Stimulus blunder Down Under sends Big Data alert

Racing to get ahead of virus-induced economic calamity, Prime Minister Scott Morrison in March rushed out one of the world’s boldest stimulus measures at the time

  
People stroll through a park in front of the Sydney Opera House amidst the easing of the coronavirus disease (COVID-19) restrictions in Sydney, Australia, May 20, 2020.

People stroll through a park in front of the Sydney Opera House amidst the easing of the coronavirus disease (COVID-19) restrictions in Sydney, Australia, May 20, 2020.

REUTERS/Loren Elliott

HONG KONG  - A happy accident in Australia provides a valuable lesson in humility for data worshippers everywhere. The country’s treasury officials just discovered that a major stimulus programme they initially expected to cost A$130 billion ($85 billion) will only set taxpayers back A$70 billion. Studying the sequence of bad inputs that led to the blunder could help others avoid less fortunate outcomes.

Racing to get ahead of virus-induced economic calamity, Prime Minister Scott Morrison in March rushed out one of the world’s boldest stimulus measures at the time. The so-called JobKeeper plan to subsidise wages, as announced, amounted to 6.5% of GDP. And it was just one piece of a costly package.

Politicians can hardly be faulted for acting aggressively under the circumstances. Australia’s relied on epidemiological estimates that so far have turned out to be exceedingly pessimistic about the country’s infection and death rates. That in turn misinformed assessments on hospital beds, unemployment and more.

Garbage data entry exacerbated the error. Of the 910,000 small businesses that requested benefits, about 1,000 botched the forms badly enough to cause considerably more than a rounding error. Many with only one employee reported that they had 1,500, confusing the headcount for the A$1,500 figure to be paid every two weeks per worker. In the end, some 3.5 million people will be eligible instead of the 6.5 million forecast. The final tab is expected to nearly halve.

Australia’s mistake was lucky; it only briefly tied up funds on paper that might have been better allocated. Officials elsewhere, however, could easily miscalculate in the other direction, as they process mountains of forms submitted by distressed executives or jobless individuals.

Noisy data distorts economic and financial models far and wide, and this pandemic is causing a din. When decisions are hurried, or based on bad numbers, the consequences can be dire. Unreliable information on private companies in China, for example, complicates estimating their stress. Countries with large informal economies such as India are even more difficult to measure. The promise of Big Data is enchanting, but the Blunder Down Under sends a clear warning signal.

CONTEXT NEWS

- Australia’s Treasury and Taxation Office said on May 22 that the cost of the country’s JobKeeper wage subsidy programme would only be A$70 billion ($45.7 billion) instead of the A$130 billion originally forecast because of recently discovered reporting errors by small businesses.

(Editing by Pete Sweeney and Sharon Lam) ((jeffrey.goldfarb@thomsonreuters.com; Reuters Messaging: jeffrey.goldfarb.thomsonreuters.com@reuters.net))

More From Economy