Gig workers boosted employment in August: Justin Smirk
EXPERT OBSERVATION
It is fair to say that the 111k or 0.9% rise in employment was a big surprise. The market had been looking for –35k while Westpac was forecasting –50k.
We had based our forecast on the observation of a –1.0% fall in Weekly Payrolls in the first two weeks of August compared to the first two weeks in July (the Labour Force Survey is conducted in the first two weeks of each month).
Even when you take seasonality into account the 0.4% rise in employment in original terms was still much stronger than the –1.0% decline reported by Payrolls.
'Gig' workers driving recovery in employment.
In the August Labour Force Survey the ABS provided a reconciliation on the difference between the two series highlighting that Weekly Payrolls measure 'employee jobs' for which a payment was reported to the Australian Taxation Office through Single Touch Payroll. This includes jobs held by employees and owner managers of incorporated enterprises (the latter of whom can be paid through the payroll of their business). However, owner managers of unincorporated enterprises are not covered by Weekly Payrolls. There is also lower coverage of small businesses (
In the Labour Force Survey employed are categorised according to their 'Status of Employment' reflecting the nature of the employer-employee relationship. The categories are:
- Employee – someone who works for an employer;
- Owner manager of an incorporated enterprise – self-employed who owns and operates their own incorporated business;
- Owner manager of an unincorporated enterprise – self-employed operating a business as a sole trader; or a,
- Contributing family worker.
Owner managers (ie. self-employed) may or may not have employees.
In original terms, the Labour Force Survey reported a 44.5k/0.4% increase in employment in August. In the release the ABS noted that the increase in original employment was due to a 50.2k increase in owner managers (self-employed mostly unincorporated sole traders) without employees. This is likely to be dominated by 'gig workers' which can be defined by anyone who is engaged as a contractor rather than an employee. This includes far more people than just those who work as the large aggregation platforms such as Uber, GoCatch, Deliveroo, Airtasker, Freelacer etc. In contrast, there was a small 2.6k lift in the number of employees which was more than offset by a –9.3k decrease in the number of contributing family workers.
Excluding 'gig workers' employment was still stronger than the Payrolls measure of jobs.
If we exclude the 50.2k rise in self-employed, total employment would have fallen –5.7k which is closer to, but still stronger than, the –1.0% fall in Payrolls. Unfortunately, the data underpinning this analysis won’t be released until Thursday so we can only note the ABS comments at this stage and add that to the data we do have.
Seasonality matters in August with employees contracting an average of -66.5k in last five years.
From the Labour Force Survey we know that August is a soft month for employment with the ABS seasonal factor falling -0.5% in the month. What we did not realise is that that seasonality is different for employees and owners/self-employed. On average through the last five years the number of employees fell -66.5k in August while the number of self-employed lifted 24.1k. That is to say it was not the strength of owners/self-employed that was surprising in the August survey but rather the lack of weakness in employees (+2.6k in the month compared to the average of -66.5k).
We can't exactly replicate the ABS season adjustment process but using eX12 we estimate that seasonally adjusted employees gained 78.8k in the month while the self-employed lifted 18.7k. The real story in August was not the surprising strength in the self-employed, at least half of which was a seasonal norm, but rather the lack of weakness in employees which would have significantly boosted the overall seasonally adjusted total.
Shutdown in Vic hit that state but the NSW recovery continued on regardless.
The shutdown in Victoria did hit the labour market in that state with employment falling –42.4k and unemployment lifting 0.3pt to 7.1%. Victorian employment is now just 9.5k/0.3% higher than the May low post the first round of shutdowns. The recovery in Victoria peaked at +51.9k/+1.6% from May to July.
The recent Weekly Payrolls data suggested the NSW recovery was running out of steam through July and into August but this did not show up in the Labour Force Survey. In August, NSW employment lifted 51.5k following on from a 56.2k gain in July and +80.3k in June. In the month, the unemployment rate fell 0.5ppt to 6.7%. The NSW labour market continued to power on in August even if the Victorian lockdown was clouding the economic environment.
The smaller states continue a robust recovery.
Elsewhere, employment rose 18k in Qld, 32.2k in WA and 18k in SA. It was a similar story for unemployment where it fell 1.3ppt in Qld and WA to 7.1% and 7.0% respectively while it was flat in SA at 7.9%.
Hours worked lagging in NSW and Victoria but rebounding elsewhere.
Even with the August surge in employment (+111.0k/0.9%) hours worked lifted just 0.1% due to a –4.8% decrease in hours worked in Victoria (employment fell –1.3%) while the rest of the country recorded a 1.8% increase in hours worked. In the month, hours worked per person fell –0.8% driven mostly by a –2.3% fall in hours worked per part-time employee. Hours worked per full-time employee fell –0.2%. Nevertheless, in August the overall weakness was less concentrated in hours worked than it was back in April as hours worked per person is still 2.9% higher than it was in April (we note that hours worked per person recovered a touch in May even though employment fell further). Hours worked per full-time employee is up 1.9% since April while hours worked per part-time employee is up 15.7% again highlighting that the hit from Covid, and recovery to date, has been concentrated more in part-time employment than full-time.
As it did in April the July/August shutdown in Victoria had a larger impact on hours worked than it had on employment. Nevertheless, given the recovery in hours worked through June and July Victorian hours worked are still 0.5% higher than the May low, a touch better than the +0.3% gain for employment. In NSW hours worked lifted 0.7% which is softer than the 1.3% rise in employment. Victorian shutdown did create some uncertainty in NSW but it was expressed more as a reduction in hours worked than a reduction in employment.
For the smaller states hours worked per person continued to recover in August. Hours worked lifted 2.4% in Qld compared to a 0.7% gain in employment, hours lifted 1.8% in SA while employment lifted 1.6% and over in WA, a 2.4% lift in employment generated a 3.8% surge in hours worked. WA stands out with the strongest recovery so far with a 5.8% increase in employment and a 8.5% increase in hours worked from the shutdown lows.
The Victorian shutdown again hit participation while a recovery continued elsewhere.
The weakness in Victoria also led to a fall in participation in that state, from 64.5% to 63.9%, which muted the rise seen elsewhere. Nationally participation rose just 0.1ppt to 64.8% (despite the outsized gain in employment) with the labour force lifting 24.5k. This muted rise in the labour force resulted in a significant correction in the unemployment rate with it falling to 6.8% from 7.5%. Back in May the unemployment rate was 7.1%; it was 6.4% in April.
Despite the employment gains in Qld participation in this state fell -0.5% to 63.9% but this was more than offset by a 0.5% gain in NSW to 65.2%, a 1.0% gain in SA to 62.8% and a 0.6% gain in WA to 67.9%. We are looking for a bounce in participation in Victoria and Qld in September but it is more unclear what to expect from the other states.
It is very interesting to note that participation in NSW and WA is almost back to the pre-Covid levels in March, down -0.10ppt om NSW and -0.06ppt in WA. In Victoria the participation rate is still -2.5ppt below its March level while in Qld it is still down -2.0ppt. The recovery in participation in NSW and WA suggest that the recovery has drawn workers back into the workforce and should we see any weakness in the labour markets in those state, there is the potential for correcting weakness in participation. Contrast that with Victoria and Qld where the economic weakness has held participation well below pre-Covid levels. As such, as these economies recover then participation is likely to pick-up preventing a more rapid fall in unemployment.
This suggests that there may be still sufficient flexibility in participation to buffer the unemployment rate from any significant change, both up and down, from moderate changes in level of employment. It will, however, depend on which states the weakness or strength appears. If the strength appears in NSW or WA, then there is less scope for participation to lift thus driving unemployment lower. On the other hand, if further weakness appears in Victoria or Qld then there is less scope for participation to correct lower and thus it is likely to drive unemployment higher.
Underemployment was flat in August.
The underemployment rate was flat a 11.2%, at two decimal places it was a 0.7ppt fall from 11.23% to 11.16%, so close to rounding down to 11.1%. From a record peak of 13.4% in April underemployment has eased back as the number of people working less hours than they normally do improved. In August almost 1,000k employees worked less than their usual hours for economic reasons with around 215k (or 22%) doing no work at all; over 760k worked some hours but fewer than usual. Following a 700k increase between March and April the number working zero hours for economic reasons fell by over 600k between April and July. However, in August the number working zero hours increased by almost 50k (+30%).
By state underemployment fell 0.9ppt in NSW to 9.8%, lifted 0.8% in Victoria to 13.8%, fell 1.3ppt in Qld to 9.3%, gained 0.4% in SA to 11.4% and fell 1.0% in WA to 9.6%. It is hardly surprising that underemployment rose in Victoria and that it is the state with the highest rate but it is interesting to note the strength in WA and that underemployment there is now lower than it is in NSW while Qld has the lowest rate for a mainland state.
The flat print on underemployment is a further sign that while August may have been a robust month for employment and unemployment, it was not such a good month for economic activity with hours worked almost flat in the month with a solid rise in the number working zero hours for economic reasons and flat underemployment. If you are looking for signs of weakness it is not the rise of self-employed but rather the lift in those working zero hours for economic reasons.
The August survey confirms a more robust recovery in the labour market than many had expected.
This update has made it much more difficult for the RBA to hit its forecast peak in unemployment of 10% by end 2020. We estimate that to achieve that forecast rise in unemployment, employment would have to collapse between -400k to -500k by December; of course it will be dependent on the size of the fall in participation as employment contracts) This would be a very extreme situation and suggests an aggressive unwinding of support packages, such as JobKeeper, a massive rise in business closures and significant job shedding. The rise in the number working zero hours is a sign that things might not be a robust as a 6.8% unemployment rate suggest and we do note that that conditions and payments for JobKeeper have been modified from end September. However, the programme has been extended to March 2021 and we do expect to see further support should conditions deteriorate further, just as we have seen Victoria as conditions deteriorated there. We are not expecting the Government to go 'cold turkey' on JobKeeper and suddenly cut the programme.
Westpac is retaining our current forecast peak in unemployment this year of 7.8% in December which implies a loss of total employment of around 130k to 170k. Without a doubt, the recovery in the rest of the country has been more robust (so far) than we thought while the hit to the Victorian labour market was not as bad as expected. However, with the rolling off or adjustment of many of support packages (in particular JobKeeper) between now and the end of the year there is clear meaningful risk of a further decline in employment and a rise in unemployment.
JUSTIN SMIRK is a senior economist for Westpac
