## LABOR MARKET ADJUSTMENT TO SHOCKS IN AUSTRALIA

## Source details

**Canonical URL:** [LABOR MARKET ADJUSTMENT TO SHOCKS IN AUSTRALIA](https://www.imf.org/-/media/files/publications/cr/2017/cr1743.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/cr/2017/cr1743.pdf.md)
- [Structured JSON version](/-/media/files/publications/cr/2017/cr1743.pdf.json)

---

### Introduction and context
- Two major shocks: the global financial crisis (GFC) and a commodity price shock.
- Commodity prices doubled between 2005Q1 and 2011Q3, driving the terms of trade to an all-time high; since 2011Q3 commodity prices declined by 40 percent, leading to a 36 percent decline in the terms of trade from the 2011Q3 peak.
- Mining investment: rose from an average of around 2 percent of GDP before 2005 to a peak exceeding 7 percent of GDP in 2013, and fell to 5½ percent of GDP by end-2015 (and declined further since).
- Real GDP growth: fell from average rates of around 3¼ percent pre-GFC to around 2½ percent post-GFC.
- Key aggregate labor market indicators:
  - Unemployment trough of 4 percent prior to the GFC.
  - Unemployment averaged 5½ percent between 2009Q1 and 2016Q2.
  - Unemployment peaked at around 6¼ percent in mid-2015 and declined to below 5¾ percent as of August 2016.
  - Multivariate-filtering estimate of the NAIRU points to an unemployment gap of around 0.6 percent.
  - Estimates of potential output indicate a negative output gap in the range of 1.25 to 1.75 percent of potential GDP.
  - Okun coefficients referenced: -0.4 and between -0.27 and -0.35.

### Main findings and conclusions
- No significant increase in structural employment in the wake of the commodity prices bust and mining investment decline.
- Increased flexibility in average hours per worker likely moderated employment reductions; elevated underemployment signals additional slack and likely weighs on wage growth.
- Reallocation of labor to expanding sectors proceeded smoothly; labor productivity growth has slowed partly due to the expansion of services.
- States’ labor markets transitioned smoothly, with migration responses aiding adjustment.

### Structural unemployment and long-term unemployment dynamics
- Long-term unemployment rose after the GFC and again following the mining investment decline:
  - Share of long-term unemployed in total unemployed: around 15 percent in mid-2009 → around 20 percent by 2011 → around 25 percent since the mining investment decline.
  - Average duration of unemployment: from a pre-GFC trough of 27 weeks → nearly 46 weeks in mid-2016.
- Evidence and interpretation:
  - Multivariate filter estimate of the NAIRU indicates a small increase beginning in 2011.
  - Fitted Beveridge curve shows a small outward shift following the terms-of-trade bust and mining investment downturn: in 2016 unemployment was around ¾ of a percent higher than rate consistent with the fitted Beveridge curve.
- Caveats:
  - The moderate outward shift is much smaller than in previous downturns and should recede as slack is eliminated; divergence is relatively small and within historical ranges.

### Cyclical adjustment dynamics: employment, hours, and underemployment
- VAR estimation (aggregate): log detrended GDP, average hours, employment; quarterly samples 1984Q4–1997Q4 and 1998Q1–2016Q1.
  - Impulse responses: the response of employment to a 1 percent GDP shock has moderated in the second sample period.
- Changing mechanics since the late 1990s:
  - Peak-to-trough declines in quarterly detrended total hours increasingly borne by reductions in average hours as well as employment.
  - Early 1990s: most cyclical decline in total hours came through reductions in employment.
  - From early 2000s onward: downturn adjustments almost equally attributed to reductions in average hours and in employment.
  - The amplitude of peak-to-trough declines in total hours moderated.
- Implications:
  - Greater flexibility in average hours likely moderated increases in unemployment and prevented larger employment losses after the mining investment downturn.
  - Elevated underemployment implies additional slack not captured by the unemployment gap alone and likely contributes to weaker wage growth.
  - Falling average hours per worker and rising part-time employment have likely contributed to elevated underemployment.
- Underemployment statistics:
  - Underemployment rate rose to around 8¾ percent in 2016Q3, above the historical average of around 7 percent.

### Wage dynamics and Phillips-curve evidence
- Aggregate private sector wage growth weaker than implied by historical relationship with the unemployment gap; the wage Phillips curve may have shifted downward since the terms-of-trade bust.
- Phillips-curve regressions (dependent variable: log private WPI Δ) include lagged unemployment gap, lagged underemployment gap, expected inflation (from 10-year indexed bonds), GDP deflator change, and labor productivity change.
  - Inclusion of underemployment measures improves fit (adjusted R-sq increases).
  - Coefficients on UEGap and lagged change in underemployment rate are negative and significant in models where included.
- Selected coefficient table entries (sample 1997:1 - 2016:1; reported values preserved):
  - Dlog PWage 0.317 0.306
  - 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 0.001 0.001 0.001 0.001 0.001
  - D4log GDPdef 0.030 0.023 0.027 0.021 0.017
  - Constant 0.003 0.005 0.004 0.002 0.003
  - R-sq (adjusted) 0.60 0.64 0.65 0.65 0.72
- Conclusion: underemployment gaps matter for wage growth and help explain weak wage growth since the terms-of-trade bust.

### Sectoral reallocation and implications for productivity
- Sectoral VARs (quarterly 1990–2016) estimate cumulative 8-quarter employment responses to a 1 percent gross value added shock by 2-digit sector.
  - Size of employment adjustment varies across sectors.
  - Transportation and storage, and miscellaneous services: peak employment impact from a 1 percent shock larger and more sustained than in manufacturing or health services.
  - Negative relationship between cumulative 8-quarter employment adjustment and sectoral share of workers with at least a Bachelor degree (measured in 2009).
  - No observed association between adjustment in average hours and education levels.
- Sectoral hours and employment shares since the GFC/terms-of-trade bust:
  - Share in hours of goods-related sectors (mining, manufacturing, utilities, construction, domestic trade) declined by nearly 5 percentage points since the GFC and mining investment downturn.
  - Share of business services and household services has risen correspondingly.
- Labor productivity:
  - Average labor productivity growth rates (1986 – 2016): household services about 0.7 percent, goods about 2 percent, business services about 1.9 percent.
  - Aggregate labor productivity growth slowed partly due to a "between" effect from allocation shifts toward lower-productivity household services and also due to slowed productivity growth within household and business services.
  - Post-GFC aggregate labor productivity growth has sustained its pre-GFC growth rate of around 1½ percent (GDP per hour worked).

### State-level adjustment and migration
- Mining boom: strong labor supply responses at state level, notably in Western Australia (WA) and Queensland (QLD); participation rates rose and unemployment declined during the upswing; working age population growth rose above long-term averages in WA and QLD.
- Following the terms-of-trade bust: labor demand in mining states fell sharply, accompanied by sharp declines in working age population growth via migration flows.
- Migration's role:
  - Aggregate: about 50-60 percent of population increase is due to international migration over the upswing.
  - Migration helped avoid labor shortages during the upswing and limited unemployment increases during the downswing.
- State-level VARs (1979–2015, with aggregate cycle removed) IRF summaries to a 1 percent employment shock (adjustment over 10 years):
  - QLD and VIC: employment and participation account for around half of the 10-year increase in employment; about half the long-term increase supported by rising population (migration).
  - NSW and WA: migration accounts for between 30-40 percent of the long-term employment increase (WA migration response larger to aggregate shocks when aggregate cycle not removed).
  - SA and TAS: migration accounts for about 10-15 percent of the long-term employment increase.
- Historical shock decompositions show migration historically key in adjusting to employment shocks, especially during the mining boom-bust cycle.
- Overall: migration is a key aspect of labor market flexibility that moderated recent shocks—preventing labor shortages and extra wage pressures during the boom and helping prevent larger rises in unemployment during the bust.

### Empirical methods and key quantitative inputs
- Aggregate VAR samples: 1983–1997 and 1998–2016 (quarterly); H-P filter smoothing parameter lambda=1600 used for detrending.
- Sectoral VAR sample: 1990–2016 (quarterly).
- Wage Phillips curve sample: 1997:1 – 2016:1 (quarterly).
- Beveridge-curve and turnover-steady-state framework:
  - Employment growth over a 1-year period expressed as (H − S)/E where H = hires and S = separations.
  - Equilibrium unemployment solved setting labor force growth g at pre-GFC average rate of "1.7 percent".
- Hiring and separations regressions (samples and reported statistics preserved exactly):
  - 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
  - Interpretation: hiring function fits well; separations equation fit weaker.

### Policy-relevant implications and summary recommendations
- Labor market flexibility (greater adjustment through average hours and part-time employment) has cushioned employment headcount during downturns; policy monitoring should account for broader slack measures (underemployment and underutilization) beyond the unemployment rate.
- Elevated underemployment is likely weighing on wage growth; policymakers should consider underemployment in assessing labor-market slack and inflationary pressures.
- Migration is an important adjustment channel at the state level; migration-sensitive policies can aid regional labor market smoothing.
- Sectoral reallocation toward lower-productivity household services is contributing to slower aggregate labor productivity growth; policies to boost productivity within expanding services and facilitate skill transitions could help raise aggregate productivity growth.

*Source: IMF staff report, chapter on labor market adjustment and productivity (excerpt).*

### References ____________________________________________________________________________ 16

### LABOR MARKET ADJUSTMENT TO SHOCKS IN AUSTRALIA

### Introduction
- Australia faced two major shocks: the global financial crisis (GFC) and a commodity price shock.
- Commodity prices doubled between 2005Q1 and 2011Q3, driving the terms of trade to an all-time high.
- Since 2011Q3, commodity prices declined by 40 percent, leading to a 36 percent decline in the terms of trade from the 2011Q3 peak.
- Mining investment rose from an average of around 2 percent of GDP before 2005 to a peak exceeding 7 percent of GDP in 2013, and fell to 5½ percent of GDP by end-2015 (and declined further since).
- Real GDP growth fell from average rates of around 3¼ percent pre-GFC to around 2½ percent post-GFC.

Key aggregate labor market indicators:
- Unemployment trough of 4 percent prior to the GFC.
- Unemployment averaged 5½ percent between 2009Q1 and 2016Q2.
- Unemployment peaked at around 6¼ percent in mid-2015 after the terms-of-trade bust and mining investment decline.
- Unemployment declined to below 5¾ percent as of August 2016.
- A multivariate-filtering estimate of the NAIRU points to an unemployment gap of around 0.6 percent.
- Estimates of potential output indicate a negative output gap in the range of 1.25 to 1.75 percent of potential GDP.
- Okun coefficients referenced: -0.4 (Ball and others (2013) post-1995) and between -0.27 and -0.35 (Tulip and Lancaster (2015)).

Summary of motivations and questions addressed:
- Has the mining investment downturn increased structural unemployment?
- How have cyclical labor market adjustment dynamics changed, and what are implications for labor market slack and wage growth?
- What does sectoral labor reallocation imply for aggregate labor productivity growth?
- How did States’ labor markets adjust to the commodity boom-bust cycle?

Main summary findings (overview):
- No significant increase in structural employment following the commodity price bust and mining investment decline.
- Increased flexibility in average hours per worker likely moderated employment reductions; elevated underemployment signals additional slack and likely weighs on wage growth.
- Reallocation of labor to expanding sectors proceeded smoothly; labor productivity growth has slowed partly due to the expansion of services.
- States’ labor markets transitioned smoothly, with migration responses aiding adjustment.

### Has Structural Unemployment Increased?
- Long-term unemployment rose after the GFC and again following the mining investment decline.
  - 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.
  - Average duration of unemployment increased from a pre-GFC trough of 27 weeks to nearly 46 weeks in mid-2016.
- Interpretation:
  - If long-term unemployment reflects skill mismatches, this provides evidence of increased structural unemployment.
  - A multivariate filter estimate of the NAIRU indicates a small increase beginning in 2011.
  - The fitted Beveridge curve for Australia shows a small outward shift following the terms-of-trade bust and mining investment downturn.
    - Observations following the terms-of-trade bust and mining investment decline appear to the right of the fitted Beveridge curve.
    - In 2016, unemployment was around ¾ of a percent higher than the rate consistent with the labor market equilibrium implied by the fitted Beveridge curve.
- Caveats:
  - The moderate outward shift is much smaller than in previous downturns, and should recede as slack is eliminated.
  - The divergence from the Beveridge curve in the post-GFC period is relatively small and within historical ranges; other advanced economies experienced much larger outward shifts.

### How Have Cyclical Adjustment Dynamics Changed? Implications for Slack and Wages
- Since the late 1990s, average hours worked per worker have become more flexible and have contributed significantly to cyclical labor adjustment.
  - Peak-to-trough declines in quarterly detrended total hours worked have been increasingly borne by reductions in average hours as well as employment.
  - In the early 1990s most cyclical decline in total hours worked came through reductions in employment.
  - Starting from the early 2000s and through to the present, downturn adjustments have been almost equally attributed to reductions in average hours and in employment.
  - The amplitude of peak-to-trough declines in total hours worked has moderated.
- Implications:
  - Greater flexibility in average hours likely moderated increases in unemployment during downturns and prevented larger employment losses in the wake of the mining investment downturn.
  - Elevated underemployment implies additional slack not captured by the unemployment gap alone and likely contributes to weaker wage growth.
  - Falling average hours per worker and rising part-time employment have likely contributed to elevated underemployment.

### Sectoral Reallocation and Productivity
- Reallocation of labor to expanding sectors proceeded smoothly following the mining investment decline.
- Aggregate labor productivity growth has slowed, partly as a result of expansion in services.

### State-Level Adjustment and the Role of Migration
- States’ labor markets transitioned smoothly through the commodity boom-bust cycle.
- Migration responses to the commodity boom-bust cycle played a key role in smoothing state-level labor market adjustment.

### Core Findings and Conclusions (bulleted)
- There does not appear to be a significant increase in structural employment in the wake of the commodity prices bust and mining investment decline.
- Increased flexibility in average hours per worker has likely moderated employment reduction in downturns and prevented a larger increase in unemployment.
- Elevated underemployment signals additional slack and is likely weighing down wage growth.
- Reallocation of labor to expanding sectors has proceeded smoothly; labor productivity growth has slowed partly due to expansion in services.
- Migration has been an important factor aiding adjustment at the state level.

*Source: IMF staff report, January 23, 2017.*

### 12.      The employment response to output shocks seems to have moderated. To illustrate the

### The employment response to output shocks seems to have moderated

### Employment response to aggregate output shocks
- VAR estimation in log detrended values of GDP, average hours, and employment using quarterly data over two samples: 1984Q4–1997Q4, and 1998Q1–2016Q1.  
- VAR ordering follows Bishop and Plumb (2016): output responds only with a lag to labor input shocks; firms adjust hours worked before employment.  
- Impulse response functions (IRFs) indicate the response of employment to a 1 percent GDP shock has moderated in the second sample period.

### Labor market features that may explain moderation in employment response
- Labor market reforms in the early 1990s increased ability for firms to bargain directly with employees, which may have allowed greater scope for firms to adjust hours worked in response to cyclical conditions.
- Tight labor markets just before the global financial crisis may have increased costs related to firms’ hiring and firing decisions.
- During the crisis, increased worker uncertainty may have raised willingness to accept lower hours for employment security; evidence also points to destruction of jobs with longer hours and replacement with jobs with shorter hours.
- Rising skill requirements and higher costs of screening and training likely made hiring and firing costlier.
- Educational attainment change: share of workers with educational 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 data).
- Occupational demand shifts (Borland, 2011): largest increases in demand in cognitive, non-routine tasks (management and professional activities) and non-cognitive non-routine tasks (personal care); largest declines in non-cognitive routine tasks (machinery and plant operation) and cognitive routine tasks (clerical and secretarial functions).

### Sectoral evidence on skills and employment adjustment
- Sectoral VARs (log detrended sectoral gross value added, average hours, employment; quarterly, 1990Q1–2016Q1) estimate cumulative 8-quarter employment responses to a 1 percent gross value added shock by 2-digit sector.
- Findings:
  - Size of employment adjustment to output shocks varies across sectors.
  - Transportation and storage, and miscellaneous services: peak employment impact from a 1 percent shock is larger and more sustained than in manufacturing or health services.
  - There is a fairly strong negative relationship between the cumulative 8-quarter employment adjustment and the share of workers with at least a Bachelor degree (measured in 2009) in the sector.
  - No observed association between adjustment in average hours and education levels.

### Underemployment, hours composition, and wage dynamics
- Employment growth since 2014 has been driven largely by part-time workers; average hours recovered to trend after the terms-of-trade bust, but full time hours show a steadily declining trend.
- Underemployment rate rose to around 8¾ percent in 2016Q3, above the historical average of around 7 percent, indicating potential slack beyond the unemployment rate.
- Elevated underemployment has likely contributed to weaker wage growth since the terms-of-trade bust.
- Aggregate private sector wage growth is weaker than implied by the historical relationship with the unemployment gap; the wage Phillips curve may have shifted downward since the terms-of-trade bust.
- Phillips curve estimation (building on Jacobs and Rush (2015)):
  - Regressions of private wage growth on measures of labor market slack (including underemployment and unemployment gap), expected inflation (inferred from inflation indexed bonds), and the GDP deflator.
  - Models fit wage growth data fairly closely; adjusted R-squared improves when underemployment measures are included along with the unemployment gap.
  - Coefficients on the underemployment gap and lagged change in the underemployment rate are negative and significant; where included, the coefficient on the lagged change in the underutilization rate is also significant (Model 2, 3 and 5 in Table A2, Appendix II).
  - Conclusion: underemployment gaps may have a negative impact on wage growth; higher underemployment since the terms-of-trade bust may partly explain wage growth weakness.
- Additional factors for unexpected wage weakness:
  - Prior periods (2000s) had wage growth above predicted levels; recent weakness may partly reflect correction of previously high wage growth.
  - Increased labor market flexibility in sectors exposed to commodity prices (mining, business services) may have added downward wage pressure.
  - Low public sector wage outcomes (benchmarks for some private wages in health and education) may have contributed to private wage weakness, though more recently public sector wage growth has exceeded private sector wage growth despite wage caps in some state governments.

### Interpretation: broader measures of slack and labor flexibility
- Increased flexibility in average hours per worker may have helped moderate the rise in unemployment.
- Growing share of part-time work and declining full-time hours suggest more labor market slack than unemployment gaps alone indicate and likely contributed to weaker wage growth.

### Sectoral shifts and labor productivity
- Sectoral reallocation can affect aggregate labor productivity via "between" effects.
  - Example cited: in the United States, decline in manufacturing and mining shares and rise in services estimated to have reduced labor productivity growth by ¼ and ½ percentage point respectively relative to a no-change counterfactual (Van Zandweghe, 2016).
- In Australia:
  - Long-term trends and post-GFC and post-terms-of-trade bust changes observed in sectoral shares of aggregate hours.
  - Manufacturing: steady decline in share of hours worked.
  - Mining: small share of aggregate hours that rose sharply over the boom and dropped sharply since the end of the boom.
  - Construction: share rose during the mining boom and stabilized at higher level.
  - Healthcare services: noticeable increase in share following the global financial crisis with further acceleration recently.
  - Retail trade and communication shares declined post-GFC, likely due to internet-enabled retail and expanded ICT use.
- Since 2011Q3, average hours worked declined across most sectors, about equally in healthcare and mining.
- Employment changes drove share changes: mining was a major contributor to job growth during the boom but employment has fallen since the terms-of-trade peak; healthcare services contributed a third of jobs added since the terms-of-trade bust, higher than a quarter of jobs added over the boom.
- Aggregate patterns since the GFC and mining downturn:
  - Share in hours of goods-related sectors (mining, manufacturing, utilities, construction, domestic trade) declined by nearly 5 percentage points since the global financial crisis and mining investment downturn.
  - Share of business services (finance, real estate services, professional services, administrative services) and household services (food and accommodation, education, healthcare, recreational, other personal services) has risen correspondingly.
  - Labor productivity (real gross value added per hour) is markedly higher in goods-related sectors and business services than in household services.
  - Average labor productivity growth rates: household services about 0.7 percent, goods about 2 percent, business services about 1.9 percent (over 1986 – 2016).
- Labor productivity growth decomposition:
  - Recent decline in labor productivity growth is partly a "between" effect from allocation shifts and also reflects slowed productivity growth within household and business services.
  - Post-GFC aggregate labor productivity growth has sustained its pre-GFC growth rate of around 1½ percent (GDP per hour worked).
  - Unlike many other advanced economies, Australia does not exhibit a clear large labor productivity level gap relative to its pre-GFC trend.

### State-level labor market adjustment and the role of migration
- Mining boom generated strong labor supply responses at state level, notably in Western Australia (WA) and Queensland (QLD), with strong private investment and labor demand (vacancies).
- New South Wales (NSW), Victoria (VIC), and South Australia (SA) saw falls in investment shares but strong labor demand growth, especially in NSW, linked indirectly to the mining boom.
- Supply response: participation rates rose and unemployment rates declined; working age population growth rose above long-term averages, particularly in WA and QLD.
- Following the terms-of-trade bust, labor demand in mining states fell sharply, accompanied by sharp declines in working age population growth via migration flows.
- Migration's role:
  - Aggregate: about 50-60 percent of population increase is due to international migration.
  - Migration helped avoid labor shortages during the upswing and limited unemployment increases during the downswing.
  - Over 2005-11, migration contributed a larger share of population growth in all states than in the previous two decades.
- State-level VARs (log changes in employment, unemployment rate, labor force participation rate) for NSW, VIC, QLD, SA, WA, TAS over 1979–2015 (aggregate cycle removed to focus on state-specific shocks) trace unemployment, participation, and population growth responses to state-specific employment shocks.
  - IRF remarks from a 1 percent employment shock (adjustment over 10 years to long-term growth rate):
    - QLD and VIC: employment and participation rates account for around half of the 10-year increase in employment; about half the long-term increase supported by rising population (migration). In QLD, participation adjusts more than employment rate relative to VIC.
    - NSW and WA: migration accounts for between 30-40 percent of the long-term employment increase.
    - WA migration response may be larger to aggregate shocks (commodity price boom) than to state-specific shocks; VARs that do not remove the aggregate cycle show WA has the largest migration response among states.
    - SA and TAS: migration accounts for about 10-15 percent of long-term employment increase.
  - Historical shock decomposition comparing actual employment growth to shocks from employment rate and participation rate:
    - In NSW, VIC, QLD, WA the sum of employment rate and participation rate shocks generally corresponds with actual employment changes, with exceptions (e.g., QLD decline in employment after 2006 exceeded what these shocks alone would cause; VIC and NSW employment increases in mid-late 2000s larger than these shocks alone).
    - Results underscore migration's historical role in adjusting to employment shocks, especially during the mining boom-bust cycle.
- Overall: migration is a key aspect of labor market flexibility that moderated recent shocks—preventing labor shortages and extra wage pressures during the boom and helping prevent larger rises in unemployment and further wage weakening during the bust.

*Source: IMF staff report, chapter on labor market adjustment and productivity (excerpt).*

### 28.      Overall, Australia’s labor markets have coped well with the adverse shocks as of late,

### cr1743 - 28.      Overall, Australia’s labor markets have coped well with the adverse shocks as of late,

### Overview and main conclusions
- Overall assessment: Australia’s labor markets "have coped well with the adverse shocks as of late, adjusting smoothly in response to the terms-of-trade bust and mining investment decline, but not without some lingering concerns."
- Unemployment: "Unemployment rose modestly following the global financial crisis and during the bust phase of the commodity prices and mining investment cycle."
- Structural unemployment: "There may have been a slight increase in structural unemployment."
- Cyclical adjustment: "The employment impact of cyclical downturns has moderated since the early 2000s and more of the cyclical adjustment in total hours worked has occurred in average hours per worker."

### Labor market adjustment, composition, and dynamics
- Adjustment mechanics:
  - Since the early 1990s (and reinforced by events including the global financial crisis), increased labor market flexibility has shifted cyclical adjustment toward average hours per worker rather than employment headcount.
  - "The slowdown in growth since the crisis and commodity price bust and mining investment decline produced significantly smaller reductions in employment than in prior downturns, and unemployment deteriorated only modestly."
- Sectoral patterns:
  - Employment growth during the recovery has been driven disproportionately by part-time work.
  - Sectoral responses differ: some sectors show bigger and more persistent employment responses to shocks, while others show lower and less persistent responses. Sectors with higher shares of workers with Bachelors degrees or higher tend to exhibit relatively less employment adjustment.

### Underemployment, hours, and wage dynamics
- Underemployment:
  - "Underemployment has played a key role in labor market slack and recent wage growth weakness."
  - Trend decline in average hours worked per worker and growth in part-time employment imply elevated underemployment and more slack than indicated by the unemployment gap alone.
- Wage developments:
  - Weak wage growth in recent periods is plausibly linked to elevated underemployment.
  - Outlook: "Wages are expected to strengthen gradually as non-mining investment picks up and the drag from declining mining investment comes to an end."
- Phillips curve behavior:
  - Evidence suggests "the Phillips curve appears to have shifted lower since the terms-of-trade bust."

### Migration and state-level adjustment
- Role of migration:
  - "Migration has played a key role in labor market adjustment at the state level," especially in response to state-specific demand shocks.
  - Migration helped avoid labor shortages during the boom and mitigated unemployment increases during the bust, contributing to the overall moderate national labor market impact over the boom-bust cycle.
- State heterogeneity:
  - Mining states experienced much stronger private investment growth during the boom, increased shares in aggregate private investment, sharper labor demand increases, stronger participation responses, steeper declines in unemployment (during the boom), and stronger working-age population growth.
  - Impulse-response analysis indicates migration plays a larger role in Queensland and Western Australia and a smaller role in New South Wales, South Australia, and Tasmania (with magnitudes varying when the aggregate cycle is included).

### Empirical methods, key parameters, and quantitative results
- Beveridge-curve and turnover-steady-state framework:
  - Employment growth over a 1-year period is expressed as (H − S)/E where H = hires and S = separations (equation (1)).
  - A Cobb-Douglas matching function and a separations equation were estimated (equations (2) and (3)) and combined with the steady-state condition to solve for equilibrium unemployment (equation (4)).
  - The equilibrium exercise evaluates the implicit function at errors set to zero and sets g (labor force growth) at its pre-GFC average rate of "1.7 percent" to solve for the equilibrium unemployment rate.
- Estimation of hires and separations:
  - Hires and separations are inferred from job-tenure data using the continuous-time proportional-rate assumption; closed-form relations for hiring and separation rates are derived from employment, employment with tenure > 1 year, and continuous-time aggregation (equations (5) and (6)).
- 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
  - Interpretation:
    - Hiring function shows a relatively good fit even with limited sample size; hires per vacancy (vacancy yield) is positively correlated with the U-V ratio.
    - Separations equation fit is weaker; the coefficient on the U-V ratio is insignificant in the shorter sample but becomes marginally significant in the longer sample.

### Figures and quantitative indicators highlighted
- Pre-GFC average labor force growth used in steady-state computations: "1.7 percent".
- NAIRU and unemployment:
  - NAIRU "has increased somewhat since 2011" (figure annotation).
- Labor-input adjustment:
  - Peak-to-trough declines and responses to 1 percent GDP shocks show smaller employment and larger hours-per-worker adjustments in the 1998-2016 period versus 1985-1997 (figure annotations and panels).
- Underemployment and hours:
  - Underemployment rates are "above historical averages" (figure annotation).
  - Average hours per quarter series for full-time and part-time workers show declines during the bust period (figure annotations).
- Wage growth:
  - Wages and earnings growth declined sharply across all sectors, "especially commodity related sectors" (figure annotation).

*Source: IMF staff chapter text and figures (cr1743).*

### 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.
  - Detrending: H-P filter, with the smoothing parameter lambda=1600.
  - Seasonal dummies included because hours worked data are non-seasonally adjusted; GDP and employment included non-seasonally adjusted as well.
  - Charts: 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 estimate response.
- Empirical findings:
  - Employment responses to a 1 percent GDP shock were larger and distinctly above zero in the earlier period (1983–1997) compared to the later period (1998–2016).
  - The response of hours worked does not appear to differ very much across the two samples, though noise in hours worked data may be a factor (Jacobs and Rush (2015) noted).

### Sectoral value added, hours worked, and employment
- VAR extension to sector-level:
  - Uses detrended values of real gross value added, sectoral average hours worked, and employment (all in logs).
  - Quarterly sample: 1990 – 2016.
  - Educational qualifications data from ABS Survey of Education and Training.
- Education and employment response:
  - Figure 4 (text reference) shows the relationship between the cumulative 8 quarter employment response and education attainments (share of sectoral labor with Bachelor degree or higher) for the year 2009.
  - Survey availability and comparability:
    - Survey available every 5 years starting from 1993.
    - 2009 survey has the most detailed sector classification; surveys for 2005 and 2001 have 2 sectors fewer than the 2009 survey.
    - In general, sectoral rankings by share of workers with at least Bachelor degrees are preserved; results would not be altered by choosing a different year of the survey.

### Wage Phillips curve estimates
- Estimated equation (as given):
  - Dependent variable: log private sector wage change (log private WPI Δ).
  - RHS variables include: lagged unemployment gap (1t URgap−), lagged underemployment rate gap (1t UERgap−), lagged expected inflation implied by 10-year indexed bonds (1 Bondinfexp t−), lagged year-on-year change in log GDP deflator (14log t GDPdefl − Δ) to proxy output prices, and lagged change in log output per worker (1 log t labprod − Δ). All RHS variables are included at first lag unless otherwise noted.
  - Parameter restrictions are imposed across several model variants (Table A2 reference):
    - Model 1 assumes β2 = β4 = 0 (underemployment gaps assumed not to impact wage growth).
    - Model 2 removes this restriction, allowing the underemployment gap and the change in the underemployment rate to exert distinct impacts.
    - Model 3 imposes equality constraints β1 = β2 and β3 = β4.
    - Models 4 and 5 consider variants of Model 1 and 3 including a lagged dependent variable term.
- Key empirical results and interpretation:
  - Output prices (GDP deflator) have a positive impact on wage growth in all models and appear robustly estimated.
  - Model 1 is closest to the RBA (2015) specification and has a fit very close to the RBA model in terms of the R-squared.
  - UGap term is not significant in Model 1, but the change in unemployment rate (DUrate) 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 (DUErate). Introducing UEGap leads to a larger (and 15% significant) negative UGap coefficient compared to Model 1.
  - Model 2 shows better fit (adjusted R-squared) than Model 1, indicating underemployment gap has a sizeable impact on wages.
  - Model 3: change in underutilization rate is significant and sizeable.
  - Including a lagged wage inflation term in Models 4 and 5 improves fit noticeably, especially in Model 5 which includes the overall underutilization rate measure.
  - Overall conclusion: underemployment gaps do matter for wage growth and may be contributing to relatively weak wage growth outcomes observed since the terms-of-trade bust.
- Table II.1. Wage Phillips Curve Estimates (selected entries exactly as presented)
  - Dependent variable: Dlog PWage. Sample 1997:1 - 2016:1.
  - Reported coefficients and statistics (values preserved exactly as in source table):
    - Dlog PWage 0.317 0.306
    - 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 0.001 0.001 0.001 0.001 0.001
    - ExpInfB10 (-2) -0.001 -0.001 0.000 -0.001 0.000
    - ExpInfB10 (-3) 0.001 0.001 0.001 0.001 0.001
    - D4log GDPdef 0.030 0.023 0.027 0.021 0.017
    - Constant 0.003 0.005 0.004 0.002 0.003
    - R-sq (adjusted) 0.60 0.64 0.65 0.65 0.72
  - Notes (verbatim): 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.

### Link to broader narrative on labor market adjustment
- Interpretation:
  - VAR results and sectoral analysis indicate a decline in the sensitivity of employment to GDP shocks in the post-1998 period relative to the pre-1998 period, while hours responses are relatively stable across periods.
  - Wage Phillips curve evidence suggests underemployment and underutilization measures provide additional explanatory power for weak wage growth since the terms-of-trade bust.
- Data and samples used (explicit):
  - Aggregate VAR samples: 1983–1997 and 1998–2016 (quarterly).
  - Sectoral VAR sample: 1990–2016 (quarterly).
  - Wage Phillips curve sample: 1997:1 – 2016:1 (quarterly).
  - H-P filter smoothing parameter: lambda=1600.
  - Education data: ABS Survey of Education and Training (survey years noted: 1993 start; 2001, 2005, 2009 availability; 2009 most detailed).

*Source: IMF staff compilation from "Appendix II. Cyclical Features of Labor Market Adjustment" in the provided content unit.*

### 12.      The RBA adopted a wait-and-see attitude and only lowered rates in early 2015. In its

### cr1743 - 12.      The RBA adopted a wait-and-see attitude and only lowered rates in early 2015. In its

### RBA policy actions and recent inflation dynamics
- The RBA indicated throughout 2014 that it projected inflation to remain in the target range and expected further lift in growth from depreciation and support to domestic demand from low interest rates.
- In February 2015 the RBA lowered the policy rate, noting that growth had remained below trend for longer than expected and that the expected pick up had not materialized, and followed up with another cut in May.
- Growth picked up in the second half of 2015, while inflation weakened further.
- After a sizeable inflation surprise on the downside in 2016Q1, the RBA lowered the policy rate in May 2016, shortly after the data had been released and then again in August 2016.
- Headline inflation has been below the lower bound of the inflation target range before, but the current episode:
  - Has been longer lasting than all but one previous episode—the deviation lasted longer during the Asian Financial Crisis.
  - Is the first time that both non-tradables and underlying inflation are below target.
- These features highlight new challenges for monetary policy in Australia.

### Current and forward-looking monetary policy challenges
- Three sets of challenges:
  - A different, more difficult economic environment: inflation undershooting the target, declining inflation expectations, persistent labor market slack, strong external disinflationary forces, and a less supportive international growth environment than during the 1990s and 2000s.
  - The lower Equilibrium Interest Rate (EIR).
  - The monetary policy time horizon and policy communication.
- The recommended strategic posture is a “prudent risk-management strategy” aiming to avoid “dark corners” where the economy could get stuck in a low inflation and low growth trap.
- A “low for longer” strategy is argued as best suited to minimize risks of entrenched low inflation and limited conventional policy room due to proximity to the ELB.
- If low inflation becomes entrenched, correcting it could be very difficult and costly, as illustrated by experiences in Japan and the euro area.

### Policy instruments and communication
- Strategy elements:
  - Use conventional policy instruments more effectively.
  - Take unconventional measures if the situation dictates.
  - Strengthen central bank communication to enhance policy effectiveness.
  - Preparatory steps for unconventional measures, such as large-scale asset purchases, funding for credit, or negative interest rates, to equip policymakers with tools to deal with adverse shocks.
- Concerns about financial-stability risks from low interest rates should be addressed via banking supervision and prudential policies to curb excessive leverage.

### The Equilibrium Interest Rate (EIR): definition and drivers
- Definition: EIR is the policy rate consistent with inflation being at the target and output at its potential level (full employment).
- EIR is unobservable and must be estimated; substantial model uncertainty implies judgment is required.
- For open economies with high capital mobility, like Australia, the domestic rate is strongly driven by global interest rates (EIRs in other major economies), with country-specific differences arising from a country risk premium (e.g., net foreign liabilities and repayment risk perceptions).

### Evidence on declining EIRs and Australian estimates
- Global/regional trend: Inflation-adjusted bond yields have declined worldwide since the early 1980s, with a renewed drop after the global financial crisis leading to downward revisions of the real EIR.
- Summers (2015) cited estimates for the U.S. neutral rate ranging between -3 and 1.75 percent; differences reflect alternative shock classifications.
- IMF staff estimates for Australia (Box 1) using multiple approaches point to a decline in the real EIR to around 1 percent.
- Box 1 methodologies and key quantitative findings:
  - Methodologies #1 and #2: Laubach and Williams (L-W) and Lubik and Matthes (L-M, TVP-VAR) are empirically-driven and subject to large statistical uncertainty.
  - Methodology #3: Indicative HP-filter univariate approach (λ=1,600) estimating trends in real 10-year, 90-day and floating mortgage interest rates.
  - L-W and L-M estimates indicate the real natural rate fell from an average of about 3 percent in the early 1990s to 0.8 percent in the second quarter of 2016, with values never turning negative.
  - Considering the lower bound of the 90 percent confidence region could point to a prevalent “tight” policy in certain periods.
  - Projecting to end of 2021, with output gap near zero and inflation close to 2.5 percent, an appropriate nominal natural cash rate projection is close to 3.0 percent (as per the latest WEO projections).
  - HP-based estimates show the real 10-year interest rate fell from over 3 percent in 2000 to just 1 percent in 2016Q2.
- Figure 5 and related staff calculations show estimated real natural rates for multiple countries (USA, UK, Japan, Korea, Australia, Indonesia, China, Malaysia) with declines evident since the 1980s.

### Implications of a low EIR for policy
- With a real EIR of around 1 percent, the RBA faces relatively greater risks of reaching the Effective Lower Bound (ELB) of about 1 percent than it would if the EIR were higher.
- Consequences:
  - Limited room for conventional monetary policy responses to negative shocks.
  - Importance of exploring scope for unconventional measures.
  - Financial-stability concerns from low rates to be managed with prudential and supervisory policies.

### Inflation-Forecast Targeting (IFT) under current conditions
- IFT principle: the central bank’s own forecast of inflation is an ideal intermediate target because it embodies all relevant information and policymakers’ preferences.
- Key aspects:
  - Policy interest rate should respond in a predictable way to shocks to minimize deviations of inflation from target while considering output implications.
  - IFT uses policy feedback via the short-term interest rate; many central banks produce an endogenous path for the interest rate in forecasting models.
  - Expectations of future policy rate movements play a crucial role under IFT, especially in Australia where variable-rate debt predominates.
- Exchange rate as shock absorber:
  - Terms of trade volatility and exchange rate flexibility have historically absorbed commodity-price shocks.
  - Terms of trade rose by 85 percent from the average of the early 2000s and reached an all-time high in late 2011, then fell 36 percent from its peak but expected to be nearing a floor at a level 25 percent above its pre-boom level.
- At the ELB, a weakened transmission mechanism can still operate via inflation expectations:
  - Under a credible regime, expectations that inflation will return to target over the medium term can lower expected future real rates, induce real depreciation, and support demand.
  - To manage expectations, the central bank could publish a forecast showing the interest rate kept low for long enough to allow inflation to rise or overshoot the target; credibility may require backing by fiscal policies or unconventional monetary measures (quantitative easing, funding for credit, or others).

### Strategy recommendation summary
- Adopt a “prudent risk-management strategy” prioritizing avoidance of prolonged low-inflation, low-growth outcomes.
- Implement:
  - Effective use of conventional instruments.
  - Preparations and conditional deployment of unconventional tools (large-scale asset purchases, funding for credit, negative rates).
  - Stronger central bank communication and explicit management of expectations.
  - Complementary fiscal or prudential measures as needed to ensure the credibility and effectiveness of monetary policy actions.

*Italic: IMF staff summary of chapter content provided in the source document.*

### 29.      Under current circumstances, an effective monetary policy should counteract the

### cr1743 - 29.      Under current circumstances, an effective monetary policy should counteract the

### Risks, overall policy prescription, and prudent risk-management
- Under current circumstances, an effective monetary policy should counteract the possible risk of persistent disinflation building up and prolonged weakness in demand settling in.
- A “prudent risk-management strategy” implies a more aggressive cut in interest rates in the future compared to a standard monetary policy reaction function to eliminate the economic slack faster, while also taking into consideration financial stability risks when deciding on the speed with which to bring inflation back to target.
- To lower risks of “dark corners”, prudent policies call for both monetary and fiscal stimuli, in the event that further contractionary shocks (lower global growth, equilibrium world real interest rates and commodity prices) were to materialize.

### Model, calibration, and limits
- The analysis uses an open-economy New Keynesian model with equations for the output gap, core inflation, the policy interest rate, and the exchange rate; it also includes headline, food and energy inflation, the commodity terms of trade, trade and financial linkages, and bond yields of various maturities.
- Expectations are forward-looking and consistent with model projections; behavioral equations embody significant lags and rigidities.
- The model exhibits nonlinearities primarily because of the ELB constraint and is calibrated with reference to historical data and existing theoretical literature. Appendix II contains model equations and parameter values.
- Caveat: real-time forecasts in practice involve substantial judgment and data analysis given known problems with simple model-driven short-term projections; nonetheless, model simulations illustrate key features of an optimal monetary policy strategy.

### Baseline scenario: initial conditions and three policy strategies
- Initial conditions (Australian economy as of 2016Q3):
  - Output gap starts at -1.3 percent (excess supply).
  - Year-on-year headline inflation rate at 1.3 percent.
  - Cash rate at 1.5 percent (assumed to be declining to 1.3 percent by end-2017 as priced by the market).

- Strategy A — Standard inflation-forecast-based (IFB) reaction function (slow recovery):
  - Interest rate set as a function of expected year-on-year inflation one year ahead and the contemporaneous output gap.
  - Policy rate is first lowered modestly and then increased gradually over time.
  - Result: depreciation of the Australian dollar that helps support the economy; gradual closing of the output gap.
  - Inflation stays below the RBA’s 2-3 percent inflation target range for several years.

- Strategy B — Optimal “low for longer” (prudent risk-management strategy):
  - Optimal path derived by minimizing a loss function placing equal weights on output gap and inflation deviations (forward-looking, not rule-based).
  - Policy rate path: cut significantly over a few quarters, falling to 0.75 percent before it increases gradually.
  - Rationale: place a premium on avoiding “dark corners” and avert possible prolonged periods of excessive economic slack.
  - Outcomes:
    - Output gap closes much faster.
    - Inflation returns inside the 2-3 percent target range shortly over the next few quarters.
    - With interest rates lower for longer, the Australian dollar depreciates by an additional 6 percent.
  - Note: Based on Lowe (2012), a below 1 percent cash rate can be regarded as the ELB in Australia.

- Strategy C — Combined monetary-fiscal response:
  - Augmenting the prudent risk-management strategy with fiscal support to close the output gap even faster.
  - Monetary policy faces a lower burden; policy rates do not decline as much.
  - Result: less depreciation of the Australian dollar and more monetary policy space retained for future shocks.

### Downside scenario: China growth surprises on the downside
- Shock assumptions and direct impacts:
  - Effective foreign output gap (export-weighted for Australia’s main trading partners) is 2.3 percentage points lower at the trough compared to the baseline.
  - Metal prices are 5.3 percent lower at the trough compared to the baseline.
- Monetary-only response (policy rate to ELB):
  - Policy rate declines to the ELB (assumed to be 0.75 percent) and stays there for 2 years under the black-line simulation.
  - Result: additional depreciation in the Australian dollar supports the economy, but a sizable negative output gap remains; headline inflation slightly overshoots the 2.5 percent mid-point of the RBA’s 2-3 percent inflation target range.
- Coordinated monetary-fiscal response:
  - Direct fiscal stimulus: direct effects equal to 0.25 percent of GDP for 2017-18.
  - Monetary policy cuts aggressively to its lower bound faster than the monetary-only case to support the fiscal expansion.
  - Result: modest overshoot of inflation above the upper bound of the 2-3 percent target range over the medium term; planned overshoot raises inflation expectations and reduces real interest rates, combined with a more depreciated exchange rate to close the output gap much faster.
  - Faster recovery allows monetary policy to normalize earlier than under the monetary-only case.

### Mechanisms: exchange rate and asset prices as absorbers or amplifiers (Appendix I key points)
- Risk-adjusted uncovered interest parity (UIP) implies that expected future changes in the exchange rate offset interest differentials and risk premia.
- Real exchange rate as shock absorber:
  - Under active, credible policy, a negative demand shock leads the central bank to reduce the policy rate and raise expected medium-term inflation; the spot exchange rate depreciates immediately to compensate for lower domestic interest rates and anticipated future appreciation, supporting demand via exports and expenditure switching.
  - At the ZLB, credible aggressive policy still lowers real interest rates via higher expected inflation, producing a real depreciation (t z ) that supports demand.
- Real exchange rate as shock amplifier:
  - At the ZLB, if policy is passive and not credible, negative demand shocks lead people to expect lower future inflation; short-term real interest rates could increase, producing a real appreciation (t z ) that reduces net exports and deepens the recession.
- Asset prices follow a similar logic:
  - A credible aggressive policy response increases asset prices (through improved profits from currency depreciation and lower real discount rates), acting as a buffer.
  - A non-credible, passive response lowers asset prices, acting as an amplifier of shocks.

### Conclusions and policy recommendations
- The RBA could use a “prudent risk-management strategy” to leverage conventional instruments to avoid “dark corners”: extending commitment to keep interest rates “low for long” would eventually raise inflation and inflation expectations, lower real interest rates, and depreciate the exchange rate.
- Fiscal policy easing may be critical to support monetary policy; fiscal support has a direct demand effect and reduces the duration the policy rate must remain at the lower bound.
- Policy credibility and anchoring inflation expectations could benefit from more forecast-oriented communications by the RBA:
  - Reorient communications to discuss explicitly how the RBA intends to use policy instruments over time to bring inflation back to target, emphasizing the “conditionality” of the path and that it is a “forecast” and not a “commitment”.
  - Experience in New Zealand, Norway, Sweden, and the Czech Republic suggests market participants learn the conditional-forecast nature of the path relatively quickly.
  - Forward guidance importance increases as policy rates approach the ELB; better forward guidance would lower risks of drifting inflation expectations at the ELB.

*International Monetary Fund — cr1743 chapter excerpt*

### Appendix II. A New-Keynesian Model for Australia

### Appendix II. A New‑Keynesian Model for Australia

### A.II.1. IS Equation
- Definition: output gap (ˆyt) = log-level of output (yt) minus potential output (yt).
- Determinants of ˆyt:
  - Past and expected future output gaps.
  - Deviations of the four‑quarter average equilibrium real interest rate (4 tr) and the real effective exchange rate (reer) from equilibrium.
  - Rest‑of‑the‑world output gap.
  - Metal price gap (Metal rp t) has a significant effect.
- Structural form (coefficients and calibration groups supplied in parentheses):
  - ˆytt = ˆytt−1 + ˆytt+1 + β1 (0.5) ˆytt−2 + β2 (0.1) 4trt−1 + β3 (0.1) rt−1 + β4 (0.01) reert−1 + β5 (0.1) Worldy t−1 + β6 (0.01) Metal rp t−1 + εt
- Equilibrium real interest rate (four‑quarter average) definition:
  - 4 tr t = (rt + rt−1 + rt−2 + rt−3)/4

### A.II.2. Phillips Curve
- Core inflation (C t π) dynamics:
  - Depends on inflation expectations (C t Eπ) and past year‑on‑year core inflation (C t−4 π), with coefficients summing to one—lagged term captures intrinsic inflation inertia.
  - Inflation expectations pinned down by the model‑consistent solution of year‑on‑year inflation one year ahead (C t+4 π).
- Additional determinants:
  - Lagged output gap enters non‑linearly.
  - Rate of real effective exchange rate depreciation and deviation of reer from equilibrium.
  - Small pass‑through from oil and food price inflation via real price of oil and food adjusted for real exchange rate effects.
- Structural form (coefficients and calibration groups supplied in parentheses):
  - C t π = (0.75) C t Eπ + (0.2) C t−4 π + (0.025) term(output gap nonlinear) + (0.025) Δ reer + (0.005) deviation(reer) + (0.005) OilFood rp terms + εt
- Expectation anchoring:
  - C t+4 Eπ = C t+4 π

### A.II.3. Policy Interest Rate: Reaction Function Options
- Linear IFB rule (Intercepted Forward‑Looking, coefficients in parentheses):
  - it* = (0.7) it−1 + (1.2) [4(πt+1)] + (0.4) yˆt − ... + εt  (policy takes into account contemporaneous output gap and model forecast of inflation three quarters ahead)
  - Projected year‑on‑year inflation uses model forecast [πt+1, πt+2, πt+3, πt+4].
  - Smoothing: nominal rate it depends on its lag to reflect gradual adjustments.
- Loss‑minimizing (risk management) strategy:
  - Minimizes discounted sum of losses from inflation deviations, output gaps, and changes in policy rate.
  - Loss function (coefficients in parentheses):
    - Loss = Σ∞i=0 βi [ω1 (πt+i − π*)^2 + ω2 yˆt+i^2 + ω3 (Δit+i−1)^2] with calibration (0.98) for β, (1.0) for ω1, (1.0) for ω2, (0.5) for ω3.
  - Quadratic penalties imply large errors weighted disproportionately and penalize sharp policy moves.
- Effective lower bound (ELB / floor i):
  - Interest rate constrained: it ≥ floor i
  - floor i assumed equal to 75 basis points in historical simulation (consistent with Lowe (2012)).

### A.II.4. Real Interest Rates and Real Exchange Rates
- Real interest rate definition:
  - rt = it − C t+1 π  (nominal interest rate minus expected core inflation)
- Bilateral real exchange rate with the United States (zt):
  - zt = C t p + s tUSC − pUS t  (constructed so an increase = AUD depreciation)
  - Decomposition: zt = z̄t + ˆzt  (equilibrium trend z̄t and deviation ˆzt)
  - Equilibrium real exchange rate trend determined by equilibrium terms of trade tot t z̄ = tot t z̄
- Real effective exchange rate (reer) entering output gap:
  - Trade‑weighted bilateral real exchange rates versus seven regions with calibration weights (in parentheses):
    - reer = 0.120 * zUS + 0.149 * zEU + 0.077 * zJA + 0.233 * zCH + 0.281 * zEA + 0.017 * zLA + 0.123 * zRC  (0.123)
- Risk‑adjusted UIP condition:
  - Links bilateral exchange rate (s) and interest rates it and iUS t with country risk premium ctry t σ and terms‑of‑trade shocks tot t σ.
  - Expected exchange rate 1t Es+ consistent with a mix of model‑consistent forward solution and backward‑looking expectations:
    - 1t Es+ = φ1 1t s+ + φ2 1t s− + adjustments including *,*US π differential de‑annualized by 1/4 and extrapolation factor 2 (details embedded in model).

### A.II.5. Relative Prices
- Headline inflation dynamics driven by relative price movements (core CPI C t p relative to headline CPI H t p).
- Long run: headline inflation = core inflation, but prolonged divergence allowed when relative prices trend.
- Relative price decomposition:
  - rp t = rp̄ t + ˆrp t
- Relative price gap determinants (coefficients in parentheses):
  - ˆrp t = ρ (ˆrp t−1) + c1 (0.5) Δ OilFood real prices adjusted for exchange rate + c2 (0.01) rp zc + c3 (0.01) rp z + εt
  - Trend growth of rp assumed autoregressive with mean zero; additional ΔΔ formulation:
    - Δ ˆrp t = (0.7) Δ ˆrp t−1 + εt

### A.II.6. Term Structure of Interest Rates
- Government bond yield with maturity k quarters (iGov k t) equals average expected short‑term rates k quarters ahead plus term component (Term k t σ), and measurement error εGov k t.
- Examples (notation preserves calibration forms shown):
  - iGov,4t = (i t + i t+1 + i t+2 + i t+3)/4 + Term,4 t σ + εGov,4 t
  - iGov,8t = (i t + i t+1 + ... + i t+7)/8 + Term,8 t σ + εGov,8 t (formulation averaged similarly)
  - iGov,20t and iGov,40t follow analogous multi‑quarter averaging with increasing term premium components and explicit calibration weights/sums.

### A.II.7. Unemployment Rate
- Unemployment rate ut decomposed into NAIRU ut and unemployment gap ˆut:
  - ut = ūt + ˆut
- Gap version of Okun’s law:
  - A one percentage point increase in ˆut is associated with approximately two percentage points of negative output gap.
- NAIRU dynamics (coefficients in parentheses):
  - ˆut = (0.25) ˆut−1 + (0.5) cyclical term + εt
  - ūt = ūt−1 + Δūt + εt
  - Δūt = (0.9) Δūt−1 + εt

### A.II.8. Potential Output
- Potential growth rate Δyt converges to steady state Δy ss in the long term but can deviate for prolonged periods.
- Dynamics (coefficients in parentheses):
  - Δyt = (0.9) Δyt−1 + (2.47) (Δy ss − Δyt−1) + εt

### A.II.9. The Rest of the World
- Rest‑of‑world output gap relevant for Australia is a weighted average of output gaps in seven regions (U.S., euro area, Japan, China, emerging Asia, Latin America, rest of world) using export shares as weights (coefficients in parentheses):
  - World yˆ t = (0.047) yˆ US t + (0.030) yˆ EU t + (0.200) yˆ JA t + (0.375) yˆ CH t + (0.270) yˆ EA t + (0.008) yˆ LA t + (0.070) yˆ RC t
- Equilibrium real interest rate relation to U.S.:
  - r t = r US t+1 − adjustments from totxtot zΔ and country risk premium ctry t σ (details retained in model equations with calibration (0.5))
- Terms‑of‑trade trend dynamics also follow autoregressive formulations with coefficients (0.95) and (0.95) for Δ and levels as shown.

### A.II.10. Commodity Terms of Trade
- Real price of oil:
  - Oil rp t = Oil p t − US p t  (real global oil price relative to U.S. CPI)
  - Oil price gap ˆOil rp t = Oil rp t − Oil rp̄ t
  - Dynamics (coefficients in parentheses):
    - ΔΔ Oil rp t = (0.95) ΔΔ Oil rp t−1 + εt
    - Oil rp t = (0.7) Oil rp t−1 + εt
- Real price of food modeled analogously to oil with same coefficient patterns (0.95) and (0.7).
- Terms‑of‑trade gap tot̂ t determined by metal price gap ˆMetal rp t with weight (0.01):
  - tot c t = (0.01) Metal rp t
- Real exchange rate depreciation consistent with terms‑of‑trade changes:
  - tot zΔ t = (0.6) (− Metal rpΔ t)
  - Equilibrium condition mirrors the same relation (0.6)
- Terms‑of‑trade premium in UIP (tot t σ) is the “surprise” component:
  - tot t σ = z t − Et−1[z t] (or modeled as 1/4(zt − Et−1 zt) in expressions shown)

### Fiscal Framework (brief from continuation of Appendix)
- Context:
  - Current framework requires reporting against a medium‑term fiscal strategy based on principles of sound fiscal management (Charter of Budget Honesty Act 1998).
  - Government’s medium‑term balance anchor: reach and maintain annual government surplus of 1 percent of GDP (historical anchor referenced).
  - Repeated misses of budget targets amid weaker‑than‑expected economy (drop in global iron ore prices, mining sector recession) raise credibility questions.
- Assessment and recommendations (from paper summary):
  - Underperformance vs targets not necessarily framework failure—weak economy can explain misses; clear communication of reasons can preserve credibility.
  - Recommend replacing medium‑term budget balance anchor with a long‑term debt anchor:
    - A flexible long‑term debt anchor reinforces debt sustainability under temporary and persistent shocks and allows fiscal policy role in short‑term stabilization.
    - Medium‑term budget balance anchor can be retained only as an intermediate objective, conditional on initial conditions and outlook to be consistent with long‑term debt anchor.
    - Achieve and maintain long‑term debt anchor via a combination of fiscal rules: 1) expenditure restrictions and 2) a flexible horizon for convergence to the anchor.
  - Caution: overly ambitious near‑term budget repair poses risks to recovery and future growth given primary downside risks from the “new mediocre.”

*Italicized attribution line: Appendix II. A New‑Keynesian Model for Australia (cr1743) — material from the source PDF.*

### 3.      There are immediate and longer term challenges spanning slower recovery and growth

### 3.      There are immediate and longer term challenges spanning slower recovery and growth

### Immediate and longer-term challenges
- Advanced-economy "new mediocre" risk: a sustained lower growth path that is likely to dampen expectations about future growth in the short term.
- Long-term demographic pressures: ageing of the population will increase spending (primarily health care, aged care and pensions) and raise long-term fiscal pressures.
- IGR 2015 finding: without policy action, spending trends related to demographic issues will lead to an unsustainable situation with a high debt-to-GDP ratio.

### Commonwealth Government’s current fiscal plans and questions
- Current plans: meet the medium-term fiscal strategy goals; budget repair required first, then renewed focus on medium/long term.
- Key policy questions identified:
  - What is the appropriate response for budget repair with the current medium-term balance anchor, in terms of speed and short-term implementation, given the current outlook, and the downside risk of the “new mediocre?”
  - Is the fiscal anchor still appropriate in the long term, given the dangers inherent in the “new mediocre” and long-term demographic pressures, starting from higher initial debt levels than in the past?
  - What are the possibilities for a new fiscal anchor and its implementation, within a better-defined long-term fiscal strategy?

### Assessment of the medium-term anchor and budget repair
- Viability: The medium-term anchor is still viable in principle.
- Risk: An ambitious budget repair strategy risks being destabilizing in the short term given the “new mediocre” downside.
- Suggested approach: a mixed spending and tax strategy tilted towards spending measures under a longer adjustment horizon, entailing measures beyond the ones already recognized in Budget 2016-17 (Commonwealth of Australia, 2016).

### Proposal for a long-term debt anchor
- Rationale:
  - A long-term debt anchor can better manage the trade-off between debt sustainability and economic stabilization.
  - A target related to the debt stock reduces uncertainty about government behavior compared to a flow-based approach with a less well-defined horizon.
- Implementation features:
  - An explicit long-term debt anchor, based on a flexible response from the tax regime.
  - A flexible rule to reach the long-term debt anchor can minimize transition costs and perform well across different shocks, provided communication of the new long-term strategy is clear.
- Communication caveat: Lack of clarity may be costlier with a long-term anchor, but a clearly stated debt anchor provides greater benefits than one based on the budget balance.

### The current fiscal framework (principles and objectives)
- Structure: principles-based, “constrained discretion” enshrined in the Charter of Budgetary Honesty Act 1998; Fiscal Strategy Statement used to specify objectives.
- Historical continuity: budget balance objectives have been central since 1996—either balanced budget over the cycle pre-FY2008/09 or return to a budget surplus of at least 1 percent of GDP in the medium term post-FY2008/09.
- Additional objectives used intermittently: (net) debt or net (financial) worth, caps on tax share in GDP, caps on expenditures; an expenditure objective limiting real expenditure growth to 2 percent was used FY2009/10 to FY2013/14.

### Government’s medium-term strategy and budget repair (as stated in Budget 2016-17)
- Medium-term fiscal strategy: achieve budget surpluses, on average, over the business cycle.
- Four policy elements (Commonwealth of Australia, 2016, Statement 3, p. 7):
  - “investing in a stronger economy by redirecting Government spending to quality investment to boost productivity and workforce participation;
  - “maintaining strong fiscal discipline by controlling expenditure to reduce the Government’s share of the economy over time in order to free up resources for private investment to drive jobs and economic growth, with: the payments-to-GDP ratio falling; stabilizing and then reducing net debt over time;
  - “supporting revenue growth by supporting policies that drive earnings and economic growth; and
  - “strengthening the Government’s balance sheet by improving net financial worth over time.”
- Projections and targets:
  - Real payments growth FY2016-17 to FY2019-20: expected to be 1.9 percent per annum on average.
  - Projected average annual real growth in payments FY2020-21 to FY2026-27: around 2.9 percent.
  - Budget repair target: deliver budget surpluses of at least 1 percent of GDP as soon as possible and maintain until achieved.
  - Budget 2016-17 medium-term projection: surplus of 0.1 of GDP in FY2019/20.
  - IMF WEO (October 2016): sees a balanced budget in percent of GDP by the end of 2021.
- Operational rule: offset any new spending measures with other spending cuts; any upward surprises in revenues/spending are to reduce the deficit.
- Horizon clarity: the horizon over which to maintain the surplus goal is contingent on sound growth and low unemployment, but is not well defined in Budget 2016-17.

### Fiscal performance since the Charter (summary findings)
- Pre-crisis (pre-global financial crisis): Charter helped maintain surpluses most of the time; net asset-to-GDP ratio peaked at 7.3 percent in FY2007/08.
- Crisis response: fiscal stimulus of roughly 4.5 percent of GDP launched at onset of global financial crisis, reversing debt trajectory from FY2008/09.
- Post-crisis repair: strategy aimed to return to a 1 percent of GDP surplus; early years achieved some deficit reversal but entailed pro-cyclicality and optimistic revenue projections.
- Later performance:
  - By FY2012/13 deficit reduced to 2.7 percent of GDP.
  - Since FY2012/13, deficit reduction has stalled; deficit stabilizing slightly above 2.7 percent of GDP, largely due to deteriorating mining-sector revenues.
  - Repeated shifts of consolidation have postponed the 1 percent surplus target further into the future.
  - Budget 2016-17 projects a deficit of 0.1 percent of GDP in FY2019/20 but relies on revenue projections with downside risk.
- Credibility concerns:
  - Repeated underperformance against plans and lengthening consolidation horizon weakened fiscal credibility.
  - Rapid increase in debt-to-GDP ratios after the crisis and high share of foreign debt financing compound concerns.
  - Rating agencies: Standard and Poor’s in 2016 placed the Australian sovereign’s AAA rating outlook at negative; Moody’s and Fitch maintained stable outlooks.

### Analytical framework: G20MOD model used for policy experiments
- Model type: G20MOD, a module of the IMF’s Flexible System of Global Models (FSGM); annual, multi-region, general equilibrium model combining micro-founded and reduced-form sectors.
- Key structural features:
  - Country/regional blocks structurally similar but with different steady-state ratios and behavioral parameters.
  - Real GDP determined by aggregate demand components short term and potential output long term; aggregate demand follows national expenditure identity.
  - Consumption and private investment based on optimizing behavior; trade semi-structural; government mostly exogenous.
  - Price dynamics modeled via CPI inflation and wage-inflation Phillips curves.
  - Commodities sector (oil, metals, food) modeled globally and by country; metals important for Australia.
  - Financial sector: simple, provides a 1-year interest rate (monetary policy instrument) and a 10-year interest rate.
  - External sector: aggregate trade with the rest of the world, not bilateral tracking.
- Key demand-side feature:
  - Consumption block based on Blanchard-Weil-Yaari overlapping generations (OLG) model.
  - OLG implies non-Ricardian properties: households treat government bonds as wealth; government debt path has significant economic implications.
  - OLG leads to endogenous national saving and demand for net foreign assets given government debt; global real interest rate adjusts in long term to equilibrate global savings.
  - Presence of liquidity-constrained (LIQ) households who consume out of current income.

*Source: International Monetary Fund, cr1743.*

### 28.      Fiscal policy is driven by a sufficiently detailed government sector that can reproduce

### cr1743 - 28.      Fiscal policy is driven by a sufficiently detailed government sector that can reproduce

### Government sector design and fiscal instruments
- Government sector reproduces simplified fiscal accounts for each country.
- Eight policy instruments:
  - Spending: government consumption; infrastructure spending; general lumpsum transfers to all households (such as pensions, aged care provisions, unemployment insurance); lumpsum transfers targeted to LIQ households (such as welfare, certain pensions).
  - Revenues: taxes on consumption (GST), personal income (PIT, on both wage and dividend income), company income (CIT), and taxes/royalties from mining and production of metals.
- For Australia the government aggregates Commonwealth and State governments; many issues focus on Commonwealth accounts and balances.

### Budget constraint, long-term deficit and debt linkage
- Deficit definition: expenditures and interest payments, less revenues.
- One instrument (general lumpsum transfers by default) is continuously adjusted so the budget constraint always holds.
- Long-term government debt target, relative to GDP, is derived from the long-term deficit target and the nominal growth rate (inflation and steady-state growth). The explicit deficit target pins down long-term government debt, which affects the global interest rate and the real exchange rate.

### Fiscal rules: definitions and G20MOD implementation
- Fiscal rule taxonomy (IMF (2009) terminology):
  - Expenditure rule (ER): restriction on some component of spending; does not directly restrict budget balance.
  - Budget balance rule (BBR): restriction to ensure the budget meets a particular target level after some horizon (balanced budget is special case).
  - Debt rule (DR): brings government debt to a particular target level after some horizon.
- Strict vs flexible:
  - ‘Strict’: target level/path met in full each year.
  - ‘Flexible’: target met on average over a horizon (for example, “over the business cycle” or “over the course of three years”).
- G20MOD: deficit target is a strict BBR with embedded short-term flexibility via automatic social transfers tied to the output gap:
  - The strict BBR translates into the deficit-to-GDP ratio ݂݁݀݃
௥௔௧
 with flexibility: ݂݁݀݃
௥௔௧
݂݁݀݃ൌ
௧௔௥
݀െ
௬௚௔௣
ݕ
௚௔௣
 (where ݀
௬௚௔௣
 is the countercyclical weight on the output gap ݕ
௚௔௣
).

### Calibration: Australian external sector and commodities
- Australia share of global GDP: approximately 1.6 percent.
- Australia metals market share (aluminum, copper, iron, lead, nickel, tin, uranium, zinc in G20MOD): 13.6 percent of global production.
- Resources share of Australian real GDP: roughly 9.1 percent.
- Metals royalties and resource taxes calibrated to 0.75 percent of GDP.
- Openness:
  - Exports: 24.6 percent of GDP.
  - Imports: 23.5 percent of GDP.
  - G-20 averages: exports 26.6 percent of GDP; imports 26.9 percent of GDP.
  - United States: exports 17.3 percent of GDP; imports 16.2 percent of GDP.
- Policy leakages and multipliers:
  - A 1 percent increase in the nominal 1-year interest rate decreases real GDP by 0.3 percent in Australia, versus 0.4 percent in the United States.
  - Two-year real GDP multiplier for government consumption: Australia 0.62; United States 0.71.
- Metals price dynamics calibrated consistent with the October 2016 WEO metals price; metals price governed by global production and demand effects proxied by output gaps (weighted by consumption and production weights).

### Calibration: domestic economy and parameters
- Australian general consolidated government net debt: 20 percent of GDP (net basis).
- Australia draws more tax income from corporate taxes as percent of GDP relative to G-20.
- Net foreign liability position is large relative to the G-20, making U.S. interest rate shocks more threatening.
- Common parameterization with other advanced economies:
  - Share of LIQ households: 35 percent.
  - Price and wage Phillips curves, financial accelerator and intertemporal elasticity of substitution follow non-European advanced-economy calibrations.

### Fiscal rules calibration (flexible rules)
- Flexible BBR: weight on the deficit gap = 0.8.
- Flexible DR:
  - Weight on debt gap = inverse of number of years (example: 0.1 for 10 years, 0.2 for 5 years).
  - Weight on deficit gap = 0.25.
  - Flexible DR uses first difference of the debt-to-GDP ratio to minimize volatility during adjustment.

### Achieving the medium-term balance anchor (budget repair)
- Government medium-term balance anchor: +1 percent of GDP.
- October 2016 WEO projects a government balanced budget (zero percent of GDP) by 2021 (FY2020/21), falling short of the +1 percent target.
- Required budget repair: +1 percent of GDP relative to the WEO baseline.
- Illustrative adjustment strategy to close the gap:
  - Achieve 25 percent of the gap by increasing the GST (revenue side).
  - Achieve 75 percent from the spending side, split evenly between government consumption and general transfers.
  - Spending adjustments phased in linearly over 30 years.
  - Two GST time horizons considered:
    - 5 years: immediate GST increase so medium-term balance anchor achieved by FY2020/21.
    - 10 years: GST increase introduced only after five years so anchor achieved by FY2025/26.

### Risk environment and the “New Mediocre” scenario
- The baseline is fairly benign but risks skewed to the downside; key risk labeled the “new mediocre” (global secular stagnation).
- “New mediocre” illustrative assumptions and impacts for Australia:
  - Permanent decline in global metals prices (negative terms-of-trade shock).
  - Australian dollar appreciates by about 1.5 percent.
  - Real GDP contracts by 1.5 percent relative to baseline after 10 years.
  - Consumption and investment decline by 2.1 and 2 percent, respectively (relative to baseline in the illustrative Box).
  - Government deficit increases by 0.5 percent of GDP by the second year, and by 0.8 percent of GDP by the tenth year.
  - Monetary response: RBA cuts the overnight cash rate by around 30 basis points at its trough in the second year.

### Budget repair: comparative outcomes for 5-year vs 10-year horizons
- Considered scenarios: budget repair completed over 5 years or 10 years, under both baseline and “new mediocre”.
- Consumption impacts of repair:
  - Consumption falls by about 0.8 percent relative to baseline under repair.
  - Consumption falls by about 1.0 percent relative to the “new mediocre”.
- Investment impacts:
  - Fall in investment is temporary, worse in the 10-year case and worse under the “new mediocre”.
  - In the fourth year the additional fall is 1.0 percent relative to the “new mediocre” versus about -0.6 percent relative to baseline.
- External adjustment and exchange rate:
  - 10-year case: real exchange rate depreciation about 0.6 percent relative to baseline.
  - 5-year case: real exchange rate depreciation about 2.6 percent relative to baseline.
- Shorter (5-year) repair tends to yield more substantive positive effects:
  - Faster consolidation → higher near-term trade balance, larger depreciation, excess demand, inflationary pressures → tighter monetary policy (higher overnight cash rate).
  - Excess demand reduces investment risk, lowering corporate interest rates relative to the 10-year repair.
- Infrastructure offset:
  - Deficit-financed infrastructure spending of 0.25 percent of GDP for each of the first two years introduced as a stimulus to real GDP and productivity.
  - This spending encourages larger real exchange rate depreciation, but given its size it does little to offset large short-term impacts if a shock like the “new mediocre” materializes.

### Conclusions on budget repair feasibility
- Budget repair is at a risky juncture:
  - Required magnitude of repair is uncertain and depends on revenue forecasts, which historically have been over-optimistic.
  - The actual required repair could be larger than 1 percent of GDP.
  - Repair relies on rebalancing from the domestic to the external sector, which may be politically or economically undesirable.
  - Shorter horizon (5-year) is preferred in the model because it leverages quicker external sector adjustment and larger depreciation that offsets consumption contraction.

*Source: IMF staff calculations and analysis in the provided chapter excerpt.*

### 49.      Regardless, budget repair under the “new mediocre” could also prove to be a much

### cr1743 - 49.      Regardless, budget repair under the “new mediocre” could also prove to be a much

### Budget repair under the “new mediocre”
- Budget repair to attain the 1 percent surplus target requires an initial adjustment equivalent to 1 percent of GDP.
- Additional offsets are required for automatic social transfers and weaker tax base effects, estimated at 0.9 percent of GDP in the illustrative scenario.
- Under the “new mediocre” environment, budget repair would prevent the Government from engaging in additional stimulus, inconsistent with historical stimulative behavior (for example, the estimated 4.5 percent of GDP fiscal impulse at the onset of the global financial crisis in FY2008/09).

### Suggesting a New Long-Term Fiscal Strategy (overview)
- Once the desired medium-term balance anchor is achieved, the paper proposes extending the fiscal framework with a long-term debt anchor.
- The analysis is structured in four parts:
  - Long-term spending pressures motivating a long-term anchor.
  - Comparison of a long-term debt anchor versus the existing medium-term balance anchor under three illustrative scenarios, using debt sustainability and economic stability metrics.
  - Ways to implement a long-term debt anchor via combinations of an expenditure-debt rule under differing speeds of adjustment, considering an optimized mix of indirect versus direct taxes.
  - Repetition of two scenarios (aggregate demand and boom-bust) to elucidate strengths of the fiscal rules, followed by examination of short-term effects of clear and credible communication on the adjustment process.

### Long-Term Spending Pressures from Demographics
- Long-term pressures are primarily demographic: health care, aged care, and pensions.
- By 2050, IGR 2015 estimates the Government will have to spend an additional 2.5 percent of GDP on such outlays.
- Assuming a linear trend starting from zero in 2021 and reaching its long-term value after 30 years:
  - Net debt in 2050 is estimated to increase by 21 percent of GDP.
  - In the very long term, net debt would stabilize after increasing by 58 percent of GDP.
- This projection is labeled the “No Policy Action Scenario.”

### Tax Switching and Fiscal Consolidation Scenarios
- Tax Switching Scenario (revenue-neutral):
  - Reduce direct taxes: CIT by 0.7 percent of GDP, PIT by 1.4 percent of GDP.
  - Offset with GST increase worth 2.1 percent of GDP (similar to Pitt (2015)).
  - Embeds the No Policy Action Scenario.
- Simple Balanced Budget Scenario:
  - Builds on the Tax Switching Scenario.
  - Government adheres to a balanced budget, aside from automatic social transfers.
  - Achieves debt control via cuts in other lumpsum transfers (outside aged care and pensions).
- Effects under No Policy Action Scenario:
  - Real GDP is stronger by less than 1 percent after 30 years.
  - Household consumption increases in the long-term by about 2 percent.
  - Private investment holds roughly at 0 percent relative to baseline in the long term (crowded out).
  - Australian net foreign liability position increases by about 20 percent of GDP.
  - Real appreciation around 2 percent relative to baseline in the long term.
- Effects under Tax Switching Scenario:
  - PIT and CIT cuts produce positive supply-side effects that outweigh the GST increase.
  - Consumption gains 0.5 percentage points; investment gains 4.0 percentage points.
  - Real GDP is about 1.4 percentage points higher relative to the No Policy Action Scenario.
  - Short-term: RBA increases the overnight cash rate by over 50 basis points to maintain inflation on target.
- Box 2 decomposition highlights:
  - Switch from CIT to PIT of 0.7 percent of GDP: investment increases, dividend income taxed more, consumption effects offset within first 5 years.
  - Further switch from PIT (wage income) to GST of 2.1 percent of GDP: dampens consumption; net effect on real GDP is relatively weak at 0.3 percent of GDP (multiplier of 0.14).
- Effects under Balanced Budget Scenario:
  - Crowding in of private investment as private saving and foreign assets are no longer diverted to government.
  - Tax switching helps arrest rebalancing toward external sector.
  - Expansionary depreciation of 1.2 percent relative to baseline in the long term.

### Benchmark and rationale for a long-term debt anchor
- The No Policy Action Scenario is the Benchmark: net government debt and net foreign liabilities deteriorate in the medium term if nothing is done.
- Risks such as sovereign risk premia raising borrowing costs are noted but not explored in the benchmark.
- Proposal: adopt a government-debt-to-GDP target as a long-term anchor and implement a long-term fiscal strategy.
- Illustrative exercise:
  - Benchmark scenario taken as given with large run-up in debt.
  - Examine implementation of long-term anchor over 2021 to 2050.
  - Maintain tax switching to better position future fiscal actions.
  - Analyze a debt norm: reduction of net government debt by 10 percent of GDP to elucidate costs and benefits.

### The Argument for a Long-Term Debt Anchor
- A long-term debt anchor provides certainty for households and firms in decision making by removing uncertainty about the real economy associated with government debt.
- Government debt influences household wealth holdings, net foreign liabilities (through current account), and capital stock (through investment).
- A long-term debt anchor (stock concept) stabilizes the economy in ways the medium-term deficit anchor (flow concept) cannot.
- Adoption can strengthen institutional commitment to fiscal responsibility and signal credibility domestically and internationally.

### Scenarios and fiscal rules evaluated
- Three illustrative scenarios covering demand and supply disturbances:
  1. Temporary but substantial increase in aggregate demand: aggregate demand stronger by about 2 percent of GDP at peak in year 3 (see Box 3).
  2. Terms-of-trade-driven boom-bust cycle: imposed path of real global metals prices matching RBA’s Index of Commodity Prices (see Box 4).
  3. The “new mediocre” (Box 1): combined secular stagnation and permanent negative terms-of-trade shock; a negative shock to future expectations.
- Three fiscal rules analyzed:
  - Flexible BBR over 7 years (flex BR) — medium-term balance anchor.
  - Strict debt rule (strict DR) — holds debt at its target.
  - Flexible debt rule consistent with a 10-year horizon (flex DR (10 yrs)).
- Evaluation metrics (based on Kinda (2015)) — five indicators:
  - Debt control flexibility: measured by RMSD from the debt norm (further from 0 implies more flexibility).
  - Debt precision: deviation from target at year 10 (2030) and absolute mean deviation (AMD) of debt from the debt norm path (closer to 0 preferred).
  - Economic stabilization: variability of real GDP over 10 years (closer to 0 preferred) and a measure of procyclicality (preference for countercyclicality).
- Definition of procyclicality measure:
  - Centered around zero, compares change in the primary surplus with change in the output gap.
  - Fiscal policy is procyclical if output gap becomes more positive (negative) and primary surplus decreases (increases).
  - Preference is for countercyclicality where primary surplus and output gap are positively correlated.
- Results summary:
  - Long-term debt anchor outperforms the medium-term balance anchor.
  - RMSD: flex DR and medium-term balance anchor show more flexibility (debt responds better to shocks).
  - Debt rules are more precise in the long term (lower AMD for both debt rules).
  - Real output is less variable under debt rules, implying more certainty for households and firms.
  - DR achieves faster convergence to desired debt-to-GDP ratio.
  - Medium-term debt anchor is more countercyclical, but flex DR also shows notable countercyclicality.
  - Given unmodeled uncertainty in the long-term end-point of the announced medium-term balance anchor, the flexible DR performs strongly.

*Italic: IMF staff summary of chapter content as provided in the source PDF.*

### 67.      To implement the long-term debt anchor requires a long-term fiscal strategy. In

### cr1743 - 67.      To implement the long-term debt anchor requires a long-term fiscal strategy. In

### Long-term fiscal strategy and rationale
- Implementing a long-term debt anchor requires a long-term fiscal strategy and a consolidation plan to support macro and fiscal stability; outcome: real GDP would be larger and debt would be expected to be more sustainable than otherwise.
- A regime based on an aggregated rule composed of explicit fiscal rules is recommended; component fiscal rules must be appropriate to macro-fiscal conditions and institutional settings.
- Legislated numerical fiscal rules may have weaker rationale in Australia given its record of responsible fiscal policies, and fiscal rules involve trade-offs between inter-temporal consistency and flexibility (scope for countercyclical policy and full operation of automatic social transfers). Trade-offs can be moderated by design elements such as escape and revision clauses.

### Benchmark, targets, and modeling setup
- Medium-term target horizon: extends between 2021 and 2050, taking the benchmark scenario as given with its large run-up in debt.
- Debt norm for analysis: a reduction in net debt of 10 percent of GDP relative to the benchmark scenario; this is treated as the “benchmark consolidation” case.
- Tax switching is implemented to better position future actions.

### Component fiscal rules used
- Expenditure rule (ER):
  - Assume government consumption and general transfers grow at only 2 percent a year for 10 years starting in 2021, roughly consistent with the average over the WEO horizon extending from 2016 to 2020.
  - Relative to Budget 2016-17, this means a reduction spending growth by 0.9 percentage points each year.
- Variants of a debt rule (DR):
  - Meet the long-term government debt target after 5 or 10 years by varying the GST rate to achieve the debt norm of 10 percent, with either a strict or flexible pursuit of the debt norm path from 2021 to 2050.

### Aggregated fiscal rules constructed
- ER + strict DR:
  - 90 percent of debt norm is achieved by 2030, with a strict adherence to the debt target each year.
- ER + flexible DR (10-year variant):
  - 90 percent of debt norm is achieved by 2030 (over 10 years) allowing for some variability in debt consistent with an average business cycle frequency of six to seven years; implies an average decrease in the deficit of around 1 percent of GDP relative to the ‘benchmark scenario’.
- ER + flexible DR (5-year variant):
  - 90 percent of debt norm is achieved by 2025 (over 5 years) allowing for some variability in debt consistent with an average business cycle frequency of six to seven years; implies an average decrease in the deficit of around 2 percent of GDP relative to the ‘benchmark scenario’.

### Key modeled outcomes and transmission
- Under the benchmark consolidation:
  - Real GDP is always stronger in the long run due to tax switching, but there are short-term downward pressures from consolidation.
  - Shorter (5-year) consolidation inflicts the largest near-term costs; more flexible rules allow real GDP to react positively to tax switching, with debt reacting more gradually.
- Domestic demand components:
  - Short term: consumption is weakened by GST increases (dominating positive effects from the PIT cut); private investment is stronger due to lower CIT and crowding in from reduced government dissaving. The 5-year DR reinforces this dichotomy.
  - Long term: private investment and higher labor income and wealth lead consumption to become a positive contributor to GDP.
- Monetary policy interactions:
  - Short term: RBA maintains inflation target, smoothing real GDP; large GST increases are disinflationary, leading the RBA to increase the overnight cash rate less on impact and then cut in the medium term—monetary policy acts as a drag except under the five-year horizon.
- External sector and exchange rate:
  - Reduction in government debt lowers domestic assets available to households, prompting demand for more foreign assets and higher current account to accumulate those assets, especially in the short term.
  - Real exchange rate depreciates as more goods are available that need to be sold abroad.
- More aggressive (5-year) debt rule:
  - Faster convergence implies rapid changes in economic flows, larger short-term consumption reduction via GST, looser monetary policy, stronger private investment, quicker real exchange rate depreciation, and larger current account surpluses on impact.

### Performance under illustrative shocks
- Temporary but substantial increase in aggregate demand:
  - More slack relative to benchmark; potential for higher revenues and lower automatic social transfers; prices and real exchange rate face upward pressures requiring monetary tightening.
  - Aggregate demand shock modifies required GST increases to meet debt rules; GST increase also dampens the shock, reducing sole reliance on monetary policy.
- Terms-of-trade-driven boom-bust cycle:
  - Real GDP is much more volatile; automatic social transfers are more variable, interacting with aggregated fiscal rules but generally constrained by the ER.
  - Real exchange rate movements dominated by commodity swings, increasing the stabilizing role for both monetary and fiscal policy.

### Rule characteristics, trade-offs, and clarity
- ER effects:
  - ER always restricts government spending; in expansions it is countercyclical (discourages spending of additional revenues), but in contractions it is procyclical (may offset automatic social transfers).
- DR effects:
  - DR always reduces government debt; in expansions it is countercyclical (absorbs fiscal gains), but in contractions it is procyclical (may offset automatic social transfers).
- Communication (clarity) matters:
  - Four concepts of clarity: Fully or immediately clear; Unclear; Stepwise clear; Partially clear.
  - Unclear communications (households/firms do not believe future path) lead to more negative effects from consolidation because forward-looking agents do not smooth consumption/investment; full clarity allows borrowing against future gains, smoothing consumption and reducing adjustment costs.
  - Figure analysis suggests unclear communications worsen short-run GDP outcomes and increase adjustment costs, while clarity improves transitions.

### Comparative evaluation and preferred rules
- Summary findings (Table 4 referenced in source):
  - ER + flexible DR rules may overshoot the debt reduction in model simulations; ER + strict DR are less flexible and restrict faster real GDP growth under alternative scenarios (more countercyclical).
  - Flexible DR rules tend to be more procyclical under some conditions but under the benchmark consolidation can be more countercyclical and allow automatic social transfers to work, causing slightly more real GDP variability that can be reduced by faster consolidation.
- Subjective ranking for sustainability and stability during transition:
  1. Preferred: flexible DR that aims to return debt to its target after 10 years — slower adjustment, more consistent on procyclicality across scenarios, closer to zero procyclicality; debt behavior is more flexible (higher RMSD) and sacrifices some precision (higher AMD); real GDP is less variable.
  2. Second choice: flexible DR returning debt to its target after 5 years.
  3. Third choice: strict DR.
  - Differences among these are not substantial; either flexible DR may be more advantageous than strict DR because they provide greater certainty (shorter adjustment horizon or definitive yearly stance).

### Policy recommendations and implementation considerations
- Recommendation: use a fiscal rule for consolidation based on a long-term debt anchor rather than a strict pursuit of the medium-term balance anchor at this time.
- Recommended rule: the flexible debt rule that attempts to achieve its debt norm over 10 years, as it best addresses risks and communication issues on average.
- Complementary measures:
  - Tax switching: shift tax composition away from taxes on factors of production (the CIT and PIT) and replace with an equivalent increase in the GST in tandem with introducing the new long-term strategy.
  - Using the GST as the adjusting instrument requires tactical planning:
    - Commonwealth and State governments could introduce a portion of the GST that accrues strictly to the Commonwealth, which the Commonwealth could use as a policy instrument.
    - Government could propose a schedule of GST increases in its tax switching and announce a schedule of tax credits that increase as needed later on.
    - Another approach: announce tax credits in the short term and offset weakening effects on revenues by broadening the tax base now.
- On the medium-term balance anchor:
  - Do not recommend strict pursuit of the medium-term balance anchor at this point; Budget 2016-17 commitment to a +1 percent of GDP Commonwealth government surplus is acknowledged but uncertainty exists on the required magnitude for budget repair and reliance on rebalancing to the external sector may be undesirable. The primary near-term risk is the “new mediocre,” where budget repair would be very costly.

*Italic: Source — IMF staff analysis and recommendations as presented in the referenced chapter section.*

### Appendix I. Overview of the Theoretical Structure of G20MOD

### Appendix I. Overview of the Theoretical Structure of G20MOD

### Model overview
- G20MOD is an annual, multi-region, general equilibrium model of the global economy combining both micro-founded and reduced-form formulations of various economic sectors.
- Structurally, each country/regional block is close to identical, but with potentially different key steady-state ratios and behavioral parameters.
- The model emphasizes: (1) demand side with an overlapping generations (OLG) framework for households and firms; (2) supply side and prices; (3) commodity sector as a global stabilizer; and (4) financial features, monetary policy, and a fiscal block.

### Consumption block and overlapping generations (OLG)
- Consumption uses a discrete-time OLG model as in Blanchard (1985), Weil (1989) and Yaari (1965).
- Utility: constant-elasticity-of-substitution utility function dependent only on household consumption.
- Key non-Ricardian implications:
  - Households treat government bonds as wealth because associated tax liabilities may fall beyond expected lifetimes.
  - Endogenous determination of national savings given the level of government debt.
  - Global real interest rate adjusts to equilibrate global supply of and demand for savings.
- Wealth tracking:
  - Stocks and flows tracked: human wealth (based on labor income) and financial wealth (based on government debt, private business capital stock, and net foreign assets).
  - Financial markets are incomplete; international financial flows tracked as net positions (net foreign assets or net foreign liabilities) denominated in U.S. dollars.
- Liquidity-constrained (LIQ) households:
  - LIQ households have no access to financial markets, do not save, and consume all income each period.
  - Presence of LIQ households amplifies non-Ricardian properties, especially for temporary fiscal shocks.

### Private investment and firm behavior
- Investment follows an updated Tobin's Q model with quadratic real adjustment costs.
- Investment is negatively correlated with real interest rates and cumulates into the private capital stock.
- Capital-to-GDP ratio is inversely related to the cost of capital, which depends on:
  - depreciation, the real corporate interest rate, the company income tax rate, and relative prices.
- Corporate interest rate:
  - A mix of the 1-year and 10-year interest rates, with a risk premium negatively correlated with the output gap to capture a financial accelerator effect as in Bernanke and others (1999).

### Government absorption and public capital
- Government absorption comprises government consumption and infrastructure investment; both are exogenous government choices.
- Government consumption affects the level of aggregate demand.
- Government investment cumulates into a public capital stock (public infrastructure) and a permanent increase in public capital permanently raises economy-wide productivity.

### External sector, trade, and current account
- Net exports respond to long-term determinants, notably the real competitiveness index (RCI), adjusting to achieve the current account balance required to support the desired net foreign asset position.
- Exports and imports are modeled as reduced-form equations:
  - Exports increase with foreign activity and with depreciation in the RCI.
  - Imports increase with domestic activity and with appreciation of the real effective exchange rate (REER).
- RCI vs REER:
  - RCI is export-weighted and accounts for third-country competition effects.
  - Real exchange rate discussions apply qualitatively to both RCI and REER.
- Aggregation and trade shares:
  - The model does not track all bilateral trade flows to limit dimensionality.
  - Time-varying trade shares depend on relative tradable and nontradable productivity within each country, enabling Balassa-Samuelson effects when tradable productivity growth exceeds nontradable productivity growth.
  - Mechanisms ensure global exports and imports sum to zero.
- Current account and net foreign assets are linked to household saving decisions; non-zero current accounts are possible in well-defined OLG steady states, allowing creditor and debtor nation analysis.

### Aggregate supply, labor market, and unemployment
- Potential output:
  - Based on Cobb-Douglas production with trend total factor productivity, steady-state labor force, NAIRU (non-accelerating inflation rate of unemployment), and actual capital stock.
  - Allows computation of output gap to track excess demand/supply; output gap is a key driver of prices.
- Labor sector:
  - Endogenous labor sector with steady-state population growth taken as exogenous.
  - Cyclical variations in participation rate and unemployment rate.
  - Participation rate behavior follows properties from GIMF and GEM: substitution effect based on real wage and income effect based on total wealth.
  - Unemployment rate varies relative to NAIRU according to an Okun's law relationship based on the output gap.

### Prices, wages, and inflation dynamics
- Core price: CPIX (consumer price index excluding food and energy) determined by an inflation Phillips curve.
  - CPIX inflation is sticky and reflects expected exchange-rate paths and the economic cycle (captured by the output gap).
  - Persistent oil price changes can leak into core inflation.
  - Degree of forward-looking behavior in inflation is country specific.
- Relative prices:
  - Mimic national accounts structure; usually weighted averages of a domestic consumer price and an import price.
  - Export prices account for third-country effects; import prices are an import-weighted average of other countries' export prices.
- Wage Phillips curve:
  - Nominal wage growth exhibits stickiness; real wage returns to equilibrium gradually depending on expected economic activity and deviation of labor supply from labor demand.

### Commodities sector
- Three commodities modeled: oil, food and metals.
- Country specifics:
  - In this context, Australia is an oil importer and a notable exporter of metals.
- Demand and supply:
  - Demand driven by global demand and relatively price inelastic in short term due to limited substitutability.
  - Supply also price inelastic in the short term.
- Households consume food and oil explicitly, allowing distinction between headline and core CPI inflation.
- Commodities as macro stabilizers/amplifiers:
  - In global excess aggregate demand, rising commodity prices (due to sluggish supply adjustment) put downward pressure on global aggregate demand.
  - In global excess supply, falling commodity prices soften deterioration.
  - Commodity-originating shocks can be highly disruptive.
  - If a country’s cycle is out of sync with global cycle, commodity prices can amplify country-specific fluctuations.

### Monetary policy and interest rates
- Nominal and real linkage:
  - In the short term, the nominal side links to the real side through monetary policy.
- Interest rate rule:
  - Monetary authorities represented by an interest-rate reaction function based on a 1-year interest rate (referred to as the overnight cash rate for Australia).
  - Standard form: an inflation-forecast-based rule operating under a flexible exchange rate (as in Australia).
- Monetary–fiscal interaction:
  - Monetary policy can work at cross-purposes with fiscal stimulus if inflation target strictly pursued, or in tandem by allowing accommodation at appropriate times.
  - Under constrained monetary policy space and rising probability of hitting the effective lower bound, overly ambitious fiscal consolidation can exacerbate economic costs.
- Term structure and rates:
  - A 10-year interest rate is featured, based on the expectations theory of the term structure plus a term premium.
  - Interest rates on consumption, investment, government debt and net foreign assets are weighted averages of the 1-year and 10-year rates, reflecting differing term structures and allowing a role for the term premium.
- Long-term interest rates affect household and firm decisions.

### Fiscal block, taxation, and budget constraint
- Government sector reproduces simplified fiscal accounts with eight policy instruments:
  - Spending-side: government consumption, infrastructure spending, general lumpsum transfers to all households (e.g., pensions, aged care, unemployment insurance), lumpsum transfers targeted to LIQ households (e.g., welfare, certain pensions).
  - Revenue-side: taxes on consumption (GST), personal income tax (PIT, on both wage and dividend income), company income tax (CIT), and taxes/royalties from mining and production of metals.
- Australia-specific features:
  - More detailed treatment of firm/company income taxation.
  - Firms pay CIT and issue dividends; households notionally pay marginal tax on dividend income as part of PIT.
  - Franked dividends: households are rebated the tax amount already paid by firms as CIT.
  - Implication: lowering CIT reduces franked dividends OLG households receive, raising the dividend-income portion of their PIT liability.
  - Tax-switch from CIT to PIT increases PIT liability on households; the burden falls on OLG households for the dividend component, while LIQ households share only the wage-income portion of PIT increases.
- Budget constraint and long-term deficit target:
  - The budget constraint is met by choosing a long-term deficit target relative to GDP.
  - Deficit = expenditures and interest payments, less revenues.
  - One instrument (general lumpsum transfers by default) is continuously adjusted to ensure the budget constraint holds.
  - The long-term government debt target relative to GDP is derived from the government deficit target and the nominal growth rate:
    - ܾ
      ௧௔௥
      =
      ሺ
      ߨ1൅
      ሻሺ
      ݃1൅
      ሻ
      ሺ
      ߨ1൅
      ሻሺ
      ݃1൅
      ሻ
      ്1
      ݂݁݀݃
      ௧௔௥
  - Variables in the equation as presented:
    - ߨ is inflation,
    - g is the steady-state growth rate,
    - ܾ
      ௧௔௥
      is the long-term debt target,
    - ݂݁݀݃
      ௧௔௥
      is the long-term deficit target.
  - The explicit deficit target pins down long-term government debt, which influences the global interest rate and real exchange rate.
- Fiscal rules:
  - Operational implementation of deficit target is complemented with fiscal rules (explained more fully elsewhere in the paper).

*Source: Appendix I. Overview of the Theoretical Structure of G20MOD (cr1743).*

---


_Source: https://www.imf.org/-/media/files/publications/cr/2017/cr1743.pdf_