The Australian Labour Market And The Early Impact Of Covid 19 An Assessment
Summary
The author manuscript of an article by Jeff Borland and Andrew Charlton, The Australian labour market and the early impact of COVID-19: An assessment, accepted for publication in The Australian Economic Review and held in the University of Melbourne's Minerva Access repository. The article reviews labour market developments in Australia from March to June 2020, the main drivers of the downturn, and how impacts differed by type of job and worker. It reports that monthly hours worked fell by 9.5 per cent from March to April and that the adjusted unemployment rate averaged 10.0 per cent for April to June. It also makes an initial assessment of the JobKeeper program. The manuscript ends with appendix tables on employment changes by age, gender, occupation and employment status, and a figure comparing hours lost with JobKeeper claims by industry.
Summary drafted by a model from the document's text below and checked by script against that text before publication. It is a navigation aid, not a reading of what the document proves. Where AI is used
Cite this document
“The Australian Labour Market And The Early Impact Of Covid 19 An Assessment.”
Original at doi.org. Archived at Pandemic Darlings, https://pandemicdarlings.com/source-documents/the-australian-labour-market-and-the-early-impact-of-covid-19-an-asses-1d34d68b98/
, SHA-256 caf062fb20c89a42cc506bf8c2f9aba04f9f272a5049757478adf0c497ad7f13.Full text
Minerva Access is the Institutional Repository of The University of Melbourne
Author/s:
Borland, J;Charlton, A
Title:
The Australian Labour Market and the Early Impact of COVID-: An Assessment
Date:
--
Citation:
Borland, J. & Charlton, A. (). The Australian Labour Market and the Early Impact of
COVID-: An Assessment. The Australian Economic Review, (), pp.-. https://
doi.org/./-..
Persistent Link:
https://hdl.handle.net//
Author Manuscript
The Australian labour market and the early impact of
COVID-19: An assessment
Jeff Borland and Andrew Charlton*
Abstract
From March to June 2020 was the most dramatic four months in the history of
the Australian labour market. Never before has a such a substantial decrease in
labour demand (and partial reversal) occurred so quickly. In this article, we
present an overview of the early impact of COVID-19: the main drivers it brought
into play and the consequent labour market developments. Aggregate effects and
how impacts differed by type of job and worker are described. We conclude with
a brief review of the main government response to COVID-19, the JobKeeper
program.
*Borland, University of Melbourne, Victoria, 3010, Australia; Charlton,
AlphaBeta, Melbourne, Australia. Corresponding author: Borland, email <
jib@unimelb.edu.au>. We are especially grateful to Ian McDonald, whose
comments have greatly improved the article. We also thank Ross Williams and
participants in webinars hosted by Social Ventures Australia, the Economic
Society of Australia, Alpha Beta, the Victorian Department of Jobs, Precincts and
Regions and the National Skills Commission, for many helpful comments. The
paper cites anonymised and aggregated data from Xero, released as part of its
Small Business Insights Program. Jeff Borland’s research has been supported by
ARC Discovery Grant DP160102269.
1. Introduction – The three graphs that tell the story
Every recession is different. Some, however, are more different than others. So it
is with the COVID-19 recession in Australia. With its origins in the onset of
COVID-19, and the rapid pace of change thereby engendered in the labour
This is the author manuscript accepted for publication and undergone full peer review
but has not been through the copyediting, typesetting, pagination and proofreading
process, which may lead to differences between this version and the Version of
Record. Please cite this article as doi: 10.1111/1467-8462.12386.
This article is protected by copyright. All rights reserved.
market, thus far it bears little resemblance to past recessions. 1 Only four months
Author Manuscript
in, already there is as much to describe, and as many questions to answer, as
might usually have accumulated after several years.
With so much to take in, and with the striking character of what is underway, the
story of the COVID-19 recession is being told through the ‘big’ graph. In that vein,
we begin our overview of developments in the Australian labour market with
three graphs to summarise the early impact of COVID-19.
Figure 1 shows changes in monthly hours worked in Australia from March to
June 2020; compared with the 1980s and 1990s recessions. Monthly hours
worked is the best way to describe changes in labour demand during the period
when the JobKeeper program is in place, because its operation is artificially
increasing the number of persons counted as employed. 2
Obviously notable are the scale and speed of decrease in monthly hours worked.
In previous recessions, hours worked gradually decreased over 18 months to
reach a trough about 6 per cent below the start of the recession. In the current
episode, hours worked fell by 9.5 per cent in just one month, from March to
April; and then fell again by 0.9 per cent from April to May. From May to June,
monthly hours worked rebounded by 3.6 per cent. Those magnitudes of increase
1 Bishop (2020) provides an overview of the economic consequences for Australia of the
1918-19 Spanish flu pandemic. In a curious coincidence, the main recent attempt to
model the consequences of a pandemic for Australia was undertaken by the current
Treasury Secretary, Steven Kennedy (and co-authors); see Kennedy et al. (2006).
Borland (2009) provides a review of what has happened in the Australian labour market
in ‘usual’ recessions.
2 Hours worked is the preferred measure of labour demand from the ABS Labour Force
Survey for the period in which JobKeeper is in place; rather than persons employed. This
is because the ABS definition of employment (following ILO convention) includes
persons who worked zero hours but are receiving pay from an employer. This group of
employed persons is artificially increased by JobKeeper. For further details, see ABS,
Labour Force Australia, 6202.0 – Classifying people during the COVID-19 period.
This article is protected by copyright. All rights reserved.
and decrease in hours worked in a single month have never been previously
Author Manuscript
observed (at least since the start of this ABS data series in 1978).
Figure 1: Changes in monthly hours worked from the start of recessions
2
Months since start of recession
Per cent change in monthly hours of work
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
-2
-4
-6
-8
Early 1980s Early 1990s Now
-10
-12
Source: ABS, Labour Force Australia, catalogue no.6202.0, Table 19. 1980s
recession start date = September 1981; 1990s recession start date = July 1990.
The same features of the COVID-19 recession stand out, even from a longer-run
historical perspective. Figure 2 shows the rate of unemployment in Australia
from the start of the twentieth century onwards, including an average rate for
April to June 2020. The average for April to June is an adjusted rate, constructed
by the Australian Bureau of Statistics (ABS) to be historically comparable (by
correcting for the impact of the JobKeeper program). That average rate was 10.0
per cent; up from 5.4 per cent in March. 3 The only extended periods with a
higher rate of unemployment have been during the Great Depression and in the
early 1990s. In neither of those episodes did the rate of unemployment rise as
quickly.
3 Data are from ABS, Labour Force Australia, 6202.0, May: i) For 2020 – Employment
and unemployment: An international perspective; and ii) Table 1. The adjustment made
by the ABS is to add to the group of labour force participants classified as unemployed
those employed persons working zero hours who indicated they had 'no work, not
enough work available or were stood down.
This article is protected by copyright. All rights reserved.
Figure 2: Long-run history of rate of unemployment
Author Manuscript
25
Rate of unemployment
20 April to June 2020: Average
adjusted rate of ue = 10.0%
15
10
5
0 1900/01
1906/07
1912/13
1918/19
1924/25
1930/31
1936/37
1942/43
1948/49
1954/55
1960/61
1966/67
1972/73
1978/79
1984/85
1990/91
1996/97
2002/03
2008/09
2014/15
Butlin series ABS LFS
a] 1900-01 to 1973/74: Butlin series: Butlin, Matthew (1977), ‘A preliminary
annual database 1900/01 to 1973/74’, RBA Research Discussion Paper 7701; b]
1966/67 to 2018/19: i] 1966-77: ABS, The Labour Force Australia, 1966-84,
catalogue no.6204.0; ii] 1978 onwards: ABS, Labour Force Australia, catalogue
no.6202.0; Original series - Average of August, November, February and May.
Figure 3 shows changes in the number of jobs in Australia from mid-March to
early July. The series displayed are from two data sources: i) ABS/Australian
Taxation Office (ATO) Payroll data – which shows employee jobs reported to the
ATO through the Single Touch Payroll system; and ii) Xero Small Business
Insights (SBI). Appendix 1 provides a brief overview of the main sources of data
used for this study.
The ABS/ATO series shows that the number of jobs in Australia fell rapidly from
21 March to 18 April, then gradually increased to 20 June, and after that again
decreased through to early July (mainly due to job losses in Victoria with the re-
emergence of COVID-19). By early July the number of jobs was about 3.5 per cent
below its level in mid-March, having regained more than half the initial decrease.
The Xero SBI series shows that the initial decrease in jobs in small business was
larger than for all businesses; and the recovery from mid-April is not evident.
Other important features of the COVID-19 recession are apparent from Figure 3.
One is that this is the first major economic downturn in Australia where (almost)
This article is protected by copyright. All rights reserved.
real-time weekly data on labour market outcomes are available. These data have
Author Manuscript
become available with online reporting to the ATO for tax collection purposes;
and recognition by the ABS and Xero/Alpha Beta of the value of those data.
Having the real-time data has allowed more precise and up-to-date assessments
of the timing and causes of labour market developments. It has also meant that
policy making appears to have been more influenced by data than ever before;
for example, the Commonwealth Treasury’s (2020) review of the JobKeeper
program cites a variety of high-frequency official and unofficial data. Policy is the
second feature evident in the figure. The timing of the stop to the decrease in the
number of jobs, following the announcement of the JobKeeper program, stands
out. More analysis of this issue is provided later in the article – Here we simply
note that fiscal policy responses by governments, and their impact on
households and businesses, has been a critical element of the COVID-19
recession.
Figure 3: Changes in number of payroll jobs, 1 February to 11 July 2020,
ABS/ATO and Xero SBI series
105
Announcement of
Number of payroll jobs (Index = 100 on
JobKeeper (30/3)
100
95
1/2/2020)
90
85
80
1-Feb 15- 29- 14- 28- 11- 25- 9-May 23- 6-Jun 20-Jun 4-Jul
Feb Feb Mar Mar Apr Apr May
ABS/ATO Xero
1] ABS/ATO: ABS, Weekly Payroll Jobs and Wages, Week Ending July 11 2020,
catalogue no.6160.0.55.001, Table 4; and 2] Xero Small Business Insights;
Analysis by Alpha Beta.
This article is protected by copyright. All rights reserved.
In the rest of the article, we elaborate on and analyse in more detail recent
Author Manuscript
labour market developments in Australia. Section 2 provides an overview of the
main drivers of labour market outcomes during the initial phase of the COVID-19
recession. Section 3 presents extra descriptive information – showing the uneven
impact on labour market outcomes between workers. Section 4 makes an initial
assessment of the central labour market policy introduced to deal with the
COVID-19 recession– the JobKeeper program. Concluding remarks are in section
5.
2. The unfolding: Main drivers and aggregate labour market outcomes
a] Main drivers
The COVID-19 recession originated with the spread of the pandemic to Australia.
The benchmark of 100 cumulative cases of COVID-19 was reached in Australia
on 10 March, 44 days after its first case was reported. Thereafter, for a time, the
caseload expanded rapidly – reaching 461 cases a week later and 2,135 cases
two weeks later. The one-day peak of 468 cases occurred on March 28. 4
COVID-19 is spread through personal contact – and some work and consumption
activities potentially involve a high degree of contact among and between
workers and consumers. The first main driver of the COVID-19 recession,
decreased consumption spending in Australia, was brought about by consumers
and governments responding to the onset of the virus by seeking to limit
activities with high levels of personal contact. First, with the onset of COVID-19,
some households are likely to have voluntarily decreased their consumption of
goods and services involving high levels of personal contact– such as eating out
and attending public events – to avoid contracting COVID-19. Second,
commencing from mid-March, commonwealth and state governments
introduced restrictions intended to limit activities with high personal contact.
4 https://www.health.gov.au/news/health-alerts/novel-coronavirus-2019-ncov-health-
alert/coronavirus-covid-19-current-situation-and-case-numbers#daily-reported-cases-
in-australia
This article is protected by copyright. All rights reserved.
Initially imposing rules for the maximum size of gatherings, ultimately (by late
Author Manuscript
March) these restrictions effectively closed designated businesses such as cafes,
restaurants, theatres and gyms, forcing a reduction in spending on those
activities. 5 As well, bans on entering Australia from March 21, and restrictions
on travel within the country, meant that aviation services were severely
curtailed; and some business – for example in the retail sector – chose to
voluntarily close in order to avoid risk to themselves and their employees.
Reflecting the impact of COVID-19, data from credit bureau illion shows that by
late March, spending in Australia on many activities involving a high degree of
personal contact had fallen dramatically. For example, in the week of 29 March
spending was 35 per cent below its pre-pandemic levels in cafés, 82 per cent
down in gyms, and 72 per cent lower in entertainment venues. The illion data
also shows that high income individuals cut back their spending more strongly
than lower income individuals. In the week of 29 March, individuals earning
more than $104,000 were spending 22 per cent below their pre-pandemic levels,
while individuals earning less than $65,000 were spending just 8 per cent less.
Lower income individuals appear to have cut their spending less because they
had a lower share of discretionary spending and were less likely to be working
from home; whereas higher income individuals previously had a larger share of
their spending on activities with high levels of personal contact. The trend of
lower income individuals cutting their spending less than higher income
individuals has persisted through the pandemic.
Early research from the United States has investigated the relative impacts of
government mandated shutdowns and voluntary decreases in household
consumption associated with the onset of COVID-19. That research indicates that
voluntary withdrawal of spending may have been most important driver of
decreases in household consumption (see Goolsbee and Syverson, 2020; Gupta et
5 https://www.mondaq.com/australia/operational-impacts-and-
strategy/929622/timeline-of-key-social-economic-covid-19-events-affecting-australian-
business
This article is protected by copyright. All rights reserved.
al, 2020; and Chetty et al., 2020, pp.28-29). The data from illion also suggest that
Author Manuscript
voluntary decreases in spending may have been important in Australia; for
example, with large decreases in spending by high income individuals on
activities involving a high degree of personal contact with suppliers and other
consumers (for a further perspective, see Ball, 2020). It is a topic for future work,
however, to determine the relative impacts of voluntary decreases in spending
and government restrictions in Australia. It seems likely that the relative impacts
would have varied by country depending on the timing of introduction of
government restrictions compared to the stage of progress of COVID-19.
As has already been noted, because it has been brought about by governments
seeking to control the spread of the virus and consumers wanting to avoid
contracting it, the early impact of the COVID-19 recession has been concentrated
on industries and businesses with high levels of personal interaction among staff
and customers, where transmission has been thought to be most likely. Various
predictions of the direct impact on employment have been made, taking into
account the likelihood that workers in those industries would be worst affected.
Those estimates ranged from a decrease of about 1 million for the policies
already introduced in late March, through to 3.6 million had the shutdown
become more intense (Borland, 2020a; Cassells et al., 2020; Coates et al., 2020;
and Wilkins, 2020).
With COVID-19 being (or appearing to be) brought under control, government
restrictions on economic activity began to be relaxed at the end of April.
However, the time at which relaxation commenced and subsequent pace of
removal differed substantially between states. 6 Therefore, by the end of June
there was considerable variation between states in regulations governing, for
example, the maximum numbers of customers allowed at cafes and restaurants
and attendance at arts venues and sporting events.
6 Relaxation of restrictions commenced from late April (WA), mid-May (NSW,
Queensland, SA and Tasmania) and early June (Victoria). See:
https://www.australia.gov.au/news-and-updates
This article is protected by copyright. All rights reserved.
Recognising the severe impact of business closures on aggregate demand and
Author Manuscript
employment, commonwealth and state governments quickly introduced policies
intended to buttress economic activity and to underpin household spending and
business viability. Government fiscal policy has therefore been the second main
driver of labour market outcomes in the COVID-19 recession. Table 1 provides a
brief summary of the main commonwealth government stimulus measures. In
the first stage of commonwealth government measures, a $17.6bn package of
business assistance and payments to welfare recipients was announced on
March 12. A second stage expanded that package with an extra round of
assistance worth $66.6bn on March 22. The third stage was the announcement
on March 30 of the $130bn JobKeeper wage subsidy program. During March the
Reserve Bank of Australia had also cut the cash rate from 0.75 per cent to 0.25
per cent; and as well, all state and territory governments introduced stimulus
measures, collectively worth $25bn (Daley et al., 2020, pages 18-21).
Table 1: Summary of main policy measures
Package 1: 12 March
1] Business investment incentives (via tax write-offs): $3.9bn
2] Cashflow assistance to business: $6.7bn – Direct support payments to
business with turnover less than $50m; $1.3bn – Wage subsidies for apprentices
3] Household stimulus: $4.8bn via a one-off $750 payment to income support
recipients
4] Assistance for severely affected regions: $1bn
Package 2: 22 March
1] Coronavirus supplement: $14.1bn – Supplement of $550 per fortnight for 6
months for selected income support recipients (eg., JobSeeker; Youth allowance)
2] Household support: $4bn – Second $750 payment for income support
recipients not eligible for coronavirus supplement
3] Early release of superannuation: Able to withdraw up to $10,000 in both
2019-20 and 2020-21
4] Increased cashflow assistance to business: $23.9bn
This article is protected by copyright. All rights reserved.
5] Coronavirus SME Guarantee: Guarantee 50 per cent of new loans issued by
Author Manuscript
eligible lenders to SMEs (Up to $20bn out of $40bn on loans)
6] Support for aviation: $715m
Package 3: March 30
JobKeeper program: $70bn
Sources: https://www.pm.gov.au/media/economic-stimulus-package;
https://www.pm.gov.au/media/supporting-australian-workers-and-business;
https://www.pm.gov.au/media/130-billion-jobkeeper-payment-keep-
australians-job
The quantum of government spending has been massive. Figure 4 shows the
impact on COVID-19 on household cashflow from January to June 2020. COVID-
19 has caused large negative impacts on household income from wages and
unincorporated businesses. That impact has been more than offset, however, by
government stimulus measures – with large fractions of that stimulus accounted
for by superannuation withdrawals and payments through the JobKeeper and
JobSeeker programs. Overall, the net impact of COVID-19 and government
policies is estimated to have increased household cashflow in the period covered
in this review by $23.9bn.
Of course, it is one thing to provide compensating sources of income – the main
question for economic activity is whether that income is spent. Figure 5 provides
a perspective on that question. It compares spending by government benefit
recipients who received one or both of a $750 stimulus payment in early April
and an increased weekly benefit payment of $550 from April to September, with
spending by individuals not receiving those payments. 7 After tracking closely
until late March, following introduction of these payments, spending by benefit
recipients was about 20 per cent higher than other individuals.
7 For further details on timing, eligibility and other conditions regarding these
payments, see:
https://www.aph.gov.au/About_Parliament/Parliamentary_Departments/Parliamentar
y_Library/FlagPost/2020/March/Coronavirus_lump-sum_payments;
https://www.aph.gov.au/About_Parliament/Parliamentary_Departments/Parliamentar
y_Library/FlagPost/2020/March/New_coronavirus_supplement
This article is protected by copyright. All rights reserved.
Figure 4: Impact of COVID-19 and fiscal policy responses on household
Author Manuscript
cashflow,, Main components, April to June 2020
Source: Analysis by Alpha Beta.
Notes: i] Lost wages: Calculated from ABS/ATO Weekly Payroll data and ABS
National Accounts; ii] Lost unincorporated income: Apply ratio of mixed income
to employee salaries (ABS, National Accounts) to calculation of lost wages; iii]
Economic support – One-off $750 payment: Payment recipient numbers
(Australian Parliament, Senate Select Committee on COVID-19) x $750; iv]
JobSeeker + COVID-19 supplement: Calculated from payment recipient numbers
(existing + additional) and amount of payments (Commonwealth Department of
Social Services); v] Incremental JobKeeper: Commonwealth Treasury (2020)
adjusted to equal payment after-tax; vi] Superannuation withdrawal – Australian
Prudential Regulation Authority; and vii] Mortgage deferrals and hardship:
Australian Bankers Association data on value of mortgage deferrals with
assumptions on interest payments and principal repayments.
This article is protected by copyright. All rights reserved.
Figure 5: Individual-level spending, Weeks ending February 10 to 22 June
Author Manuscript
2020
160
Weekly household spending
140
120
100
80
Received stimulus
60
Did not receive stimulus
40
Source: illion; Analysis by Alpha Beta.
In future (and perhaps to some degree already) a third driver of the COVID-19
recession will be the extent to which the initial impact of COVID-19 causes a
longer-term and broader-based decline in aggregate demand. Such a decline
could happen via a decrease in household spending (especially discretionary)
due to reduced income and/or wealth and uncertainty about the future; a
decrease in business investment due to lowered expectations of profitability; or
a decrease in demand for Australian exports due to the impact of COVID-19 on
our main trading partners and global trading networks.
b] Aggregate labour market outcomes
Labour market outcomes in Australia in the initial phase of the COVID-19
recession reflect the push and pull of decreased household consumption and
fiscal policy responses. Table 2 shows the key aggregate labour market outcomes
from March to June. Several main features are evident.
This article is protected by copyright. All rights reserved.
Table 2: Changes in labour market aggregate outcomes (sa), March to June
Author Manuscript
2020
March to April April to May May to June
Monthly hours -9.5 -0.9 +3.6
(per cent
compared to
March)
Employment -4.7 -2.0 +1.2
(per cent
compared to
March
EMP/POP rate -3.0 -1.3 +1.0
(ppts)
LFP rate (ppts) -2.3 -0.9 +1.3
Rate of UE (ppts) +1.1 +0.7 +0.4
Source: ABS, Labour Force Australia, 6202.0, Tables 1 and 19.
First, there was a large decrease in aggregate labour demand, evident in both
monthly hours worked and employment. The decrease was concentrated from
March to April. Labour demand was then relatively stable from April to May; and
reversed from May to June as the threat of COVID-19 appeared to be receding
and relaxation of restrictions on business activity commenced. Using the
monthly hours worked measure, about one-third of the decrease in labour
demand from March to May was reversed between May and June.
Second, monthly hours worked responded more than employment to changes in
labour demand from March to June. No doubt this mainly reflects the impact of
the JobKeeper program, which subsidises employers to retain their existing
workforces. The ‘one-in all-in’ rule effectively forces employers to adjust to the
closing and reopening of economic activity by decreasing and increasing the
This article is protected by copyright. All rights reserved.
hours of their workforces, rather than laying off and then hiring workers. 8 In
Author Manuscript
previous recessions, by contrast, decreases in aggregate hours worked are
accounted for almost entirely by lay-offs (Borland, 2011, p.201; and Bishop et al.,
2016).
Third, the initial phase of decrease in employment brought a much larger flow
out of the labour force and less flow into unemployment than in previous
recessions. Figure 6 shows changes in the employment/population and Labour
Force Participation (LFP) rates – from March to June and in the 1990s recession.
In the 1990s a decrease in the employment/population rate of about 4.0ppts was
associated with a decrease in the LFP rate of 1.3ppt. Whereas the decrease in
employment/population rate from March to June of 3.3 ppts has been associated
with a decrease in the LFP rate of 1.9ppt.
Figure 6: Changes in EMP/POP rate and LFP rate, Now and 1990s recession,
By month
Months since start of recession
0
0 1 2 3 4 5 6 0 1 2 3 4 5 6 7 8 9 10 11 12
-1
Change (per cent)
-2
-3
-4
-5 Employment - Now
-6 Labour Force - Now
Employment - 1990s
-7
Labour Force - 1990s
-8
Source: ABS, Labour Force Australia, 6202.0, Tables 1 and 19.
8 The one-in all-in rule means that employers who claim JobKeeper for any employee
have to claim the payment for all their employees. Hence, laying off employees is not an
option for businesses receiving the JobKeeper subsidy. Appendix Table 2.1 presents data
on the number of persons employed working zero hours from March to June 2020.
This article is protected by copyright. All rights reserved.
The larger flow out of the labour force is likely to reflect two main influences:
Author Manuscript
first, a strong discouraged worker effect due to the scale of the decrease in
labour demand from March to May (that is, workers who have lost their jobs
believe they have no prospect of finding employment, and hence do not search
for work); and second, workers (mainly females) withdrawing from the labour
force due to the need to look after children as school closures were
implemented. 9
3. Who is being affected?
In this section we review changes in several main labour market outcomes -
hours of work, employment and labour force status - for different types of jobs
(by industry; occupation; and employment status) and different types of workers
(by gender; age; country of birth; and location of residence). 10
a] By industry
Changes in employment by industry reflect that the initial impact of COVID-19
was to cause a decrease in spending on activities with high levels of personal
contact; and then the partial resumption of that spending with the virus
appearing to have been brought under control.
Changes in employment for selected industries from February to May 2020 – the
period when total employment declined substantially - are shown in Table 3.
Two measures of employment are reported using data from the ABS Labour
Force Survey: first, the percentage decrease in persons employed; and second,
9 It has also been suggested that the larger flow to OLF might reflect removal of
job search requirements on recipients of the JobSeeker payment. But a
comparison of gross flows for March to April 2020 against the average of the
flows for the same months in the previous three years finds that only about 10
per cent of the increase from previous years to this year in the change in OLF
stock from March to April is explained by increased flows from UE to OLF.
10 Wilkins (2020) forecasts impacts by types of workers based on assumptions on the
industries most likely to be affected by decreases in consumption due to COVID-19.
This article is protected by copyright. All rights reserved.
the share of the decrease in aggregate hours worked accounted for by each
Author Manuscript
industry. Accommodation & food services and arts & recreation services
experienced the largest percentage decreases in employment. As well, those two
industries, together with retail trade, accounted for 18 per cent of employment
in February 2020, but 42 per cent of the decrease in employed persons to May. 11
Table 3: Changes in employment in selected industries, February to May
2020
Per cent decrease in Share of decrease in
employment (%) (sa) hours (%) (original)
Accommodation& -31.2 24.4
food services
Transport, postal & -9.8 11.6
warehousing
Retail trade -4.5 9.9
Other services -10.1 9.6
Manufacturing -7.4 8.5
Professional, scientific -4.9 8.4
and technical services
Arts and recreation -35.4 7.8
services
Administration & -10.4 7.3
support services
11 Other industries, that might be regarded as less directly affected by the shutdown, did
however account for sizable proportions of the decrease in aggregate hours worked;
such as manufacturing and professional, scientific and technical services.
This article is protected by copyright. All rights reserved.
Health care and social -3.3 7.2
Author Manuscript
assistance
Education & training -7.9 6.6
Information, media & -15.9 2.5
telecommunications
Source: ABS, Labour Force Australia, Detailed – Quarterly, 6291.0.55.003, Table 4
and EQ11.
Using data from the ABS/ATO Payroll series it is possible to describe changes in
the number of jobs by industry for the period to early July, which encompasses
the initial recovery in total employment. 12 Figure 7 shows series for total jobs
and for the number of jobs in the three industries with high levels of personal
contact. Consistent with COVID-19 being the main cause of changes in
employment, strong increases in the number of jobs occur in the three industries
from mid-April as the threat of COVID-19 receded and government restrictions
began to be relaxed.
12 Changes in employment by industry using data from the ABS Labour Force Survey and
the ABS/ATO Payroll series are highly correlated. See Appendix Figure 2.1.
This article is protected by copyright. All rights reserved.
Figure 7: Changes in number of jobs, Selected industries, 14 March to 11
Author Manuscript
July 2020
110.0
105.0
Number of jobs (Index = 100 on
100.0
95.0
90.0
85.0
14/3/2020)
80.0
75.0
70.0
65.0
60.0
14-Mar-20 14-Apr-20 14-May-20 14-Jun-20
All Retail trade
Accommodation & food services Arts & recreation services
Source: ABS/ATO: ABS, Weekly Payroll Jobs and Wages, Week Ending July 11
2020, catalogue no.6160.0.55.001, Table 4
With the impact of COVID-19 on employment being so concentrated by industry,
changes in the industry composition of employment are an important window to
understand the broader labour market impacts of COVID-19. Young people,
females and workers in occupations such as community and personal services
and sales are relatively concentrated in the most affected industries. That
concentration by industry explains why their labour market outcomes have been
disproportionately adversely affected by the impact of COVID-19. Appendix
Tables 2.2a to 2.2c show how hypothetical rates of change in employment by
occupation, age and gender – predicted using just the impact of the rate of
change in employment by industry from February to May – are highly correlated
with actual rates of change.
At the same time as emphasising how the impact of COVID-19 has had a
pronounced impact on the industry composition of employment, it should not be
ignored that there is also likely to be substantial reallocation of employment
occurring within industries due to COVID-19 (see Barrero et al., 2020). For
example, spending on food services has switched from seated venues to take-
away; purchases of books and clothing have moved from physical stores to on-
This article is protected by copyright. All rights reserved.
line suppliers; and tourism within Australia is substituting for international
Author Manuscript
travel.
b] By other job characteristics
Changes in employment from February to May 2020 varied substantially
between occupations. For example, employment of community and personal
services workers decreased by 29.1 per cent; whereas the decrease for
professionals was only 3.2 per cent (see Appendix Table 2.3). Figure 8 shows the
percentage changes in actual hours worked in occupation sub-groups; where
occupations are classified into deciles ordered by average weekly earnings. 13
Decreases in hours worked by occupation have been strongly correlated with
average weekly earnings – with low earners experiencing much larger decreases
than high earners. Remarkably, workers in occupations in the bottom decile
(ordered by average weekly earnings) experienced a decrease in hours worked
of 27.5 per cent – compared to a decrease of only 3.2 per cent for workers in the
top decile. Sixty per cent of the decrease in total hours worked is accounted for
by workers in the bottom three deciles of the distribution of weekly earnings.
Figure 8: Changes in employment by occupation ordered by average
weekly earnings, February to May 2020
Decile in distribution of weekly earnings
0
1 2 3 4 5 6 7 8 9 10
Per cent change in actual weekly
-5
-10
-15
hours worked
-20
-25
-30
13 Average weekly earnings are for full-time employed persons. Deciles are
created based on total actual hours worked in each occupation sub-group.
This article is protected by copyright. All rights reserved.
Source: a] Actual hours worked by occupation sub-group: ABS, Labour Force
Author Manuscript
Australia, Detailed – Quarterly, 6291.0.55.003, EQ07a; b] Average weekly
earnings by occupation sub-group: ABS, Characteristics of Employment 2015 to
2019, Tablebuilder.
The incidence of the impact of COVID-19 has also varied between jobs by their
employment status. Owner/managers without employees have experienced the
largest decreases in employment and hours worked. Employees also had a large
decrease in their employment. By contrast, employment for owner/managers
with employees slightly increased (see Appendix Table 2.4). Within the category
of employees, casual employees have experienced a much larger negative impact
than permanent employees. Figure 9 shows the evolution of the number of
casual and permanent jobs from February to late May. The number of casual jobs
lost is much larger than permanent jobs: about 20 per cent compared to 5 per
cent.
Figure 9: Number of jobs by employment status, Casual and permanent,
Weeks ending 2 February to 24 May 2020
110
Number of jobs (Index = 100 2/2/20)
105
100
95
90
85
80
75
70
65
60
2-Feb 9-Feb 16-Feb 23-Feb 1-Mar 8-Mar 15-Mar 22-Mar 29-Mar
5-Apr 12-Apr 19-Apr 26-Apr 3-May 10-May 17-May 24-May
Casual Permanent FT Permanent PT
Source: Xero Small Business Insights; Analysis by Alpha Beta.
c] By gender, age and country of birth
Labour market outcomes by gender have been a much remarked-on feature of
the impact of COVID-19. Females were more adversely impacted by males by the
This article is protected by copyright. All rights reserved.
decrease in labour demand following the onset of COVID-19; although the gap in
Author Manuscript
outcomes narrowed somewhat as recovery commenced. Females were also more
likely than males to withdraw from the labour force. Table 4, for example, shows
that monthly hours worked decreased from March to May by 11.8 per cent for
females compared to 9.5 per cent for males; with those decreases being
(respectively) 7.3 per cent and 6.5 per cent by June.
Table 4: Changes in labour market outcomes (sa), By gender, March to June
2020
March to May May to June
Monthly hours Females -11.8 +4.5
(per cent
relative to
March)
Males -9.5 +3.0
EMP/POP (ppts) Females -4.5 +1.1
Males -4.0 +0.8
LFP (ppts) Females -3.6 +1.4
Males -2.9 +1.2
Rate of UE (ppts) Females +1.8 +0.4
Males +1.9 +0.4
Source: ABS, Labour Force Survey, 6202.0, Tables 1 and 19.
This article is protected by copyright. All rights reserved.
Figure 10: Change in employment/population rate, By gender, During
Author Manuscript
recessions
Quarters since start of recession
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
-1
Change in EMP/POP rate (ppts)
-2
-3
-4
-5
Females (Average: 1970s/1980s/1990s)
-6 Males (Average: 1970s/1980s/1990s)
Females (COVID-19)
-7 Males (COVID-19)
Source: a] 1966-77: ABS, Labour Force Australia, Historical Summary 1966 to
1984, catalogue no.6204.0, Table 2; b] 1978 onwards: ABS, Labour Force
Australia, catalogue no.6202.0, Table 1.
The impact by gender in the present recession is very different to previous
recessions in Australia, which have seen substantially larger negative impacts on
male than female employment – as shown in Figure 10. Previous recessions have
been concentrated periods of decreased employment in male-dominated
industries – especially manufacturing and construction. By contrast, the early
impact of COVID-19 was on industries employing above-average shares of
females. Further to the gender impact of COVID-19, there is also evidence that
females have taken on a disproportionate share of home-based schooling
following school closures. 14
14 For more details on labour market outcomes by gender in Australia, see Borland
(2020b). For a general overview on COVID-19 and labour market-related gender
impacts in the United States and United Kingdom see respectively Alon et al. (2020) and
Hupkau and Petrongolo (2020).
This article is protected by copyright. All rights reserved.
Young people (aged 15-24 years) experienced much greater decreases in
Author Manuscript
employment and movement out of labour force during the initial stages of
COVID-19 than prime age and older workers. Table 5, for example, shows that
the employment/population rate of the young decreased by 9.8 ppts between
March and May, compared to 3.8ppts and 1.4ppts respectively for prime age and
older populations. As has already been described, this partly reflects that the
most affected industries employ above-average shares of young workers; as well,
young employees made up almost 50 per cent of the workforce in short-term
casual jobs, which are excluded from the JobKeeper program (ABS,
Characteristics of Employment, 2015 to 2019, 6333.0, Tablebuilder). What is also
implied by the differential impact of COVID-19 by industry is that, with the
commencement of reopening of economic activity, young people have benefited
most. For example, from May to June, the employment/population ratio of the
young rebounded by 3.3 ppts, compared to only 0.7ppt and 0.2ppt for the prime
age and older populations.
While the deterioration in young people’s labour market outcomes may have
reversed somewhat, it is important however to note that the stage of recession
during which young people are most adversely affected has not arrived yet. In
past recessions, the negative impact on young people comes progressively as
cohorts ending their education seek to make the transition to work and find that
– with lower rates of job creation – they are less likely to find employment.
Hence, any short-term recovery in employment of the young due to reopening of
economic activity is likely to be followed by a second wave of impact as young
people completing their full-time education struggle to find employment in a
depressed labour market. 15
15 It is also important to note that this is against the background of increasingly difficult
labour market conditions for the young in the decade post-Global Financial Crisis; see
for example, Rayner (2016); Dhillon and Cassidy (2018), Borland and Coelli (2020) and
Productivity Commission (2020). For a review of potential long-term scarring impacts
from entering the labour market during a macroeconomic downturn, see Borland
(2020c).
This article is protected by copyright. All rights reserved.
Table 5: Labour market aggregate outcomes (sa), By age, Australia, March
Author Manuscript
to June 2020
EMP/ LFP Actual hours
rate worked in all
POP rate (ppts) jobs
(ppts) (Weekly)
(%)
March to May to March May to March to May to
May June to May June May June
Age
15-24 -9.8 +2.7 -9.4 +3.4 -19.2 +5.1
25-54 -3.8 +0.7 -2.6 +0.9 -7.7 +1.1
55 plus -1.4 +0.2 -1.3 +0.5 -6.5 +1.1
Source: ABS, Labour Force Survey, Detailed – Electronic Delivery, 6291.0.55.001,
Table 01 and EM1a.
Employment outcomes by country of birth tell a striking story – shown in Table
6. The Australian-born population experienced a decrease in its
employment/population rate rate from March to June of 2.2 ppts. The size of
decrease for immigrants who have been in Australia for 5 years or more has
been similar: 2.8 ppts. But for immigrants in Australia for less than 5 years the
average decrease in the employment/population rate was 6.0 ppts. Undoubtedly,
the difference incidence reflects the exclusion from JobKeeper of immigrants
who are temporary visa holders.
Table 6: Change in employment/population rate, By country of birth and
years since arrival, March to June 2020, ppts
March to May May to June
Australia -2.8 +0.6
This article is protected by copyright. All rights reserved.
Immigrant
Author Manuscript
Less than 5 -9.2 +3.3
years
5 years plus -4.0 +1.2
Source: ABS, Labour Force Survey, Detailed – Electronic Delivery, 6291.0.55.001,
LM4.
d] By state
No state has been exempt from labour market effects of COVID-19. But there has
been considerable variation. All states saw large decreases in monthly hours
worked from March to April of around 10 per cent. Thereafter, as shown in
Figure 11, experiences between states have differed somewhat – and perhaps
not surprisingly, in a way that that appears correlated with COVID-19 being
brought under control and with the timing and extent of relaxation of
government restrictions. As a guide, Table 7 presents a brief overview of changes
to restrictions from mid-May to mid-June.
Figure 11: Monthly hours worked by state, March to June 2020
100
Per cent change in monthly hours worked
98
96
94
92
90
88
86
84
March April May June
Axis Title
NSW VIC QLD SA WA TAS
This article is protected by copyright. All rights reserved.
Source: ABS, Labour Force Australia, 6202.0, Table 19.
Author Manuscript
In South Australia and Western Australia, which saw relatively few cases of
COVID-19 and where relaxation began relatively early and proceeded fastest,
hours worked increased from April onwards. In NSW and Tasmania, which have
also been more advanced in reopening, hours worked continued to decrease to
May, but then rebounded very strongly from May to June – by about 4 ppts. By
comparison, Victoria and Queensland, which were slower and less advanced in
reopening by mid-June, saw continued decreases in hours worked from April to
May, and then smaller increases from May to June than other states. 16
Table 7: Relaxation of restrictions on economic activity by June 15 2020
NSW 15/5: cafes etc: 10 customers in seated areas
1/6: cafes etc: 50 customers in seated areas; museums, zoos
etc open; camping grounds etc open
VIC 1/6: cafes etc: 20 customers
QLD 1/6: cafes etc: 20 customers in seated areas; museums, zoos
etc open; camping grounds etc open
SA 11/5: cafes etc: 10 customers in seated areas; tourist sites
open; museums, zoos etc open
1/6: cafes etc: 80 customers in seated areas
WA 18/5: cafes etc: 20 customers in seated areas; unrestricted
travel in state
1/6: cafes etc: 2 sqm rule in seated areas; museums, zoos etc
open
16 Subsequently, with the renewed break-out of COVID-19 in Victoria (and to a lesser
extent in NSW), the number of jobs declined in those states from late June to early July,
while remaining steady in other states – see ABS, Weekly Payroll Jobs and Wages, Week
Ending July 11 2020, catalogue no.6160.0.55.001, Table 4
This article is protected by copyright. All rights reserved.
TAS 18/5: cafes etc: 10 customers in seated areas
Author Manuscript
5/6: cafes etc: 40 customers in seated areas; museums, zoos
etc open; camping grounds etc open
Source: State government websites accessed via https://www.australia.gov.au/
4. JobKeeper: A preliminary assessment
In this section we provide a brief overview and evaluation of the JobKeeper
program; the major policy used by the commonwealth government to deal with
the decrease in labour demand in the early stages of the COVID-19 recession.
a] Main details of the program
The initial phase of the JobKeeper program – from 30 March to 27 September –
has involved the commonwealth government making a $1500 per fortnight per
worker payment to eligible employees (via their employers) and to eligible self-
employed individuals. The main objectives of the program are to: (i) seek to
ensure viability of businesses through the shutdown of economic activity; (ii)
provide financial support to households; and (iii) preserve connections between
employers and their workforces (Australian Government, 2020; and
Commonwealth Treasury, 2020, pp.14-16).
Eligible employees – who are nominated by their employer - include full-time
and part-time permanent employees; and casual employees who (at 1 March)
had been working with their employer on a regular basis for more than 12
months. Employees who had been retrenched after 1 March but then re-attached
after announcement of the program are included in the scope of the program. 17
For an employer to be able to claim JobKeeper for their employees, they must
anticipate a reduction in turnover relative to a comparable period a year ago: a
17 Eligibility also requires an employee to be an Australian citizen, the holder of a
permanent visa, a Protected Special Category Visa Holder, a non-protected
Special Category Visa Holder who has been residing continually in Australia for
10 years or more, or a Special Category (Subclass 444) Visa Holder.
This article is protected by copyright. All rights reserved.
15 per cent reduction for a registered charity; a 30 per cent reduction for
Author Manuscript
business with turnover less than $1billion; and a 50 per cent reduction for a
business with turnover more than $1billion. The reduction in turnover must be
for a minimum of one month during the six months in which the JobKeeper
program is in operation. Self-employed individuals are eligible to receive the
JobKeeper payment where they have suffered or expect to suffer a 30 per cent
decline in turnover relative to a comparable a period a year ago. Eligibility for
JobKeeper has also been restricted via ruling some types of organisations to be
outside the scope of the program – for example, universities.
The $1,500 (before tax) payment per fortnight is a flat rate. Employees who have
been stood down receive $1,500 per fortnight, regardless of their normal salary.
Similarly, employees who are still working, but whose earnings would be less
than $1,500, receive the full payment. Employees whose work hours are such
that their total earnings are above $1,500 must have the difference between the
subsidy and their total earnings paid by the employer.
Commonwealth Treasury (2020, p.7) has reported that JobKeeper payments
were made for about 3.5m individuals in April-May 2020. There has been large
variability across industries in the take-up of JobKeeper. For example, JobKeeper
was paid for about 60 per cent of employed persons in the private sector
working in the arts and recreation services industry, but to below 5 per cent in
mining. Young people are under-represented in receipt of JobKeeper payments
relative to their share of the overall decrease in private sector employment
decrease; and males are slightly over-represented (Commonwealth Treasury,
2020, pp.17-21).
b] Impact of JobKeeper – Theory
Since its implementation, three main issues of interest have been raised
regarding the impact of JobKeeper on labour market outcomes. These are:
i] its impact on employment;
ii] its impact on labour incomes; and
This article is protected by copyright. All rights reserved.
iii] the overall benefit-cost of the program.
Author Manuscript
To investigate the potential effects of the JobKeeper program on these outcomes,
it is useful to define profit-maximising levels of weekly hours of work for
different scenarios:
• ℎ𝑝𝑟𝑒𝐶 = profit-maximising hours of work for an employee or sole proprietor
prior to COVID-19;
• ℎ𝑝𝑜𝑠𝑡𝐶 = profit-maximising hours of work after the onset of COVID-19 in the
absence of JobKeeper; and
• ℎ 𝐽𝐾 = profit-maximising hours of work after the onset of COVID-19 and with
JobKeeper.
We begin by considering the impact of JobKeeper on employment. JobKeeper
will cause an increase in employment where it increases hours of work
compared to the post-COVID-19 state without JobKeeper (that is, ℎ 𝐽𝐾 > ℎ𝑝𝑜𝑠𝑡𝐶 ).
This case can arise because the subsidy element of JobKeeper reduces the cost to
a business of retaining its existing workforce. Specifically, a business can employ
a worker at zero cost up to a number of weekly hours for each worker equal to
($750/worker’s hourly wage rate). Where sufficient demand exists for a
business’s output, that decrease in cost may induce the business to offer a higher
number of hours of work than otherwise. By inducing an increase in hours
worked, the JobKeeper program is promoting what might be referred to as a
work connection; that is, due to the cost of their labour being subsidised, the
worker preserves a substantive connection to employment.
As an example, suppose that with the onset of COVID-19 and restrictions on
activity, a café owner/chef has potential revenue of $250 per day from selling
take-away coffee and food. To do this it is necessary to have a food sales worker
working for 3 hours a day – which would have a usual wage cost of $75 per day.
Other costs of operating the café and inputs to making the coffee and food are
$300 per day. In the absence of the JobKeeper program the café would not open;
but with the program, which reduces the cost of employing the food sales worker
This article is protected by copyright. All rights reserved.
to zero, the café will open and there will be a positive effect on hours worked by
Author Manuscript
both the owner and food sales worker. 18
It is also important to note that an alternative type of impact on employment
(but not hours of work) from JobKeeper can arise. Eligible employees can receive
the JobKeeper payment, even where they are working zero hours. That is,
JobKeeper does not affect hours of work, but does preserve a tacit connection
between and employee and their employer; what can be referred to as a job
connection. Preserving a job connection increases the count of persons
employed in the ABS Labour Force Survey, but does not affect the measure of
hours worked. The motivation for wanting to preserve job connections is that
breaking the connection between an employer and employee means losing the
job-worker match-specific surplus (such as the worker’s firm-specific human
capital); and makes necessary expensive job search and hiring respectively by
the employee and employer post COVID-19, thereby prolonging the length of
adjustment in the labour market.
Second, we consider the impact of JobKeeper on labour income. Either pre or
post COVID-19 labour income could be used as the benchmark for comparison.
Most relevant to JobKeeper’s objective of underpinning household incomes is a
comparison with post-COVID-19 labour income in the absence of the program.
For this comparison, the effect of JobKeeper is to increase labour income
whenever ℎ𝑝𝑜𝑠𝑡𝐶 ∙ 𝑤 is less than $750 (where w denotes the worker’s hourly
wage rate). An increase in income can happen either because JobKeeper causes a
worker to be offered extra hours of work; or as an income transfer. An income
transfer can substitute for a decrease in hours of work experienced due to
COVID-19; or can happen because prior to COVID-19 a worker earned less than
$750 per week. Comparison with pre-COVID-19 income addresses the issue,
which has received wide coverage, of whether JobKeeper has allowed workers to
18 This example assumes that the food sales worker is eligible to receive JobKeeper.
Where the café owner is also eligible to receive JobKeeper, there will be a further
financial incentive for the business to remain open.
This article is protected by copyright. All rights reserved.
increase their weekly earnings relative to their ‘usual’ earnings. This will occur
Author Manuscript
whenever ℎ𝑝𝑟𝑒𝐶 ∙ 𝑤 is less than $750.
Third, we describe the incidence of benefits derived from JobKeeper. The
component of JobKeeper received by an employer as a subsidy is equal to
ℎ𝑝𝑜𝑠𝑡𝐶 ∙ 𝑤, up to a limit of $750. This is because ℎ𝑝𝑜𝑠𝑡𝐶 ∙ 𝑤 is the amount that a
profit-maximising employer would have been prepared to pay to a worker, even
in the absence of JobKeeper. The component of JobKeeper that represents an
increase in income to a worker is equal to the difference between $750 and the
employer subsidy. Two polar examples can illustrate. First, suppose that ℎ𝑝𝑜𝑠𝑡𝐶 =
0. That is, the profit-maximising hours post-COVID-19, both with and without
JobKeeper, equal zero. Hence, the entire JobKeeper payment represents an
income transfer to the worker. Second, suppose that ℎ𝑝𝑜𝑠𝑡𝐶 ∙ 𝑤 > $750. That is,
the employer would have been willing to pay weekly earnings above $750, even
in the absence of JobKeeper. Therefore, the entire JobKeeper payment represents
a subsidy to the employer.
c] Impact on employment
It has already been noted (with regard to Figure 3) that data on the number of
jobs in Australia and the timing of announcement of the JobKeeper program is
suggestive of an impact of the program – in slowing or stopping the decline in
jobs. However, that cannot be taken as causal evidence. For example, there are
other reasons – such as implementation of government restrictions on business
activity being completed at around the same time as announcement of JobKeeper
– that might explain the apparent association. 19
19 On the other hand, there are reasons why the impact of the JobKeeper program may
not coincide exactly with its date of announcement. For example, it was not until the
first week of May that the first tranche of JobKeeper payments were made to employers
and self-proprietors; and therefore only if employers were able to fund wage payments
through to that time would the announcement be expected to have an immediate impact
on employment.
This article is protected by copyright. All rights reserved.
It is possible to move a little closer to establishing causal evidence for impact of
Author Manuscript
the JobKeeper program on employment using two types of evidence obtained
from the Xero SBI database. First, we compare the change in the number of jobs
held by short-term casual employees who were not eligible for JobKeeper
because they had been with their employer for less than 12 months, with the
change for long-term casual employees who were eligible due to having been
with their employer for at least 12 months. Second, we compare the impact of
the introduction of wage subsidy programs for Australia and New Zealand. This
comparison allows us to make inferences on the relation between the timing of
introduction of a wage subsidy program and its impact on employment.
Figure 12 shows the evolution of the number of non-eligible (short-term) and
eligible (long-term) casual jobs from before to after the announcement of
JobKeeper. Prior to the introduction of JobKeeper, the number of short-term
casual jobs had decreased by about 12.5 per cent and long-term casual jobs by
16.5 per cent. But then in the three weeks following the announcement of the
program, short-term casual jobs decreased by about 9 per cent, compared to an
increase of 2 per cent for long-term casual jobs. We take the comparison of job
outcomes for non-eligible (short-term) and eligible (long-term) casual
employees as suggesting a causal impact of JobKeeper on employment. 20
20 After three weeks, the series converge – which is likely to reflect that all new hires of
casual employees will by definition be short-term. See also the comparison of changes in
the number of jobs held by employees who were eligible and ineligible for JobKeeper in
Commonwealth Treasury (2020, p.24).
This article is protected by copyright. All rights reserved.
Figure 12: Number of casual jobs, By tenure with current employer, Weeks
Author Manuscript
ending 2 February to 24 May 2020
110
105
Introduction of
100 JobKeeper
95
Number of jobs
90
85
80
75 Long-term casual
70
Short-term casual
65
60
2-Feb 9-Feb 16-Feb 23-Feb 1-Mar 8-Mar 15-Mar 22-Mar 29-Mar
5-Apr 12-Apr 19-Apr 26-Apr 3-May 10-May 17-May 24-May
Source: Xero Small Business Insights; Analysis by Alpha Beta.
More support for this finding of a causal impact of JobKeeper comes from a
comparison of Australia with New Zealand. Figure 13a provides information on
numbers of jobs in each country – compared against the timing of announcement
of the respective wage subsidy policies. It shows that, in each country, after the
introduction of the policy, there is no further substantial decrease in
employment.
Of course, differences between countries in the size of job losses could be
explained by a different impact of COVID-19. To address this issue, Figure 13b
shows changes in total jobs and revenue for small business in each country. It
shows that, in fact, there was a larger decrease in small business revenue in New
Zealand than Australia. Greater ‘undoing’ of the revenue impact on employment
in New Zealand than Australia may relate to details of their wage subsidy policies
– although the policies appear relatively similar in design. 21 Hence, it raises the
21 New Zealand’s wage subsidy program is similar to Australia: It pays employers
$585.80 pw for an employee working more than 20 hours or more and $350.00pw for
an employee working less than 20 hours. Employers are required to make ‘best
This article is protected by copyright. All rights reserved.
interesting possibility that it is the timing of introduction of the policy that
Author Manuscript
matters for its impact. That is, impact of policies such as JobKeeper are mainly
prospective – stopping future job losses that would otherwise have occurred.
Figure 13a: Number of jobs, Australia and New Zealand, 2 February to 24
May 2020
105
Number of jobs (Index = 100 on 2/2/2020)
100
95
17/3:
90
NZ wage subsidy
announced 30/3:
85 JobKeeper
announced
80
2-Feb 9-Feb 16-Feb 23-Feb 1-Mar 8-Mar 15-Mar 22-Mar 29-Mar
5-Apr 12-Apr 19-Apr 26-Apr 3-May 10-May 17-May 24-May
Australia New Zealand
Source: Xero Small Business Insights; Analysis by Alpha Beta.
endeavours’ to pay at least 80 per cent of regular weekly earnings. Eligibility requires
proof of a 30 per cent decrease in revenue for any month relative to a comparable
month in 2019; https://www.ey.com/en_au/covid-19/all-you-need-to-know-about-the-
new-zealand-wage-subsidy
This article is protected by copyright. All rights reserved.
Figure 13b: Change in revenue and jobs, Australia and New Zealand,
Author Manuscript
February to May 2020
0
-45 -40 -35 -30 -25 -20 -15 -10 -5 0 5
April
Change in number of jobs (Per cent
New Zealand -2
M
-4
Australia
-6
compared to February)
M h
M
-8
-10
M h
April
-12
Change in revenue (Per cent compared to February)
Source: Xero Small Business Insights; Analysis by Alpha Beta.
While our analysis suggests that JobKeeper has affected employment, this is very
much a preliminary conclusion. Our evidence is only descriptive; and it is with
regard to a specific type of employee working in small business, raising the issue
of external validity. Furthermore, the operation of other stimulus programs
raises the possibility that those programs increased the demand for labour by
business, and that JobKeeper simply gave an incentive for that increase in
demand to be directed to towards employees eligible for the program. 22
d] Impact on labour income
The JobKeeper program has as a major objective to underpin the labour income
of employees and sole proprietors in businesses adversely affected by COVID-19.
The Commonwealth Treasury JobKeeper report (2020, p.28) states that one-
22 This identification problem could be potentially be overcome by comparing similar
firms which differ in their proportions of eligible and non-eligible casual employees.
Making such a comparison for small businesses in the hospitality industry finds that the
size of job losses experienced by a business were inversely correlated with the
proportion of its casual employees who were eligible to receive JobKeeper.
This article is protected by copyright. All rights reserved.
quarter of JobKeeper payments in April-May represented an increase in labour
Author Manuscript
income. Expressed in ‘person equivalents’ this suggests that 875,000 workers
have had their incomes increased by JobKeeper (relative to the post-COVID-19
outcome in the absence of JobKeeper).
There has also been great interest in the distributional implications of
JobKeeper. One source of this interest has been exclusion from eligibility of
short-term casual employees and some categories of migrants. The exclusion of
these groups implies an increase in income inequality (at least in the short-term)
between them and the eligible population of workers (Cassells and Duncan,
2020a, provide further analysis). The other interest regarding distributional
impacts is those workers for whom the JobKeeper payment represents a
significant increase compared to their usual earnings. The Commonwealth
Treasury JobKeeper report (2020, pp.27-28) states that overall, about one-
quarter of JobKeeper recipients experienced an increase in income compared to
pre-COVID-19 weekly earnings, with the average increase being $275 per week.
This implies that a little over one-third of the total increase in labour income has
been as above-usual earnings. 23
e] Social welfare implications
JobKeeper may have increased employment and provided income support, but
that spending will need to be repaid by taxpayers. Hence, the relative size of
benefits and costs of the program is important question. At this stage, it is only
possible to provide a partial and preliminary answer. What is apparent, however,
23 Such an impact has been most evident for the population aged less than 20 years.
Between mid-March and late-June their jobs fell by 5.2 per cent, yet the wages paid to
this group increased by 11.6 per cent. What this roughly implies is that the average
worker aged less than 20 years – who has retained their job – was receiving about a 20
per cent increase in earnings. ABS, Weekly Payroll and Jobs in Australia, Week ending 27
June 2020, 6160.0.55.001, Table 4. Changes are for 14 March to 13 June. Cassells and
Duncan (2020b) also provide a detailed description of replacement rates of JobKeeper
by occupation.
This article is protected by copyright. All rights reserved.
is that the answer will depend critically (as in most cases) on how benefits are
Author Manuscript
defined.
A narrow approach to defining benefits would focus on the short-term impact of
the program on workers. Using this approach, it would have to be concluded that
that the JobKeeper program has been relatively expensive. The Commonwealth
Treasury JobKeeper report (2020, p.28) states that one-quarter of JobKeeper
payments in April-May represented an increase in labour incomes; implying that
three-quarters have gone to employers as a subsidy, paying wages for workers
that would have been paid for in any case by an employer. 24 Moreover, this is
likely to be the peak share of JobKeeper payments going to labour, since the
share would be expected to decline as economic activity recovers. Consistent
with the Treasury JobKeeper report, Breunig and Watson (2020) estimate that
the cost per job retained by JobKeeper for 6 months is between $70,000 and
$102,000. They note that this is high by comparison with international
programs; for example, the American Recovery and Reinvestment Act of 2008
had an average cost per job saved of $US47,619 (Chodorow-Reich, 2019). 25
A broader approach to defining benefits is to also include longer-term effects of
the JobKeeper program on the viability of businesses (which of course then
24 This is roughly the size of transfer to labour that emerges from a back-of-the-
envelope calculation. First, the component of JobKeeper that is paid to workers
whose hours have been reduced relative to what would have occurred can be
estimated as the difference between the decreases from March to May in hours
worked and persons employed (adjusted for the full-time/part-time composition
of the decrease in employment) – equal to 5.0 per cent. Second, the component of
JobKeeper that is paid to workers whose earnings are usually below $750pw can
be estimated as: (Share of workforce receiving JobKeeper)*(Share of employees
whose AWE are below $750pw)*(Average fraction by which their AWE are
below $750pw) = 2.5 per cent. This implies a total transfer to labour equivalent
to 7.5 per cent of the workforce in March 2020, equal to 975,000 persons. This
represents 27.8 per cent of workers for whom JobKeeper is being paid.
25 See also Wilkins and Borland (2020). The issue of targeting is being reflected
in media discussion of issues such as payment of bonuses to management of
businesses that have received JobKeeper (for example, Powell, 2020).
This article is protected by copyright. All rights reserved.
becomes an indirect source of benefits to workers). There does seem reasonable
Author Manuscript
evidence that, in general, JobKeeper payments have gone to support businesses
that have experienced the largest decreases in revenue. Evidence from the
Commonwealth Treasury JobKeeper report (2020, p.32) indicates that
businesses receiving payments experienced average declines in revenue from
April 2019 to April 2020 that were 35 per cent larger than businesses not
receiving payments. 26 What remains to be resolved – and will be critical for
measuring the benefits of JobKeeper using a definition that includes the impact
on business viabaility - is the extent to which the subsidy to business has allowed
them to increase their level of activity and employment, and to remain viable for
the longer-term, relative to if JobKeeper was not available.
f] Impacts on labour market dynamics
An emerging theme in discussions of whether to continue the JobKeeper
program beyond its initial end-date (at the end of September 2020) is its effect
on efficiency (for example, Commonwealth Treasury, 2020, pp.29-33).
Employers only receive JobKeeper payments for employees in their existing
workforce on March 1; and employees only receive JobKeeper if they remain
connected to their employer at that time. Hence, it was always understood that
the JobKeeper program would ‘freeze’ the labour market: reducing the extent of
worker transfers between jobs (for example, from low-productivity to high-
productivity firms). As well, only businesses that existed on March 1 are eligible
for JobKeeper payments: thereby underpinning the survival of businesses that
26 Another approach to investigate this question is to compare the take-up of
JobKeeper by industry with ABS Labour Force Survey data on decreases in hours
worked by industry. This is done in Appendix Figure 2.2 by comparing industry
shares of decreases in hours worked and recipients of JobKeeper. A well-targeted
program would imply a close correspondence between those shares. The figure
shows that for the most part there appears to be a close relation. The main
exceptions are accommodation and food services where the share of JobKeeper
recipients is much less than the share of loss in hours of work; and construction
where the reverse applies.
This article is protected by copyright. All rights reserved.
would, even in regular economic conditions, have ceased to exist; and acting as a
Author Manuscript
disincentive for new businesses to begin operation.
These costs of missing productivity growth due to freezing of labour mobility
and reduced firm entry and exit, however, seem likely to be negligible. Annual
labour productivity growth from 2012 to 2019 in Australia was 1.1 per cent per
annum (ABS, Australian System of National Accounts, catalogue no.5204.0, Table
1). The share of productivity growth that can be attributed to employment
reallocation – based on reviews of microeconomic studies of firm-level
productivity - is about one-quarter to one-half (Foster et al., 1998, Table 4).
Hence, the cost of missing productivity growth is equal to about 0.25 to 0.55 per
cent of GDP per annum. At the time of writing in late July, the costs of
underutilised labour due to COVID-19 therefore overwhelm the efficiency costs
of JobKeeper. The JobKeeper program would only need to achieve a small
increase in employment in order to outweigh any efficiency costs it might
impose.
5. Conclusion
The impact of the COVID-19 recession on labour market outcomes in Australia is
hard to keep up with. After the largest one-month decrease in hours worked
from March to April, this was followed in May to June by the largest one-month
increase (in the period following 1978 for which data on monthly hours are
available). At the same time, labour market adjustment has been very different to
previous recessions in Australia.
In this article, we have argued that what happened in the Australian labour
market from March to June is best understood by looking at the main drivers –
first, the impact of COVID-19 on household spending via government business
restrictions and consumers voluntarily withdrawing from activities in which
they perceive themselves to be at risk of contracting the virus ; and second, the
government’s fiscal response, and especially the JobKeeper program.
This article is protected by copyright. All rights reserved.
Understanding the impact of COVID-19 on household spending means, to begin
Author Manuscript
with, understanding the concentrated industry-level impact. That impact is
evident from large differences in changes in the amount of work by industry.
Because some types of workers and jobs are concentrated in the most affected
industries, the industry impact also becomes a lens to understand, for example,
why the young have been particularly hard hit during the early stages of the
COVID-19 recession.
Recognising the role of the government fiscal response is essential for
understanding the path followed by aggregate labour market outcomes. Take the
example of employment. Both through the impact of increased household
spending supported by government payments and related initiatives, and via
incentives for employers to retain employees through the JobKeeper program
and other business support, government fiscal policy appears to have had a
strong influence on the path of aggregate employment. The same policies are
also needed to understand specific developments that occurred; for example,
large decreases in employment for short-term casual employees and recent
migrants appearing to be associated with eligibility conditions of the JobKeeper
program.
Beyond June, developments with COVID-19 and government health and fiscal
policy will continue to affect labour market outcomes. A third driver will
probably also have a substantial influence on the length and depth of recession
that Australia experiences. Income losses for households and businesses, and
their levels of confidence about the future, are likely to cause further decreases
in household spending and business investment. The COVID-19 recession could
then become a longer-term and broader-based type of recession (see for
example McDonald, 2020). At the same time, structural changes in the labour
market unleashed by COVID-19 (such as increased working from home, an
acceleration in the shift to online buying and more extensive use of tele-health)
promise to make for a more disrupted path to recovery (see Autor et al., 2020;
and Barrero et al., 2020).
This article is protected by copyright. All rights reserved.
References
Author Manuscript
Alon, Titan M., Matthias Doepke, Jane Olmstead-Rumsey and Michele Tertilt
(2020), ‘The impact of COVID-19 on gender equality’, National Bureau of
Economic Research, Working Paper no.26947.
Australian Government (2020), ‘Supporting businesses to retain jobs’, Fact
Sheet; accessed at:
https://www.qta.com.au/resources/Fact%20Sheets/Coronavirus%20Fact%20S
heets/JobKeeperPackage_Summaries%20for%20Employers_Employees.pdf
Autor, David and Elisabeth Reynolds (2020), ‘The nature of work after the
COVID-crisis: Too few low-wage jobs’, Hamilton Project Essay 2020-14; accessed
at: https://www.brookings.edu/wp-
content/uploads/2020/08/AutorReynolds_LO_FINAL.pdf
Ball, Jarrod (2020), ‘Real time economic data is in big demand but short supply’,
CEDA Blog; accessed at: https://www.ceda.com.au/Digital-hub/Blogs/CEDA-
Blog/April-2020/COVID-19-Real-time-economic-data-is-in-high-demand-but-
short-supply
Barrero, Jose Maria, Nick Bloom and Steven Davis (2020), ‘COVID-19 is also a
reallocation shock’, forthcoming, Brookings Papers on Economic Activity.
Bishop, James (2020), ‘Economic effects of the Spanish flu’, RBA Bulletin June;
accessed at:
https://www.rba.gov.au/publications/bulletin/2020/jun/pdf/economic-effects-
of-the-spanish-flu.pdf
Bishop, James, Linus Gustafsson and Michael Plumb (2016), ‘Jobs or hours?:
Cyclical labour market adjustment in Australia’, Reserve Bank of Australia, RDP
16-06.
Borland, Jeff (2009), ‘What happens to the Australian labour market in
recessions?’, Australian Economic Review, 42(2): 232-42.
Borland, Jeff (2011), ‘The Australian labour market in the 2000s: The quiet
decade’, in Hugh Gerard and Jonathan Kearns (eds.) The Australian Economy in
the 2000s (Reserve Bank of Australia).
Borland, Jeff (2020a), ‘COVID and the Australian labour market: The immediate
impact’, Labour Market Snapshot 56, March; accessed at:
https://drive.google.com/file/d/1azG0Sg7ezMtssRh_p05xFk0WspTJmpde/view
Borland, Jeff (2020b), ‘Female and male employment during recessions in
Australia’, Labour Market Snapshot #62; accessed at:
This article is protected by copyright. All rights reserved.
https://drive.google.com/file/d/1W0AQS7H9gUemgHHBOKj6j7OAaGO1lQMS/v
Author Manuscript
iew
Borland, Jeff (2020c), ‘Scarring effects: A review of main international and
Australian literature, forthcoming, Australian Journal of Labour Economics.
Borland, Jeff and Michael Coelli (2020), ‘It is dog days for the young in the
Australian labour market?’; accessed at:
https://drive.google.com/file/d/1HW_6fwNxfP3VvRcol83t9A14SzciuiCA/view
Breunig, Robert and Tim Watson (2020), ‘Strengthening JobKeeper’; accessed at:
https://www.policyforum.net/strengthening-jobkeeper/
Cassells, Rebecca and Alan Duncan (2020a), ‘Short-term and long-term casual
workers: How different are they?’, Bankwest Curtin Economics Centre Research
Brief COVID-19 #4; accessed at: https://bcec.edu.au/covid-19-data-insights-4-
short-term-and-long-term-casual-workers/
Cassells, Rebecca and Alan Duncan (2020b), ‘JobKeepers and JobSeekers: How
many workers will lose and how many will gain?’, Bankwest Curtin Economics
Centre Research Brief COVID-19 #3; accessed at: https://bcec.edu.au/covid-19-
data-insights-3-job-keepers-and-job-seekers-how-many-workers-will-lose-and-
how-many-will-gain/.
Cassells, Rebecca, Alan Duncan, Mike Dockery, Daniel Kiely and Astghik
Mavisakalyan (2020), ‘Potential job losses in the COVID-19 pandemic’, Bankwest
Curtin Economics Centre Research Brief COVID-19 #2; accessed at:
https://bcec.edu.au/covid-19-data-insights-2-potential-job-losses-in-the-covid-
19-pandemic/.
Chetty, Raj, John N. Friedman, Nathaniel Hendren, Michael Stepner and The
Opportunity Insights Team (2020), ‘How did COVID-19 and stabilization policies
affect spending and employment? A new real-time economic tracker based on
private sector data’, National Bureau of Economic Research, Working Paper no.
27431.
Chodorow-Reich, Gabriel (2019), ‘Geographic cross-sectional fiscal spending
multipliers’, American Economic Journal: Economic Policy, 11(2): 1-34.
Coates, Brendan, Matt Cowgill, Tony Chen and Will Mackey (2020), ‘Shutdown:
Estimating the COVD-19 employment shock’, Grattan Institute Report; accessed
at: https://grattan.edu.au/report/shutdown-estimating-the-covid-19-
employment-shock/
Commonwealth Treasury (2020), ‘The JobKeeper payment: Three-month
review’; accessed at: https://treasury.gov.au/publication/jobkeeper-review
This article is protected by copyright. All rights reserved.
Daley, John, Danielle Wood, Brendan Coates, Stephen Duckett, Julie Sonnemann,
Author Manuscript
Marion Terrill and Tony Wood (2020), ‘The recovery book: What Australian
governments should do now’, Grattan Institute Report; accessed at:
https://grattan.edu.au/report/recovery-book/
Dhillon, Zoya and Natasha Cassidy (2018), ‘Labour market outcomes for young
people’, RBA Bulletin, June. Accessed at:
https://www.rba.gov.au/publications/bulletin/2018/jun/labour-market-
outcomes-for-younger-
people.htmlhttps://www.rba.gov.au/publications/bulletin/2018/jun/pdf/labou
r-market-outcomes-for-younger-people.pdf
Foster, Lucia, John Haltiwanger and C.J. Krizan (1998), ‘Aggregate productivity
growth: Lessons from microeconomic evidence’, National Bureau of Economic
Research, Working Paper no. 6803.
Goolsbee, Austan and Chad Syverson (2020), ‘Fear, lockdown and diversion:
Comparing drivers of pandemic economic decline 2020’, National Bureau of
Economic Research, Working Paper no.27432.
Gupta, Sumedha, Kosali Simon and Coady Wing (2020), ‘Mandated and voluntary
social distancing during the COVID-19 epidemic’, forthcoming, Brookings Papers
on Economic Activity.
Hupkau, Claudia and Barbara Petrongolo (2020), ‘COVID-19 and gender gaps:
Latest evidence and lessons from the UK’, VOX CEPR Policy Portal, 22 April;
accessed at: https://voxeu.org/article/covid-19-and-gender-gaps-latest-
evidence-and-lessons-uk
Kennedy, Steven, Jim Thomson and Peter Vujanovic (2006), ‘A primer on the
macroeconomic effects of an influenza epidemic’, Treasury Working Paper 2006-
01; accessed at: https://treasury.gov.au/publication/2006-01-a-primer-on-the-
macroeconomic-effects-of-an-influenza-pandemic/2006-01-a-primer-on-the-
macroeconomic-effects-of-an-influenza-pandemic
McDonald, Ian (2020), ‘Macroeconomic policy to aid recovery after social
distancing for COVID-19’, forthcoming, Australian Economic Review.
Productivity Commission (2020), ‘Why did young people’s incomes decline?’,
Research Paper; accessed at:
https://www.pc.gov.au/research/completed/youth-income-decline
Powell, Dominic (2020), ‘JobKeeper could boost company exec bonuses’, The Age
Business, July 4, p.1.
Rayner, Jennifer (2016), Generation Less: How Australia is Cheating the Young
(Redback Quarterly 9, Schwartz Publishing).
This article is protected by copyright. All rights reserved.
Wilkins, Roger (2020), ‘Who’s hardest hit by the economic effects of COVID-19?’,
Author Manuscript
Melbourne Institute Research Insights, No. 10/20.
Wilkins, Roger and Jeff Borland (2020), ‘That estimate of 6.6 million Australian
workers on JobKeeper, it tells us how it can be improved’, April 29; accessed at:
https://theconversation.com/that-estimate-of-6-6-million-australians-on-
jobkeeper-it-tells-us-how-it-can-be-improved-137237
Appendix 1: Data sources – Important details for interpreting the labour
market impact of COVID-19
1] The Australian Bureau of Statistics: Labour Force Survey
The monthly Labour Force Survey (LFS) covers approximately 0.32% of the
Australian civilian population aged 15 years and over. The survey occurs across a
2-week period centred on the Sunday that falls between the 5th and 11th of each
month; and therefore mainly reflects developments in the first half of each
month. In March 2020, that date was March 8 so that the LFS covered March 1-
14. For that reason, the March LFS does not reveal any significant impact of
COVID-19, since major impacts of COVID-19 occurred after that date.
The ABS classification of employment/unemployment follows conventions
established by the International Labor Organisation. A person is classified as
employed if they… ‘were away from their job for any reason (e.g. they were stood
down), and were paid for some part of the previous 4 weeks (which could
include wages subsidised through the JobKeeper scheme); or were away from
their job for four weeks or less for any reason, without pay, but believe they still
have a job to go back to (e.g. they were stood down, with no pay)’ (ABS, Labour
Force Australia, 6202.0 – Summary: Classifying People During the COVID-19
Period). Hence, the JobKeeper program, and other arrangements between firms
and their employees, are likely to have caused an increased proportion of
persons classified as employed but who worked zero hours.
2] The Australian Bureau of Statistics and Australian Taxation Office: Weekly
Payroll Jobs and Wages
The ABS/ATO Payroll data encompasses employee jobs for which a payment was
reported to the ATO through the Single Touch Payroll (STP) system. The data
have been released as a weekly series. Being restricted to paid employee jobs,
the Payroll data excludes owner/managers of unincorporated enterprises and
contributing family workers, about 8.9 per cent of employment in April 2020
(ABS, Labour Force Australia, 6202.0 – Summary: Understanding differences
between labour force employment statistics and weekly payroll jobs). Around 99
per cent of businesses with 20 or more employees, and 80 per cent of businesses
with less than 20 employees, are estimated to report through the STP system
This article is protected by copyright. All rights reserved.
(Commonwealth Treasury, 2020, p.4). The Payroll data include all jobs, so
Author Manuscript
persons who have multiple jobs are counted multiple times. About 6.7 per cent of
persons employed have multiple jobs (ABS, Labour Account Australia, Quarterly
Experimental Estimates, 6150.0.55.003, December 2019, Table 1). All jobs where
an employee was paid are included in the Payroll data. Hence, jobs occupied by
an employee receiving JobKeeper and working zero hours are included in the
number of jobs in the Payroll data.
3] Xero Small Business Insights data
The Xero Small Business Insights program provides analysis on the sector’s
health, with its metrics based on anonymised, aggregated data drawn from
hundreds of thousands of subscribers. Xero has produced a series of specialised
monthly metrics, providing a week by week view of the impact of the COVID-19
event. The data series for Australia covers employees at about 115,000
businesses who receive weekly payslips – accounting for about 54 per cent of
employees on the Australian Xero database. A variety of exclusion restrictions
are applied to arrive at the final data set for estimating the number of jobs: (i)
Very new firms and firms which ceased operations after the onset of COVID-19
are excluded; (ii) Trusts, self-managed superannuation funds and unclassified
businesses are dropped; and (iii) Payslips with less than one hour of work, more
than 31 days of work or a rate of pay less than $10 per hour are excluded.
4] illion data
Illion is one of Australia’s leading credit bureaus with access to the anonymised
and aggregated bank transactions of more than a million Australians. To track
the impact of COVID19, illion and AlphaBeta have used a weekly sample of
250,000 individuals to describe spending trends across Australia by category.
The illion sample is weighted by age, geography, gender and income to be
representative of the Australian population.
Appendix 2: Extra tables and figures
Appendix Table 2.1: Employed persons who worked less hours than usual
working zero hours, March to May 2020
March to April to May to March to
April May June June
No work, not enough 690.8 -396.6 -137.9 153.3
work available, or
stood down
This article is protected by copyright. All rights reserved.
Annual leave, 42.2 -234.1 31.2 -160.7
Author Manuscript
holidays, flexitime or
LSL
Began, left or lost a 334.8 -182.6 -94.5 57.7
job during the week
Source: ABS, Labour Force Australia, Detailed – Electronic Delivery,
6291.0.55.001, EM2a.
Appendix Figure 2.1: Percent change in employment by industry, February
to March 2020
30
20
ABS/ATO Payroll 10
Labour Force Survey
0
-40 -30 -20 -10 -10 0 10
-20
-30
-40
-50
-60
-70
Source: a] LFS – ABS, Labour Force Australia, Detailed – Quarterly,
6291.0.55.003, Table 1; b] ABS/ATO – ABS, Weekly Payroll Jobs and Wages in
Australia, 6160.0.55.001, Table 4.
Appendix Table 2.2a: Change in persons employed (per cent), By
occupation, February to May 2020
Actual change Predicted change: Based
on change in
employment by industry
Managers -9.0 -10.8
Professionals -3.2 -8.5
This article is protected by copyright. All rights reserved.
Technicians and trade -13.6 -12.1
Author Manuscript
workers
Community and personal -29.1 -16.0
service workers
Clerical and administrative -4.8 -9.2
workers
Sales workers -16.6 -13.2
Machinery operators and -9.4 -12.5
drivers
Labourers -17.3 -13.4
Source: ABS, Labour Force Australia, Detailed – Quarterly, 6291.0.55.003, EQ09.
Appendix Table 2.2b: Change in persons employed (per cent), By age,
February to May 2020
Actual change Predicted change: Based
on change in
employment by industry
15-24 -22.7 -16.0
25-54 -9.4 -10.1
55 plus -7.7 -9.5
Source: ABS, Labour Force Australia, Detailed – Quarterly, 6291.0.55.003, EQ12.
This article is protected by copyright. All rights reserved.
Appendix Table 2.2c: Change in persons employed (per cent), By gender,
Author Manuscript
February to May 2020
Actual change Predicted change: Based
on change in
employment by industry
Females -7.1 -6.9
Males -5.6 -5.7
Source: ABS, Labour Force Australia, Detailed – Quarterly, 6291.0.55.003, EQ12.
Appendix Table 2.3: Changes in employment by occupation, March to May
2020
Per cent decrease in
hours worked
Managers -9.0
Professionals -3.2
Technicians & trade workers -13.6
Community & personal service -29.1
workers
Clerical & administrative workers -4.8
Sales workers -16.6
Machinery operators & drivers -9.4
Labourers -17.3
This article is protected by copyright. All rights reserved.
Appendix Table 2.4: Per cent change in employment and hours worked by
Author Manuscript
employment status in main job, March to June 2020
Share of Per cent change Per cent change
employment – in employment in actual hours
March 2020 worked in main
job
Employee 83.9 -5.1 -6.9
Owner/manager 5.6 +2.8 +0.1
with employees
Owner/manager 10.2 -6.1 -18.1
without employees
Source: ABS, Labour Force Australia, Detailed – Electronic Delivery,
6291.0.55.001, EM6.
Appendix Figure 2.2: Changes to hours worked and claims for JobKeeper,
By industry
30
Share of decrease in hours worked
25
20 Accommodation & food services
15
(March to May 2020)
10
5 Construction
0
0 5 10 15 20 25 30
Share of employees and eligible business participants receiving
JobKeeper
Sources: a] JobKeeper take-up: Commonwealth Treasury (2020, Appendix Table
4); b] Share of decrease in hours worked: ABS, Labour Force Australia, Detailed –
Quarterly, 6291.0.55.003, EQ11.
This article is protected by copyright. All rights reserved.
File and source
- File
- the-australian-labour-market-and-the-early-impact-of-covid-19-an-assessment.pdf
- Size
- 562,181 bytes
- SHA-256
- caf062fb20c89a42cc506bf8c2f9aba04f9f272a5049757478adf0c497ad7f13
- Original
- doi.org