Yesterday’s long term chart from UBS’ Peter Lee started a conversation with a friend who is a (pure) technical trader about long term trend lines.

Some of the more bearish technical traders use a trend line that connects the 1929 low and the 1943 low — projected forward 60 years, it forecasts SPX 400.

That trend derived forecast has several problems: First, its so old, that small errors are greatly magnified. Like a satellite off by a few degree, by the time it reaches Pluto its nowhere near where its supposed to be. Same issue with projecting out decades. Also, the post-1929 lows might be considered aberrational.

Perhaps most importantly,  trend lines are validated by the number of touches they make — the more the better — and so tagging not only 1942 (and near ’49), but ’73, ’82 and 2009 increases the validity of the trend line.

Funny thing is, I found a chart that does that — again, by UBS Peter Lee:

>

click for ginormous graph

>

Source:
2011 Technical Market Outlook
Peter Lee – Chief Technical Strategist
Wealth Management Research
December 2010

Category: Markets, Technical Analysis

Please use the comments to demonstrate your own ignorance, unfamiliarity with empirical data and lack of respect for scientific knowledge. Be sure to create straw men and argue against things I have neither said nor implied. If you could repeat previously discredited memes or steer the conversation into irrelevant, off topic discussions, it would be appreciated. Lastly, kindly forgo all civility in your discourse . . . you are, after all, anonymous.

17 Responses to “Long Term Trend Lines (SPX 1929-2010)”

  1. machinehead says:

    Another way to confirm a trendline is to fit a regression line to the data, then draw upper and lower channel lines above and below the regression line (but parallel to it) so as to touch extremes.

    The white channel line comes closer to that approach than the green line, although its slope still looks a little deficient.

  2. Jack Damn says:

    Nice. Another great looking chart. I’m not really a fan of trend lines (too subjective at times), but I’m a big fan of mean line reversion. Good stuff.

  3. Super-Anon says:

    I’m wearing my DOW 6800 cap today!

  4. constantnormal says:

    You guys really ought to take a look at dshort’s approach to this (hat tip to veneziano for the link in the earlier discussion) — he presents an inflation-adjusted view of this data, runs an Excel regression line through the middle of it, and examines the over- and under-shoots made around the line.

    MUCH less subjectivity to that approach, and the overshoots and undershoots all (except for the last dip) fit within a similar band of values.

    The question is, does washing out the noise of inflation increase the reliability or not? Another question might be: Should we be artfully drawing lines connecting peaks and valleys, or should we be drawing a single regression line through all the data and examining the overshoots and undershoots? Which way results in more accurate guesses at the near-term future? It is possible to back-test these methods and arrive at a conclusion, but I have never seen anyone try to do so.

    Of course the intent behind nearly all the drawers of these charts to to present a picture that their already-formed notions can agree with. Objective analysis has little to do with it, or we would see lots of predictions and tests of hypotheses. Instead, we just see lots of pictures.

  5. Mike in Nola says:

    Has anyone considered that this trendline is constructed to cover a period when the US became the dominant military and economic power in the world, starting from the point that it was the only major power not having its homeland devastated by WWII the “arsenal of democracy (with the attendant economic benefits) and contiuing threw the “sole superpower” status of the 1990′s and early 2000′s?

    I think most of us agree that that situation is changing with the rise of China and the Chinese taking us to the cleaners in manufacturing, outsourcing of other high tech jobs to India, and the bubbles which have sapped our financial strength and sent our accumulated treasure overseas. And, more recently, the revelation that, for example, Chinese ballistic missles and keep our enormously expensive aircraft carriers out of its sphere of influence if a showdown comes.

    Assuming that the trendline was valid during during the prior period, will it still be?

  6. Woof says:

    Really interesting to see over long time period, but the trend lines drawn don’t represent ‘mean reversion’ since they connect the lows.

    I could only find data back to 1950 from Yahoo! Finance and I’m sure the data may have other limitations I don’t understand. However, when I let Excel draw a trendline on on the semi-log chart, it suggests the we are currently at or below the ‘mean.’

    Movement lower would be movement away from the mean trend, not towards it.

    Sorry if I’ve unintentionall demostrated my ignorance, unfamiliarity with empirical data, etc., etc.

  7. Jack Damn says:

    @constantnormal – I have post dshort’s stuff here also. Good stuff. His “regression to trend” charts are excellent.

  8. machinehead says:

    ‘Assuming that the trendline was valid during during the prior period, will it still be?’

    Nothing guarantees that a linear trendline must continue forever. Several stock markets in the 20th century actually closed their doors and went dark — due to war, revolution, etc. (or in the case of the Nasdaq, because squirrels gnawed through a data cable in Trumbull, Connecticut one day in the Nineties).

    Mystical sacred trendlines didn’t stop their demise.

  9. Woof says:

    One thing Barry can ding me for is that he never said the trendlines represented regression to the mean.

    That came out of the comments. So put me down for “creat(ing) straw men and argu(ing) against things (Barry) neither said nor even implied”

    Sorry.

  10. seana0325 says:

    This is the ‘king of all’ chart porn!

    I can barely contain my erotomania…

  11. dead hobo says:

    If you mean by all of this that what has happened most often in the past is likely to occur in the future, assuming no convoluted logic or is involved, then you are probably correct.

    To me, I see we are entering a spot on average period, historically, and will probably remain average for a long time unless economic conditions change. Thus, confirmation of my ‘beginning of a buy and hold period’ is above.

    I personally see the market congestion of late as withdrawal symptoms from the first absurd Fed pump that allowed the markets to rise when horrible news was constant and frequent. I believe that pump was just a back door method to re-capitalize the banks, and Wall Street fucked over the Fed by keeping the cash for commissions rather than put it back into the loans.

    Wall Street crooks are confused because the really easy money has left the building. This pump is being managed in some manner to minimize the most egregious skimming. Stock picking by professionals is now required to make a good buck and all successful markets have both winners and losers. The herd is mostly looking for leadership, aka milling around aimlessly hoping for a fad to follow. Again, this is also known as a healthy market. Remember , most of the people you see and read are chiefly salesmen with little more insight into money than anyone else, and only because of an affection for investments. For each and every breakout, some clown will say the end of the world is coming.

    Drawing lines using mathy ideas and reasonable sounding thought for support is helpful if kept simple, and idiotic when it is used to claim the future is predetermined.

  12. egghead168 says:

    Great chart.

    I added another line to the chart and it looks even more interesting.

    http://egghead168.blogspot.com/2011/01/s-long-term-trend.html

  13. JimRino says:

    Let’s go super-macro.
    - Population: What happens when we reach “Peak Population”. Typically populations go Boom and Bust just like markets, and we are in the midst of a 110+ year boom, that may last 25 more years. At some point, there will be a bust.

    This trend line may just be a bunch of points on a graph, with reality bearing down on us soon.
    What is wealth? In a crisis you can’t sell gold coins. If we ignore peak population, peak water, peak oil and climate change, in a crisis you won’t need gold, you’ll need a .32.

  14. dead hobo says:

    Woof Says:
    January 7th, 2011 at 12:51 pm

    Sorry if I’ve unintentionall demostrated my ignorance, unfamiliarity with empirical data, etc., etc.

    reply:
    ———–
    Not at all! You did great!. Never let anyone who makes a living by investing tell you that their ideas are better than yours IF your ideas come through effort, analysis, and intelligent attempts to figure things out. You are attempting to know the unknowable. This also means your ideas are as good, and maybe better, than people who refer to themselves as Wall Street Pros, although the wall streeters will ignore you dismissively. Screw them. You should always think of anyone connected with Wall Street or investing in general as someone who wants to separate you from your money. Not all do, but enough work that way to make them all look suspicious.

  15. dead hobo says:

    BR stated:

    Funny thing is, I found a chart that does that — again, by UBS Peter Lee:

    observation:
    ——————–
    Excellent example of confirmation bias.

    In reality, nobody on earth nows how good this chart is, although it probably implies a lot of reasonable ideas. (supports my confirmation bias, too)

  16. JustOne says:

    There are a number of ways to view long term data. Removing the bias of improved technology, inflation, population growth, etc. might prove interesting. What if you normalized the data to precious metals, or a commodity price blend, or energy costs, or GDP, or Global GDP (many of the companies in the S&P are international), or the integral of free cash flow to OPEC nations, or ??? There are a lot of underlying trends that are folded into the price of equities.

    Guess that proves that any fool can ask questions faster than a thousand wise men can answer. ;-) Hope some of the experts here can provide links to other views of this data.

  17. cognos says:

    So… is this line and chart useful?

    In 1990 it would’ve provided some “justification” for saying… “were at 400 on SPX, we should come back to 200-300 and touch the line in the next 5 years”.

    Oops!

    I suspect the outcome will be very similar from here.