6302.0 - Average Weekly Earnings, Australia, May 2014 Quality Declaration 
ARCHIVED ISSUE Released at 11:30 AM (CANBERRA TIME) 14/08/2014   
   Page tools: Print Print Page Print all pages in this productPrint All

FEATURE ARTICLE: AVERAGE WEEKLY EARNINGS AND WAGE PRICE INDEX – WHAT DO THEY MEASURE?


INTRODUCTION

The ABS publishes a variety of information on wages and salaries (often referred to as ‘earnings’) from both household and employer surveys. These data have many uses including economic analysis, social research, policy formulation and evaluation, and research by employer and employee associations. The decision on which data to use should depend on the purpose and type of analysis to be undertaken.

The six monthly Average Weekly Earnings (AWE) and quarterly Wage Price Index (WPI) collections both measure the wages and salaries of employees. These collections have different purposes and, as a result, use different methodologies.

This article begins by outlining the purpose and key uses of AWE and WPI. The article then briefly describes the AWE and WPI methodologies, and uses examples based on hypothetical labour market conditions to demonstrate how the two surveys can respond differently to economic events.


BACKGROUND

Examining changes in wages and salaries assists in identifying inflationary pressures in the economy as well as highlighting structural changes in the labour market. As wages and salaries paid to employees represents a significant component of operating costs for businesses, changes in wages and salaries can highlight cost pressures facing businesses. Changes in wages and salaries can reflect the impact of the economic cycle on the labour market or sections within the labour market.

The AWE and WPI collections aim to measure different, albeit related, concepts. The AWE is part of the suite of statistics designed to capture employee remuneration (for more detail please see the Feature Article 'Understanding Earnings in Australia Using ABS Statistics' in the July 2014 issue of Australian Labour Market Statistics, cat. no. 6105.0). AWE is designed to measure earnings, which consist of payments-in-cash and payments-in-kind such as fringe benefits (Labour Statistics: Concepts, Sources and Methods, 2013, cat. no. 6102.0.55.001). In practice, however, it is only practical for ABS earnings series to include wages and salaries in cash as well as salary sacrifice arrangements (which are in-kind payments that are at the discretion of the employee). The WPI is designed to measure inflationary pressures associated with the Compensation of Employees (CoE), as outlined by the System of National Accounts (2008). Theoretically, WPI would include all elements of CoE, but for practical reasons it focuses on wages and salaries payments in cash, as well as salary sacrifice payments. Thus, despite differences in the underlying aims of the two collections, there is considerable commonality in the scope of the two collections. For the sake of simplicity, the term 'wages and salaries' has been used to refer to the scope of WPI and AWE throughout this article.

The WPI measures changes in the wages and salaries paid by employers for a unit (i.e. hour) of labour where the quality and quantity of labour are held constant. It has the dual purpose of monitoring wages and salaries inflation in the economy and supporting the compilation of the Australian System of National Accounts. To achieve this, the WPI uses a Laspeyres index methodology (where the price in a particular period is compared to that in a previous fixed period) designed to produce a measure of pure price change in wages and salaries independent of compositional factors (i.e. the quantity and quality of labour are held constant). ‘Quantity’ refers to compositional factors such as the effect of changing hours paid for and number of employees. ‘Quality’ refers to changes in job specifications or job holder characteristics such as employee performance or relative level of experience. These factors are held constant by ensuring that jobs are matched between quarters with no change in job specifications and by holding weekly hours constant between quarters. Adjustments to remove changes in quality and quantity are made during the statistical production phase of the WPI survey.

In contrast, the AWE is designed to provide an accurate estimate of the current average value of wages and salaries paid to employees by an employer over a specified period. The emphasis placed on producing a contemporary measure of average wages and salaries means that the AWE reflects structural changes that occur over time (such as changes in hours paid for and employment). The AWE achieves this by collecting payroll data for a specified period. This method allows quantity and quality (i.e. compositional effects) to be included in the AWE outputs. The examples in the next section demonstrate how AWE and WPI will be affected differently by real world economic events.


EXAMPLES DEMONSTRATING THE EFFECT OF LABOUR MARKET CHANGES ON AWE AND WPI GROWTH RATES

To illustrate the way in which the AWE and WPI respond to various changes in the labour market the following simplified examples using hypothetical data are provided.

Consider an initial population (period 0) of three businesses with a combined total of eight occupied jobs. The four examples that follow contain period 1 data with a change in either: (1) the number of employees; (2) the weekly hours paid for (hereafter referred to as 'hours'); (3) market based changes to the hourly rates; or (4) non-market based changes to the hourly rates. For illustrative purposes the examples require a number of assumptions to be made:

  • All jobs/businesses are collected from the hypothetical population. In reality, the AWE and WPI are both stratified sample based surveys and do not sample all possible jobs or businesses within the economy;
  • The examples ensure that changes are only applied to one variable (hours, employment, or hourly rate) at a time to allow the impact on the respective estimates to be isolated. In the real economy, these changes occur concurrently, and separately identifying the impact of these changes is difficult;
  • Changes to employment and hours are applied uniformly and in a consistent direction across the population in these examples. In reality, employment or hours can change in different directions and by varying magnitudes within different sections of the economy, which can result in complex distributional effects;
  • The WPI index methodology incorporates expenditure value data to combine elementary aggregates (groupings of similar jobs) in the aggregation process (see Wage Price Index: Concepts, Sources and Methods, 2012, cat. no. 6351.0.55.001). There is only one elementary aggregate in this example, so expenditure values are not required; and
  • Period 0 is assumed to be the index reference period for WPI, resulting in an index number of 100.0 being used in this example.

Consider period 0 data below:

Period 0 — Data used for Examples 1-4

Business ID Job ID
Weekly Hours
Hourly Rate
Weekly Wages and Salaries


Business A Job 1
38
24.50
931.00
Job 2
38
24.71
938.98
Job 3
38
23.83
905.54


Business B Job 4
38
44.64
1 696.32
Job 5
38
45.25
1 719.50
Job 6
38
52.13
1 980.94


Business C Job 7
38
19.85
754.30
Job 8
38
32.74
1 244.12


Total 8
10 170.70



Average weekly earnings can be calculated using the following formula:

Equation 1:

Equation: Equation 1 (12)

In period 0:

Equation: In period 0

As stated earlier, the WPI in period 0 is 100.0. To calculate wages growth in the WPI (and the change in the AWE), period 1 data are required.


Example 1: Employment Change

Period 1 — AWE and WPI comparison, with the addition of two jobs

Business ID Job ID
Weekly Hours
Hourly Rate
Weekly Wages and Salaries


Business A Job 1
38
24.50
931.00
Job 2
38
24.71
938.98
Job 3
38
23.83
905.54


Business B Job 4
38
44.64
1 696.32
Job 5
38
45.25
1 719.50
Job 6
38
52.13
1 980.94


Business C Job 7
38
19.85
754.30
Job 8
38
32.74
1 244.12
Job 9
38
19.85
754.30
Job 10
38
19.85
754.30


Total 10
11 679.30



In this example, business C has hired two additional employees between period 0 and period 1. The weekly hours and hourly rates of the existing jobs are held constant between the two periods. First, AWE is calculated in period 1 using equation 1:

Equation: using equation 1

Therefore, the percentage change in the AWE from period 0 to period 1 is:

Equation: AWE from period 0 to period 1 is

In this case, the addition of the two new employees causes a drop in the AWE, since the weekly earnings of both employees are below the period 0 average. If the new employees received weekly earnings above the period 0 average, the AWE would show a rise.

This can be compared to the effect on the WPI. In the absence of expenditure values, the change in the WPI is calculated as a ratio of weighted average prices, using the following formula:

Equation 2:

Equation: Equation 2

Where: Rt is known as the current period (period 1 in this case) 'price relative'; pt is the current period hourly rate; p0 is the index reference period hourly rate; and h0 is the weekly hours which are held constant from the index reference period.

The WPI measures price changes to constant quantity and quality, so only jobs that are matched between periods will contribute to the index. Since jobs 9 and 10 are not matched between period 0 and 1, they will not contribute to the index until two periods of data are obtained for these jobs.

Therefore:

Equation: Therefore

To calculate the new WPI index number (It), the index number from the index reference period (I0) is multiplied by the price relative in the current period (Rt):

Equation 3:

Equation: Equation 3

In period 1:

Equation: In period 1

In this case, there is no change to the WPI as a result of the increase in the number of employees in Business C as they are not yet included in the index calculation. The exclusion of the new employees from the index calculation means that there has been no change in the hourly rate between periods 0 and 1.


Example 2: Change in Hours

Period 1 — AWE and WPI comparison, with a uniform increase in hours worked for all jobs

Business ID Job ID
Weekly Hours
Hourly Rate
Weekly Wages and Salaries


Business A Job 1
40
24.50
980.00
Job 2
40
24.71
988.40
Job 3
40
23.83
953.20


Business B Job 4
40
44.64
1 785.60
Job 5
40
45.25
1 810.00
Job 6
40
52.13
2 085.20


Business C Job 7
40
19.85
794.00
Job 8
40
32.74
1 309.60


Total 8
10 706.00



In this example, hours worked have increased from 38 to 40 for all jobs in the population. This has resulted in a rise in weekly wages and salaries for all jobs. AWE can be recalculated in period 1 using equation 1:

Equation: calculated in period 1 using equation 1

Therefore, the percentage change in the AWE from period 0 to period 1 is:

Equation: from period 0 to period 1 is (eq)

A uniform increase/decrease in hours worked will always result in a rise/fall in the AWE when hourly rates and the number of employees are held constant. In reality, changes in hours worked are rarely uniform across the economy and distributional effects will also affect the results. For example, hours may increase in relatively high paid sectors of the economy and decrease in relatively low paid sectors. In this case, there may be no change in aggregate hours worked, but average weekly earnings would still rise.

Equation 2 for the WPI price relative (Rt) shows that the hours worked (h0) are held constant between periods. The current period hours do not enter the formula. Therefore, the WPI price relative is calculated as follows:

Equation: relative is calculated as follows

There is no change in the WPI regardless of the magnitude or distribution of the change in hours worked within the population (assuming no change in the hourly rate).


Example 3: Market based pay rise

Period 1 — AWE and WPI comparison, with a ‘Market based’ pay rise

Business ID Job ID
Weekly Hours
Hourly Rate
Pay change
Weekly Wages and Salaries


Business A Job 1
38
25.48
4%
968.24
Job 2
38
25.70
4%
976.60
Job 3
38
24.78
4%
941.64


Business B Job 4
38
44.64
0%
1 696.32
Job 5
38
45.25
0%
1 719.50
Job 6
38
57.34
10%
2 178.92


Business C Job 7
38
19.85
0%
754.30
Job 8
38
32.74
0%
1 244.12


Total 8
10 479.64



In this example, there has been a rise in the hourly rate for jobs 1, 2, 3 and 6 between periods 0 and 1. It is assumed that these rises are purely ‘market based’ (i.e. rises are determined solely by market based factors, such as broad based CPI increases, Enterprise Agreement rises or minimum wage rises). AWE can be recalculated in period 1 using equation 1:

Equation: recalculated in period 1 using equation 1

Therefore, the percentage change in the AWE from period 0 to period 1 is:

Equation: 0 to period 1 is (eq)

As the increases are considered ‘market based’ and the quantity and quality of labour are held constant the WPI price relative is calculated, using equation 2, as follows:

Equation: using equation 2, as follows

From equation 3, the new index number can be calculated as 103.0 in period 1.

Equation: calculated as 103.0 in period

The period 0 and period 1 indexes can be used to calculate a quarterly movement of 3.0%.

Equation: quarterly movement of 3.0%

In this example, AWE and WPI produce the same result.

Example 4: Performance based pay rise

In this example, it is assumed that the pay changes that occur between period 0 and period 1 described in example 3 are based on factors unrelated to the market, such as good performance or the relative level of employee experience in the job. In other words, the pay changes occur due to changes in the ‘quality’ of the jobs.

This distinction does not impact on the calculation of the AWE. The change between periods 0 and 1 is still 3.0%. However, this change in quality is removed from the WPI during processing and would result in no movement being observed for the WPI under this example.


CONCLUSION

The above examples are highly simplified and provided for illustrative purposes only. They should not be taken as hypotheses for historical divergences between the series. In reality, the changes to employment, pattern of hours, and non-market changes described above do not occur in isolation and are seldom spread uniformly across the economy. In practice, it is virtually impossible to disentangle these effects and pinpoint the precise cause of any given movement in the AWE series. Movements in the AWE result from a complex interrelationship between distributional influences and changes in hours, employment levels and pay rates that can often be pulling in different directions.

The choice of using WPI or AWE growth rates will be dictated by the purpose of analysis. If analysis is focused on current value of average wages and salaries that reflects contemporary structural change in the labour market (e.g. changes in employment in particular industries), then AWE should be the preferred measure. If analysis is concerned with the inflationary pressure associated with wages and salaries, then users should consider using the WPI.