COVID-19: Real-time economic data is in high demand but short supply

Amidst the daily barrage of dismal economic headlines, it is incredible to think that we won’t officially know whether Australia has entered a recession until September when the June national accounts are released.

Of course, the die has already been cast and the monthly jobs data will be as good an economic indicator as any of our economic health prior to September. But even then, the full story will only gradually reveal itself. The March survey out in mid-April will reflect the first two weeks of March before significant restrictions on social gatherings started. The April job numbers will provide a better picture, including what kind of an impact the government’s $130 billion wage subsidy is having on keeping people attached to employers.

None of this is to take anything away from the Australian Bureau of Statistics, who are already responding with a suite of additional more regular statistical releases to measure the economic impacts of COVID-19. Last week it released new business indicators showing that around half of all businesses had been adversely impacted by coronavirus and 86 per cent of businesses are expected to be adversely impacted in coming months. It also confirmed that accommodation and food services businesses will be amongst the hardest hit, with over three-quarters already having been impacted.

These sorts of real-time indicators that are available paint a sobering picture of the decline in daily economic activity and the implications it is already having on Australians’ livelihoods. Long before the complete closure of restaurants to seated diners, Australian restauranteurs were seeing a precipitous decline in patronage according to OpenTable data. According to data from the City of Melbourne, pedestrian foot traffic in popular retail spots has slumped. At midday last Saturday, the usually bustling northern end of the Bourke St Mall in Melbourne saw 551 pedestrians an hour, around a fifth of what would usually be expected.
 

The sharp decline in economic activity resulting from important restrictions is hitting Australian workers hard, evident in the long queues at Centrelink last week. While the immediate volume of these queues clearly took some by surprise, Google Trends data foreshadowed the spike with interest in the search term “newstart allowance centrelink” growing twenty-fold in the first three weeks of March. The hoarding we’ve seen at supermarkets also belies a more fundamental funk for consumers, with the ANZ-Roy Morgan weekly consumer confidence dropping to its lowest level since 1973 when inflation hit double digits.

Having reliable sources of real-time or even near-time data is critical in a crisis. Just as health officials need disaggregated detail on coronavirus cases to combat its spread, policymakers need to understand the progression of adverse economic impacts into the community. This allows economic policies to be targeted and gives real-time feedback on whether existing policies are going to be enough.

As we have seen with the government’s three stimulus packages announced over the space of just 18 days, perfect is the enemy of good when it comes to responding to an unfolding crisis. There will inevitably be unintended consequences and other wrinkles in policies revealed – the key is to identify them as quickly as possible and remedy them.

We can look to examples here in Australia and overseas for encouragement on how we can better exploit data that is already collected, for more practical and timely insights on how the economy is performing and how policy must respond.

The Commonwealth Bank’s Household Spending Intentions series combines the bank’s customer spending data with publicly available Google Trends search activity to identify prospective spending behaviours of households. Their February release provided early evidence of a sharp decline in travel and education spending intentions.

But better data doesn’t just give us a better handle on how the economy will perform in coming months. We can also use it better to test economic theories and policies. In 2018 as debate raged over the merits of corporate tax cuts, AlphaBeta used anonymised data from Xero to examine the short-term effects of the already legislated tax cuts for small business. It found that firms receiving tax cuts hired additional workers, but that there was little impact on wages and only a slight impact on investment.

As many Australian economists have pointed out in the last week, the United States releases weekly updates of unemployment insurance claims. Their release for the week ending 21 March showed claims up over three million from the previous week.

In Australia similar numbers are released on a monthly basis, with the latest data available from the Department of Social Security for December 2019. It is therefore unsurprising that there is a growing appetite for more regular labour market indicators in the current climate. As CEDA Director and labour market economist Professor Jeff Borland recently suggested in a research note, there may be value in providing extra funding to the ABS to conduct a twice‐monthly Labour Force Survey given the pace of change we are likely to see in Australia’s labour market over coming months.

The digital world is now believed to contain 44 trillion gigabytes of data. Unlocking a bit more of that data in a timely fashion would go a long way to better anticipating events and quickly assessing the effectiveness of our policy responses. In the short-term the government, private sector, not-for-profit sectors and researchers should work together wherever they can to share real-time data in this crisis. 

Beyond that, we all have an obligation to get behind government efforts to unlock data for public benefit, including through the national data commissioner and the legislative frameworks being established by the Federal Government.