whenever I went to tight
    with my timing rules, I avoided 2018 in the backtest, but lost monster performance in other times) → project is not
    dead, but until then I need to work on my mental toughness to get through the DDs (so far sucessfull!)

  • improving ranking system by adding industry momentum that will catch the different market regimes of GDP rate of change and inflation
    rate of change → successfull, got a much better ranking system now, that catches agressive stocks in risk on regimes and
    more defensive stocks in risk off times (+ 10% Performance, live since Summer 2019 with good results.)

  • improving the ranking system by diversifiying the ranking system: sucessfull, I only had value and momentum in it, I added
    quality, fundamental momentum, accumulation / distribution and NOA ((DbtTotQ + IsNA(NonControlIntQ, 0) + IsNA(PfdEquityQ, 0) + ComEqQ - IsNA(CashEquivQ, 0)) /AstTotQ) and got very, very good results, higher total return and better distributed through the years (+5%, live
    since October, so far with good results)

So my two cents: between Georg and Yuval, try to stay out of disasters, but not trying to avoid up to 25% DDs (since they
hurt total performance big time, at least in my backtests).

What helped my thinking is, that the article gave me a good backround on why my trading system works especially
good in hard times (after hard times) and that it might be good to let the DDs come, bc. that is actually the way
to outperform (afterwards) with a highly diversified ranking system.

Best Regards
Andreas

27

here my macro chart on this subject

https://www.portfolio123.com/indicators?uiu=1616 (hope you can see it, settings on puplic)

28

Andreas,

I tend to agree more and more with the belief that drawdowns can actually be good and if we are trying to maximize returns they need to be tolerated to some degree. Yes the initial drawdowns are unpleasant but the recovery presents extremely profitable opportunities which tend to occur very quickly in front of the ability of a market timing system to anticipate. When I examine backtests the greatest outperformance comes out of recovery from a drawdown. Pairing with risk off makes those drawdowns more shallow and with some rebalancing between strategies you can begin to shift capital from risk off over to risk on value strategies. If you can’t tolerate drawdowns I tend to think diversification or constant hedging is going to be a better strategy than market timing. You will sacrifice alpha in both scenarios, but you can know with greater confidence the diversification will reduce drawdowns and prevent drawdown lock-in.

Jeff

29

More evidence that avoiding recessions can be harmful to your bottom line:

from https://compoundadvisors.com/2020/how-to-think-about-investing-cash-on-the-sidelines

And if you’re NOT perfect about timing recessions, it gets worse.

30

Yuval,

I’ve always questioned why it would be any use to accumulate cash when the markets get supposedly overpriced. The “pros” say you should do that. But again it’s market timing and markets can be apparently irrational for long periods of time. Sure you won’t have as much of a drawdown, but you are missing out on extra profits at the top and likely waiting too long to get back in because you won’t know the bottom is in until hindsight.

I’ve just recently been reading up on and investigating some tail risk hedging strategies where you buy deep OTM puts on SPY. Suffice to say in backtesting it doesn’t look like the strategy really works out as well as advocates would have you believe. You get some decent payouts during market instabilities, but the drag on your portfolio is worse than the benefit. I guess there is no free lunch for safety.

Jack Bogle had some good wisdom in that area which seems more and more reasonable to me as time goes on. He was a strong advocate of diversification and rebalancing which I agree can go a long way to smoothing out a portfolio. Additionally, he said always be buying. Everybody knows about dollar cost averaging but when you employ market timing strategies you lose out on that benefit because you are more often buying high and missing out on low price buying opportunities.

Jeff

31

Yuval, there is no daily value of the S&P composite available from 1929 until 1950. So the analysis in the article you reference is probably using the Shiller monthly average of S&P and Biello’s numbers are probably incorrect. Also who would move to cash when the much better alternative is the 10-year Treasury bond, or similar.

So I don’t buy your or Charlie Biello’s argument. It is always prudent to be out of equities during recessions and in Treasuries during those periods.

If you followed my MAC-REC system
https://www.advisorperspectives.com/articles/2020/02/10/market-timing-with-the-s-p-500-golden-cross-and-a-recession-indicator
your annualized return from 1966 to now would have been 11.7%, and this is only by going to SHV (cash with interest), not Treasuries during recessions. You can add at least another 1% to that if one went to 10-year Treasuries. So your annualized return would have been about 13% versus 9.14% for buy and hold the S&P500. That is not bad for doing nothing most of the time.

So please don’t tell us to “press on regardless” trading stocks when the prudent strategy is to be out of equities during recessions.

1 Like
32

Yuval, thanks for the URL to the detailed article. Here’s the points that stood out for me.

  1. The study is very comprehensive for both the time periods covered (8 crisis periods from 1974 to 2016) and the range factors considered (over 400, pages 9 and 10). I also greatly appreciate the detail the study writer shares which allows one to confirm or qualify their summary statements.

  2. Their multifactor model has some elements similar to what many of us use in our P123 models plus some elements that are new to me. Here are the factors in their model. (pages 16-21).
    . . Asset Turnover
    . . Positive Net Income
    . . Volume (volume / shares outstanding)
    . . Value Composite (EV/EBITA, P/B, P/E, FCF yield)
    . . Positive Cash Flow
    . . Decreasing Leverage
    . . Net Debt / EV
    Note: The study looked at some 400 variables (page 9) but only gives a suggestive list of what they were (bottom of page 9 and top of 10). Presumably the items in their multifactor model were the best of the 400.

  3. I found the introductory summary promising: “Bull markets are all bullish in their own ways… But crises are alike. A recent academic study found that standard models for predicting returns in equity markets are 8x as predictive during recessions as during expansions.” (page 5)

This sounded very promising to me. But I was disappointed to see that their multifactor model failed to do as well as the market during 12 month test period for 2 of the 8 crisis periods. It is hard for me to accept the statement that “Crisis are alike” when their best model underperforms for 1/4 of the crisis periods. Oh well, I guess that is some marketing PR talk.

Crisis Year — Multifactor vs General Market (12 months gain), page 25.
1974 = 6 %
1980 = 11 %
1986 = 28 %
2000 = 24 %
2008 = minus 5 %
2010 = minus 4 %
2012 = 42 %
2016 = 24 %
Their multifactor model did less well than the market for 1/4 of the periods (still profitable, but underperforming the general market).

  1. As far as I can see the methodology of the study doesn’t provide enough info to confirm or reject the view that market timing is helpful. Here is their methodology.
    . . Look for “crisis” periods where economic participants are extraordinarily fearful as indicted by a specific credit spread being over 6.5%
    . . 3 months after the 6.5% spread appears, invest funds (page 9)
    . . measure results after 12 and 24 months

For a few, but not all, of the crisis periods, the study notes how the market segments did for the prior 12 months. We would need this info for all of the periods in order make a determination of whether it is best to use Market Timing vs always being in.

  1. The study does not comprehensively address the issue of whether their diverse multifactor approach is better than other more focused multifactor approaches. They did give results for one focused multifactor approach – ie their value only composite (EV/EBITA, P/B, P/E, FCF yield). But the underperformance of this value is not conclusive for two reasons.

First, my experience is that value factors need to be matched with an additional factor to exclude value traps (something like short interest or price momentum, etc.).

Second their value factor model is for the best decile (1/10 of all stocks) but the results of their multifactor model is for a more far more selective group (ie just the 50 best stocks). Looks like PR to me (make their multifactor approach look good by using in an appropriate comparison).

Oh, I didn’t give any attention to the chapter dealing with bonds since my interest is in equities.

Regards,
Brian
…

…
.