No. of Recommendations: 13
There are always 2 sides to the coin - especially in markets - that's what makes them function.
I have been following and tracking these for a long time - and here's my take and some technical/mathematical critique
- these indicators are all based on the concept of Conditional Expectations from statistics in any case.
So I will start with one of the most venerated members and Statistician Jim's Bear indicator : DBE.
DBE is "passage of time" class indicator - there's of course the No new high check ( price based) - but its basically a long long noose for the market to try hang itself in some sense. Geometrically (although not assumed) - and for the visually minded - based on my analysis and Jim's past comments it relies on some sort of a "inverse/reverse mirror image" Long tick mark behavior ( I wish I could draw it) - ie a Sharpish drop and then a comeback , but not enough to create a new high - the 100 days ( ~ 5 months) essentially you gives you an opportunity to leave for the door without too much damage often within 5-7% of the peak.
GTT is essentially setup trying to do the same thing. Why?
(i) Fidelity is monthly - very very key. It acts as a smoother. Daily check on something like this is futile. Min should be weekly ( Jim uses an actual smoothing technique IIRC)
(ii)Then comes the Economic indicator check - these are published monthly too - typically in the 3rd week. This means - very likely you would be holding till EOM ( unless you use weekly)
See the pattern - tendency is "Hold...hold....hold..." - Constant second guessing. Rationale should be obvious too - that table from Zee illustrates that easy enough.
All this Recession talk etc to me almost missing the point. If I had a choice for words.....
Anyway, these are the basic premises
(a) Stock fall in a recession. Zee was once kind enough to run this. Absolute truth.
(b) As Elaan and others said - Stock's have been one of the best leading indicators for Recession.... Is it?
As has been mentioned many a time here - the best US Recession indicator has been the Yield Curve model - the ONLY one I know of that signalled prior to COVID ( Q3 2019 - but would you have trusted it sitting right there - without seeing the right side to come?). But its a terrible timer - It went off this cycle in 2022. No recession thus far. It has come back again ( at least my implementation) - so now what?
The point is - you really don't care much for whether it predicts the recession or not - all you need is the following to be true
"Once it signals, will the forward drop be significant enough to take action and thus improve long term draw-down without sacrificing too much returns?"
(1) The 200MA first strike point is literally TERRIBLE in the short term. Its the first part of the inverse tick - ie you will be missing the marginal comeback (not guaranteed of course) and taking the full FIRST BLOW in the chin.
(2) The issue of course is that if 200MA (which is a Statistical slope change indicator) is your only tool in your arsenal - you will be taking the whipsaws as evident from Zee's table.
(3) So therein, comes in the GTT Economic overlay - which basically they claimed ( a bit on this later) to have smoothed out majority of the whipsaws.
Basically, from the point of the confirmation - there was ENOUGH of downside REMAINING in Market ie price action for the overlay to be effective. What it did as a recession indicator is a moot point.
KEY POINT : Why did I use the word CLAIM? GTT was published in 2015-16 IIRC and in-sample was till 2013 or something.
Folks... Unfortunately whether we like it or not - every major Bear catcher which is a combinatorial ruleset has to accomplish only 4 things - it needs to FIT the following periods as best it can - 1929 Crash, 1970s stagflation, 2000s Tech Crash and 2008 Financial crisis - ie 4 time periods. Jim has mentioned this and as has Robbie - simple statistical fact - very low sample incidence.
So really the game ( mathematically) is to actually optimize or minimize the errors on the other side - ie ensure it doesn't go off too much in the Bullish periods.
A good illustration is the ONLY post-discovery recessionary sample we have ie COVID. 200DMA signaled last week of Feb 2020. You had to wait till 3rd week of March for the helluva Economic prints to come out ( I think they might actually have been delayed - which I will discuss in my next point) - and market bottomed what, like 3/21-3/23!
SAHMs indicator ( which I personally think is like a centerfold illustration of an overfit rule) - is based on the Slope of the Unemployment Rate, also failed ( Cant blame her) - initial U/E print came out as 15% in March and I think was revised down to 12.5ish? Taking in all COVID assistance it was close to 25%!!
But - from an economical standpoint you should be able to get behind all of these - they make sense. Key is what do they say about market direction?
And that brings me to my last point - this I refer to as "Wonderful basket of misleading delights" - aka US Economic Releases. The very nice and very defensible thing about them - they are constantly REVISED - some every month! This to accurately reflect the truth - except it embeds one of the biggest bane with market analysis- Look forward bias. When I first read GTT - that was one of the things I checked - source. It uses what essentially you and I can download from FRED.
So Q: That sure shot thing on RRSFS YoY % change - was that of the initial release available say last week of Oct, 1987? or actually the final revised one from Jan , 1988- ahem?
NET :
(1) 200DMA First signal is typically a terrible exit point
(2) The 2 economic overlays from GTT are defensible theoretically, just be cognizant of their biases.
(3) To me DBE is simpler - but that really depends on a "roundish, longish top" assumption - and did TERRIBLY during COVID too.
DBE did worse than GTT in 2022 Bear - DBE exit was June and GTT I think either Mar or April cycle.
Either way - this is not the week to run for cover. Wait it out!
Best
