New research helps predict stock market

February 20, 2009,

stock market
(PhysOrg.com) -- Researchers from Massey University have developed a new way to predict stock markets that has been recognised with an award from New Zealand finance specialists.

Professor Ben Jacobsen, Associate Professor Ben Marshall and Dr Nuttawat Visaltanachoti have found that analysing data on a daily basis or other shorter intervals - rather than monthly - offers a much higher success rate of stock market predictions.

Traditionally, institutional investors, such as hedge funds or mutual and pension funds, try to predict stock markets one month ahead relying on information also measured at monthly levels.

The academic researchers have moved away from that convention with what they have described as “amazing results”.

The research was awarded best investments paper by the Institute of Financial Professionals at the New Zealand Finance Colloquium this month.

The team studied how changes in prices of energy, such as oil, and industrial metals, such as aluminium and zinc, strongly influence world stock markets.

Dr Marshall says the paper, Return Predictability Revisited, is an eye-opener both practically and academically. “While the change of intervals seems innocent enough, this new approach suggests that stock market returns are much more predictable than previously thought. Economic variables that seemed unimportant now may warrant a second look, but measured at a different interval.

“Professional investors in Europe are already starting to implement this approach. Theoretically, the study is of interest because it shows how an innocuous change in observation intervals has a dramatic impact on the way we think about financial markets in the academic world.

“With hindsight, it seems surprising that many professional investors and academics alike have overlooked the possibility that the interval of observation would make a huge difference in trying to predict markets.”

Dr Marshall says the traditional approach to use monthly observations to predict monthly returns is a “crude” approach based on convention rather than theory. “When predicting March, we are saying look at the last week or even the last day of February because that could be more relevant,” he says. “Using a full month of data to predict may increase noise levels and understate - or even fully mask, the actual predictability present.”

It is the third time Professor Jacobsen has received the investments paper award and the second time for Dr Visaltanachoti.
Researchers from Massey University have developed a new way to predict stock markets that has been recognised with an award from New Zealand finance specialists.

Professor Ben Jacobsen, Associate Professor Ben Marshall and Dr Nuttawat Visaltanachoti have found that analysing data on a daily basis or other shorter intervals - rather than monthly - offers a much higher success rate of stock market predictions.

Traditionally, institutional investors, such as hedge funds or mutual and pension funds, try to predict stock markets one month ahead relying on information also measured at monthly levels.

The academic researchers have moved away from that convention with what they have described as “amazing results”.

The research was awarded best investments paper by the Institute of Financial Professionals at the New Zealand Finance Colloquium this month.

The team studied how changes in prices of energy, such as oil, and industrial metals, such as aluminium and zinc, strongly influence world stock markets.

Dr Marshall says the paper, Return Predictability Revisited, is an eye-opener both practically and academically. “While the change of intervals seems innocent enough, this new approach suggests that stock market returns are much more predictable than previously thought. Economic variables that seemed unimportant now may warrant a second look, but measured at a different interval.

“Professional investors in Europe are already starting to implement this approach. Theoretically, the study is of interest because it shows how an innocuous change in observation intervals has a dramatic impact on the way we think about financial markets in the academic world.

“With hindsight, it seems surprising that many professional investors and academics alike have overlooked the possibility that the interval of observation would make a huge difference in trying to predict markets.”

Dr Marshall says the traditional approach to use monthly observations to predict monthly returns is a “crude” approach based on convention rather than theory. “When predicting March, we are saying look at the last week or even the last day of February because that could be more relevant,” he says. “Using a full month of data to predict may increase noise levels and understate - or even fully mask, the actual predictability present.”

It is the third time Professor Jacobsen has received the investments paper award and the second time for Dr Visaltanachoti.

Provided by Massey University

Explore further: Detroit auto show, and industry, prepare for transition

Related Stories

AI can help retailers understand the consumer

January 14, 2019

Consumer brands and retailers often struggle to fully understand ever-changing customer needs. That is why you mostly find XL sizes in your favorite fashion store and no M sizes. That is why you have to spend hours looking ...

Could Tesla price cuts mean demand is slowing?

January 3, 2019

Tesla made about 9,300 more vehicles than it delivered last year, raising concerns among industry analysts that inventory is growing as demand for the company's electric cars may be starting to wane.

Recommended for you

Frog choruses inspire wireless sensor networks

January 21, 2019

If you've ever camped by a pond, you know frogs make a racket at night; but what you might not know is how functional and regulated their choruses really are. Frogs communicate with sound, and amid their ruckus is an internally ...

3 comments

Adjust slider to filter visible comments by rank

Display comments: newest first

Soylent
5 / 5 (2) Feb 20, 2009
The recursive nature of stock markets makes them exceedingly difficult to predict. Your model needs to be designed to include the behaviour of other stock market participants. If other stock market participants begin using your model their actions will distort the stock-market so that your predictions become less accurate. It's red queen's race, need to keep running as fast as you can just to stay in the same place.
googleplex
4 / 5 (1) Feb 20, 2009
I agree.
It is like a technological war between large trading houses. Yesterdays models won't work today. They hire the brightest academic minds to build the models.
Another point is that if the model is so good how come they aren't billionaires already instead of wasting time writing research papers.
It is no suprise to me that a 30 day moving average is worse at forecasting tomorrow than a 7 day moving average. Any serious investor knows that moving averages are not designed for daily predicting a spot price. They are designed to signal a change in trend and figure out support levels/trading ranges etc. Trend is your friend.
menkaur
not rated yet Feb 21, 2009
The recursive nature of stock markets makes them exceedingly difficult to predict.

not difficult.... it's impossible...
once research on some regularity in market becomes public, it's useless
so, i'd call the article bullshit...

Please sign in to add a comment. Registration is free, and takes less than a minute. Read more

Click here to reset your password.
Sign in to get notified via email when new comments are made.