What a simple statistical investment tool can tell us about the climate
Guest essay by Eric Worrall
One of the simplest statistical tools used by investors is a moving average plot. If you plot the average share price of a company, or other investment product, with different smoothing periods, on the same graph, a crossover between the different plots can provide early warning of an imminent change in trend – a buy or sell signal.
From Wikipedia:-
“A moving average, as a line by itself, is often overlaid in price charts to indicate price trends. A crossover occurs when a faster moving average (i.e., a shorter period moving average) crosses a slower moving average (i.e. a longer period moving average). In other words, this is when the shorter period moving average line crosses a longer period moving average line. In stock investing, this meeting point is used either to enter (buy or sell) or exit (sell or buy) the market.”
http://en.wikipedia.org/wiki/Moving_average_crossover
The interesting thing about moving averages is they can provide useful, actionable information, without requiring any knowledge of the nature of the underlying commodity.
So what happens if we create a moving average plot of global temperature (in this case Hadcrut4)?
Source: http://www.woodfortrees.org/plot/hadcrut4gl/mean:60/plot/hadcrut4gl/mean:120/plot/hadcrut4gl/mean:240/plot/hadcrut4gl/mean:360
What is immediately apparent is we may be in the early stages of a significant inflection point – it is too early to tell for sure, but the beginning of an inflection which appears to centre on the early 2000s seems very similar to the inflection which occurred in the 1940s, heralding decades of cooling temperatures.
If the numbers I was plotting was the value of an investment, I would interpret the chart as a strong “sell” signal – a warning that a substantial drop could be imminent.
UPDATE: A previous post on WUWT by David Dohbro also comes to the same conclusion, and has a more detailed analysis:
http://wattsupwiththat.com/2013/10/01/if-climate-data-were-a-stock-now-would-be-the-time-to-sell/
Anthony,
Kindly keep the Wall Street chartists where they belong: in a halfway house for recovering snake oil peddlers.
REPLY: Kindly run your own blog, and I’ll run mine with things I think are interesting. -Anthony
Applying different averages like that is simply a linear filter of some form. Unfortunately with the stock market, a lot of people are applying any and every form of mathematical analysis, and if any new form of mathematical analysis is found to work, it is immediately applied to maximal extent, such that it no longer works.
Using moving averages in investing has the distinction that changes in price trajectory will automatically be missed until the algorithm catches up. Many people claim predictive ability with technical indicators, but remember that if price is going up today, there is a probability it will go up tomorrow unless some piece of random negative information hits the market.
Matt is correct in his observations.
I am not sure why Doug Huffman listed Taleb’s books, but Taleb will tell you real quick that you cannot predict the economy or stock prices. Statistical algorithms have no predictive value in those complex systems.
If you look a the sea level chart, you’ll see a similar “sell” pattern. It looks like the sea level is potentially in a dropping trend…
Agreed. Statistical algorithms in themselves have limited, if any, predictive value. However, they can be very powerful when used in combination with other analytic tools and modeling engines.
Bring on the cold! Would love to see a mini-ice-age hit New York, Washington D.C., Chicago and Washington state along with Harvard and Yale where this stuff is being taught. Put up giant billboards saying “I TOLD YOU SO!” everywhere.
fhhaynie says: @ur momisugly January 30, 2014 at 1:45 pm
Warren Buffet is a buddy of Obama, but if he is like Dwayne Andreas (Archer-Daniels-Midland Company) and perhaps America’s champion all-time campaign contributor, he has politicians from both sides in his pocket. (ADM is a big winner in the corn==> biofuel scam.) Believe me the big money bags are going to get advanced notice BEFORE the plug is pulled on the CAGW scam so they can get out before the rush.
Here is an example with some of the same players. I expect a repeat of the same game plan.
Technical Analysis of stocks is of a chaotic market. No one can predict the peak or the bottom. What it does is try to identify trends and crossover points where the trend changes. Nothing to say the trend will not change again.
Daryl Guppy’s Multiple Moving Average though gives unequivocal turning points because the volatility is covered by the widest bands. When they narrow or cross, there is something of interest happening.
Bob says: January 30, 2014 at 12:56 pm “I am not sure why Doug Huffman listed Taleb’s books, but Taleb will tell you real quick that you cannot predict the economy or stock prices. Statistical algorithms have no predictive value in those complex systems.”
That is precisely why! And he had much more to say, shouting over the blurring line between mathematical logic and quantitative analysis. Everyone here should read at least his popularized books and, if one’s technical background is strong enough, his academic papers.
He quotes his collaborator Benoit Mandelbrot that reality is fractally complex, beggaring “complex systems.”
Thanks for that. Here are a couple of comments I made on that thread, about 4 months ago:
Doug Hoffman: I agree, and did not articulate my thoughts well. I have read two of the Taleb books, and have two more in Kindle form. I will read them as I get the chance. Taleb does not mince words!
I have taken a look at a couple of his papers, but don’t have math background. It will take a while to “book” it through the papers.
What about charting the share price moving averages of the listed green technology companies … along with their subsidy funding … 😉
I would bet on technical analysis over climate modeling any day!
I dabbled in technical analysis of stocks some years ago. I learned they have no real predictive value. In other words, a great way to lose your shirt.
However some of the techniques are good tools for visualizing trends and noise in data. As visualization tools they can have some value. I use a variant on Bollinger Bands for example, in some process control metrics.
Just don’t hope to predict tomorrow.
Matt says:
January 30, 2014 at 6:44 am
Matt, there is a reason that technical (chart) analysts wear bad suits. If their analysis worked they would be rich.
Doug Huffman says:
January 30, 2014 at 2:34 pm
——————————————————-
Taleb’s Black Swans don’t really have a place in climate science. The climate is not interested in one time events (well it wasn’t before extreme weather, but we digress). The big problem with mathematical analysis in climate is the data. It is incomplete, it has been tampered with and we don’t understand the underlying physics (well enough) to attribute theory to observation.
For example, the use of Fourier transforms to describe non-periodic events. When the mathematical analysis becomes so complex that even the scientists using it disagree with each other – well there must be a problem with the data.
“So what happens if we create a moving average plot of global temperature (in this case Hadcrut4)?”
In this case, you are working with a chart with feloniously exaggerated data points (jail time if this was the TSX). The Sell order was probably years before that. Don’t forget that Hansen in 1998 believed we had to be going to finally beat the mid1930s/early 40s all time high in the US. When it didn’t happen, he re-rigged the earlier data downward in a few tranches by several tenths C.
In Canada, he forgot to re-rig the record at Yellow Grass Saskatchewan where the July 5th, 1937 record still stands:
“In more than 120 years of official meteorological observation, Saskatchewan has captured some remarkable weather records. The highest temperature ever recorded in Saskatchewan—indeed in all of Canada—was observed at Yellow Grass and Midale on July 5, 1937, when the temperature hit 45.0°C. The lowest temperature ever recorded in the province is -56.7°C, observed at Prince Albert on February 1, 1893.
http://esask.uregina.ca/entry/meteorology.html
Wow – I sure hope this guy isn’t handling other people’s money.
[snip]
Don’t pooh-pooh this too fast. Whenever there’s a change of direction occurring or developing, the short-term average responds faster than the long-term, necessarily. That shows up as a crossover.
Eric – thanks for the article. No disrespect, and no hard feelings but I already did a MACD analyses of Hadcrut4 data last year and it got posted on WUWT here: http://wattsupwiththat.com/2013/10/01/if-climate-data-were-a-stock-now-would-be-the-time-to-sell/
and a quick search -to do your due diligence and homework – and referencing it in your article would have been appropriate. Btw, the MACD is still on a sell since 2007 for all databases (GISS, NCDC, Hadcrut)
To all those commenting here thinking that Technical Analysis is bogus and doesn’t work; you obviously don’t know what you are talking about and I am 1001% sure you don’t even trade. What would you recommend for trading and investing? Just buy and hold? Good luck… When do you buy, when do you sell? Oh and nothing is random in the market. Try Fibonacci extensions for a change, the market loves those… Fibs are not random. The whole universe, our world, everything is Fib related. So are the financial markets.
MACD, FSTO, MFI, EWT, S/R, Pivots, BBs, RSI, MAs, etc, etc are all KEY to investing. You need to be able to track, follow and understand them all to be successful. TA, and TI are KEY to investing. ALL traders and investors use them; from minutes to yearly time frames. None-traders, dumb money, retail investors, moms and pops don’t use them because they haven’t learned to understand them. You must fall under one of those categories, otherwise you wouldn’t state what you stated. Please step aside and leave the real trading up to the pros. No in fact please trade, I’ll take all your money; ’cause if somebody looses money another one pockets it… So please, keep on investing… Saying TA is voodoo-magic you may then very well side with AGWers… Sorry but you really don’t know what you are talking about.
You can see what you want in those curves. How is the plot now much different from what you would have seen in 1980 or 1985 with a ‘crossing’ followed by a sharp up-tick? It’s because we know there are other factors all moving in the right direction to support the downturn it is more credible now than those previous cases. Nevertheless, it is still a bet. In fact, as others have said, we don’t know what will happen.
david dohbro sorry for missing your article – I would have referenced it had I recalled it at the time of writing my post. It is a fascinating field, and I thoroughly recommend anyone interested in more depth view your post – it is a far more detailed analysis than my WFT effort.
http://wattsupwiththat.com/2013/10/01/if-climate-data-were-a-stock-now-would-be-the-time-to-sell/
But the AMO was negative all through the 1980s and up to the mid-1990s.
Why don’t you contact James Annan to see if he is still taking climate bets.