## 1. Key Real Sector and Labor Market Developments

## Source details

**Canonical URL:** [1. Key Real Sector and Labor Market Developments](https://www.imf.org/-/media/files/publications/wp/2017/wp17124.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2017/wp17124.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2017/wp17124.pdf.json)

---

### Introduction: shocks and aggregate outcomes
- Two major shocks: the global financial crisis (GFC) and the prolonged rise and subsequent sharp drop in commodity prices.
- Prices of Australian resource exports nearly doubled between 2005 and 2011, driving the terms of trade to an all-time high.
- Mining investment:
  - rose from an average of around 2-3 percent of GDP before 2005 to a peak exceeding 9 percent of GDP in 2013;
  - fell to 4½ percent of GDP by 2016 after a 35 percent decline in commodity prices since 2011.
- Real GDP growth:
  - fell from average rates of around 3¼ percent pre-GFC to around 2½ percent post-GFC.
- Unemployment:
  - reached a trough of 4 percent prior to the GFC;
  - rose to an average of 5½ percent between 2009 and 2016;
  - peaked at around 6¼ percent in mid-2015;
  - since then fluctuated around 5¾ percent.
- Labor market characteristics:
  - relatively mild unemployment fluctuations (within ¾ of a percent of the 2000-16 average of 5½ percent) attributed to flexible labor markets aiding rebalancing.
  - long-term unemployment and underemployment have remained elevated since the GFC and the terms-of-trade collapse.
- Estimates and gaps:
  - the unemployment gap in 2016 is estimated at just under 1 percent (based on a NAIRU jointly estimated with the output gap using a small Bayesian model and Kalman filter).
  - estimates of potential output indicate a negative output gap (excess capacity) in the range of 1.25 to 1.75 percent of potential GDP.
  - Okun coefficient references: Ball and others (2013) estimate -0.4 (post-1995); Tulip and Lancaster (2015) estimate between -0.27 and -0.35.

### Does higher long-term unemployment indicate more labor market frictions?
- Long-term unemployment developments:
  - percent share of long-term unemployed in total unemployed rose from around 15 percent in mid-2009 to around 20 percent by 2011, and to around 25 percent since the mining investment decline.
  - duration of unemployment increased from a pre-GFC trough of 27 weeks to a high of 50 weeks in 2016.
  - long-term unemployment rate has risen to levels seen in the early 2000s (but well below 1990s peaks).
- NAIRU and Beveridge curve evidence:
  - estimate of the NAIRU in Australia has risen by a small amount since 2011.
  - simple Beveridge curve shows outward shifts in the 1980s and the 1990s; broadly stable since then aside from a small outward shift following the terms-of-trade and mining investment decline.
  - a fitted Beveridge curve (Hobijn and Sahin (2012) method) indicates 2016 unemployment was around ¾ of a percent higher than the rate consistent with the labor market equilibrium implied by the curve.
- Interpretation:
  - the ¾ of a percent gap is consistent with the NAIRU-based gap and does not necessarily signal a major structural deterioration.
  - the recent outward shift is small relative to earlier downturns and much smaller than post-GFC outward shifts in some other advanced economies (Portugal, Spain, Sweden, U.K., U.S.).

### Implications of greater labor market flexibility for slack and wages
- Hours flexibility and reforms:
  - early 1990s reforms made it easier for firms to bargain directly with employees, facilitating adjustment of hours worked.
  - since the early 2000s a relatively greater share of adjustment in aggregate hours worked has been borne by hours per worker rather than aggregate employment.
- Peak-to-trough dynamics:
  - in the early 1990s, cyclical declines in total hours worked were mainly via reductions in employment.
  - from the early 2000s onward, downturn adjustment has been almost equally through reductions in average hours and in employment; amplitude of peak-to-trough declines in total hours worked has decreased.
- VAR/IRF evidence:
  - impulse response functions (VAR in log detrended GDP, hours per worker, employment) over 1984Q4–1997Q4 and 1998Q1–2016Q1 indicate the response of employment to a 1 percent GDP shock has moderated in the later period.
- Conjunctural and structural factors:
  - tight labor markets before the GFC may have increased hiring/firing costs; during the GFC increased worker uncertainty may have raised willingness to accept lower hours for job security.
  - rising skill requirements have likely increased screening and training costs for firms.
  - educational attainment: share of workers with qualifications up to a Bachelor degree or higher rose from 18 percent in 1993 to 26 percent in 2009 (ABS Survey of Education and Training).
  - occupational demand shifts: increases in cognitive non-routine tasks (management and professional activities) and non-cognitive non-routine tasks (personal care); declines in non-cognitive routine tasks (machinery and plant operation) and cognitive routine tasks (clerical and secretarial) (Borland, 2011).
- Sectoral heterogeneity:
  - sectoral VARs (1990Q1–2016Q1) show cumulative 8 quarter employment response to a 1 percent gross value added shock varies across sectors.
  - transportation and storage, and miscellaneous services show larger and more sustained employment impacts compared with manufacturing or health services.
  - negative relationship between cumulative employment adjustment magnitude and the share of workers with at least a Bachelor degree; no similar association for average hours.
- Underemployment and wage implications:
  - since 2014 employment growth pickup has been driven by part-time workers.
  - underemployment rate increased to above 8½ percent, compared to a historical average of around 7 percent before the decline in mining investment.
  - elevated underemployment suggests more slack than the unemployment rate indicates.
  - wage Phillips curve estimates (building on Jacobs and Rush (2015)):
    - models fit wage growth data closely and fit better when underemployment measures are included (higher adjusted R-squared).
    - coefficients on the underemployment gap and lagged change in the underemployment rate terms are negative and significant in relevant models.
    - broader measures of slack (including underemployment) are important for gauging slack and pressures on wage growth.

### Have sectoral shifts in labor affected labor productivity?
- Sectoral reallocation effects:
  - aggregate labor productivity can be affected by shifts in the sectoral allocation of labor.
  - U.S. example cited: decline in share of manufacturing and mining hours and rise in services hours estimated to have reduced labor productivity growth by ¼ and ½ percentage point respectively (relative to unchanged shares), given higher productivity in mining and manufacturing versus services.
- Services productivity:
  - slower pace of labor productivity growth in services could exert a drag on aggregate labor productivity growth going forward (Van Zandweghe, 2016).

### Key findings and policy-relevant conclusions
- Labor markets have adjusted smoothly overall despite major shocks.
- Structural unemployment does not appear to have increased significantly despite elevated long-term unemployment.
- Increased flexibility in labor input (hours per worker) has moderated employment declines during downturns; adjustment to output shocks has shifted partly from employment to hours.
- Rising part-time employment and elevated underemployment (above 8½ percent) imply more slack than unemployment rates alone indicate and have likely contributed to weaker wage growth.
- Rebalancing toward services has been accompanied by a transitory decline in aggregate labor productivity growth, partly due to weaker productivity in services.
- Migration has played a key role in smoothing state-level labor market adjustments.

---

### Sectoral shifts, productivity, and state-level adjustment

### Sectoral shifts and employment contributions
- Manufacturing share has declined steadily over time.
- Mining share in aggregate hours:
  - Rose sharply over the commodity price boom.
  - Dropped sharply since the end of the boom.
- Construction share:
  - Rose steadily during the mining boom.
  - Stabilized at that higher level since.
- Services:
  - Healthcare services share increased following the global financial crisis, with further acceleration more recently.
  - Retail trade and communication shares declined post-GFC, likely due to increasing adoption of internet enabled retail services and expanded use of ICT technology.
- Job contributions since the terms-of-trade peak in late 2011:
  - The mining sector was a major contributor to job growth during the boom but its share has fallen since late 2011.
  - Healthcare services have contributed a third of the jobs added since the peak, higher than its share of one-quarter of total jobs created over the boom.
- Aggregate reallocation effects:
  - The share in hours worked of goods related sectors (mining, manufacturing, utilities, construction, and domestic trade) has declined by nearly 5 percentage points since the global financial crisis and following the mining investment downturn.
  - The share of business services (finance, real estate services, professional services, and administrative services) and particularly household services (food and accommodation, education, healthcare, recreational, and other personal services) has risen correspondingly.

### Labor productivity levels and growth by sector
- Level differences (real gross value added per hour):
  - Labor productivity is markedly higher in goods-related sectors and in business services compared to household services.
- Growth rates over 1986 – 2016:
  - Household services: about 0.7 percent.
  - Goods: 2 percent.
  - Business services: 1.9 percent.

### Aggregate labor productivity dynamics and transition effects
- Overall performance:
  - Labor productivity growth in Australia has sustained its historical average rate through the transition.
  - Post-GFC labor productivity growth has maintained its pre-GFC growth rate of around 1½ percent (in terms of GDP per hour worked).
  - Unlike many other advanced economies exhibiting large labor productivity level gaps relative to their pre-GFC trend, Australia does not appear to exhibit such a gap.
- Short-term developments:
  - Labor productivity growth did slow in 2015 due to shifts in sectoral allocation of labor — described as a transitional “between” effect — and recovered in 2016.
- Risks and offsets:
  - It remains to be seen whether lower labor productivity growth in services sectors will drag aggregate labor productivity growth down to a lower average rate once productivity gains from new mining capacities coming on-stream are fully realized.
  - Increasing competition in some services such as retail should provide continued support to productivity growth.

### State-level adjustment, migration, and labor market flexibility
- Mining boom and states:
  - Mining boom drove a strong pickup in private investment growth in Western Australia (WA) and Queensland (QLD) with strong labor demand (measured by vacancies).
  - New South Wales (NSW), Victoria (VIC), and South Australia (SA) saw investment growth and states’ shares in total investment fall, but experienced strong labor demand growth particularly in NSW.
  - Working age population rose above long-term average growth rates, particularly in WA and QLD, accompanied by rising participation rates and declining unemployment.
  - With the terms-of-trade decline, labor demand in mining states fell sharply, accompanied by sharp declines in working age population growth reflecting reversal in migration inflows.
- Importance of migration:
  - On aggregate, about 50-60 percent of the increase in population is due to international migration.
  - Migration helped avoid labor shortages on the upswing and big increases in unemployment on the downswing in the mining states.
- VAR analysis (state-level VARs over 1979-2015 in log changes in employment, unemployment rate, and labor force participation rate):
  - Following a 1 percent state-specific shock to employment (IRFs):
    - QLD and VIC: employment and participation rates account for around half of the increase in employment over 10 years; thus about half the long-term increase in employment is supported by rising population (migration). In QLD, the employment rate has a smaller role in adjustment relative to participation compared to VIC. In all states, participation rates do more adjusting than employment rates.
    - NSW and WA: migration accounts for between 30-40 percent of the long run increase in employment over 10 years.
    - SA and TAS: migration accounts for only about 10-15 percent of the long run increase in employment.
  - Historical shock decomposition:
    - In NSW, VIC, QLD, and WA, the sum of employment rate and participation rate shocks correspond reasonably well with actual changes in employment, with exceptions at certain points (e.g., QLD decline after 2006 exceeded what would be caused by these shocks alone; VIC and NSW increases in the mid-to-late 2000s larger than those two shocks alone; WA showed positive employment shocks during the mining boom not explained by employment rate and participation shocks alone).
  - Aggregate versus state-specific shocks:
    - WA shows a larger migration response when aggregate cycle is not removed from the data; VARs that do not remove the aggregate cycle show WA has the largest migration response among all states.
- Implications:
  - Migration is a key aspect of labor market flexibility that helped moderate the impact of recent shocks.
  - Over the boom, migration likely prevented labor shortages and additional wage cost pressures.
  - Over the subsequent decline in commodity prices and mining investment, the decline of migration into mining states likely helped prevent unemployment from rising higher and likely prevented wage growth from weakening further.

---

### Cyclical features, Beveridge curve fitting, and wage Phillips curve evidence

### Macroeconomic and measurement context
- Terms of trade and mining investment have reversed sharply (figure captions).
- Commodity prices index referenced to "2013/14 = 100" shown in figures (no numerical series reproduced here).
- Pre-GFC average labor force growth used in steady-state Beveridge calculations set at 1.7 percent.
- Unemployment increased modestly, but long term unemployment has risen (figure captions).
- Underemployment rates are above historical averages (figure captions).
- Recent employment growth has been driven mainly by part-time workers (figure captions).
- Average quarterly hours worked are estimated from monthly hours and employment status (note).

### Beveridge curve and steady-state framework
- Steady-state framework and matching/separations equations:
  - Employment growth over a 1-year period: g_{t,t+1} = (H_{t} - S_{t}) / E_{t}.
  - Cobb-Douglas matching function and separations equation estimated as:
    - ln(H/V) = h_μ + h_α ln(U/V) + ε_{h,t}
    - ln(S/V) = s_μ + s_α ln(U/V) + ε_{s,t}
  - Steady-state implicit condition combines (1)–(3) evaluated at ε_{·,t} = 0 and vacancy rate at observed rates to solve for equilibrium unemployment.
  - Hires and separations inferred from job tenure data following Hobijn and Sahin (2012).
- Regression estimates (Table I.1):
  - Sample: 1986-2008
    - log H/V: Constant 2.1***; log U-V ratio .58***; R-sq 0.89
    - log S/V: Constant -1.5***; log U-V ratio 0.04; R-sq 0.09
  - Sample: 1986-2011
    - log H/V: Constant 2.0***; log U-V ratio .6***; R-sq 0.89
    - log S/V: Constant -1.6***; log U-V ratio .06*; R-sq 0.15
  - Summary: hiring function shows a relatively good fit even with limited sample size; separations equation fit is much weaker.

### VAR specification and cyclical features
- Aggregate VAR:
  - three-variable VAR in log detrended GDP, employment, and average hours worked.
  - two quarterly samples: 1983–1997 and 1998–2016.
  - variables detrended using H-P filter with lambda=1600; seasonal dummies included.
  - empirical findings: “Employment responses were larger, and distinctly above zero in the earlier period, as compared to the later period.” “The response of hours worked does not appear to differ very much across the two samples.”
- Sectoral VAR:
  - excludes agriculture; excludes public administration and safety, education and training, and health and social assistance (public or public-adjacent sectors).
  - sample quarterly 1990–2016.
  - Figure 4 relationship: cumulative 8-quarter employment response versus share with Bachelor degree or higher (2009 survey).

### Wage Phillips curve estimates and interpretation
- Estimated variables and definitions:
  - logprivate t WPI is private sector wage (log change).
  - 1t URgap is the lagged deviation of unemployment from its sample average (sample runs from 1998Q1 – 2016Q1).
  - 1t UERgap is the lagged underemployment rate gap similarly calculated.
  - 1 Bondinfexp t is lagged expected inflation implied by 10-year indexed bonds.
  - 14log t GDPdef   is lagged year-on-year change in log of the GDP deflator.
  - 1 log t labprod   is the lagged change in log output per worker.
- Model variants and key results (Table II.1 and text):
  - Model 1 (underemployment gaps constrained out) similar in fit to RBA (2015); UGap not significant but change in unemployment rate is negative and significant.
  - Model 2 (underemployment gap included): UEGap coefficient negative and significant; lagged change in underemployment rate also negative and significant; Model 2 has higher adjusted R-squared.
  - Model 3: change in underutilization rate significant and sizeable; R-squared larger than Model 1.
  - Models 4 and 5: including lagged wage inflation improves fit, especially Model 5 that includes overall underutilization rate.
  - Overall conclusion: “underemployment gaps do matter for wage growth and may be having some impact in relatively weak wage growth outcomes observed since the terms-of-trade decline.”
- Selected coefficient and fit indicators (as presented):
  - Dlog PWage: 0.3170.306
  - ExpInfB10: 0.0010.0010.0010.0010.001
  - D4log GDPdef: 0.0300.0230.0270.0210.017
  - R-sq (adjusted): 0.600.640.650.650.72

### Appendix III: role of migration in state adjustment (methodology)
- Identity and log decomposition:
  - E - emp - p - wap relationship used to infer migration responses.
- Specification choices:
  - Australian data cast doubt on stationarity in levels of unemployment rate and participation rate; variables entered in first differences and VARs run in log differences.
- Data and VAR sets:
  - employment, unemployment rate, and participation rate at state level from ABS; sample 1979–2015 at annual frequency.
  - first set: state variables “acyclic” to aggregate (aggregate cycle removed); second set: aggregate cycle not removed.
- Unit-root testing:
  - Table III.1 reports ADF tests with rejection of null (unit root) at 5% for changes in certain series across states.

*Source: wp17124 - 1. Key Real Sector and Labor Market Developments*

### 1. Key Real Sector and Labor Market Developments  _____________________________ 14

### 1. Key Real Sector and Labor Market Developments

### Introduction: shocks and aggregate outcomes
- Two major shocks: the global financial crisis (GFC) and the prolonged rise and subsequent sharp drop in commodity prices.
- Prices of Australian resource exports nearly doubled between 2005 and 2011, driving the terms of trade to an all-time high.
- Mining investment rose from an average of around 2-3 percent of GDP before 2005 to a peak exceeding 9 percent of GDP in 2013, then fell to 4½ percent of GDP by 2016 after a 35 percent decline in commodity prices since 2011.
- Real GDP growth:
  - fell from average rates of around 3¼ percent pre-GFC to around 2½ percent post-GFC.
- Unemployment:
  - reached a trough of 4 percent prior to the GFC;
  - rose to an average of 5½ percent between 2009 and 2016;
  - peaked at around 6¼ percent in mid-2015;
  - since then fluctuated around 5¾ percent.
- Labor market characterized by relatively mild unemployment fluctuations (within ¾ of a percent of the 2000-16 average of 5½ percent), attributed to flexible labor markets aiding rebalancing.
- Long-term unemployment and underemployment have remained elevated since the GFC and the terms-of-trade collapse.
- Estimates and gaps:
  - the unemployment gap in 2016 is estimated at just under 1 percent (based on a NAIRU jointly estimated with the output gap using a small Bayesian model and Kalman filter).
  - estimates of potential output indicate a negative output gap (excess capacity) in the range of 1.25 to 1.75 percent of potential GDP.
  - Okun coefficient references: Ball and others (2013) estimate -0.4 (post-1995); Tulip and Lancaster (2015) estimate between -0.27 and -0.35.

### II. Does higher long-term unemployment indicate more labor market frictions?
- Long-term unemployment developments:
  - percent share of long-term unemployed in total unemployed rose from around 15 percent in mid-2009 to around 20 percent by 2011, and to around 25 percent since the mining investment decline.
  - duration of unemployment increased from a pre-GFC trough of 27 weeks to a high of 50 weeks in 2016.
  - long-term unemployment rate has risen to levels seen in the early 2000s (but well below 1990s peaks).
- NAIRU:
  - estimate of the NAIRU in Australia has risen by a small amount since 2011.
- Beveridge curve evidence:
  - simple Beveridge curve shows outward shifts in the 1980s and the 1990s; broadly stable since then aside from a small outward shift following the terms-of-trade and mining investment decline.
  - a fitted Beveridge curve (Hobijn and Sahin (2012) method) indicates 2016 unemployment was around ¾ of a percent higher than the rate consistent with the labor market equilibrium implied by the curve.
- Interpretation:
  - the ¾ of a percent gap is consistent with the NAIRU-based gap and does not necessarily signal a major structural deterioration.
  - the recent outward shift is small relative to earlier downturns and much smaller than post-GFC outward shifts in some other advanced economies (Portugal, Spain, Sweden, U.K., U.S.).

### III. Implications of greater labor market flexibility for slack and wages
- Labor market reforms and hours flexibility:
  - reforms in the early 1990s made it easier for firms to bargain directly with employees, facilitating adjustment of hours worked.
  - average hours worked per worker have become more flexible; since the early 2000s a relatively greater share of adjustment in aggregate hours worked has been borne by hours per worker rather than aggregate employment.
- Peak-to-trough dynamics:
  - in the early 1990s, cyclical declines in total hours worked were mainly via reductions in employment.
  - from the early 2000s onward, downturn adjustment has been almost equally through reductions in average hours and in employment; amplitude of peak-to-trough declines in total hours worked has decreased.
- VAR/IRF evidence:
  - impulse response functions (VAR in log detrended GDP, hours per worker, employment) over 1984Q4–1997Q4 and 1998Q1–2016Q1 indicate the response of employment to a 1 percent GDP shock has moderated in the later period.
- Conjunctural and structural factors:
  - tight labor markets before the GFC may have increased hiring/firing costs; during the GFC increased worker uncertainty may have raised willingness to accept lower hours for job security.
  - rising skill requirements have likely increased screening and training costs for firms.
  - educational attainment: share of workers with qualifications up to a Bachelor degree or higher rose from 18 percent in 1993 to 26 percent in 2009 (ABS Survey of Education and Training).
  - occupational demand shifts: increases in cognitive non-routine tasks (management and professional activities) and non-cognitive non-routine tasks (personal care); declines in non-cognitive routine tasks (machinery and plant operation) and cognitive routine tasks (clerical and secretarial) (Borland, 2011).
- Sectoral heterogeneity in employment adjustment:
  - VARs at 2-digit sectoral level (1990Q1–2016Q1) show the cumulative 8 quarter employment response to a 1 percent gross value added shock varies across sectors.
  - example: transportation and storage, and miscellaneous services, show larger and more sustained employment impacts compared with manufacturing or health services.
  - negative relationship observed between cumulative employment adjustment magnitude and the share of workers with at least a Bachelor degree; no similar association for average hours.
- Underemployment and wage implications:
  - since 2014 employment growth pickup has been driven by part-time workers.
  - underemployment rate increased to above 8½ percent, compared to a historical average of around 7 percent before the decline in mining investment.
  - elevated underemployment suggests more slack than the unemployment rate indicates.
  - wage Phillips curve estimates (building on Jacobs and Rush (2015)), regressing private wage growth on underemployment and unemployment gap measures, expected inflation (from inflation-indexed bonds), and the GDP deflator, find that:
    - models fit wage growth data closely and fit better when underemployment measures are included (higher adjusted R-squared).
    - coefficients on the underemployment gap and lagged change in the underemployment rate terms are negative and significant in relevant models.
    - broader measures of slack (including underemployment) are important for gauging slack and pressures on wage growth.

### IV. Have sectoral shifts in labor affected labor productivity?
- Sectoral reallocation effects:
  - aggregate labor productivity can be affected by shifts in the sectoral allocation of labor.
  - U.S. example cited: decline in share of manufacturing and mining hours and rise in services hours estimated to have reduced labor productivity growth by ¼ and ½ percentage point respectively (relative to unchanged shares), given higher productivity in mining and manufacturing versus services.
- Services productivity:
  - slower pace of labor productivity growth in services could exert a drag on aggregate labor productivity growth going forward (Van Zandweghe, 2016).

### Key findings and policy-relevant conclusions
- Labor markets have adjusted smoothly overall despite major shocks.
- Structural unemployment does not appear to have increased significantly despite elevated long-term unemployment.
- Increased flexibility in labor input (hours per worker) has moderated employment declines during downturns; adjustment to output shocks has shifted partly from employment to hours.
- Rising part-time employment and elevated underemployment (above 8½ percent) imply more slack than unemployment rates alone indicate and have likely contributed to weaker wage growth.
- Rebalancing toward services has been accompanied by a transitory decline in aggregate labor productivity growth, partly due to weaker productivity in services.
- Migration has played a key role in smoothing state-level labor market adjustments (noted in paper’s scope; methodological detail in Appendices).

*Source: wp17124 - 1. Key Real Sector and Labor Market Developments*

### 22.      In Australia, changes in the share of aggregate hours worked across sectors reflect

### 22. In Australia, changes in the share of aggregate hours worked across sectors reflect

### Sectoral shifts and employment contributions
- Manufacturing share has declined steadily over time.
- Mining share in aggregate hours:
  - Rose sharply over the commodity price boom.
  - Dropped sharply since the end of the boom.
- Construction share:
  - Rose steadily during the mining boom.
  - Stabilized at that higher level since.
- Services:
  - Healthcare services share increased following the global financial crisis, with further acceleration more recently.
  - Retail trade and communication shares declined post-GFC, likely due to increasing adoption of internet enabled retail services and expanded use of ICT technology.
- Job contributions since the terms-of-trade peak in late 2011:
  - The mining sector was a major contributor to job growth during the boom but its share has fallen since late 2011.
  - Healthcare services have contributed a third of the jobs added since the peak, higher than its share of one-quarter of total jobs created over the boom.
- Aggregate reallocation effects:
  - The share in hours worked of goods related sectors (mining, manufacturing, utilities, construction, and domestic trade) has declined by nearly 5 percentage points since the global financial crisis and following the mining investment downturn.
  - The share of business services (finance, real estate services, professional services, and administrative services) and particularly household services (food and accommodation, education, healthcare, recreational, and other personal services) has risen correspondingly.

### Labor productivity levels and growth by sector
- Level differences (real gross value added per hour):
  - Labor productivity is markedly higher in goods-related sectors and in business services compared to household services.
- Growth rates over 1986 – 2016:
  - Household services: about 0.7 percent.
  - Goods: 2 percent.
  - Business services: 1.9 percent.

### Aggregate labor productivity dynamics and transition effects
- Overall performance:
  - Labor productivity growth in Australia has sustained its historical average rate through the transition.
  - Post-GFC labor productivity growth has maintained its pre-GFC growth rate of around 1½ percent (in terms of GDP per hour worked).
  - Unlike many other advanced economies exhibiting large labor productivity level gaps relative to their pre-GFC trend, Australia does not appear to exhibit such a gap.
- Short-term developments:
  - Labor productivity growth did slow in 2015 due to shifts in sectoral allocation of labor — described as a transitional “between” effect — and recovered in 2016.
- Risks and offsets:
  - It remains to be seen whether lower labor productivity growth in services sectors will drag aggregate labor productivity growth down to a lower average rate once productivity gains from new mining capacities coming on-stream are fully realized.
  - Increasing competition in some services such as retail should provide continued support to productivity growth.

### State-level adjustment, migration, and labor market flexibility
- Mining boom and states:
  - Mining boom drove a strong pickup in private investment growth in Western Australia (WA) and Queensland (QLD) with strong labor demand (measured by vacancies).
  - New South Wales (NSW), Victoria (VIC), and South Australia (SA) saw investment growth and states’ shares in total investment fall, but experienced strong labor demand growth particularly in NSW, likely linked indirectly to the mining boom.
  - Working age population rose above long-term average growth rates, particularly in WA and QLD, accompanied by rising participation rates and declining unemployment.
  - With the terms-of-trade decline, labor demand in mining states fell sharply, accompanied by sharp declines in working age population growth reflecting reversal in migration inflows.
- Importance of migration:
  - On aggregate, about 50-60 percent of the increase in population is due to international migration.
  - Migration helped avoid labor shortages on the upswing and big increases in unemployment on the downswing in the mining states.
- VAR analysis (state-level VARs over 1979-2015 in log changes in employment, unemployment rate, and labor force participation rate):
  - Following a 1 percent state-specific shock to employment (IRFs):
    - QLD and VIC: employment and participation rates account for around half of the increase in employment over 10 years; thus about half the long-term increase in employment is supported by rising population (migration). In QLD, the employment rate has a smaller role in adjustment relative to participation compared to VIC. In all states, participation rates do more adjusting than employment rates.
    - NSW and WA: migration accounts for between 30-40 percent of the long run increase in employment over 10 years.
    - SA and TAS: migration accounts for only about 10-15 percent of the long run increase in employment.
  - Historical shock decomposition (employment rate and participation rate shocks versus actual employment):
    - In NSW, VIC, QLD, and WA, the sum of employment rate and participation rate shocks correspond reasonably well with actual changes in employment, with exceptions at certain points (e.g., QLD decline after 2006 exceeded what would be caused by these shocks alone; VIC and NSW increases in the mid-to-late 2000s larger than those two shocks alone; WA showed positive employment shocks during the mining boom not explained by employment rate and participation shocks alone).
  - Aggregate versus state-specific shocks:
    - WA shows a larger migration response when aggregate cycle is not removed from the data; VARs that do not remove the aggregate cycle show WA has the largest migration response among all states.

### Implications for adjustment and labor market outcomes
- Migration as flexibility:
  - Migration is a key aspect of labor market flexibility that has helped moderate the impact of recent shocks.
  - Over the boom, migration likely prevented labor shortages and additional wage cost pressures.
  - Over the subsequent decline in commodity prices and mining investment, the decline of migration into mining states likely helped prevent unemployment from rising higher and likely prevented wage growth from weakening further.
- Broader labor market adjustment patterns:
  - Employment impacts of cyclical downturns have moderated since the early 2000s.
  - More of the cyclical adjustment in total hours worked has occurred in average hours per worker, likely due to increased labor market flexibility following reforms in the early 1990s that enabled firms to adjust labor input without reducing employment.
  - The slowdown in growth since the global financial crisis and the commodity price and mining investment downturn produced significantly smaller reductions in employment than in previous cyclical downturns; the unemployment rate rose only slightly.
- Remaining concerns:
  - With persistent economic slack since the global financial crisis, long-term unemployment has risen, though likely not due to structural deterioration in labor markets.
  - Increasing share of part-time employment in total employment has raised underemployment, likely accounting for some ongoing weakness in wage growth.
  - The on-going rebalancing of the economy has thus far exerted only a transitory drag on labor productivity growth, but weaker productivity growth rates in some expanding services sectors, particularly in human services, may have longer lasting effects on aggregate productivity growth.

*Source: IMF working paper content unit wp17124 (chapter/section provided).*

### 0.75 percent on average...

### 0.75 percent on average...

### Macroeconomic and commodity context
- Terms of trade and mining investment have reversed sharply (figure caption).
- Commodity prices index referenced to "2013/14 = 100" shown in figures (no numerical series reproduced here).
- Pre-GFC average labor force growth used in steady-state Beveridge calculations set at 1.7 percent.

### Labor market outcomes and composition
- Unemployment increased modestly, but long term unemployment has risen (figure caption).
- Underemployment rates are above historical averages (figure caption).
- Recent employment growth has been driven mainly by part-time workers (figure caption).
- The share of part-time employment has increased (figure caption).
- Deviation of unemployment and underemployment from long-run averages shown in figures (no additional numeric series reproduced here).
- Average quarterly hours worked are estimated from monthly hours and employment status (note).

### Slack, wages, and Phillips curve
- Wages and earnings growth declined sharply across all sectors, especially commodity related sectors (figure caption).
- The Phillips curve appears to have shifted lower since the terms-of-trade decline (figure caption).
- Wage Phillips Curve samples shown as: 1998Q3-2011Q3 and 2011Q3-2016Q1 (figure captions).

### Structural unemployment and Beveridge curve
- NAIRU has increased somewhat since 2011 (figure caption).
- Beveridge curve indicates some outward shifting after the global financial crisis and terms-of-trade decline (figure caption).
- The Beveridge/steady-state framework follows:
  - Employment growth over a 1-year period: g_{t,t+1} = (H_{t} - S_{t}) / E_{t} (equation (1) in text).
  - Cobb-Douglas matching function form and separations equation estimated as:
    - ln(H/V) = h_μ + h_α ln(U/V) + ε_{h,t}  (equation (2))
    - ln(S/V) = s_μ + s_α ln(U/V) + ε_{s,t}  (equation (3))
  - Steady-state implicit condition combining (1)–(3) evaluated at ε_{·,t} = 0 and vacancy rate at observed rates to solve for equilibrium unemployment (equation (4) and text).
  - Parameter and steady-state evaluations use hires/separations inferred from job tenure data following Hobijn and Sahin (2012) methodology (text).
- Pre-downturn peaks dated 1989Q4, 2000Q3, 2008Q3, and 2011Q3 for labour-input HP-detrended series (note under Figure 3).

### Cyclical adjustment in labor input
- Peak-to-trough declines in aggregate hours worked are documented across downturns (figure captions).
- Employment and hours responses to a 1 percent GDP shock are compared for 1985-1997 and 1998-2016 samples (figures).
- In some sectors employment response is bigger and more persistent; in others lower and less persistent (figure captions).
- Employment adjusts relatively less in sectors with high educational qualifications (figure caption); no relationship observed for hours adjustment (figure caption).

### Sectoral patterns and shares
- Average hours have declined across most sectors (figure caption).
- The contribution of mining to employment has shrunk following the decline (figure caption).
- Figures show sectoral shares in aggregate hours for Goods related, Business services, and Household services, and output per hour indicators (figure captions).
- State-level patterns: private investment growth much stronger in mining states during the boom, increasing their share in aggregate private investment (figure captions).
- Labor demand and participation responses stronger in mining states; steeper declines in unemployment and stronger working age population growth observed in those states (figure captions).

### Migration and regional employment shocks
- Impulse responses to a 1% state-specific employment shock indicate:
  - Migration plays a bigger role in Queensland and Victoria than in New South Wales and Western Australia, with the least role in South Australia and Tasmania (Figure 11A caption).
  - Including the aggregate cycle, migration shocks in Western Australia are much larger, still sizeable in Queensland, and play a smaller role in New South Wales, Victoria, Tasmania, and the least in South Australia (Figure 11B caption).

### Historical decomposition of employment growth
- Figures present historical decompositions of employment growth by state (NSW, VIC, QLD, WA, SA, TAS) with series for actual employment and combined unemployment-rate and employment shocks (figure captions).

### Appendix I — Regression estimates (fitting a Beveridge curve for Australia)
- Hires and separations regressions inferred from ABS job-tenure data; time-aggregation formulas shown:
  - h = ln(E_{t+1} / E_{t+1}^{τ>1}) (text; continuous-time derivation summary).
  - s = ln(E_{t}^{τ>1} / E_{t+1}^{τ>1}) (text; continuous-time derivation summary).
  - Time-aggregated hires and separations expressed in equations (5) and (6).
- Table I.1. Regression Estimates (values reproduced exactly as in source):
  - Sample: 1986-2008
    - log H/V: Constant 2.1***; log U-V ratio .58***; R-sq 0.89
    - log S/V: Constant -1.5***; log U-V ratio 0.04; R-sq 0.09
  - Sample: 1986-2011
    - log H/V: Constant 2.0***; log U-V ratio .6***; R-sq 0.89
    - log S/V: Constant -1.6***; log U-V ratio .06*; R-sq 0.15
- Text summary: hiring function (2) shows a relatively good fit even with limited sample size; hires per vacancy (vacancy yield) is positively correlated with the U-V ratio. The separations equation (3) fit is much weaker and the coefficient on the U-V ratio is insignificant in the shorter sample. Results presented in the text are based on the shorter sample up to 2008 and are in line with Hobijn and Sahin (2012) for Australia; a longer sample up to the terms-of-trade decline after 2011 yields very similar parameter values.

*Source: wp17124 - 0.75 percent on average... (IMF staff figures, Haver Analytics database, and IMF staff calculations).*

### Appendix II. Cyclical Features of Labor Market Adjustment

### Appendix II. Cyclical Features of Labor Market Adjustment

### Aggregate GDP, hours worked, and employment
- VAR specification:
  - Three-variable VAR in log detrended GDP, employment, and average hours worked.
  - Two quarterly samples: from 1983 to 1997, and 1998 to 2016.
  - Variables detrended using the H-P filter with smoothing parameter lambda=1600.
  - Seasonal dummies included because hours worked data are in non-seasonally adjusted form; GDP and employment data are also included in non-seasonally adjusted form.
- Impulse-response presentation:
  - Left panels show response of log employment in the two sub-periods; right panels show response of average hours per worker.
  - Error bands show two standard errors above and below the estimated response.
- Key empirical findings:
  - “Employment responses were larger, and distinctly above zero in the earlier period, as compared to the later period.”
  - “The response of hours worked does not appear to differ very much across the two samples; noise in hours worked data may be a factor as noted in Jacobs and Rush (2015).”

### Sectoral value added, hours worked, and employment
- Extension of VAR framework to sector-level:
  - Excludes agriculture; excludes public administration and safety, education and training, and health and social assistance (public or public-adjacent sectors).
  - Data include detrended values of real gross value added, sectoral average hours worked, and employment (all in logs).
  - Sample: quarterly from 1990 – 2016.
- Education data:
  - Educational qualifications taken from ABS Survey of Education and Training.
  - Figure 4 (text) shows relationship between cumulative 8-quarter employment response and education attainments (share of sectoral labor with Bachelor degree or higher) for the year 2009.
  - Survey available every 5 years starting from 1993; comparability over time limited with expanded sectoral classification, with the 2009 survey having the most detailed sector classification.
  - Surveys for 2005 and 2001 have 2 sectors fewer than the 2009 survey.
  - “In general, the sectoral rankings in terms of the share of workers with education attainments of at least Bachelor degrees are preserved, and the results shown in the text would not be altered by choosing a different year of the survey.”

### Wage Phillips curve estimates
- Estimated equation (as presented):
  - 11213141
    516 27 38191
    log
    BondinfexpBondinfexpBondinfexp4 loglog
    private
    ttttt
    tttttt
    WPIURgapUERgapURUER
    GDPdefllabprod
    
       
    
       
    
      
- Variable definitions (as in source):
  - logprivate t WPI is private sector wage (log change).
  - 1t URgap is the lagged deviation of unemployment from its sample average (sample runs from 1998Q1 – 2016Q1).
  - 1t UERgap is the lagged underemployment rate gap similarly calculated.
  - 1 Bondinfexp t is the lagged expected inflation term implied by 10-year indexed bonds.
  - 14log t GDPdef   is the lagged year-on-year change in log of the GDP deflator (proxy for changes in output prices).
  - 1 log t labprod   is the lagged change in log output per worker.
- Parameter restrictions and model variants (Table A2 / text):
  - Model 1 assumes 24 0 (underemployment gaps assumed not to impact wage growth).
  - Model 2 removes this restriction; underemployment gap and change in underemployment rate may exert distinct impact on wage growth.
  - Model 3 imposes equality constraints on 12  and 34 .
  - Models 4 and 5 consider variants of Model 1 and 3 that include a lagged dependent variable term.
  - All models: output prices (GDP deflator) have a positive impact on wage growth and this effect “appears robustly estimated.”
- Table II.1. Wage Phillips Curve Estimates (selected entries as presented)
  - Dependent variable: quarterly log difference in private wage. Sample 1997:1 - 2016:1.
  - Model coefficients and statistics (as shown):
    - Dlog PWage: 0.3170.306
    - UGap: -0.010-0.031-0.015
    - UEGap: -0.064
    - UUGap: -0.025-0.024
    - DUrate: -0.234-0.188-0.198
    - DUErate: -0.099
    - DUUrate: -0.138-0.122
    - ExpInfB10: 0.0010.0010.0010.0010.001
    - ExpInfB10 (-2): -0.001-0.0010.000-0.0010.000
    - ExpInfB10 (-3): 0.0010.0010.0010.0010.001
    - D4log GDPdef: 0.0300.0230.0270.0210.017
    - Constant: 0.0030.0050.0040.0020.003
    - R-sq (adjusted): 0.600.640.650.650.72
  - Notes: PWage = private wage, U = unemployment, UE = underemployment, UU = underutilization, ExpinfB10 = inflation expectations inferred from 10 year inflation indexed bonds, GDPdef = GDP deflator. All RHS variables are included at first lag unless otherwise noted. Bold figures are significant at 10% or higher.
- Interpretations and findings:
  - Model 1 is closest in specification to RBA (2015) and the fit is very close to the RBA model in terms of the R-squared.
  - The UGap term is not significant in Model 1, but the change in unemployment rate has a negative and significant effect on wages.
  - In Model 2, the coefficient on the underemployment gap variable UEGap is negative and significant, as is the lagged change in the underemployment rate.
  - Introducing UEGap leads to the sign on UGap becoming larger (and significant at 15%) compared to Model 1.
  - Model 2 has a better fit in terms of adjusted R-squared and shows that the underemployment gap has a sizeable impact on wages.
  - In Model 3, the change in underutilization rate is significant and has a sizeable impact; R-squared is larger than in Model 1.
  - Including a lagged wage inflation term in Models 4 and 5 improves fit noticeably in Model 5, which includes the overall underutilization rate measure.
  - Overall conclusion: “underemployment gaps do matter for wage growth and may be having some impact in relatively weak wage growth outcomes observed since the terms-of-trade decline.”

### Appendix III. Role of Migration in States’ Labor Market Adjustment
- Methodological framework:
  - Follows Bayoumi and others (2006), building on Blanchard and Katz (1992).
  - Identity noted: (1) ** tttt EURLFPRWAP , where E is employment, (1)UR is the employment rate, LFPR is the participation rate, and WAP is the working age population.
  - Taking logs and rearranging gives eemp p wap where e is log employment, emp is log employment rate, and p is log participation rate.
  - From impulse-responses to employment shocks in a VAR involving employment, employment rate, and participation rate, one can infer the role played by wap in adjustment to employment shocks (capturing migration response of potential workers in working age groups).
- Stationarity considerations and specification:
  - In Blanchard-Katz, the VAR for the United States is implemented with employment rate and participation rate as level stationary.
  - Australian data cast doubt on stationarity in levels of unemployment rate and participation rate; these are entered in first differences.
  - Table A3 (unit root test results) shows that in individual states’ samples, unemployment rate and participation rate may be non-stationary.
  - Consequently, specification in this paper includes all variables in log differences:
    - ' (,, ) tt tt yeempp      and ' (,   , ) tetemptpt       .
  - Implication: unemployment rate and participation rate thus have a role in long term adjustment to employment shocks.
- VAR ordering and interpretation:
  - Shocks to employment are interpreted as labor demand shocks; supply responses occur through shocks to the unemployment rate and to participation.
  - Employment shocks are ordered first, and supply responses feed through to employment with a lag.
  - Two lags of each variable are included in the equations (as in Bayoumi and others (2006)).
- Data:
  - Employment, unemployment rate, and participation rate at the state level from ABS.
  - Sample runs from 1979 to 2015 at annual frequency.
  - First set of VARs: each state variable is “acyclic” to the aggregate economy (influence of aggregate shocks removed by regressing the state-level variable on the national variable and a constant; residuals used).
  - Second set of VARs: aggregate cycle is not removed from state-level data.
- Unit-root testing summary:
  - Table III.1. Individual Unit Root Test: ADF Test with AIC (Change in employment; Employment rate; Change in employment rateLFPR; Change in LFPR) reported for states (NSW, VIC, QLD, SA, WA, TAS) with rejection of null: unit root at 5%. Sample: 1979 - 2015 (annual).

*Appendix II. Cyclical Features of Labor Market Adjustment — wp17124 (IMF Working Paper).*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2017/wp17124.pdf_