sign uplog in
Want to go ad-free? Find out how, here.

Massey's world-first, machine-learning GDP predictor "GDPLive" got an intense lesson on how to handle the unexpected in 2020. Christoph Schumacher looks at how it has handled the challenges

Massey's world-first, machine-learning GDP predictor "GDPLive" got an intense lesson on how to handle the unexpected in 2020. Christoph Schumacher looks at how it has handled the challenges
Photo by from Pexels

When the government releases their previous quarter GDP estimates, we always check to see how close our machine-learning predictions were. Prior to the pandemic hitting our shores and our economic fortunes, our forecasts have been very close to official figures, and in many cases spot on.

This time, however, our Q4, 2020 quarterly and annual growth predictions of around -4% and -3.7% respectively were more pessimistic than the just-released values of -1.0% (quarterly) and -2.9% (annual). What could have caused this discrepancy? There are three points worth highlighting.

First, GDPLive is powered by a machine-learning ensemble algorithm that is very good in making predictions when there is sufficient past data for the algorithm to learn from.

Things, however, get difficult if there isn’t enough training data.

The lockdown happened in Q2, 2020 and until today, the government has only released Q2 and Q3 quarterly and annual GDP growth rates. GDPLive therefore only has two historic values to learn from since the economy was shut down – not enough for consistently good predictions. From a research point of view, it highlights areas were machine-learning needs further development as more traditional human-based forecasting models might still have the upper hand in predicting very sudden, unusual, data-poor events.

Second, during the lockdown, Kiwi’s consumption patterns changed almost overnight. One of the big changes was the large increase online spending brought about by retail closures and the need for social distancing. This trend is likely to continue even though we are, for most of the time, back to business-as-usual. GDPLive is partly powered by in-store sales data which means that a change in spending behaviour could negatively influence the prediction quality.

Third, the figures released by the government are estimates of GDP and often get revised at a later stage. These revisions can be substantial and are not that well-advertised. This means that we only know how good or bad our predictions are when the revisions have been made.

GDPLive is a world-first, machine-learning GDP predictor. Like us, it is trying to understand the impact of the events that unfolded in 2020. As more data becomes available, the more it can learn and the better the predictions will be.

*Christoph Schumacher is Professor in Innovation and Economics at Massey University and Director of the Knowledge Exchange Hub. This article is a post from the GDPLive blog, and is here with permission. The New Zealand GDPLive resource can also be accessed here.

We welcome your comments below. If you are not already registered, please register to comment.

Remember we welcome robust, respectful and insightful debate. We don't welcome abusive or defamatory comments and will de-register those repeatedly making such comments. Our current comment policy is here.


My guess is the machine is likely to be on more on track than the government estimates. Just a hunch though!

To use the famous analogy, even a dart throwing chimp could forecast or estimate better than the Govt.
While machine learning and AI has surpassed human domains of complete information (eg chess) it is still second best to talented humans in all areas of incomplete information. That is why every game of complete information has basically been "solved", whereas despite numerous efforts spanning many years poker is nowhere near solved and AI is still a distant second to (the best) humans.

Actually there are several cases where machine learning algorithms have proven to be more accurate than trained doctors at detecting diseases from various sources of medical imaging.

The reason being? - diseases, such cancer cells growth is linear, progressively up and can even be accelerated with more detriment factors, such as ageing, drugs habit, alcohol/regular toxin injection etc. so the robotic scan of the regent comparative analysis with other variables? this will improve much of the predictability/calculation, alleviate the 'human error factor', regent directly being scrutinized by machine.. mostly progressive, hardly produce regressive results, BUT if that to happen?/as per treatment aka towards remission.. say after the chemo/radio therapy, periodical reduction treatment will also be calculated in predictable way slow/intensity down. As per lab test monitoring regime.
The real challenge of AI, is to be self autonomous reacting to unpredictability, random behavior of the up, down, left, right, pause, etc.
Most in current NZ govt? do listen to scientist/trained medical fraternity for C19 (Dr, Prof etc.) But for Economy? nope, as the stake is too high when listened to the Prof, Phd of macro/micro Economics - That's why if not mistaken only two Phd employed with the current govt staffing structure..none of them in Economics (do correct me if this is wrong).. one in Geography & one in Theology? - greed & fear are their fearsome ruler..(ok let's name them, the OZ banks cartel) bhua ha hah..

Perhaps they could try offering the machine a lifetime supply of chocolate

No surprises here. Artificial intelligence is not real intelligence. It's essentially just a whole bunch of if statements that can be fed huge data sets, then reproduce the same patterns in a predictive format. Works well with fixed variables liked language and chess, works terribly with human behaviour where the variables are infinite and illogical.

It cannot predict human nature, not can the RB trying to manipulate markets