UnThreaded | Threaded | Whole Thread (1) | Ignore Thread Prev | Next
Author: GrandpaRalph Big red star, 1000 posts Old School Fool Add to my Favorite Fools Ignore this person (you won't see their posts anymore) Number: of 68  
Subject: Fun with Numbers Part 7: Analysts Report Card Date: 8/22/2001 11:57 PM
Post New | Post Reply | Reply Later | Create Poll . Report this Post | Recommend it!
Recommendations: 3
 
The ANALYSTS! Everyone just loves to hate 'em.

Why is that? Well, I'm sure you've heard the more popular reasons...
- They give advice which moves prices in the direction most beneficial
to their own financial interests.
- They have secret conspiracies designed to churn the market, thus
generating more commissions.
- They don't really understand valuations, they just use impressive
sounding metrics.
- They play the momentum game.
- They give the investor whiplash, with downgrades just two weeks
after an upgrade, or vice versa.
- The advice "today" should have been given "yesterday", BEFORE the
company guided lower.
- From our "friends" at CNBC, this dilemma: What do you mean you are
upgrading from "sell" to "hold"? What am I going to hold if I
followed your sell advice? And if I didn't follow your
previous advice, why should I now?
- ad infinitum

But does their advice hold any value? If I follow their advice, can
I do better because of their "insight"? Or maybe I can play contrarian
and go against their advice, if I really believe they are crooks,
intentionally giving "bad" advice.

Those who have followed this "fun with numbers" series over the past
6 months or so are aware that I have found some fairly significant
statistical data showing undeniable correlation between various
fundamental business (i.e. RuleMaker) metrics and valuation. I have
not yet been successful in using these relationships to develop a
market beating investment strategy....but at least I'm getting more
realistic in my expectations, and learning the difference between
substance and vapor.

"Vapor" here means something which will quickly dissipate, with no
long term impact on anything. Short term, it may be visible, and even
cause some price movement; but it dissipates so rapidly, and has such
a short term influence, that it somewhat reminds me of the ozone
after a lightning strike. If you have seen the lightning, it's already
too late to run. All you can do is smell the vapor and marvel at
how fast and devastating the lightning was. The ozone vapor results
from the lightning, but is harmless in small quantities. Analyst
recommendations are vapor:
There is no statistical evidence that they actually correlate with
anything in the long run. You can't side with them and do better.
You can't side against them and do better. The best thing that can
be done is to just ignore them, thus saving personal time and
energy.

So OK, here comes the data to support my claims that the analyst
recommendations are meaningless. Below are a series of "scatterplots"
showing price movement for a given time period versus the analyst
recommendations that were current at the beginning of that time period.
(Recall the point scale 1 thru 5 = Strong Buy through Strong Sell)
Each scatterplot also has a "best fit" line drawn through it, and a
statistical correlation factor associated with how well those points
correlate. The "ending time period" is Aug 17 in all cases, and
successive graphs go further and further back in time to see if there
is a difference between "one month" and "longer period" results.

Period: July 1 to Aug 17; Statistical Correlation 0 (Truly Random)
Sample Size: 88 companies (listed at end of post)
Clicking on the link below will show the scatterplot of price
percent change for the period 7/1 to 8/17 as a function of the
analyst recommendation as of 7/1 for all 88 companies.
http://people.mw.mediaone.net/ryoungbl/anal71817.gif
It seems fairly intuitive to me that there is no correlation.

At this point, I feel obligated to show a graph that DOES have some
statistical correlation, (for those who haven't followed my prior posts
on valuation) so I will use Net Margin vs. Price/Sales ratio for those
same 88 companies, (except ebay, for reasons I've explained
in previous posts) as of 8/17/01:
http://people.mw.mediaone.net/ryoungbl/nm081701.gif
(Note how this scatterplot does not look quite so random, and that the
points do indeed have a relationship which is partially captured by
the blue line.) Statistical correlation is .55, as opposed to 0
(random) or 1 (all points on line).

OK, so there was no correlation between analyst recs given July 1 and
stock price performance for the period July 1 to August 17. Maybe
we should look at a longer term horizon....

Period: June 16-Aug 17; Statistical correlation .01 (Virtually random)
Sample size: Same 88 companies.
http://people.mw.mediaone.net/ryoungbl/anal616817.gif

Did you notice that this graph had a slight positive slope? At a
higher statistical correlation, this would indicate the analysts were
wrong more often than right, and you could do better by going counter
to their advice. But alas, the correlation is too low to really infer
that this would be a good move.

OK, let's look at roughly 3 months...
Period: May 24-Aug 17; Statistical correlation .02 (Still near random)
Sample size: 76 (I added some companies after 5/24, and don't have
the historical analyst recs, so only 76 of the 88 had complete data)
http://people.mw.mediaone.net/ryoungbl/anal524817.gif

Again, a positive (and even more steep) slope, but still of no real
statistical significance that would cause me to "short" the analysts.

I have more time periods, but let's go to the longest term I have:
Period: Mar 2 to Aug 17; Statistical correlation .01
Sample size 74 companies (There were 2 added between 3/2 and 5/24.)
http://people.mw.mediaone.net/ryoungbl/anal32817.gif

I also have multiple 1 month time frames in the period 3/2/01 to
8/17/01, and they all pretty much look the same; so I guess you can't
argue for a "monthtrading" mentality either...

Bottom Line

I believe this data presents compelling evidence that the "average"
analyst recommendation is a useless piece of data, and that you can
spend your time doing other sorts of Due Diligence that will yield
better results (or at least better education). I have NOT decided
that EVERY analyst's recommendation is useless, just that the "average
of ALL analysts" is. I have seen certain analysts that have good track
records over certain periods of time, but have yet to find one that
consistently does well. Lacking time of my own, this can be an
assignment to the "interested students".

Rear View Mirror?

Having cut this data every which way I could looking for some value
to analyst recommendations, (I didn't do this hoping to actually find
any... I did it to try and convince myself with data that what I felt
to be true could be statistically demonstrated to be true.) I decided
to look at it one more way...
I plotted price change for the period 3/2 to 8/17 as a function of
analyst recommendation on 8/17: (i.e. looking backwards)
http://people.mw.mediaone.net/ryoungbl/anal81732.gif
The correlation was .15, still not real strong, but fairly convincing
to me that at least SOME analysts are basing their recommendations
based on looking at price movement (relative strength) in the rear
view mirror. So now consider this: If some analyst's recommendations
are based on momentum, but correlation drops like a rock when you switch
from looking backward to looking forward, what does that say about
the value of so called "momentum investing?" Yup, as a corrollary
to my theory that "average" analyst recs are useless, I suspect
momentum investing is also "on average" a strategy that will only
generate commissions and tax liabilities, but neither beat nor get
beaten by the market before those costs.

And another thought...

All those "earnings estimates" for everything from next quarter to 2
years out are developed by these same analysts, and are partially the
basis for the buy/hold/sell recommendations. Would you buy a PEG ratio
from one of these earnings estimate vendors?

Darn. Valuation is a tough nut to crack.

The 88 companies currently in my spreadsheet:

AA ABI ABT ADBE ADI ADP AEOS AFCI ALTR AMAT AMGN AOL APCC BA BARZ BBBY
BMC BMY BRCD BRCM BUD CA CAKE CCL CHV CMVT COH CPB CSCO DELL EAT EBAY
EK ELY EMC F GE GIS HDI IBI IBM IDTI INTC ITWO JDSU JNJ JNPR JNY KKD
KLAC KO LIZ LLTC LLY LSCC LTD MACR MDT MERQ MMM MSFT MXIM NOK NOVL NT
NTAP ORCL OSI PAYX PEP PFE PMCS QCOM RATL SBC SBUX SEBL SGP SNPS SUNW
SYMC TJX TLB TYC WWY XLNX XOM YHOO

Why these 88? Just the result of an admittedly imperfect screening
process that identifies companies with good fundamentals that I might
be interested in at the right price. I am in search of finding the
right prices.

Ralph
Post New | Post Reply | Reply Later | Create Poll . Report this Post | Recommend it!
Print the post  
UnThreaded | Threaded | Whole Thread (1) | Ignore Thread Prev | Next

Announcements

Post of the Day:
Value Hounds

Mylan Transitioning from Generics
What was Your Dumbest Investment?
Share it with us -- and learn from others' stories of flubs.
When Life Gives You Lemons
We all have had hardships and made poor decisions. The important thing is how we respond and grow. Read the story of a Fool who started from nothing, and looks to gain everything.
Community Home
Speak Your Mind, Start Your Blog, Rate Your Stocks

Community Team Fools - who are those TMF's?
Contact Us
Contact Customer Service and other Fool departments here.
Work for Fools?
Winner of the Washingtonian great places to work, and "#1 Media Company to Work For" (BusinessInsider 2011)! Have access to all of TMF's online and email products for FREE, and be paid for your contributions to TMF! Click the link and start your Fool career.
Advertisement