Australia: Selected Issues

International Monetary Fund. Asia and Pacific Dept

doi:10.5089/9798400230660.002.A001

Author:

International Monetary Fund. Asia and Pacific Dept

International Monetary Fund. Asia and Pacific Dept
Search for other papers by International Monetary Fund. Asia and Pacific Dept in
Current site
Google Scholar
PubMed

Publication Date:: 01 Feb 2023

eISBN:: 9798400230660

ISBN:: 9798400230660

Language:: English

Keywords:: IMF staff calculation; inflation in Australia; Phillips curve relationship; wage pressure; agreement share; Labor markets; Labor market institutions; Unemployment

Download PDF (896.5 KB)

Australia, like other advanced economies, is witnessing high inflation that has become broad-based, though wage pressures in Australia have so far been contained. Although wages in Australia are less sensitive to labor market tightness compared with other economies, the tight labor market conditions are likely to feed into higher wages in the future, as evidenced by a robust Phillips curve relationship and business surveys pointing to emerging pressures. The RBA needs to remain vigilant, continuing to tighten policy to rebalance demand and supply, and ensuring that inflation expectations and wage pressures remain contained.

Abstract

Inflation and Wage Dynamics in Australia^¹

A. Introduction

1. Like other advanced economies, Australia is witnessing high inflation that has become broad-based, though wage pressures in Australia appear contained so far. Headline and core measures of inflation have accelerated and are well above the target range, with a large share of the CPI basket experiencing price gains of over 3 percent per year. Energy inflation in Australia remains significantly below that of other OECD countries, though core inflation, which excludes volatile energy and food components, shows Australian inflation well within the range of other OECD economies. That said, despite a tight labor market, wage pressures appear subdued so far.

2. This paper analyzes inflation dynamics in Australia, with a special focus on wages.

Inflation: The paper finds inflation dynamics in Australia are similar to what is seen in most advanced economies in recent quarters, with inflation primarily driven by cyclically sensitive components. While external shocks following the Russian invasion of Ukraine and supply and shipping bottlenecks stemming from COVID-19 played a significant role in the initial spike in inflation, a positive output gap amid a historically tight domestic labor market is likely to play an increasingly important role going forward.
Wages: Wage pressures are likely to pick up. Cross-country comparisons suggest that wages in Australia are less sensitive to labor market tightness than in other countries, potentially due to labor market institutions (collective bargaining and length of contracts). Nevertheless, labor market tightness is an important driver of wage growth, and tight labor market conditions are likely to feed into higher wages in the future. Static Phillips curve models predict higher wage growth than currently observed, indicating that continued labor market tightness could add upward pressure to wage dynamics in the period ahead. Evidence from business surveys also point to emerging pressures on labor costs. On the other hand, reopening of borders may mitigate some of the wage pressures.
Policy: Monetary policy needs to continue tightening to rebalance demand and supply, with the pace and extent of tightening depending on the evolution of inflation, inflation expectations, and wage growth. The RBA will need to be especially vigilant to inflation expectations and wage developments given upside risks.

3. The rest of the paper is structured as follows. Section B takes a deeper look at inflation developments in Australia in an international context, including the role of cyclical components. Section C focuses on wage dynamics, comparing recent wage growth to model predictions, explores evidence from sectoral Phillips curves, and uses cross-country data to analyze the role of labor market institutions in shaping the unemployment-wage nexus across countries. Section D concludes.

B. Inflation Developments

4. Inflation in Australia has accelerated and become broad-based. Headline inflation reached 7.3 percent in the third quarter of 2022. Measures of underlying inflation, including trimmed mean and weighted median inflation, have also accelerated and are now well-above the RBA’s target range of 2 to 3 percent. While there is a significant external component to inflation, with tradable inflation increasing to 8.7 percent, non-tradable inflation has also accelerated in recent quarters, reaching 6.5 percent in 2022Q3. Overall, inflation pressures appear broad-based, with about 90 percent of the CPI basket experiencing price gains of over 3 percent per year, and also reflect strong domestic demand and rising input prices, although wage pressures remain subdued (see below).^²

5. Australia’s inflation developments are similar to those in other advanced economies, though with lower energy inflation. Inflation has increased in most advanced economies in recent quarters, with headline inflation in Australia being on the lower end of the spectrum compared to other OECD countries. The spike in global energy prices following the Russian invasion of Ukraine has resulted in a steep rise in energy inflation around the world, especially in Europe. Energy inflation in Australia has also increased to close to 15 percent, though remains significantly lower than in other OECD countries. A comparable measure of core inflation across countries, which excludes volatile energy and food components, shows Australian inflation at close to the OECD median.^³

Cyclical Versus Non-Cyclical Inflation

6. Overall inflation can be broken down into cyclical and non-cyclical inflation to better understand the dynamics. Following Stock and Watson (2019), we can estimate the cyclically sensitive component of inflation based on the association between inflation and economic slack. While there are different measures for economic slack, for this exercise, we measure slack by using the unemployment gap—the difference between the observed unemployment rate and trend unemployment (see Lian and Freitag, 2022).^⁴ Correlations are calculated for a panel of 30 advanced economies for which breakdown of CPI components are available for the 12 one-digit COICOP components, over the period 2000Q1 to 2022Q2.^⁵ While a more detailed breakdown of inflation is available for Australia, using the 12-digit COICOP allows us to compare across countries and also estimate the sensitivity of sectors in a panel setting, avoiding potential endogeneity concerns. But this comes at a cost of having relatively broad and heterogenous components. We find that cyclical sensitivity varies significantly across CPI components and overall, the classification we obtain is similar to that from earlier studies. For the most cyclical components, the correlation between their inflation and unemployment gap is negative for our sample of advanced economies, allowing us to break down overall inflation between cyclical and non-cyclical components. For headline CPI components, housing, transportation, furnishing, hotels and restaurants, food and miscellaneous items are classified as cyclical components. We also construct an alternate measure which excludes food and energy (from transportation and housing) to arrive at a “core” measure of the cyclical component.

7. Our analysis suggests that in Australia, like in other advanced economies, current inflationary pressures are driven by cyclically sensitive components. This does not necessarily mean that inflation in Australia is driven primarily by domestic factors—for example, while furnishing and housing are cyclically sensitive, their prices may still be driven by higher global prices of inputs and supply and shipping bottlenecks. Also, the analysis does not take into account one-off factors such as severe weather and flooding that affected food prices in Australia. However, the increase in cyclical inflation globally does suggest a role for the synchronized pickup in demand after the pandemic, particularly in the advanced economies, in the presence of continuing supply disruptions, including labor. This has resulted in a tight labor market globally and hence an increase in the cyclically sensitive inflation component, i.e., the component correlated with unemployment gap. Although these results do not necessarily imply that inflation is primarily driven by domestic factors, the fact that cyclical components that are historically sensitive to economic slack are driving inflation suggests a role for domestic demand management policies. These results are robust to differences in CPI weights across countries and continue to hold if weight differences are controlled for by using Australia CPI weights when comparing across countries. In addition, stripping out volatile components such as food and energy from the analysis does not change the overall result.^⁶

Changing Importance of External and Domestic Factors

8. The initial inflation spike in Australia was driven by surging international food and energy prices, along with supply chain and shipping disruptions. But pressures from food and commodity prices are easing, and there are signs that capacity is beginning to catch up with the strong demand for goods. Supplier delivery times have returned to levels last observed 18 months ago, and global shipping costs have started to ease, albeit from high levels. Aggregate demand is expected to moderate as central banks across the globe tighten policy, but the risks of further supply disruptions cannot be ruled out given uncertainty surrounding China’s economy.

9. Wages are likely to play an increasing role in determining the persistence of inflation in Australia. While the initial spike in inflation was driven by external factors, domestic demand is likely playing an increasing role. Labor markets in Australia are at historically tight levels, with the employment-to-population ratio and participation rate at record highs. Job vacancies are at very high levels, and underemployment and underutilization have come down as hiring workers has become increasingly challenging. How quickly and persistently these tight labor market conditions feed into wages, and in turn prices, will play a dominant role in determining the future path and persistence of inflation.

C. Wage Dynamics

10. Wage pressures appear to be subdued so far. Wage inflation in Australia, as measured by the year-on-year change in the Wage Price Index (WPI), declined to a low of 1.4 percent during the COVID crisis. Wage inflation has increased to 3.1 percent in 2022Q3, with private sector wages picking up to 3.4 percent. Other metrics of wage inflation based on national accounts data have also seen an uptick but remain moderate by historical standards. Sectoral wage data also shows wage growth remains contained so far, even in sectors such as construction and manufacturing, where business surveys tend to show labor costs and labor shortages as major areas of concern for companies. While wages growth in most sectors is higher than a year ago, overall increases appear contained, though it is possible that they may increase in the future as wages catch up with labor shortages.

11. Wage inflation in Australia has remained low in international comparison. To assess wage pressures in Australia in an international context, and to analyze the drivers of wage inflation, we assemble data on wage inflation for a group of 24 advanced economies. As harmonized wage price indices are not available for all countries, we use various national sources, including the Wage Price Index in Australia, Labor Cost Indices for Europe and New Zealand, and other national sources where available.^⁷ Most advanced economies have higher wage growth than Australia, with median wage inflation in peers at about 4 percent as of 2022Q2. Wage growth is generally higher in countries with tight labor markets, indicating an important role for domestic labor market conditions in explaining wage dynamics.

Wage-Unemployment Relation in Australia and Internationally

12. The wage-unemployment relation varies across countries and within countries over time. In Australia, aggregate wage inflation has been low since about 2015, with the relation between wage growth and unemployment, the wage Phillips curve, seemingly shifting downwards and becoming flatter. Wage growth in the post-COVID period seems significantly lower than what a historical Phillips curve would suggest, though more in line with a lower and flatter Phillips curve. By contrast, in the US and in New Zealand, wage growth in recent quarters seems consistent with longer-term Philips curves.

13. Australia’s pre-pandemic low wage growth can be empirically explained by an augmented wage Phillips curve, though recent wage growth is low compared to model estimates. A standard wage Phillips curve relationship, which incorporates unemployment gap and one-year ahead inflation expectations, suggest that wage growth in Australia has been lower than what the historical empirical relationship suggests.^⁸ However, such an estimate does not take into account persistent underemployment in Australia before the pandemic, where underemployment is defined as part-time employees who want to work more hours but are not able to. In addition, the standard Phillips curve relationship does not control for the secular decline in labor productivity since the mining boom of the early 2010s and shocks to the terms of trade.^⁹ Controlling for these factors in an augmented wage Phillips curve relationship with additional control variables results in a much better fit with the observed dynamics of the WPI before the pandemic (see Annex for regression tables; see also IMF, 2020).^¹⁰ This analysis aims to establish the long-run relationship between economic slack and wage dynamics, as opposed to short-term forecasting, and implies that recent wage growth in Australia is below what economic indicators would imply. This suggests upside risks to wage inflation in the period ahead if the labor market remains tight.^¹¹

Role of Labor Market Tightness in Determining Wage Dynamics: Australia compared to Peers and Evidence from Australian Sectoral Data

14. Given the tightness of labor markets, we further explore the wage-unemployment relation in an international context. We estimate wage Phillips curves using quarterly data for 23 countries from the period 2000 to 2019:^¹²

\begin{matrix} π_{i, t}^{w} = α_{i} + α_{t} + β_{1} u_{i, t}^{g a p} + β_{2} π_{i, t}^{e} + β_{3} g_{i, t}^{L P} + ε_{t, i}^{h} & (1) \end{matrix}

where $π_{i, t}^{w}$ is the y-on-y percent change in wages in country T at time $u_{i, t}^{g a p}$ is the.unemployment gap, computed as the deviation between the unemployment rate and an Hodrick-Prescott (HP) trend; $π_{i, t}^{e}$ is 1-year ahead inflation expectations from the consensus forecasts; and $g_{i, t}^{L P}$ is trend labor productivity growth.^¹³ All regressions include country and time fixed effects to control for unobservable time-invariant country characteristics and global shocks respectively and standard errors are clustered at the country level. We also consider dynamic versions of the equation, which includes 4 lags of the dependent variable.

15. Labor market tightness plays an important role in explaining wage dynamics, though Australian wages are somewhat less sensitive to unemployment compared to peers. Table 1 reports results for the baseline estimates of the wage Phillips curve. As predicted by theory, higher unemployment relative to trend is associated with lower wage growth, while higher inflation expectations are associated with higher wage growth. Lower trend productivity growth is also associated with lower wage growth (column 1). Column 2 repeats the same specification but only for Australia. Unemployment gap remains negative and significant, though the coefficient is smaller than for the cross-country panel, suggesting that Australian wages are less sensitive to labor market tightness than the average among advanced economies. Similarly, the smaller coefficient on inflation expectations suggests a much smaller reaction in Australian wages to inflation expectations than in the overall sample. These results are robust to including lags of the dependent variable as additional controls (column 3 and 4), using q-on-q growth rates of wages as the dependent variable, and using NAIRU instead of a HP trend to compute unemployment gap (Annex Table 2)

^{Table 1.}

Australia: Wage Phillips Curve – Australia Compared to Peers

Source: IMF staff estimates. Notes: Data at the country-quarter level. Reports results for estimates of equation 1. Dependent variable in each regression is growth rate of wages. Columns 1 is for the full sample of 23 countries while column 2 is for Australia only. Columns 3 and 4 repeat column 1 and 2 respectively but include lag values of the dependent variable. Standard errors are clustered at the country level for columns 1 and 3. *, **, and *** indicate significance at the 10, 5 and 1 percent level respectively.

^{Table 1.}

Australia: Wage Phillips Curve – Australia Compared to Peers

	(1)	(2)	(3)	(4)
	Baseline	Baseline	With lags	With lags
VARIABLES	Full Sample	Australia	Full Sample	Australia
Unemployment gap	-0.944*** (0.284)	-0.554*** (0.207)	-0.272** (0.0973)	-0.160*** (0.0505)
Expected inflation	1 244*** (0.270)	0.309 (0.191)	0.282*** (0.0956)	0.0448 (0.0649)
Trend labor productivty growth	0.890** (0.375)	0.602 (0.374)	0.324* (0.158)	0.0472 (0.119)
Lag 1, dependent variable			0.694*** (0.0813)	1.207*** (0.0940)
Lag 2, dependent variable			0.185*** (0.0397)	-0.261 (0.159)
Lag 3, dependent variable			0.0986** (0.0367)	-0.122 (0.184)
Lag 4, dependent variable			-0.337*** (0.0254)	0.105 (0.102)
Constant	-0.505 (0.807)	1.602*** (0.597)	0.119 (0.252)	0.0336 (0.1000)
Observations	1,660	80	1,604	80
R-squared	0.703	0.300	0.885	0.959

Source: IMF staff estimates. Notes: Data at the country-quarter level. Reports results for estimates of equation 1. Dependent variable in each regression is growth rate of wages. Columns 1 is for the full sample of 23 countries while column 2 is for Australia only. Columns 3 and 4 repeat column 1 and 2 respectively but include lag values of the dependent variable. Standard errors are clustered at the country level for columns 1 and 3. *, **, and *** indicate significance at the 10, 5 and 1 percent level respectively.

^{Table 1.}

Australia: Wage Phillips Curve – Australia Compared to Peers

	(1)	(2)	(3)	(4)
	Baseline	Baseline	With lags	With lags
VARIABLES	Full Sample	Australia	Full Sample	Australia
Unemployment gap	-0.944*** (0.284)	-0.554*** (0.207)	-0.272** (0.0973)	-0.160*** (0.0505)
Expected inflation	1 244*** (0.270)	0.309 (0.191)	0.282*** (0.0956)	0.0448 (0.0649)
Trend labor productivty growth	0.890** (0.375)	0.602 (0.374)	0.324* (0.158)	0.0472 (0.119)
Lag 1, dependent variable			0.694*** (0.0813)	1.207*** (0.0940)
Lag 2, dependent variable			0.185*** (0.0397)	-0.261 (0.159)
Lag 3, dependent variable			0.0986** (0.0367)	-0.122 (0.184)
Lag 4, dependent variable			-0.337*** (0.0254)	0.105 (0.102)
Constant	-0.505 (0.807)	1.602*** (0.597)	0.119 (0.252)	0.0336 (0.1000)
Observations	1,660	80	1,604	80
R-squared	0.703	0.300	0.885	0.959

Source: IMF staff estimates. Notes: Data at the country-quarter level. Reports results for estimates of equation 1. Dependent variable in each regression is growth rate of wages. Columns 1 is for the full sample of 23 countries while column 2 is for Australia only. Columns 3 and 4 repeat column 1 and 2 respectively but include lag values of the dependent variable. Standard errors are clustered at the country level for columns 1 and 3. *, **, and *** indicate significance at the 10, 5 and 1 percent level respectively.

16. Sectoral data for Australia also suggests a significant relationship between wage growth and labor market tightness. We proxy unemployment at the sector level using advertised job vacancies in a particular sector, normalized by employment in that sector. We then calculate a measure of the vacancy gap using an HP filter on the vacancy rate by sector. At the aggregate level of all industries, we observe a positive correlation between wage growth and the vacancy rate, except in the aftermath of the last mining boom. Using the latest available data (2022Q3), we again find that sectors with a higher vacancy gap tend to have higher wage growth.

17. Formal analysis using a sectoral Phillips curve confirms the role of labor market tightness in explaining wage growth. Empirical estimates of a wage Phillips curve relationship using industry level data, using a panel setting, confirms that a higher vacancy gap is associated with faster wage growth. For our baseline specification (Table 2, columns 1–4), we restrict the sample to the pre-COVID period (2000Q1 to 2019Q4) to avoid contaminating the regressions with pandemic related noise. In the simplest specification (column 1) we find that higher vacancy rates are associated with higher wage growth. This result continues to hold when we add 1-year ahead inflation expectations (column 2, instead of time fixed effects), labor productivity growth (column 3, though labor productivity is not statistically significant), and lagged wage growth (column 4). The results also hold if we include data from the pandemic years (column 5 and 6). Results in columns 1–6 are based on private sector wage growth, which is likely to be more sensitive to labor market conditions. However, we get similar, but weaker, results when we use total wages (which includes the public sector) in columns 7 and 8. Finally, the results are robust to using q/q growth in wages instead of y/y growth.

^{Table 2.}

Australia: Wage Phillips Curve – Sector Level Analysis

Source: IMF staff estimates. Notes: Data at the sector-quarter level. Dependent variable in each regression is y/y growth rate of wages, except columns 9 and 10 where q/q rates are used for robustness. Columns 1 to 4 use data for private sector wages and restricts the sample to the per-COVID period (2000Q1 to 2019Q4). Columns 5 and 6 report results for the full sample until 2022Q2 (including COVID period). Columns 7 and 8 use total WPI (private and public sector). All regressions include country and time fixed effects, except column 2 which excludes time fixed effect and instead uses 1-year ahead inflation expectations for the Australian economy (does not vary by sectors). Standard errors are clustered at the country level. *, **, and *** indicate significance at the 10, 5 and 1 percent level respectively.

^{Table 2.}

Australia: Wage Phillips Curve – Sector Level Analysis

VARIABLES	Private WPI	Private WPI	Private WPI	Private WPI	Private WPI	Private WPI	Total WPI	Total WPI	Q-on-Q	Q-on-Q
Vacancy gap	0.195* (0.101)	0.381** (0.177)	0.168** (0.0697)	0.0783** (0.0356)	0.159* (0.0852)	0.0809** (0.0319)	0.180 (0.108)	0.0907* (0.0493)	0.0874** (0.0307)	0.0926*** (0.0299)
Inf Exp, 1 year ahead		0.00571*** (0.000828)
Trend labor productivity growth			-0.112 (0.0738)
Lag 1, dependent variable				0.765*** (0.0359)		0.751*** (0.0414)		0.632*** (0.0705)		-0.177*** (0.0331)
Lag 2, dependent variable				-0.0199 (0.0310)		-0.0104 (0.0306)		0.000205 (0.0289)		-0.171*** (0.0397)
Lag 3, dependent variable				0.0674** (0.0240)		0.0650** (0.0229)		0.00775 (0.0275)		-0.114*** (0.0358)
Lag 4, dependent variable				-0.143*** (0.0431)		-0.152*** (0.0409)		-0.0125 (0.0348)		0.427*** (0.0575)
Constant	0.0303*** (1.88e-05)	0.0154*** (0.00216)	0.0319*** (0.00111)	0.01000*** (0.00167)	0.0288*** (2.42e-07)	0.0100*** (0.00164)	0.0306*** (2.00e-05)	0.0116*** (0.00193)	0.00757*** (5.71 e-06)	0.00782*** (0.000895)
Observations	1,323	1,323	1,323	1,323	1,593	1,533	1,323	1,323	1,323	1,323
R-squared	0.703	0.278	0.713	0.869	0.719	0.877	0.722	0.841	0.416	0.589

^{Table 2.}

Australia: Wage Phillips Curve – Sector Level Analysis

VARIABLES	Private WPI	Private WPI	Private WPI	Private WPI	Private WPI	Private WPI	Total WPI	Total WPI	Q-on-Q	Q-on-Q
Vacancy gap	0.195* (0.101)	0.381** (0.177)	0.168** (0.0697)	0.0783** (0.0356)	0.159* (0.0852)	0.0809** (0.0319)	0.180 (0.108)	0.0907* (0.0493)	0.0874** (0.0307)	0.0926*** (0.0299)
Inf Exp, 1 year ahead		0.00571*** (0.000828)
Trend labor productivity growth			-0.112 (0.0738)
Lag 1, dependent variable				0.765*** (0.0359)		0.751*** (0.0414)		0.632*** (0.0705)		-0.177*** (0.0331)
Lag 2, dependent variable				-0.0199 (0.0310)		-0.0104 (0.0306)		0.000205 (0.0289)		-0.171*** (0.0397)
Lag 3, dependent variable				0.0674** (0.0240)		0.0650** (0.0229)		0.00775 (0.0275)		-0.114*** (0.0358)
Lag 4, dependent variable				-0.143*** (0.0431)		-0.152*** (0.0409)		-0.0125 (0.0348)		0.427*** (0.0575)
Constant	0.0303*** (1.88e-05)	0.0154*** (0.00216)	0.0319*** (0.00111)	0.01000*** (0.00167)	0.0288*** (2.42e-07)	0.0100*** (0.00164)	0.0306*** (2.00e-05)	0.0116*** (0.00193)	0.00757*** (5.71 e-06)	0.00782*** (0.000895)
Observations	1,323	1,323	1,323	1,323	1,593	1,533	1,323	1,323	1,323	1,323
R-squared	0.703	0.278	0.713	0.869	0.719	0.877	0.722	0.841	0.416	0.589

Role of Labor Market Institutions in Shaping the Unemployment-Wage Relation: Lessons from Cross-Country Analysis

18. Labor market institutions can have varying effects on the cyclical response of wages to economic conditions. The literature on how labor market institutions impact the responsiveness of wages to shocks remains inconclusive. The “corporatism” view of Bruno and Sachs (1985) and Calmfors and Drifill (1988) suggests that higher collective bargaining coverage could moderate wage demands in response to domestic and external shocks, as unions internalize the macroeconomic implications of higher wage increases, especially in response to a supply shock. A more recent literature argues that with higher bargaining power, workers are more likely to negotiate larger wage increases in tight labor market conditions, with the recent decline in union density and collective bargaining cited as one of the reasons for a flattening Phillips curve (BIS 2022, Lombardi and other 2020, and Ratner and Sim 2022, Suthaharan and Bleakley 2022). The theoretical literature is clearer on the impact of contract lengths on wage responsiveness: in countries where employment contracts are set for longer periods of time, wages are more sticky, reacting less to domestic and external conditions.

19. To assess the role of labor market institutions in wage dynamics, we modify the standard Phillips curve to allow for heterogeneity based on labor market institutions. In particular, we introduce interaction terms to the standard Phillips curve described in equation 1, estimating:

\begin{matrix} π_{i, t}^{w} = α_{i} + α_{t} + β^{1} u_{i, t}^{g a p} + γ (X_{i} * u_{i, t}^{g a p}) + β^{2} π_{i, t}^{e} + β^{3} g_{i, t}^{L P} + ε_{t, i}^{h} & (2) \end{matrix}

where X_i is a variable which captures labor market institutions in country i. We consider 3 labor market characteristics from OECD: (i) average share of workers under collective bargaining agreements, (ii) average length of collective bargaining agreements; and (iii) trade union density.^¹⁴

20. Among the different labor market characteristics, longer contract length is most robustly associated with wages being less sensitive to unemployment and inflation expectations. Table 3 reports results for equation 2, where the unemployment gap is interacted with various measures of labor market institutions. The coefficient on the interaction term of collective bargaining with the unemployment gap is positive, which combined with the negative coefficient on unemployment gap itself, indicates that wages are less sensitive to movements in unemployment in countries with higher levels of collective bargaining. The same is true for length of contracts (column 2), with countries with longer contracts generally having less sensitive wages, while results for trade union density (column 3) are somewhat weaker. As the various measures of labor market institutions are correlated, in column 4 we include all three interactions together in the same regression. Consistent with theory, we find that length of contract is most robustly associated with wages being less sensitive to unemployment.^¹⁵ Quantitatively, about a 9-month longer average contract length reduces the sensitivity of wages to unemployment by half. Table 4 also reports results for equation 2, but interacting the institutional variables with inflation expectations. The results again point to length of contracts being most robustly associated with less sensitivity of wages to inflation developments. Therefore, length of contracts is likely the most salient feature of labor market institutions which matter for the cyclical response of wages to economic conditions.^¹⁶

^{Table 3.}

Australia: Wage Phillips Curve – Unemployment Interacted with Labor Market Institutions

Source: IMF staff estimates. Notes: Data at the country-quarter level. Reports results for estimates of equation 2. Dependent variable in each regression is year-on-year wage growth. Columns 1–4 include lag dependent variables as controls, while columns 5–8 do not. Column 1 and 5 interact collective bargaining share with unemployment gap, columns 2 and 6 interact average length of collective agreements with the unemployment gap, columns 3 and 7 interact trade union density with unemployment gap, while columns 4 and 8 include all three interactions in the same regression. All regressions include country and time fixed effects. Standard errors are clustered at the country level. *, **, and *** indicate significance at the 10, 5 and 1 percent level respectively.

^{Table 3.}

Australia: Wage Phillips Curve – Unemployment Interacted with Labor Market Institutions

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
	Lags of dependent variable included				Lags of dependent variable not included
VARIABLES	Collective bargaining	Length of agreement	Trade union density	All together	Collective bargaining	Length of agreement	Trade union density	All together
Unemployment gap	-0.515*** (0.160)	-0.715*** (0.227)	-0.364** (0.140)	-0.720*** (0.232)	-1.608*** (0.328)	-2.030*** (0.395)	-1.274*** (0.388)	-2.027*** (0.394)
Coll Bargaining*U gap	0.00485** (0.00177)			0.00156 (0.00191)	0.0155*** (0.00436)			0.00657 (0.00552)
Avg length of contract*U gap		0.0189*** (0.00641)		0.0148* (0.00785)		0.0535*** (0.0135)		0.0349* (0.0189)
Trade union density*U gap			0.00522 (0.00311)	0.00108 (0.00254)			0.0205* (0.0102)	0.00516 (0.00815)
Expected inflation	0.292*** (0.0849)	0.322*** (0.0860)	0.286*** (0.0932)	0.323*** (0.0850)	1.110*** (0.201)	1.135*** (0.189)	1.206*** (0.248)	1.112*** (0.188)
Trend labor productivty growth	0.324* (0.160)	0.321* (0.161)	0.325* (0.160)	0.324* (0.164)	0.804** (0.322)	0.754** (0.316)	0.867** (0.370)	0.757** (0.316)
Lag 1, dependent variable	0.672*** (0.0780)	0.655*** (0.0788)	0.688*** (0.0799)	0.653*** (0.0789)
Lag 2, dependent variable	0.177*** (0.0395)	0.172*** (0.0382)	0.183*** (0.0403)	0.171*** (0.0385)
Lag 3, dependent variable	0.0947** (0.0365)	0.0936** (0.0357)	0.0974** (0.0366)	0.0930** (0.0356)
Lag 4, dependent variable	-0.342*** (0.0272)	-0.343*** (0.0277)	-0.338*** (0.0260)	-0.344*** (0.0284)
Constant	0.225 (0.241)	0.276 (0.247)	0.143 (0.250)	0.283 (0.250)	-0.112 (0.594)	0.00298 (0.546)	-0.387 (0.741)	0.0407 (0.543)
Observations	1,604	1,444	1,604	1,444	1,660	1,500	1,660	1,500
R-squared	0.887	0.889	0.885	0.889	0.729	0.744	0.710	0.747

^{Table 3.}

Australia: Wage Phillips Curve – Unemployment Interacted with Labor Market Institutions

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
	Lags of dependent variable included				Lags of dependent variable not included
VARIABLES	Collective bargaining	Length of agreement	Trade union density	All together	Collective bargaining	Length of agreement	Trade union density	All together
Unemployment gap	-0.515*** (0.160)	-0.715*** (0.227)	-0.364** (0.140)	-0.720*** (0.232)	-1.608*** (0.328)	-2.030*** (0.395)	-1.274*** (0.388)	-2.027*** (0.394)
Coll Bargaining*U gap	0.00485** (0.00177)			0.00156 (0.00191)	0.0155*** (0.00436)			0.00657 (0.00552)
Avg length of contract*U gap		0.0189*** (0.00641)		0.0148* (0.00785)		0.0535*** (0.0135)		0.0349* (0.0189)
Trade union density*U gap			0.00522 (0.00311)	0.00108 (0.00254)			0.0205* (0.0102)	0.00516 (0.00815)
Expected inflation	0.292*** (0.0849)	0.322*** (0.0860)	0.286*** (0.0932)	0.323*** (0.0850)	1.110*** (0.201)	1.135*** (0.189)	1.206*** (0.248)	1.112*** (0.188)
Trend labor productivty growth	0.324* (0.160)	0.321* (0.161)	0.325* (0.160)	0.324* (0.164)	0.804** (0.322)	0.754** (0.316)	0.867** (0.370)	0.757** (0.316)
Lag 1, dependent variable	0.672*** (0.0780)	0.655*** (0.0788)	0.688*** (0.0799)	0.653*** (0.0789)
Lag 2, dependent variable	0.177*** (0.0395)	0.172*** (0.0382)	0.183*** (0.0403)	0.171*** (0.0385)
Lag 3, dependent variable	0.0947** (0.0365)	0.0936** (0.0357)	0.0974** (0.0366)	0.0930** (0.0356)
Lag 4, dependent variable	-0.342*** (0.0272)	-0.343*** (0.0277)	-0.338*** (0.0260)	-0.344*** (0.0284)
Constant	0.225 (0.241)	0.276 (0.247)	0.143 (0.250)	0.283 (0.250)	-0.112 (0.594)	0.00298 (0.546)	-0.387 (0.741)	0.0407 (0.543)
Observations	1,604	1,444	1,604	1,444	1,660	1,500	1,660	1,500
R-squared	0.887	0.889	0.885	0.889	0.729	0.744	0.710	0.747

^{Table 4.}

Australia: Wage Phillips Curve – Inflation Expectations Interacted with Labor Market Institutions

Source: IMF staff estimates. Notes: Data at the country-quarter level. Reports results for estimates of equation 2. Dependent variable in each regression is year-on-year wage growth. Columns 1–4 include lag dependent variables as controls, while columns 5–8 do not. Columns 1 and 5 interact collective bargaining share with expected inflation, columns 2 and 6 interact average length of agreements with expected inflation, columns 3 and 7 interact trade union density with expected inflation, while columns 4 and 8 include all three interactions in the same regression. All regressions include country and time fixed effects. Standard errors are clustered at the country level. *, **, and *** indicate significance at the 10, 5 and 1 percent level respectively.

^{Table 4.}

Australia: Wage Phillips Curve – Inflation Expectations Interacted with Labor Market Institutions

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
	Lags of dependent variable included				Lags of dependent variable not included
VARIABLES	Collective bargaining	Length of agreement	Trade union density	All together	Collective bargaining	Length of agreement	Trade union density	All together
Unemployment gap	-0.286*** (0.0976)	-0.312*** (0.107)	-0.272** (0.0979)	-0.319*** (0.108)	-0.955*** (0.260)	-1.009*** (0.270)	-0.941*** (0.278)	-1.007*** (0.259)
Expected inflation	0.408** (0.159)	0.599*** (0.207)	0.301** (0.123)	0.565** (0.200)	1.759*** (0.376)	2.149*** (0.493)	1.529*** (0.400)	2.122*** (0.464)
Coll Bargaining*Inf exp	-0.00345 (0.00224)			-0.00327* (0.00182)	-0.0156** (0.00632)			-0.0145** (0.00512)
Avg length of contract*Inf exp		-0.0195** (0.00886)		-0.0182** (0.00805)		-0.0632** (0.0228)		-0.0436** (0.0178)
Trade union density*inf exp			-0.00107 (0.00245)	0.00736* (0.00378)			-0.0165 (0.0117)	0.0132 (0.00859)
Lab prod growth	0.316* (0.152)	0.263 (0.157)	0.320* (0.161)	0.288* (0.150)	0.819** (0.323)	0.610* (0.347)	0.821** (0.375)	0.690** (0.318)
Lag 1, dependent variable	0.687*** (0.0823)	0.678*** (0.0817)	0.694*** (0.0812)	0.676*** (0.0828)
Lag 2, dependent variable	0.181*** (0.0389)	0.179*** (0.0397)	0.185*** (0.0397)	0.178*** (0.0392)
Lag 3, dependent variable	0.0961** (0.0369)	0.0956** (0.0356)	0.0983** (0.0368)	0.0951** (0.0358)
Lag 4, dependent variable	-0.335*** (0.0255)	-0.338*** (0.0255)	-0.337*** (0.0254)	-0.338*** (0.0257)
Constant	0.270 (0.242)	0.506* (0.265)	0.141 (0.248)	0.461* (0.261)	0.190 (0.598)	0.861 (0.609)	-0.161 (0.678)	0.824 (0.606)
Observations	1,604	1,444	1,604	1,444	1,660	1,500	1,660	1,500
R-squared	0.885	0.887	0.885	0.887	0.713	0.722	0.705	0.727

Source: IMF staff estimates. Notes: Data at the country-quarter level. Reports results for estimates of equation 2. Dependent variable in each regression is year-on-year wage growth. Columns 1–4 include lag dependent variables as controls, while columns 5–8 do not. Columns 1 and 5 interact collective bargaining share with expected inflation, columns 2 and 6 interact average length of agreements with expected inflation, columns 3 and 7 interact trade union density with expected inflation, while columns 4 and 8 include all three interactions in the same regression. All regressions include country and time fixed effects. Standard errors are clustered at the country level. *, **, and *** indicate significance at the 10, 5 and 1 percent level respectively.

^{Table 4.}

Australia: Wage Phillips Curve – Inflation Expectations Interacted with Labor Market Institutions

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
	Lags of dependent variable included				Lags of dependent variable not included
VARIABLES	Collective bargaining	Length of agreement	Trade union density	All together	Collective bargaining	Length of agreement	Trade union density	All together
Unemployment gap	-0.286*** (0.0976)	-0.312*** (0.107)	-0.272** (0.0979)	-0.319*** (0.108)	-0.955*** (0.260)	-1.009*** (0.270)	-0.941*** (0.278)	-1.007*** (0.259)
Expected inflation	0.408** (0.159)	0.599*** (0.207)	0.301** (0.123)	0.565** (0.200)	1.759*** (0.376)	2.149*** (0.493)	1.529*** (0.400)	2.122*** (0.464)
Coll Bargaining*Inf exp	-0.00345 (0.00224)			-0.00327* (0.00182)	-0.0156** (0.00632)			-0.0145** (0.00512)
Avg length of contract*Inf exp		-0.0195** (0.00886)		-0.0182** (0.00805)		-0.0632** (0.0228)		-0.0436** (0.0178)
Trade union density*inf exp			-0.00107 (0.00245)	0.00736* (0.00378)			-0.0165 (0.0117)	0.0132 (0.00859)
Lab prod growth	0.316* (0.152)	0.263 (0.157)	0.320* (0.161)	0.288* (0.150)	0.819** (0.323)	0.610* (0.347)	0.821** (0.375)	0.690** (0.318)
Lag 1, dependent variable	0.687*** (0.0823)	0.678*** (0.0817)	0.694*** (0.0812)	0.676*** (0.0828)
Lag 2, dependent variable	0.181*** (0.0389)	0.179*** (0.0397)	0.185*** (0.0397)	0.178*** (0.0392)
Lag 3, dependent variable	0.0961** (0.0369)	0.0956** (0.0356)	0.0983** (0.0368)	0.0951** (0.0358)
Lag 4, dependent variable	-0.335*** (0.0255)	-0.338*** (0.0255)	-0.337*** (0.0254)	-0.338*** (0.0257)
Constant	0.270 (0.242)	0.506* (0.265)	0.141 (0.248)	0.461* (0.261)	0.190 (0.598)	0.861 (0.609)	-0.161 (0.678)	0.824 (0.606)
Observations	1,604	1,444	1,604	1,444	1,660	1,500	1,660	1,500
R-squared	0.885	0.887	0.885	0.887	0.713	0.722	0.705	0.727

Source: IMF staff estimates. Notes: Data at the country-quarter level. Reports results for estimates of equation 2. Dependent variable in each regression is year-on-year wage growth. Columns 1–4 include lag dependent variables as controls, while columns 5–8 do not. Columns 1 and 5 interact collective bargaining share with expected inflation, columns 2 and 6 interact average length of agreements with expected inflation, columns 3 and 7 interact trade union density with expected inflation, while columns 4 and 8 include all three interactions in the same regression. All regressions include country and time fixed effects. Standard errors are clustered at the country level. *, **, and *** indicate significance at the 10, 5 and 1 percent level respectively.

21. Prevalence of enterprise wage agreements in Australia with long contract lengths may have moderated the increase in wage growth. As of 2022Q2, about 1.74 million employees (13 percent of employees) were covered under current federally registered enterprise agreements,^¹⁷ with over 35 percent of employees (as of May 2021) covered when considering both federal and state registered enterprise agreements. This figure does not include awards, which accounted for 23 percent of employees in 2021. Therefore, while lower than in the past, collective wage agreements continue to cover a substantial portion of the workforce, with significant heterogeneity across sectors. This, coupled with the fact that the average duration of wage agreements has increased to around 3 years in 2022, in part explains why wages in Australia have remained contained.^¹⁸ By design, wage increases under agreements with long durations are slow moving, and this avoided a sharper drop in wages during the pandemic, while in the current conjuncture, it has also resulted in a slower upward response of wages to labor market tightness. This is also borne out by wage increases granted in newly approved enterprise bargaining agreements, which have also remained subdued.

D. Conclusion and Policy Discussion

22. Inflation in Australia has risen and has become broad-based, with core inflation within the range of other advanced economies. This is despite lower pressure from energy in Australia compared with other advanced economies. Cyclical inflation in Australia has also increased in line with what has been observed in other economies. The initial spike in inflation was more external, driven by a spike in international food and energy prices due to the war in Ukraine and supply chain and shipping disruptions, but these pressures are now easing. Commodity prices have come off their highs earlier in the year, and supply-side indicators such as global shipping costs, supplier delivery times and inventory-to-sales ratios are all easing, although second-round effects via higher utility prices (given the increase in wholesale costs) are expected to continue to exert pressures. But increasingly over time, these non-labor-cost external pressures are being replaced by domestic factors given an output gap that is expected to remain positive over the next couple of years and historically tight labor market with unemployment rates near a 50-year low.

23. Although wage growth has remain subdued, given robust evidence of the Phillips curve relationship in Australia, it is expected to pick up. Institutional features such as the prevalence of wage agreements with long contract lengths may have helped keep wages in check. Recent developments, including the Fair Work Commission (FWC) decision to increase award and national minimum wages by 4.6 and 5.2 percent, respectively, are likely to impact wage growth through the indirect impact on wages growth in other pay-setting methods, even though the direct impact is small. Average wage growth in enterprise agreements is expected to lag other pay setting methods, reflecting the multiyear duration of the outstanding stock of enterprise agreements and the low wage cap policies still in place across most state government employers. However, if inflation expectations increase and wage agreements lock in high wage growth, this could add to the persistence of inflation in the future. Survey evidence points to this risk, but the reopening of the border could, over time, help to alleviate laboadjustor'sses, easing pressures.

24. Inflation is expected to remain above target through 2024, necessitating continued monetary policy tightening to rebalance demand and supply. Broad-based inflationary pressures are expected to gradually decline, with inflation reaching the inflation target band by the end of 2024, subject to some uncertainty. While wages have been slow to pick up, partly due to institutional features, the RBA needs to remain vigilant and ensure that supply and demand rebalance quickly so that current, high inflation does not lead to dis-anchoring of inflation expectations and higher wage demands, making inflation even more persistent. The RBA should communicate its policy intentions clearly and stand ready to recalibrate its forward guidance in response to evolving economic circumstances.

References

Bank of International Settlement. (2022). Annual Economic Report.
- Search Google Scholar
- Export Citation
Bruno, Michael and Jeffrey Sachs. (1985). Economics of Worldwide Stagflation. Harvard University.
- Search Google Scholar
- Export Citation
Calmfors Lars and John Driffill. (1988). Bargaining structure, corporatism, and macroeconomic performance. Economic Policy 3(6):13–62.
- Search Google Scholar
- Export Citation
IMF (2020). Australia Article IV Staff Report.
- Search Google Scholar
- Export Citation
IMF. (2022). Wage Dynamics Post–COVID-19 and Wage-Price Spiral Risks. World Economic Outlook, Chapter 2.
- Search Google Scholar
- Export Citation
Lian, W., & Freitag, A. (2022). Inflation Dynamics in Advanced Economies: A Decomposition into Cyclical and Non-Cyclical Factors. IMF Working Papers, 2022(091). Bishop and Chan (2019)
- Search Google Scholar
- Export Citation
Lombardi, M, M Riggi and E Viviano. (2020). “Bargaining power and the Phillips curve: a micro-macro analysis”, BIS Working Papers, no 903, November.
- Search Google Scholar
- Export Citation
Ratner, D and J Sim. (2022). Who killed the Phillips curve? A murder mystery, Finance and Economics Discussion Series, no 2022–028, Board of Governors of the Federal Reserve System.
- Search Google Scholar
- Export Citation
Suthaharan, N., & Bleakley, J. (2022). Wage-price Dynamics in a High-inflation Environment: The International Evidence| Bulletin–September 2022.
- Search Google Scholar
- Export Citation
Stock, J. H., & Watson, M. W. (2020). Slack and cyclically sensitive inflation. Journal of Money, Credit and Banking, 52(S2), 393–428.
- Search Google Scholar
- Export Citation

Annex I. Additional Results

Annex Table 1 shows the regression tables for the baseline and augmented Phillips curves for both the WPI and different AENA measures of wage growth. The fit is less tight for the AENA measure, particularly for the augmented model, given higher volatility of the series. In addition, for the AENA augmented models, expected inflation enters with a wrong sign, even though it is statistically insignificant.

^{Annex Table 1.}

Australia: Wage Phillips Curves

Source: IMF staff estimates. Notes: Data at the quarterly frequency with robust standard errors. *, **, and *** indicate significance at the 10, 5 and 1 percent level respectively.

^{Annex Table 1.}

Australia: Wage Phillips Curves

	(1)	(2)	(3)	(4)	(5)	(6)
VARIABLES	Standard	Augmented	AENA Average Earnings per Employee	Augmented AENA Average Earnings per Employee	AENA Non-Farm Average Earnings per Employee	Augmented AENA Non-Farm Average Earnings per Employee
Unemployment gap	-0.418** (0.193)	-0.238 (0.249)	-1.510*** (0.499)	-0.628 (0.632)	-1.444*** (0.510)	-0.609 (0.639)
Expected inflation	0.560*** (0.207)	0.0587 (0.142)	1.009** (0.418)	-0.538 (0.567)	1.020** (0.417)	-0.565 (0.558)
Trend labor productivity growth		0.252 (0.385)		0.749 (1.064)		0.907 (1.052)
Underemployment		-0.681*** (0.161)		-1.058*** (0.353)		-1.014*** (0.365)
Change in terms of trade		0.114 (0.421)		5.941** (2.304)		6.061** (2.371)
Constant	1.695*** (0.542)	7.800*** (1.665)	0.640 (1.080)	11.57*** (3.311)	0.604 (1.073)	11.17*** (3.418)
Observations	80	76	80	80	80	80
R-squared	0.283	0.827	0.283	0.529	0.271	0.517

Source: IMF staff estimates. Notes: Data at the quarterly frequency with robust standard errors. *, **, and *** indicate significance at the 10, 5 and 1 percent level respectively.

^{Annex Table 1.}

Australia: Wage Phillips Curves

	(1)	(2)	(3)	(4)	(5)	(6)
VARIABLES	Standard	Augmented	AENA Average Earnings per Employee	Augmented AENA Average Earnings per Employee	AENA Non-Farm Average Earnings per Employee	Augmented AENA Non-Farm Average Earnings per Employee
Unemployment gap	-0.418** (0.193)	-0.238 (0.249)	-1.510*** (0.499)	-0.628 (0.632)	-1.444*** (0.510)	-0.609 (0.639)
Expected inflation	0.560*** (0.207)	0.0587 (0.142)	1.009** (0.418)	-0.538 (0.567)	1.020** (0.417)	-0.565 (0.558)
Trend labor productivity growth		0.252 (0.385)		0.749 (1.064)		0.907 (1.052)
Underemployment		-0.681*** (0.161)		-1.058*** (0.353)		-1.014*** (0.365)
Change in terms of trade		0.114 (0.421)		5.941** (2.304)		6.061** (2.371)
Constant	1.695*** (0.542)	7.800*** (1.665)	0.640 (1.080)	11.57*** (3.311)	0.604 (1.073)	11.17*** (3.418)
Observations	80	76	80	80	80	80
R-squared	0.283	0.827	0.283	0.529	0.271	0.517

Source: IMF staff estimates. Notes: Data at the quarterly frequency with robust standard errors. *, **, and *** indicate significance at the 10, 5 and 1 percent level respectively.

Annex Table 2 below show robustness results for Equation 1. Columns 1 and 2 repeat the baseline specifications from Table 1, columns 3 and 4 use data on NAIRU from the World Economic Outlook Database to construct unemployment gap instead of an HP trend, while columns 5 and 6 use quarter-on-quarter growth rate of wages instead of year-on-year.

^{Annex Table 2.}

Australia: Robustness for Wage Phillips Curve – Australia Compared to Peers

Source: IMF staff estimates. Notes: Data at the country-quarter level. Reports results for estimates of equation 1. Dependent variable in each regression is growth rate of wages. Columns 1 and 2 repeat the baseline from Table 1 in the main text, with column 1 being for the full sample of 23 countries while column 2 is for Australia only. Columns 3 and 4 repeat column 1 and 2 respectively but use NAIRU from the World Economic Outlook Database to construct unemployment gap instead of an HP trend, while columns 5 and 6 use quarter-on-quarter growth rate of wages instead of year-on-year. Standard errors are clustered at the country level for columns 1 and 3. *, **, and *** indicate significance at the 10, 5 and 1 percent level respectively.

^{Annex Table 2.}

Australia: Robustness for Wage Phillips Curve – Australia Compared to Peers

	(1)	(2)	(3)	(4)	(5)	(6)
	Baseline	Baseline	NAIRU ugap	NAIRU ugap	Q-on-Q	Q-on-Q
VARIABLES	Full Sample	Australia	Full Sample	Australia	Full Sample	Australia
Unemployment gap	-0.272** (0.0973)	-0.160*** (0.0505)	-0.312** (0.115)	-0.170*** (0.0340)	-0.194** (0.0719)	-0.0976** (0.0482)
Expected inflation	0.282*** (0.0956)	0.0448 (0.0649)	0.142 (0.0954)	0.0674 (0.0564)	0.129* (0.0709)	0.0434 (0.0448)
Trend labor productivty growth	0.324* (0.158)	0.0472 (0.119)	0.643*** (0.200)	-0.0845 (0.0937)	0.221** (0.0991)	0.0369 (0.0836)
Lag 1, dependent variable	0.694*** (0.0813)	1.207*** (0.0940)	0.586*** (0.0806)	1.137*** (0.0929)	-0.167 (0.103)	0.360*** (0.129)
Lag 2, dependent variable	0.185*** (0.0397)	-0.261 (0.159)	0.198*** (0.0357)	-0.235 (0.159)	0.0980 (0.0635)	0.0826 (0.156)
Lag 3, dependent variable	0.0986** (0.0367)	-0.122 (0.184)	0.117*** (0.0350)	-0.111 (0.179)	0.0298 (0.0673)	0.184* (0.101)
Lag 4, dependent variable	-0.337*** (0.0254)	0.105 (0.102)	-0.340*** (0.0406)	0.0874 (0.103)	0.158 (0.142)	0.125 (0.149)
Constant	0.119 (0.252)	0.0336 (0.1000)	0.319 (0.188)	0.324*** (0.120)	0.114 (0.173)	0.0261 (0.0727)
Observations	1,604	80	1,388	80	1,646	80
R-squared	0.885	0.959	0.819	0.963	0.353	0.688

Source: IMF staff estimates. Notes: Data at the country-quarter level. Reports results for estimates of equation 1. Dependent variable in each regression is growth rate of wages. Columns 1 and 2 repeat the baseline from Table 1 in the main text, with column 1 being for the full sample of 23 countries while column 2 is for Australia only. Columns 3 and 4 repeat column 1 and 2 respectively but use NAIRU from the World Economic Outlook Database to construct unemployment gap instead of an HP trend, while columns 5 and 6 use quarter-on-quarter growth rate of wages instead of year-on-year. Standard errors are clustered at the country level for columns 1 and 3. *, **, and *** indicate significance at the 10, 5 and 1 percent level respectively.

^{Annex Table 2.}

Australia: Robustness for Wage Phillips Curve – Australia Compared to Peers

	(1)	(2)	(3)	(4)	(5)	(6)
	Baseline	Baseline	NAIRU ugap	NAIRU ugap	Q-on-Q	Q-on-Q
VARIABLES	Full Sample	Australia	Full Sample	Australia	Full Sample	Australia
Unemployment gap	-0.272** (0.0973)	-0.160*** (0.0505)	-0.312** (0.115)	-0.170*** (0.0340)	-0.194** (0.0719)	-0.0976** (0.0482)
Expected inflation	0.282*** (0.0956)	0.0448 (0.0649)	0.142 (0.0954)	0.0674 (0.0564)	0.129* (0.0709)	0.0434 (0.0448)
Trend labor productivty growth	0.324* (0.158)	0.0472 (0.119)	0.643*** (0.200)	-0.0845 (0.0937)	0.221** (0.0991)	0.0369 (0.0836)
Lag 1, dependent variable	0.694*** (0.0813)	1.207*** (0.0940)	0.586*** (0.0806)	1.137*** (0.0929)	-0.167 (0.103)	0.360*** (0.129)
Lag 2, dependent variable	0.185*** (0.0397)	-0.261 (0.159)	0.198*** (0.0357)	-0.235 (0.159)	0.0980 (0.0635)	0.0826 (0.156)
Lag 3, dependent variable	0.0986** (0.0367)	-0.122 (0.184)	0.117*** (0.0350)	-0.111 (0.179)	0.0298 (0.0673)	0.184* (0.101)
Lag 4, dependent variable	-0.337*** (0.0254)	0.105 (0.102)	-0.340*** (0.0406)	0.0874 (0.103)	0.158 (0.142)	0.125 (0.149)
Constant	0.119 (0.252)	0.0336 (0.1000)	0.319 (0.188)	0.324*** (0.120)	0.114 (0.173)	0.0261 (0.0727)
Observations	1,604	80	1,388	80	1,646	80
R-squared	0.885	0.959	0.819	0.963	0.353	0.688

Source: IMF staff estimates. Notes: Data at the country-quarter level. Reports results for estimates of equation 1. Dependent variable in each regression is growth rate of wages. Columns 1 and 2 repeat the baseline from Table 1 in the main text, with column 1 being for the full sample of 23 countries while column 2 is for Australia only. Columns 3 and 4 repeat column 1 and 2 respectively but use NAIRU from the World Economic Outlook Database to construct unemployment gap instead of an HP trend, while columns 5 and 6 use quarter-on-quarter growth rate of wages instead of year-on-year. Standard errors are clustered at the country level for columns 1 and 3. *, **, and *** indicate significance at the 10, 5 and 1 percent level respectively.

Prepared by Pragyan Deb and Siddharth Kothari (both APD), with contributions from Yosuke Kido (BoJ). The chapter benefited from valuable comments from the Reserve Bank of Australia and the Commonwealth Treasury of Australia.

Business Indicators data suggest an increase in non-wage non-profit shares of total sales, and PMI indices show significant input price pressures in Australia. National accounts data however suggests that wage and profit shares have not shifted significantly in recent quarters through 2022Q2.

In Australia, owner-occupied housing is a key driver of non-food non-energy inflation and the methodology used to measure owner-occupied housing inflation in Australia is different relative to other countries such as Canada, euro area, New Zealand, the United Kingdom, and the United States. If owner- occupied housing is excluded, core inflation in Australia is also at the lower end of the spectrum.

Trend unemployment is measured using an HP filter with a smoothing parameter of 14400. Results are robust to alternate specifications, including a smoothing parameter of 1600.

Countries in our sample are Australia, Austria, Belgium, Canada, Czech Republic, Denmark, Estonia, France, Germany, Finland, Greece, Iceland, Ireland, Israel, Italy, Japan, Korea, Latvia, Lithuania, Luxembourg, Netherlands, Norway, New Zealand, Spain, Portugal, Slovenia, Spain, Sweden, Switzerland, the United Kingdom and the United States.

Energy component is removed from the transportation and housing sectors. Food is omitted from the analysis.

In particular, we use Labor Cost Indexes covering industry, construction and services for all European countries, the Employment Cost Index for all civilians for the US, compensation per hour worked for Canada, and the Labor Cost Index covering public and private sector for New Zealand. For Japan, we compute hourly wages by dividing contractual earnings by total hours worked.

We calculate unemployment gap by estimating trend unemployment using an HP filter with a smoothing parameter of 14400. Results are robust to alternate specifications, including a smoothing parameter of 1600. Results are robust to alternative measures of labor market slack, such as the unemployment gap calculated relative to the NAIRU.

The downward shift in the wage Phillips curve in Australia can also reflect other structural factors not captured by the augmented Phillips curve, such as impact of globalization or changes in the relative bargaining power of labor. Bishop and Chan (2019) found that lower trade union density in Australia is not responsible for lower wage growth.

Results are similar when a national account (AENA) based measure of wages, such as compensation per employees is used. However, given higher volatility of the AENA measure, the fit is less tight, especially for the augmented model (see Annex)

Models using lagged wages as explanatory variables generally provide a better fit. But these models suggest very high persistence in wages, which in turn drives the results. While useful for forecasting, such persistence makes it harder to link wage dynamics to labor market slack (unemployment gap or underemployment). Hence the specification in this analysis does not include lagged wages.

We exclude Greece from our sample when estimating the Phillips curves as the wage formation process in Greece was likely severely affected by the crisis in the country for a large part of our sample period.

For trend unemployment, to ensure cross-country comparability, in the baseline we use HP filtered unemployment rate with a smoothing parameter of 14400. As a robustness we also consider $u_{i, t}^{g a p}$ measured as a deviation from NAIRU from the WEO Database. Trend labor productivity is also computed using an HP filter.

Australia and the United States get dropped from the sample when considering regressions with average length of collective bargaining contracts, as OECD does not have data on average contract lengths for these countries due to comparability issues.

As an extension, we also considered adding an interaction of the lagged dependent variable with length of contract. The interaction terms were insignificant.

Other structural characteristics can also impact sensitivity of wages to economic conditions. IMF (2022) finds that economies with more stringent employment protection laws and greater market power in product markets (as proxied by the average price markup in the economy) have wages that are more sensitive to changes in unemployment and inflation expectations.

The data excludes expired agreements that have not yet been replaced.

Harmonized data for average length of contracts for Australia is not available from OECD, making comparisons more difficult. Among OECD countries, while contract lengths are relatively short at 12 months in several countries, they extent to as long as 36 months in Denmark, Spain, Italy, and Sweden. In Australia, while the average length of enterprise agreements is long, awards are reset every year, suggesting heterogeneity in responses for different groups.

Other Publishers

Save
Cite
Email this content

Share Link

Copy this link, or click below to email it to a friend
Email this content
or copy the link directly:

https://www.elibrary.imf.org/view/journals/002/2023/051/article-A001-en.xml
The link was not copied. Your current browser may not support copying via this button.

Link copied successfully

Collapse
Expand

Australia: Selected Issues

Author:

International Monetary Fund. Asia and Pacific Dept

Volume/Issue:: Volume 2023: Issue 051

Publisher:: International Monetary Fund

ISBN:: 9798400230660

ISSN:: 1934-7685

Pages:: 37

DOI:: https://doi.org/10.5089/9798400230660.002

Keywords
Inflation and Wage Dynamics in Australia1
References
Annex I. Additional Results

View raw image

Inflation Has Increased in Australia

(CPI inflation, percent change, y/y)

View raw image

Tradable Inflation Has Risen Sharply, Though Non-tradable Inflation Has Also Increased.

(Percent change, y/y)

View raw image

Inflation Has Also Become Broad-based

(CPI items rising by more than 3 percent, sum of weights)

View raw image

High Inflation Partly Reflects Tightness in the Domestic Economy

(Trimmed mean inflation, y/y vs. unemployment gap)

View raw image

Headline Inflation Has Increased in Most Advanced Economies

(Headline CPI inflation, percent change, y/y)

View raw image

The Rise in Australian Inflation is on the Lower End of OECD Peers

(CPI, percent change in 2022Q3, y/y)

View raw image

Though Still High, Energy Inflation in Australia Is Lower than in Other OECD Countries

(CPI energy, percent change in 2022Q3, y/y)

View raw image

Non-Food Non-Energy Inflation in Australia Is on Par with the OECD Average

(CPI non-food non-energy, percent change in 2022Q3, y/y)

View raw image

Some Inflation Components Are More Sensitive to Cyclical Conditions

(Sensitivity to unemployment gap)

View raw image

Cyclical Components Are Currently Driving Headline CPI in Australia

(Headline CPI, percent, y/y)

View raw image

Cyclical Components Are also the Main Driver of Inflation in Other Advanced Economies

(Cyclical headline CPI, percent, y/y)

View raw image

Results Are Robust to Excluding Energy and Food

(Cyclical component, percent, y/y)

View raw image

Global Energy and Food Prices Have Come Off their Peaks

(USD per Barrel; USD per Bushel)

View raw image

Supply Chain Pressure Have Eased

(Global Supply Chain Pressure Index)

View raw image

Labor Market Indicators

(Percent seasonally adjusted)

View raw image

Wage Inflation in Australia Remains Moderate

(Wage inflation, percent change, y/y)

View raw image

Wage Pressures in Sectors Facing Worker Shortages Remain Moderate

(Wage inflation, percent change, y/y)

View raw image

Wage Inflation in Australia Remains Low by International Standards

(Wage inflation, percent change in 2022Q2, y/y)

View raw image

Countries with Tighter Labor Markets Tend to Have Higher Wage Inflation

(WPI y/y percent change vs. unemployment gap)

View raw image

In Australia, the Wage Phillips Curve Remains Low and Flat

(Australia, Wage Phillips Curve)

View raw image

In the US, Wage Growth Is Now Above Levels Implied by Historical Trends

(USA, Wage Phillips Curve)

View raw image

New Zealand Also Has Higher Wage Growth Than Historical Trends

(NZL, Wage Phillips Curve)

View raw image

Augmented Wage Phillips Curve Can Explain Low Pre-pandemic Wage Growth in Australia

(Percent, y/y)

View raw image

Latest Wage Growth Is Well Below Predictions Based on Phillips Curves

(WPI growth, y/y, percent, 2022Q3)

View raw image

Wage Growth and Vacancies Are Correlated at the Aggregate Level

(Percent)

View raw image

Sectors with Higher Vacancy Rates Are Exhibiting Higher Wage Growth

(Percent, 2022Q3)

View raw image

Coverage under Federal Wage Agreements Has Declined but Remains Significant

(Share of employed under federal wage agreement, percent)

View raw image

Federally Registered Wage Agreement Shares Differ Significantly across Sectors

(Share of employed under federal wage agreement, percent)

View raw image

The Duration of Agreements Has Remained High Since the 2010s

(Wage agreement duration, years)

View raw image

Wage Increases Under Federal Agreements Have Fallen but Stabilized in Recent Years

(Percent, y/y)

View raw image

Wages and Business Survey of Labor Costs (Quarter on quarter change: percent seasonally adjusted)

View raw image

AENA Wage Measure Is More Volatile, but Phillips Curve Relationship Holds

(Percent, y/y)

View raw image

Latest Wage Growth Is Again Below Standard Phillips Curve Prediction

(WPI growth, y/y, percent, 2022Q3)

Foresee

^{Annex Table 2.}

Australia: Robustness for Wage Phillips Curve – Australia Compared to Peers

	(1)	(2)	(3)	(4)	(5)	(6)
	Baseline	Baseline	NAIRU ugap	NAIRU ugap	Q-on-Q	Q-on-Q
VARIABLES	Full Sample	Australia	Full Sample	Australia	Full Sample	Australia
Unemployment gap	-0.272** (0.0973)	-0.160*** (0.0505)	-0.312** (0.115)	-0.170*** (0.0340)	-0.194** (0.0719)	-0.0976** (0.0482)
Expected inflation	0.282*** (0.0956)	0.0448 (0.0649)	0.142 (0.0954)	0.0674 (0.0564)	0.129* (0.0709)	0.0434 (0.0448)
Trend labor productivty growth	0.324* (0.158)	0.0472 (0.119)	0.643*** (0.200)	-0.0845 (0.0937)	0.221** (0.0991)	0.0369 (0.0836)
Lag 1, dependent variable	0.694*** (0.0813)	1.207*** (0.0940)	0.586*** (0.0806)	1.137*** (0.0929)	-0.167 (0.103)	0.360*** (0.129)
Lag 2, dependent variable	0.185*** (0.0397)	-0.261 (0.159)	0.198*** (0.0357)	-0.235 (0.159)	0.0980 (0.0635)	0.0826 (0.156)
Lag 3, dependent variable	0.0986** (0.0367)	-0.122 (0.184)	0.117*** (0.0350)	-0.111 (0.179)	0.0298 (0.0673)	0.184* (0.101)
Lag 4, dependent variable	-0.337*** (0.0254)	0.105 (0.102)	-0.340*** (0.0406)	0.0874 (0.103)	0.158 (0.142)	0.125 (0.149)
Constant	0.119 (0.252)	0.0336 (0.1000)	0.319 (0.188)	0.324*** (0.120)	0.114 (0.173)	0.0261 (0.0727)
Observations	1,604	80	1,388	80	1,646	80
R-squared	0.885	0.959	0.819	0.963	0.353	0.688

Source: IMF staff estimates. Notes: Data at the country-quarter level. Reports results for estimates of equation 1. Dependent variable in each regression is growth rate of wages. Columns 1 and 2 repeat the baseline from Table 1 in the main text, with column 1 being for the full sample of 23 countries while column 2 is for Australia only. Columns 3 and 4 repeat column 1 and 2 respectively but use NAIRU from the World Economic Outlook Database to construct unemployment gap instead of an HP trend, while columns 5 and 6 use quarter-on-quarter growth rate of wages instead of year-on-year. Standard errors are clustered at the country level for columns 1 and 3. *, **, and *** indicate significance at the 10, 5 and 1 percent level respectively.

Topics

Countries

Series

About

Resources

Abstract

Inflation and Wage Dynamics in Australia1

A. Introduction

B. Inflation Developments

Cyclical Versus Non-Cyclical Inflation

Changing Importance of External and Domestic Factors

C. Wage Dynamics

Wage-Unemployment Relation in Australia and Internationally

Role of Labor Market Tightness in Determining Wage Dynamics: Australia compared to Peers and Evidence from Australian Sectoral Data

Role of Labor Market Institutions in Shaping the Unemployment-Wage Relation: Lessons from Cross-Country Analysis

D. Conclusion and Policy Discussion

References

Annex I. Additional Results

Other IMF Content

Other Publishers

Asian Development Bank

Inter-American Development Bank

International Labour Organization

The World Bank

Share Link

Table of Contents

Headings

Figures

Metrics

Foresee

Inflation and Wage Dynamics in Australia^¹