Data Science goes Whooooo!

The only function of economic forecasting is to make astrology look respectable.

John Kenneth Galbraith

Putting the above statement to the test, at the start of this year, BBC Business Daily program chose to forego the opinion of the world’s leading economists and visited the flourishing business of financial astrology. Listen to their program here.

The show started with a fascinating interview with Cathleen Skinner. The uncertainty generated in 2016 boosted demands for her services as a financial astrologer. Check her website here. She has published a book where she discusses the correlation of stars to movements of the world stock markets.

Being data driven, she studied the 10 greatest stock market scenarios from the 20th century and their corresponding astrological situations. One thing that stood out was that whenever one of the outer planets moved past a particular zodiac constellation, the markets got a big “wobble”. This was true in 1907, 1929, 1957, 1987 and so on. In 2017, Pluto is going past the same “sensitive” part of the sky, oh dear. To make matters worse, at the same time Saturn will spend 3 months passing through the Galactic Center (whatever that is) and that is also correlated with poor stock market performance.

But surely, one opinion is certainly not enough. Let us look at global pseudo-scientific research. According to the Chinese calendar, this is the year of the fire rooster, always an uncertain sign.

Skeptics point out that if you want to see a pattern, you will see a pattern in almost anything. Isaac Asimov claimed that he could make any trend fit any set of arbitrary dates and demonstrated this with the dates corresponding to the sighting of Halley’s comet making it a portent of doom.

To confuse matters, there are generally accepted to be real cycles in the economy. Business cycles are well-studied if not necessarily well understood. People have been finding periodic rises and falls for a long time. Short term cycles of 4-8 years have entered popular culture and spawned research groups that specialize in them. More controversially, there is also some evidence for longer term cycles.

In 1939, Soviet economist Nikolai Kondratiev published an observation that all sectors of the world economy underwent cycles of expansion, stagnation and recession over a 30-40 year period. Nobody knows why this happens. Explanations range from technological revolutions to demographics to debt deflation. But the phenomenon itself was dubbed by Schumpeter as “Kondratiev waves”. Here is how that idea has progressed since then.

kondratievCycle.png

Kondratiev Waves

On the other hand, pseudo-science adopts the trappings of science without the spirit of skepticism and self-criticism. A famous case of pattern imagining was reported by Richard Feynman in the book Surely You are Joking Mr.Feynman where he spoke of a scientific paper in the early  20th century which claimed that the human sense of how quickly time passed was linked with iron in the body. What the author had done was to ask his fever ridden wife to count while he measured her temperature. He then plotted how her sense of timing changed with her temperature and compared it with the properties of various elements from a chart. Some propertied of iron seemed to fit the “data”, so iron was determined to be the cause.

Final thought, whenever confronted with any data driven result, go through the fine print. Astrology is bad data science codified.

Advertisements

One response to “Data Science goes Whooooo!

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s