Barron’s Mike Santoli looks at one of our favorite pet peeves: The use and abuse of data. His column this week, The Limits of History, notes that this has become especially prevalent of late:
"People who try to handicap the markets for a living practice the art of the plausible. Many trudge from conference room to lunch table to banquet hall lugging PowerPoint decks full of unobjectionable statistical touchstones for commission-wielding clients. At times of investor confusion and market dissonance, such as now, their art is often reduced to carving a slice out of economic history that ratifies their existing outlook."
Mike then proceeds to look at a "dog’s breakfast of the kinds of historical analogies making the rounds." What is especially amusing about his list is the number of "Every time X happens, Y has occurred" that collectively produce all manners of mutually exclusive results. Since "X" occurred we will definitely have/avoid a recession; Stocks are undervalued/overvalued; Markets must rally/fall.
What is an investor to do? Whenever you are confronted with an "incontrovertible proof" based on historical data, prior to taking any action, I suggest asking yourself this short list of questions:
• Do we have enough historical examples? Is the data sample statistically significant?
• Causation or Correlation? Does "X" cause "Y" to occur? Or, are we presented with two things that may have the same underlying causes? Is there even interaction between X & Y?
• Coincidence? How possible is it that these two items are utterly unrelated (i.e., proof-we-are-clueless Superbowl indicator).
• Look for differentiating elements in different time periods: What factors are similar? What factors are different?
• Compare interest rates, inflation, dividend yield, P/E contraction or expansion, sentiment, overall market trend, business cycles — across different eras. Might that account for potentially different outcomes?
• Any recent market environmental changes (regulation shift, financial innovation, etc.) have an impact? What might these specific changes do to the data? Consider Decimalization, ETFs, online trading, change in dividend tax, etc.
• Subjective versus Objective measures: Are the factors under discussion hard numerical data, squishy or somewhere in between? Percentage of stocks over 200 day moving average is objective; I find some chart pattern readings subjective. Earnings at time have been rather subjective; official inflation measures somewhere in the middle.
• Consider things in terms of probabilities, not outcomes: Assume a causative factor resulted in a specific event (X –> Y) 7 out of 9 times. The most you can say is that when "X" occurred in the past, it has resulted in "Y" approximately 78% of the time.
• There is a difference between historical occurrence and future likelihood. In the example above, this does not necessarily even mean that since "X" has just occurred, there is a 78% that "Y" will happen. Consider: was the first X/Y occurrence really a 100% or zero? Did the second one become 100% or 50%, then next a 66% or 33%?
• Contextualize data: Sometimes a single data point — even a mean or median — only tells half a story. Any data point can be trending or reversing. Going higher, lower, topping, bottoming. Each of these may have differing implications for what comes next. Inflation is high, but coming down. Gold is high — and going up. It helps to think of data not as a still photograph, but as a frame in an ongoing film.
I’m sure there are more — that short list is off the top of my head.
Any others? Please make your suggestions below. If we get enough good ones, I’ll try to massage this into a more formal column.
>
Source:
The Limits of History
Mike Santoli
Barron’s, MONDAY, JANUARY 28, 2008
http://online.barrons.com/article/SB120130827516518451.html
Weather happens every day (so we have a huge database of historical measurements from which to project), but we still don’t forecast too well (doppler has made a big difference, but it won’t work on WS – too much tin foil).
If past history was all there was to the game, the richest people would be librarians.
– Warren Buffett
I respect Mike Santoli a lot. He writes a thoughtful line. Ben Stein on the otherhand, makes me want to scream! After this morning’s NYTs piece, “Can Their Wish Be the Market’s Command?” I may have to skip anything with his name on it.
Like any good lawyer knows:
When comparing things, first find all the similarities and differences.
As any good gambler or statistician knows:
Take all the similarities and determine probabilities based on conditional logic.
As any good physicist or historian knows:
Take all the differences and simply watch because anything can happen.
Thus, in our current situation we can make comparisons to many past markets. We can then figure all types of fancy probabilities by scribbling equations over history. Then, after all the quants have mathematically determined their 6 standard deviation events as nearly impossible (e.g., occurring once every billion years), the physicists and historians will laugh last when today’s unprecedented financial engineering creates a new unique market experience that we will later add to the database for future comparison.
It is not difficult to make a convincing argument about unkowable events.
Knowledge is not wisdom.
Quantitative data translated into “info-porn” does not connote good judgment.
Barry, thanks for the post. It is a timely message, and especially useful perspective to place against many (but not all) bloggers, media pundits, and commenters here at TBP that spout historical data and generally distractive “noise” without any apparent use of logic or sound judgment.
Thanks for the wisdom…
Using good data to support or refute a premise is always a good idea. But data from whom and how gathered is being more and more politicized.
About 30 years ago, good economists used to look at paperboard production, intercity truck tonnage, coal car loadings and many other objective stastics to determine the health of the economy. As an example of a changing world, we look at the Baltic Dry, sales and income tax receipts and other stastics that were not available or even relevent in years past.
One thing I have learned is that you pick good objective people and follow their lead. The old saying, bet on the jockey and not the horse is my simple way of assessing the reality.
One decipline that has been dumbed down over the years is accounting. I’ll bet many analysts can’t tell you if the companies they follow use Fifo, Lifo or average cost inventory accounting. How many inventory ‘turns’ per year. How depreciation is calculated. All I seem to get are stock buyback intentions and linear graphs. Balance sheet accounting sucks. Everyone seems hung up on EBITDA. But what is real E? Whatever the company tells you it is.
Sorry for the semi-highjack. Information overload on Sunday morn.
because people did not do their jobs.
Being a dynamic, non-linear system, it seems to me that chaos theory would be more applicable to markets. Point being is that what seems insignificant at the time can have huge ramifications later.
Not the greatest example, but how many of us grasped early on the systemic significance of the first few no document mortgage loans?
Thanks to Mike Santoli for the article and you, Barry, for the amplification. This is a column that should probably be repeated annually (something like Vermont Royster’s Christmas editorial in the WSJ)
People generally just seem to want to believe whatever the talking heads say. I can’t understand why, since many are intelligent and would never base business decisions on such claptrap.
The book “How to Lie with Statistics” is obviously someone’s bible.
Another statistical story. High School students with larger feet get higher SAT scores. Is this a causal relationship? Nooooo!. Students with the largest feet are generally seniors. They have probably taken the SATs at least once, if not twice. Each time they take it they usually do better.
BTW, any way of getting the complete Mike’s column without being a subscriber?
I regard all statistical analysis and forecasting thusly: A man with one foot in a bucket of ice and the other foot in a bed of hot coals is, on average, very comfortable.
Has the indicator turned into a self fulfilling prophecy and how likely are market participants to fade the potential?
I’d add three:
1. Consider the source. Even totally neutral sources (if there is such a thing) introduce some degree of normal human bias.
2. Consider positive and negative feedback loops embedded in the argument, and whether and when inflections might turn the argument on its head.
3. Consider that if the pattern was formerly unobserved, but is now widely observed, people are likely to act in anticipatory ways which change the timing, if not the outcome itself.
This site is always great, but this article was especially poignant. Thanks!
“Statistics are used much as a drunk uses a lamp post: for support, not illumination.”
— Vin Scully
“I know of someone who drowned crossing a river with an average depth of 6 inches.”
— Unknown
If an indicator has, in the past, been such a good predictor of short term moves in the market as to reach the level of “statistical significance”, and if that indicator were to become widely recognized as having reached that elusive threshold of statistical significance, the indicator would cease to be effective once the majority recognized its prior effectiveness. Thus, anyone waiting for statistical certainty is certain to be disappointed.
One example of some relevance is the VIX. It has become so widely watched that a level which previously would have indicated a market turnaround is unlikely, at present, to result in that turnaround.
Marcus Aurelius,
It seems to me that instead of Doppler radar, the Holy Grail of the market pundit’s quest is a doppleganger radar that displays historical twins.
Here’s one that’s important for quants:
* Observe all points in your analysis where you assumed that events were independent (in the probabilistic sense). Re-examine whether this assumption is still valid given the current market environment. If it is not, how does this effect your calculations?
Historical market data relevance is skewed by the rapidly increasing amount of information available, the expanded universe of investment vehicles, and the speed with which information is disseminated.
Predictable physiological reactions to events may take unpredictable forms in the market.
Three things to do to discover the direction of the markets.
1) Watch the Godfather trilogy on Sunday
2) Jump down turn around lift a bail of cotton
3) Analyize the entrails of owls
Works every time
AlB,
This is a very interesting topic, because I am a fascinated observer of those who inject opinion with a degree of certainty. That alone often works to distort an analysis of cause and effect.
So, I want to ask you a few questions and observe your response. I’m curious to know that if you are able to be objective, rational and unbiased, if I might change your mind about something.
Let’s proceed:
Is it rational and logical that, on average, those with larger feet would have also grown taller as well?… and that, given that their feet are larger than some other persons of the same age with rather similar or identical genetic lineage, that there must have been some reason that they alone are taller with bigger feet?
Could it be that they may have been better nourished as growing children, and thus had better brain development?… Fewer fever days (fever hampers growth) because of better health and possibly the economic status of their parents who might have better afforded health care than the parents of little-feets?… and also likely better afforded higher quality educations for their children?
If identical twins are separated at birth and then mysteriously reunited just as they’re taking the SAT, on average would you expect the child with the larger feet to score higher, or the one with the smaller?
Has any of this information caused you to have any doubts about your theory that it’s only because the children are repetitive testers?
What if I added that it is very unlikely that repetitive testers will, on average, score higher than singular testers who are more intelligent, better educated and take the test once with larger feet?
Have I changed your mind at all? You are my philosophical experiment for the day.
Barry, I have argued for a long time that it’s nearly impossible to determine what has been a direct correlation due to economic policy and what has been merely coincidental as a result of that policy.
2 primary examples are:
1) The economy soared in the 1990’s after Bill Clinton’s economic programs took effect, effectively the most prosperous of times in our nation’s history. The Clintons (and many of my Democrat friends) will argue that it was their policy that was responsible for the economic improvement (i.e. fiscal responsibility, balanced budgets etc etc). I argue that it was merely coincidental, that the economy soared despite Bill Clinton’s policies not because of them, when 3 new world-changing technologies (the internet, the PC in every home, the rollout of the cell phone)and the economy would have been great. Microsoft and Intel did more to stimulate the economy than Bill Clinton did. Not to mention the peace dividend at the time, with the cold war being over. A monkey could have been President and the economy would have done well in the 1990’s.
2. Most if not all of the base of the Republican Party will say that the economy post-9/11 did very well because of the Bush tax cuts. I argue that the economic strength could have been due entirely to the bubble-producing policies of Alan Greenspan’s 1% Fed funds, and to the huge economic gains in BRIC which helped the exporting sector of our economy immensely. It’s entirely possible the Bush tax cuts benefitted only the top 1-2% of the wealthiest Americans and did very little to stimulate the economy. I’m wondering if the Gates, Buffetts, Rockefellers and Waltons of the world really did go out and spend just so, so much more money stimulating the economy just because they had a big tax cut? A monkey could have been President and things may have worked out just the same. My own opinion is a monkey would be a better President than George W. Bush has been.
Mere coincidences? I think that’s highly likely.
The Limits of History
STREETWISE
By MICHAEL SANTOLI
PEOPLE WHO TRY TO HANDICAP THE MARKETS for a living practice the art of the plausible. Many trudge from conference room to lunch table to banquet hall lugging PowerPoint decks full of unobjectionable statistical touchstones for commission-wielding clients. At times of investor confusion and market dissonance, such as now, their art is often reduced to carving a slice out of economic history that ratifies their existing outlook.
Here’s a dog’s breakfast of the kinds of historical analogies making the rounds, drawn from a non-scientific culling of the old e-mail inbox from the past couple of weeks.
• Every time the U.S. unemployment rate has risen by at least 0.3% in a month, as it did in December, a recession has occurred. So a recession is a sure thing.
• Prior to the onset of recession, there’s typically at least a 25% one-year rise in weekly unemployment claims. The increase in claims in December was less than 7%. So a recession is not imminent.
• But, then: The average decline in the S&P 500 from a pre-recession peak to a trough since 1945 has been 25%, just a few percent more than the index had lost from its 2007 peak to its intraday low Wednesday of last week. So, maybe the market has mostly discounted a typical recession scenario?
• Hold on, because a bearishly inclined forecaster suggests that in recessions, 70% of the prior bull market’s upside is undone, implying another 20% or so downside risk from today’s levels. Scared yet?
• Not so fast, because every one of the 23 times since 1987 that the ratio of bears to bulls in the weekly American Association of Individual Investors poll has exceeded two (as now), the market was up 12 months later, by an average of 21%.
Such evidence of history’s malleability is almost enough to make us all post-modern, and deny that any objective reality exists. More to the point, it’s a little different every time, and the key is to figure out that difference.
A lot looked familiar about the selloff and ferocious rebound in stocks Tuesday and Wednesday. The S&P touched a supposedly key level — 1275 — and fear hit a crescendo. Purchases of protective options surged to gigantic levels, and the resulting bounce was so violent that only a precious few lucky-more-than-smart traders could have caught it.
By recent standards — if you’ll pardon our own stab at historical analysis — the market appears to have hit a pretty respectable low that won’t give way quickly without a heaping helping of nasty news. This column suggested in recent weeks that because stocks’ prior range was breached, the panic and oversold conditions that exist before a durable low arrives might have to be even more extreme than that seen in 2006-’07. Only in some respects was this standard met.
DOW JONES REPRINTSThis copy is for your personal, non-commercial use only. To order presentation-ready copies for distribution to your colleagues, clients or customers, use the Order Reprints tool at the bottom of any article or visit: http://www.djreprints.com. • See a sample reprint in PDF format • Order a reprint of this article now.
On the brighter side, it became commonplace later in the week to call it a mere bounce, a bear-market rally, mostly short-covering, nothing that would lead to anything interesting on the upside, and so forth. These are defensible conclusions, for sure, but it’s positive that the prevailing mood is one of caution and mistrust of the rally.
Public investors, at any rate, are in a dark mood; they yanked some $13 billion from stock mutual funds in the past week. The sharpies at “quant” hedge funds have suffered another setback (coincident with interim market lows last year). But corporate executives continue to refuse to sell their shares in important amounts.
None of these clues will matter next week if stocks buckle under from an expected cavalcade of important news reports. Some 200 major companies report earnings. Downward earnings revisions for the second half of 2008 appear inevitable.
But stocks have already been repriced to reflect expectations that published estimates are too high. The current S&P 500 forecast could be chopped by 10% and still market valuations wouldn’t appear extreme at current prices. But the market’s real-time response to muted corporate outlooks will deliver the true verdict on whether the market has built in enough fear.
The Fed’s next interest-rate decision will focus much anxiety on a single moment Wednesday afternoon. There’s quite an argument raging on the Street as to whether the unpredictable Bernanke is behind the curve or already has cut too much. History (again!) tells us we’ll only know in retrospect.
Consider how this headline captures the present moment: “A ‘Global Margin Call’ Rocks Markets, Banks — Stocks Drop World-Wide; There’s Sober News, Too, About the U.S. Economy — Bond Yields Keep Dropping.”
It appeared Aug. 28, 1998, in the Wall Street Journal. We now know that the Fed’s response to that crisis generated excess liquidity and an eventual bubble. At times late last year the markets seemed to be betting on something similar happening, with “global reflation” plays leading. If only it were more plausible.
I doubt there is enough historical data, but there certainly seems to be correlation between “the shit hitting the fan” and articles in the Dallas Morning News:
Quote: “Zale Corp. said Tuesday that it will close 60 stores in the next 90 days, joining a growing list of retailers announcing post-holiday store consolidations.
The just-completed holiday shopping season showed the worst sales growth in five years. Since then, retailers have slashed earnings outlooks, seen their already depressed stock prices collapse further and announced a steady stream of store closings.
Zale also is cutting its capital spending to $85 million this year from a planned $100 million, he told the Cowen & Co. annual consumer conference in New York.
In addition to Zale, home furnishings retailer Ethan Allen Interiors Inc. said Tuesday it will close 12 stores and two service centers, cutting operating costs amid a slowdown in the housing market.
Last week, Liz Claiborne Inc. said it’s closing its 54-store Sigrid Olsen chain.
On Jan. 4, Talbots Inc. said it’s closing 66 Talbots Kids and 12 Talbots Men’s stores.
On the same day, Pacific Sunwear of California Inc. said it’s shutting the remaining 154 stores in its demo chain.
Three days after Christmas, Macy’s Inc. said it was closing nine additional stores, including one in Dallas at Valley View Center, and some analysts predict more are likely.
The latest closures follow a slowdown last year in new-store expansions by Wal-Mart Stores Inc. and Home Depot Inc. Wal-Mart cut its plans for 2008 growth in square footage to 4 percent from 7 percent.” End Quote.
However, try as I may, I have not been able to find a causitive link between tightening credit standards and debt-expanding economic recoveries, declining asset values and inflation, or hopes for a second half rebound and reality alterations.
Barry,
Coudln’t agree more with most of these. However, I take issue with one statement; 200 day moving averages are not as neutral as you imply.
~~~
BR: You misread the details — I was referencing the % of NYSE stocks over their 200 day moving average. Its a sentiment measure. (The moving avg is actually irrelevant)
~~~
(Back to Byno)
You discuss statistical significance, but there’s nothing to suggest that buy and hold over long time periods is statistically different than using moving averages of any kind. In fact, the evidence is to the contrary; on the whole, moving averages are inferior to buy and hold by an order of magnitude over anything like a thirty year time horizon.
Furthermore, there’s no reason to believe that the 200 DMA, 50 DMA, 50 EDMA, 50 SDMA or any other such measure is statistically different than a 219 DMA, 167 DMA, or the like.
The one constant is that moving averages work during periods of high PE ratios. Otherwise, moving averages, MACD, RSI, etc are bunk. And, when taxes, transaction costs, and slippage are factored in, you’d be better putting your money in the Wilshire 5000 index fund and going to sleep till you retire than trying to game the market during anything other than high PE (when PEs rise above nineteen and stay there till they break below ten again) cycles.
I have yet, in all my years of analyzing the markets, discovered any indexing strategy that beats buy and hold unless the markets are at PE ratios as detailed in the previous paragraph. Not inflation, not treasury spreads, not dividend yields, not AAII/AIMR/IBD/Marketvane/Whatever sentiment ratios; not anything.
Truth be told, if we are at the beginning of the end of Bubble v 2.0, and P/E ratios dip below 10 again in the next few years, you’re going to have to find much more value-added content for the site, because otherwise, the difference between buying a basket of Vanguard funds and gaming the market will have nothing but delitirious effects for individual investors.
“It isn’t what we don’t know that gives us trouble, it’s what we know that isn’t so.” – Will Rogers, quoted by Gerald Weinberg in “An Introduction to General Systems Thinking.
A summary of Weinberg’s summary of what to about this (“the Systems triumvirate”):
1) Why do I see what I see?
2) Why do things stay the same?
3) Why do things change?
It’s a denser book than say “How to Lie with Statistics” or “How To Solve It”, but it has helped me in my attempt to uncover my own bias in evaluating things.
Remember that chartism is always of limited validity.
Why? Because charts compress (more technically project) information from an infinite dimensional space onto a finite (usually 1) dimensional space, and hence lose information.
Another reason (related) charts only tell you about what has happened in the past, if new phenomena occur there is nothing the chart can do to help you understand the effect of that phenomema.
So, for example, 9/11 (my turn to play Rudy Guilliani :) ) No chart anywhere ‘predicted’ 9/11, and certainly not any financial chart. All the chartists were busy extrapolating curves and then all of a sudden… wham. The chart changes. Of course the humans know why — there was this uh event. But to the chart its just another weird curve shift. One which if you now incorporate into your statistical model will to some extent give you bad predictions because the chart, it’s derivative, it’s statistics, is not causative of the phenomena. The chart only reflects the phenomena.
So in the current era does the chart tell us anything? Well it’s going to reflect ‘local’ information, and local of course will depend on the kind of real world phenomena that affect the price — some may have longer lead times than others — but all of it will only indicate what someone knew or believed in the (recent past). If we are entering a period where the fundamentals are changing, if the stock run-up of the last 5 years has been due to phenomena that stop occuring (fraud), more generally if the position of the US in the world, and our economic fundamentals have changed in big way, if the political realities of a war, external debt, oil dependence etc. have not yet been ‘discounted’, if they suddenly reach their ‘critical point’ where the equilibrium of the past is destabilized, then past price information cannot tell us anything about future prices.
All these stats and charts are linear guesstimates that assume an otherwise unchanging world. And guess what?
The world, she just keeps changing.
Byno,
You make some excellent points and the only one I would question is the level of PE multiples that you find significant for differentiating the best strategy of buy and hold vrs moving averages.
I’d prefer to call moving averages something else, maybe active participation, or management… but it’s still the same thing… buy and hold vrs some form of input of intellectual skill.
You’ve identified the PE break at 19 and generally that this level or lower ought to cause one to opt for buy and hold, certainly over a 30 year period you say.
However, aren’t you being a bit short-sighted about the history of PE multiples?
For example, for me, were PEs to be under 12, I might play the index and make like Rip Van Winkle, but I have no such confidence in under-19 for that strategy as you do.
These discussions on TBP usually end up in a debate about the potential for, and probabilities of, market PEs of 12 or under, 9 or even as low as 7.
19 down to 12 in so short a time as to not be counterbalanced by earnings increases over the same period would be quite painful to a contented buy and hold investor. And what of 9?… To be absolutely confident that we’ll not see a 9 anytime in the immediate future might just represent optimistic willfulness for its own sake.
I would take a line from Andy Grove/INTC that used to be quoted,it feels like, more than today:
1. Only the paranoid survive. Consider not only the many things that you may not know, but also the many things that others are actively doing that will upset your view.
technical analysis is not a crystal ball.it’s more like an instrument panel or the dash board in your car.distance,speed rpm,temp,volts,etc.it’s a very good tool if it’s the right way.
Several years ago one of the Federal Reserve Banks published a paper by one of its statisticians/analysts. I think it was Kansas, but I no longer remember exactly.
It is important to emphasize that this paper was an academic one, NOT an FRB position paper. The FRB published it as UC Berkeley Press publishes scholarly books regardless of what UC officially thinks.
The gist of the paper was that there are too few stock market statistics for mathematical reliability. The author (sorry, I don’t recall her name) said that such statistics are based on Ibottson data, some 70 years worth at the time. She asserted that a sample size of 70 was insufficient for statistics.
My impression is that market “statistics” are too much like sports data. Neither means much though both do make you sound knowledgeable.
Barry,
A great posting for me since I try to use a methodical, analytical, empirically based approach and it’s a lot easier to make a mistake than to be correct.
I’ll repeat what I learned in my 400-level Sociology statistics class a long time ago ( maybe it was the Pleistocene, maybe the Jurassic, I forget ). To prove causality, you need three things:
co-variation ( aka, correlation )
time order ( to cause B, A has to happen first. This is the only thing you and the other posters have omitted. It so obvious that analysts almost always forget to confirm it but sometimes it is not present. If you are going to assert that higher unemployment leads to lower sales of widgets, make sure unemployment has gone up in the past before the widget sales go down. )
non-spuriousness ( which is impossible to prove; it is what most of your points are concerned with by discussing types of spuriousness, eg. the Super Bowl indicator )
Byno,
I meant to say “no confidence in anything at or higher than 19.”
In other words, I consider 19 to be an historically high PE multiple.
I’d add three more:
1. Whether correlation or causation, how strong is it w.r.t not just the direction but also amount of movement ? If a 5% change in X is correlated/causatively to a 15% change in Y, check on correlation between a 10% change in X with the change in Y ? Does it taper off ? Does it invert ? Does it non-linearly increase ?
2. Beware of “ceteris paribus” i.e. “all other things being equal. A historical correlation/causation between 2 variables ( or a n to 1 pattern ) exists in the context of other excluded variables having specific values. Those variable may have quite different values today and screw up the correlation.
3. The market is reflexive in the George Soros sense of the word – my way of putting it is that the market “learns”/ market participants “learn”. Just for that reason alone, be extremely beware of applying past historical patterns to current data and predict forward. The market also “forgets”.
-K
My fault Barry.
Not trying to bust your balls, as this is one of the five or so websites I read daily (and, at that, more than once a day).
If I misread it, that’s on me. I’m someone who, if you’re not hearing a peep out of them, you’re running on all cylinders.
Eclectic: there’s nothing magical about a PE of 19. The only reason I even use that as a benchmark is that it is one standard deviation away from the 140 year + average of 14.5.
What I will say is that using a moving average strategy – whether it be 50 day, 73 day or 42 day exponential – keeps you out of all hell breaking loose when the markets do get wacky in terms of valuations.
The moral of using the moving averages is that there is no bell that rings when the markets hit PEs of 19 and up. Only that, you can sit on really good stocks during bull markets that begin in the single digit PE range, but you must be especially vigilant in high PE markets.
STatistically, there’s no reason to believe a 20 DMA is better than a 50 DMA. But, using either would have kept you on the sidelines as the Great Depression broke, the 60’s and 70’s bear market decimated portfolios (especially in inflation-adjusted terms), 1987, 1998, 2000, and the most recent debacle.
Again, Barry, my apologies for misreading the gist of your post. IMWorthlessO, best blog on the net, just trying to make things as airtight as possible. Byno
No linkfest this weekend? Hope you had a good one Barry.
>> However, try as I may, I have not been able to find a causative link between tightening credit standards and debt-expanding economic recoveries, declining asset values and inflation, or hopes for a second half rebound and reality alterations.
>> Posted by: Winston Munn | Jan 27, 2008 3:21:15 PM
You must try harder, Grasshopper! ;-) (Just kidding. Loved your sarcasm.)
>> 2 primary examples are:
Todd, “bravo!” (applause). Great examples.
Great post. A couple of comments:
1. Various ways to measure central tendency – mean, median and mode. Each has pros and cons and can be “biased” depending on the …
2. Distribution – most stats are based on a Bell Curve or normal distribution. If you can’t provide a basis for the normative assumptions, you must use a distribution-free or non-parametric test; otherwise your conclusions are dubious
3. Most data patterns are either linear, curvilinear or non-linear – like costs in accounting, it depends on the time period of interest (costs could be fixed, variable or mixed, e.g.)
4. Look at outliers … very much related to central tendency – similar to non-recurring costs
5. Extraordinary claims need extraordinary evidence
6. The key is to know what someone’s interests are – are they trying to sell you something – they will naturally frame their presentation to compel a certain action. Key question: What’s missing? What one thing would have to be true for them to be wrong?
7. present:not present (if X is present; when is Y not present – find an example). not present:present (if X is not present; when is Y present)
8. What kind of logic is being used? Inductive, deductive, analogic, etc. There are assumptions w/ each. Inductive can always be wrong – the “black swan” issue. Deductive – what are the premises? Analogic – do the compares have materially relevant similarities?
9. Case studies are often wrong. Many stories are built on them and often have compelling detail, but they are often part of a selling frame that is not generalizable to other situations.
I have more, but those are top of mind.
MP
Thanks Byno,
Plus or minus 1 standard deviation from the mean ought to represent approximately 67% of all distributed observations of magnitude of random events.
However, the upper one-half of that 1 S.D. may have no negative consequences relative to the occurrence of events that vary from the mean.
What I mean by that, in reference to your theory, is that if one uses a buy and hold strategy for an indexed broad market investment when the PE multiple is *at* 14.5, then one-half of 67% is about 34%, and this means that in approximately 34% of all times, the PE will reside in the range of 14.5 PLUS 1 S.D.
Such upper range will be complimentary to the strategy, and even more complimentary when it exceeds 1 S.D.
However, the other 34% of times will result in a PE that is within the range of 14.5 MINUS 1 S.D., and those will be quite detrimental to the strategy, and even devastating to it if greater than -1 S.D.
Math and statistics teacher by profession and I’ll just skip to my personal 3 as I learned them and do believe
1) Among who? Every single data point could potentially predict quite well some future behavior. What if the data only tells me what poor southern women will vote. Doesn’t help
2) Barry touched on this one, how often? If the possibility of it happening is low, why am I betting on it? Likely cause I’m stupid.
3) How much? If all of Barry’s list are true and all of the additions, what is the actual consequence? You could have “perfect” data, but the consequence 1 additional cold in Wisconsin.
I think too many people care first and ask questions later esp. chart porn. I assert you should ask the question, do I really give a shit first. That’s the hardest question to ask in this information age.
Todd- It could be argued that if Clinton had used the first Trade Center bombing as impetus to re-invade Iraq, the 90s economy would have looked very different. It did matter who was in the White House.
…and more on-topic from my previous post: always take a high-pass view; was it Tukey who said something like, “it’s better to be sort-of right than exactly wrong”?
The Use and Abuse of Data: Two Excellent Commentaries
In the current issue of Barron’s Mike Santoli has some special insight for those trying to interpret the many historical analogies making the rounds. He writes: • Every time the U.S. unemployment rate has risen by at least 0.3% in