Can I still use regular statistical releases in a Covid-19 world?
Amid the economic and demographic uncertainty caused by the spread of COVID-19, we continue to receive statistical updates and releases from the ABS and other sources. Many people are wondering which of these releases can still be used with confidence, and which we should take with a grain of salt. In this blog, Glenn provides some guidance around two significant releases we’ve seen recently – Overseas arrivals and departures, and Labour force statistics – both of which have been significantly impacted.
We often get people lamenting the fact that statistics take time to collate and publish, to the point where they are no longer valid by the time they’re available. In most cases, this isn’t true. We use Census data a lot, and despite being more than 3 years old, many demographic characteristics, particularly for local areas, are still very relevant, and this is the best source of information we have for those communities at the moment. Having said that, there are some statistics at some times, that seem very out of date by the time they are published. The ABS in the last week have put out two of those.
In the current COVID-19 pandemic, the goalposts are changing on a daily basis, so these two publications can’t really be taken as representing the current reality, even though they’re only a month or so old.
Overseas arrivals and departures
The first was the monthly “Overseas Arrivals and Departures” collection.
This was released last week for the month of February. It shows a substantial drop in short-term visitor arrivals to Australia, down 26% from the previous February.
Source: ABS, Overseas Arrivals and Departures, Feb 2020 (3401.0)
But, really, the big story here is the country of origin. February was when Covid-19 was mainly in China, particularly Wuhan, and the Federal Government closed our borders to anyone who’d been in China, while other countries were still allowed in. This is very evident from this graph.
Source: ABS, Overseas Arrivals and Departures, Feb 2020 (3401.0)
Small drops from most other countries, but a 90% fall in Chinese arrivals, reflecting the time when the travel bans took place.
Now as we know, in March, Australia took the unprecedented step of closing our borders entirely to international migrants, so expect the March figures, due for release on May 12, to reflect a much larger drop (borders were still open in early March), followed by the April figures which should be close to a 100% drop.
ABS has acknowledged this and there are some interesting articles on their website about the drop in international students, among other things.
Some other datasets are being released in this area, not all by the ABS. For instance, the Department of Home Affairs have a quarterly update on the number of temporary Visa holders in Australia.
Labour Force statistics
The second publication is the monthly Labour Force release (Labour Force, Australia, 6202.0), where we get the national unemployment rates. This survey of 25,000 households per month is a critical economic and social indicator and forms the basis for the quarterly LGA unemployment estimates which are included in our economic profiles.
This shows national unemployment steady at 5.2% in trend terms, and only a very slight increase in seasonally adjusted terms. This might be perplexing, given the barrage of news in late March about economic shutdowns, particularly in tourism and airline industries, as well as retail, and huge job losses associated with these. The important thing to remember with the labour force survey is the reference period. It’s conducted every month by phone interview, in the second week of the month, with the reference period being the previous week (either first or second week of the month depending on the interview date). So this all relates to the first week of March, when, though the coronavirus was a concern, most businesses were still open, and the major shutdowns were yet to occur. So the major effects will show up in the April release, due out on May 14th.
The ABS have been quite up-front about this too. When the social and economic landscape is changing so quickly, it’s hard for the stats to keep up! This is where economic modelling can come to the fore (for those who missed it, you can watch a recording here of the webinar presented by .id’s consulting economist Rob Hall, sharing some modelling that team are working on to help councils understand the impact of Covid-19 on local economies).
The ABS are also conducting a number of smaller, additional surveys, specifically assessing the impact of Covid-19 on households and businesses. These have a much smaller sample size (~1,000 households rather than the 25,000 in the labour force survey), and are designed to produce rapid statistics at a national level – they can’t include any local information (even for states), but are being released much more rapidly than anything the ABS has done before. One of these came out on Monday. Covid-19 impacts on Australian households. This shows, among other things, a 3% reduction of the proportion of population who had a job in early April (not the unemployment rate, as this is calculated differently).
How to stay up-to-date with these releases
More of these rapid release small surveys are coming out regularly now – the best way to keep up to date with them is to stay subscribed to our blog for analysis from our teams. You can also subscribe to our New data and features email list to be notified when new data and features are published in our public resources.
What are ‘Trend terms’?
Seasonally adjusted figures make adjustments for effects in the data which occur regularly every year. Examples of this include an increase in the workforce in agriculture during fruit picking season, or a reduction in the number of student arrivals in Australia during the Christmas school and university holidays. The trend series takes this smoothing out a lot further, and is used to emphasise long term trends over short term random variability (the irregular component).
The ABS use a technique called the “Henderson Moving Average” to weight the last 13 months (that’s more than a year) of data, with more weighting on the last few months, relative to the earlier months, to derive a smooth trend over the long-term. This means the most recent months are subject to review, as new months of data come into the mix, and you can see trends which are currently in the future.
The ABS recently abandoned the trend series for overseas migration – because of the sharp shock of the current border closures, the long-term trend becomes largely irrelevant, and would put too much weight on the past to have any relevance. Trend series are good at removing random effects and seeing the pattern in slow, longer term change, but not good with rapid fundamental change in a characteristic.
Glenn Capuano - Census Expert
Glenn is our resident Census expert. After ten years working at the ABS, Glenn's deep knowledge of the Census has been a crucial input in the development of our community profiles. These tools help everyday people uncover the rich and important stories about our communities that are often hidden deep in the Census data. Glenn is also our most prolific blogger - if you're reading this, you've just finished reading one of his blogs. Take a quick look at the front page of our blog and you'll no doubt find more of Glenn's latest work. As a client manager, Glenn travels the country giving sought-after briefings to councils and communities (these are also great opportunities for Glenn to tend to his rankings in Geolocation games such as Munzee and Geocaching).
