Case study Unscheduled absence

Paid unscheduled absence (PUA) is a measure of paid working hours lost due to sickness or caring for an immediate family or household member. As an indicator of both the productivity and health of a workforce, PUA is a key metric for effective workforce monitoring and planning.
The importance of managing PUA in the NSW public sector was reinforced by the Public Service Commissioner in his foreword to the 2016 State of the Sector Report. The Commissioner noted a progressively increasing rate of PUA and appealed to the Sector to better understand and manage this.

Part One begins with an analysis of PUA in the public sector over time - showing the rate has reduced for the first time since 2012 - followed by a comparison by Service, occupation and age. In Part Two, the analysis undertaken last year of job mobility and PUA is continued in more detail.

Part One - Key metrics and time series comparisons by service, occupation & age

Paid unscheduled absence

64.1 hours per FTE for NSW public sector employees in 2017

63.1 Hours per FTE of PUA in 2017 (male)

64.8 Hours per FTE of PUA in 2017 (female)

<50

Age under 50 years
56.4 Hours per FTE of PUA in 2017

50+

Age 50 years+
76.5 Hours per FTE of PUA in 2017

<$100k

Salary under $100k
67.5 Hours per FTE of PUA in 2017

$100k+

Salary +$100k
53.4 Hours per FTE of PUA in 2017

Paid unscheduled absence (PUA) – Total Sector

In 2017, the total number of hours of paid unscheduled absence (PUA) in the NSW public sector was 19,548,742, a decrease of 4.7% from the previous year. The number of hours of PUA per FTE was 64.1, a decrease of 3.0 hours per FTE from the previous year (see Table 9.1)1.

These decreases can be partly explained by an atypically long annual reference period in 2016 compared to 2015 and 2017 (27 fortnights instead of the standard 262).  However, even after adjusting for the duration of the reference period, PUA per FTE in 2017 was down on what it was in 20163

A comparison of PUA between 2017 and 2015 – both reference periods were the standard 26 fortnights in duration – revealed a similar decrease in PUA.  It can be seen in Table 9.1 that PUA declined by 1.0 hours per FTE from 2015 to 2017, amounting to a reduction of 300,583 total PUA hours across the sector between the two years. This reduction is estimated to have a value of approximately $10,000,0004.

Table 9.1: Paid unscheduled absence (total hours & hours per FTE) by Total public sector, 2012 - 2017

2012 2013 2014 2015 2016 2017
Paid unscheduled absence (hrs) 18,790,501 19,014,400 19,484,099 19,878,307 20,512,427 19,548,742
Change from previous year % - 1.2% 2.5% 2.0% 3.2% -4.7%
Hours per FTE (non-casual) 61.4 62.3 63.7 65.1 67.1 64.1
Change from previous year (hrs) - 0.9 1.5 1.4 2.0 -3.0

In the following sections, 2017 data are presented alongside 2015 data, instead of 2016 data, to provide a more accurate indication of change across time.

PUA by Service

As is often the case, sector-level statistics provide an overly simplistic picture of what is occurring in the sector. For example, the decrease in PUA per FTE from 2015 to 2017 was driven by substantial decreases in External to Government Sector agencies (-10.2%), State owned corporations (-8.0%), the Public Service (-4.0%), and Other Crown Services (-3.3%). The NSW Health Service also experienced a decrease, though of a smaller magnitude (-1.7%). On the other hand, three Services experienced non-trivial increases: The NSW Police Force (4.6%), the Teaching Service (3.4%), and the Transport Service (3.4%). Table 9.2 shows the changes across time in each of the Services.  The reasons behind changes to PUA are complex and inter-related, and include the composition and characteristics of the workforce, the types of functions performed and the culture within workplaces.

Table 9.2: Paid unscheduled absence by Service, 2015 - 2017

Service 2015 FTE 2017 FTE 2015 Hrs per FTE 2017 Hrs per FTE % change, Hrs per FTE
Public Service 60,616 58,461 71.7 68.8 -4.0%
NSW Health Service 102,202 106,540 62.8 61.7 -1.7%
NSW Police Force 19,298 19,264 62.2 65.0 4.6%
Teaching Service 57,199 59,697 58.5 60.6 3.4%
Transport Services 12,544 13,172 60.3 62.3 3.4%
Other Crown Services 34,549 34,970 73.8 71.3 -3.3%
State owned Corporations 18,243 11,567 66.8 61.4 -8.0%
External to Government Sector 821 1,132 55.1 49.5 -10.2%
Total Public Sector 305,471 304,801 65.1 64.1 -1.4%

PUA by Occupation

The direction of change in PUA per FTE from 2015 to 2017 was not consistent across occupation groups (see Table 9.3).  Decreases occurred for Technicians and Trades Workers (-6.5%), Labourers (-3.7%), Clerical and Administrative Workers (-2.9%), and Professionals (-0.2%), while the converse was true for Managers (1.3%), Machinery Operators and Drivers (.5%), and Community and Personal Services Workers (0.3%).

Table 9.3: Paid unscheduled absence by ANZSCO, 2015 - 2017

ANZSCO Major Group 2015 FTE 2017 FTE 2015 Hrs per FTE 2017 Hrs per FTE % change, Hrs per FTE
Technicians and Trades Workers 19,030 15,013 74.2 69.3 -6.5%
Professionals 145,103 148,456 59.0 58.9 -0.2%
Managers 17,251 18,919 51.8 52.5 1.3%
Machinery Operators and Drivers 9,772 10,056 83.4 83.8 0.5%
Labourers 12,822 12,169 76.6 73.8 -3.7%
Community and Personal Service Workers 52,038 50,406 73.6 73.8 0.3%
Clerical and Administrative Workers 49,118 49,566 68.4 66.4 -2.9%
Total Public Sector 305,471 304,801 65.1 64.1 -1.4%

NB – Sales Workers was a small group and excluded (due to this Total Public Sector FTE is greater than the sum of ANZCO Major Groups)

PUA by Age

While differences in the extent and direction of change in PUA emerged across the Services and occupation groups, this was generally not the case for different age bands. Table 9.4 shows that PUA per FTE decreased in ten of the eleven five-year age bands (the 65 plus group was relatively small and it is possible that a few extreme cases had a large impact on the results for this group).

Table 9.4: Paid unscheduled absence by Age, 2015 - 2017

Age Bands 2015 FTE 2017 FTE 2015 Hrs per FTE 2017 Hrs per FTE % change, Hrs per FTE
15 to 19 211 217 40.7 38.6 -5.2%
20 to 24 8,826 8,544 45.9 43.2 -5.9%
25 to 29 29,345 30,288 49.5 48.3 -2.5%
30 to 34 33,109 34,907 57.6 56.1 -2.7%
35 to 39 33,157 34,275 60.2 59.7 -0.8%
40 to 44 39,638 37,388 61.1 59.3 -2.9%
45 to 49 39,789 41,806 61.2 59.9 -2.2%
50 to 54 43,599 39,689 65.1 64.7 -0.5%
55 to 59 43,244 41,413 75.1 74.6 -0.7%
60 to 64 24,575 25,336 88.1 87.3 -0.9%
65 plus 9,939 10,899 100.2 101.6 1.4%
Total Public Sector 305,471 304,801 65.1 64.1 -1.4%

Part Two - paid unscheduled absence, engagement & job mobility

Background

The link between paid unscheduled absence (PUA) and various factors has been explored in previous Workforce Profile (WFP) Reports.  It was found that PUA tended to increase with age and decrease with income, and these effects appeared to be independent of each other.  It also has been shown that rates of PUA vary across Services and occupations, after controlling for the effects of age and income. 

Last year the link between job mobility and PUA was explored.  Some evidence was found to support the notion that high job mobility was associated with lower rates of PUA.  Amongst employees who had joined the public  sector at least 10 years ago, those who had joined their current agency in the past two years had lower rates of PUA than those who had joined their current agency less recently.  This analysis is extended in the following section, beginning with an examination of the link between employee engagement and PUA, followed by a proposition that job mobility increases employee engagement which, in turn, decreases PUA in the NSW public sector.

Employee engagement and paid unscheuled absence

A relationship between the level of employee engagement and sickness absence rates has been identified in several studies in Australia and overseas.

A meta-analysis of the effect of employee engagement on key performance outcomes found that of the 23,910 business units investigated, those in the top half of employee engagement scores had a lower probability of high absenteeism5.  Business units with more highly engaged employees took an average of 2.7 sick days per year, compared to those with disengaged employees, who took an average of 6.2 days5.  Similarly, a report by the United States (US) Merit Systems Protection Board found a significant correlation between the average level of employee engagement in an agency and the amount of sick leave used by employees6.  Employees in the most engaged US agencies took an average of 9.0 sick days per year, compared to the least engaged agencies where employees took an average of 12.0 days6.  Gallup studies also indicated that engaged employees had  less absenteeism than disengaged employees (up to 27% lower rates), while another study found substantial linkages between employee engagement and the number of sick days taken, after controlling for the effects of demographics and prior health conditions (including Body Mass Index)5

Australian Public Service Commission (APSC) employee census results consistently show that employees with higher levels of all types of engagement (determined by measuring the relationship employees have with their job, team, immediate supervisor and agency) are less likely to report having taken any sick leave in the preceding fortnight than those with lower levels of engagement7.

Engagement in the NSW public sector is measured via The People Matters Employee Survey (PMES).  Like the WFP, this survey is intened to provide a census of all employees (all employees are invited to participate).  In 2017, 140, 000 responses to the PMES were received.  While it is not possible to link individaul records between the PMES and the WFP, correlations between average engagement and PUA per FTE can be calculated at an agency level. 

The correlation between mean engagement and PUA per FTE for 68 NSW public sector agencies is presented in Figure 9.18.  An inverse relationship can ben seen - PUA tended to be low when engagement was high (correlation -.36). 

 

The stability of the results depicted in Figure 9.1 were tested by partitioning the data by age and income for each agency (this was done because it has been shown  in previous WFP reports that age and income influence PUA, and the composition of age and income varies by agency).

Looking at Figure 9.2 and Figure 9.3, it can be seen that the relationship between engagement and PUA remained.  Further, the correlation co-effieicent improved for three of the four partitions: under 50 years old (correlation:-.47, n=58), 50+ years old (correlation: -.41, n=47), under $110,000 remuneration (correlation: -.52, n=54), and $110,000+ remuneration (correlation: -.28, n=43)8.

It also can be seen that PUA tended to be higher in the older age group (see Figure 9.2) and lower in the higher income group (see Figure 9.3).  This observation is consistent with findings in previous WFP Reports at an overall Sector level, providing further confidence in the agency level results.

 
 

Job mobility and employee engagement

PMES 2017 data comparing tenure in current organisation and tenure in current role is presented in Figure 9.4 below.  The general pattern noticed here was that shorter tenure in current role was associated with higher level of employee engagement, irrespective of length of service within the current organisation.  Engagement also appeared to progressively decline the longer employees remained in the same role within the same organisation.

Looking at employees who had been in their current organisation for 2 to 5 years, those who had been in their current role for less than 1 year had higher engagement than those who had been in their current role for 2-5 years (71 vs 64 or 11% difference). 

Amongst employees who had been in their current organisation for 5 to 10 years, those who had been in their current role for less than 1 year had higher engagement than those who had been in their current role for 5 to 10 years (69% vs 60% or 15% difference). 

For those who had been in their current organisation for 20+ years, the difference in engagement between those who had been in their current role for less than 1 year and those who had been in their current role for 20+ years was 18% (71% vs 60%).

 

Table 9.5: Number of respondents - mean engagement and tenure in current role & organisation, 2017

Number of respondents current role less than 1 year current role 1 - 2 years current role 2 - 5 years current role 5 - 10 years current role 10 - 20 years current role more than 20 years
current org less than 1 year 11,586 na na na na na
current org 1 - 2 years 1,708 8,397 na na na na
current org 2 - 5 years 2,643 2,681 13,946 na na na
current org 5 - 10 years 2,164 2,287 4,426 13,261 na na
current org 10 - 20 years 2,002 2,044 4,272 5,122 14,930 na
current org more than 20 years 1,073 1,026 2,354 2,925 3,898 9,035

The stability of the relationship observed in the preceding chart was further tested by dividing the data into various partitions.  Looking at Figure 9.5 and Figure 9.6, the pattern remained the same – short tenure in current role resulted in higher engagement than longer tenure in current role, irrespective of the length of time in the current organisation (2-5 years versus 20+ years) and across all selected sub groups  (gender, age, income, service delivery role vs non service delivery role).

 

Table 9.6: Number of respondents - Mean engagement and tenure in current role (shortest v highest tenure) of staff employed in current organisation for 2-5 years (selected sub groups), 2017

Number of respondents current role less than 1 year current role 1 - 2 years current role 2 - 5 years
male 861 860 4,781
female 1,754 1,776 8,831
under 50 yrs 2,261 2,262 10,723
50+ yrs 352 392 2,988
under $110k 1,800 1,953 10,367
$110k+ 676 549 2,491
non service delivery 1,492 1,495 6,010
service delivery 933 963 6,529
 

Table 9.7: Number of respondents - Mean engagement and tenure in current role (shortest v highest tenure) of staff employed in current organisation for 20+ years (selected sub groups), 2017

Number of respondents current role less than 1 year current role 1 - 2 years current role 2 - 5 years current role 5 - 10 years current role 10 - 20 years current role more than 20 years
male 400 368 961 1,176 1,555 3,341
female 661 643 1,365 1,702 2,292 5,507
under 50 yrs 397 413 800 871 1,008 1,989
50+yrs 665 608 1,527 2,023 2,840 6,890
under $110k 478 476 1,069 1,530 2,446 6,808
$110k+ 540 472 1,110 1,177 1,155 1,478
non service delivery 413 427 895 968 1,254 2,440
service delivery 559 482 1,184 1,644 2,215 5,649

The analysis of tenure by engagement using PMES data followed a similar logic to the analysis conducted last year where WFP data regarding tenure in current organisation was analysed for employees who had been in the public sector for more than 10 years against PUA.  A similar finding was presented then, and the direction of this finding remained when the same analysis was repeated with 2015 and 2017 WFP data - shorter tenure in current role (a proxy for higher job mobility) was associated with lower PUA after controlling for a number of factors. 

Job mobility, employee engagement and PUA

In the absence of a single data set or the ability to directly link records across the WFP and PMES data sets, multiple data sources and reasoning have been used to formulate the following proposition: high job mobility leads to higher engagement which in turn leads to lower paid unscheduled absence (PUA) in the NSW Government Sector9.  This proposition and the data sources used to support it are depicted in figure 9.7.

Figure 9.7: Proposed link between job mobility, engagement and paid unscheduled absence in the NSW Government Sector9

The link between job mobility and engagement was demonstrated with PMES data for 2017.  It was shown that mean engagement was higher for those who had been in their current roles for the shortest periods of time and that mean engagement appeared to progressively decline as length of time in current role increased.  This general pattern remained irrespective of length of time in the current organisation and after repeating the analysis for various sub-groups.

The link between engagment and paid unscheduled absence (PUA) was demonstrated by correlating mean engagement scores (PMES data) and paid unscheduled absence per FTE (WFP data) at an agency level.  A negative correlation was found (when engagement was high paid unscheduled absence was low, and vice versa).  This correlation remained (and tended to improve) after the data was divided into age and income sub-groups (groups that are known to modify variability in PUA rates).  Further, the results at agency level were consistent with those that have been observed at an overall level – the high income group tended to exhibit lower rates of PUA compared to the low income group, and the older age group tended to exhibit higher rates of PUA compared to the younger age group.  A relationship between engagement and sickness absence also has been found in several overseas studies and by the Australian Public Service Commission. 

The preceding analysis continued the line of inquiry from last year, introducing and testing additional data sources to provide further support for the notion that high levels of job mobility within the NSW public sector have a positive outcome for employees and employers in the form of higher engagement & lower paid unscheduled absence.   

End Notes

1 -The number of pay periods varies across the Sector in any given year. In 2017 a number of agencies had a longer reporting period than usual while some agencies had shorter ones. The rate of PUA is influenced by the length of the reference period.
2 - Every 11 years, an extra pay cycle occurs within the financial year for some agencies. This means the reference period used for calculating PUA per FTE will be longer. This longer reference period effects calculations for individuals who worked the entire year and had an FTE of 1.0 (but not those with an FTE of less than 1.0 and / or who did not work across the entire reference period).
3 - After making a broad assumption that all agencies had an extra pay cycle and that calculations for every individual were impacted by this, a decrease remained. These assumptions were conservative for reasons mentioned in footnote 1. This also means that producing an accurate adjusted figure is difficult and therefore not recommended.
4 - The median salary of $83,689 was divided by 36.5 hours (unweighted average of 38 hours and 35 hours) and 52 weeks to estimate an average hourly rate of $44.09 which was then multiplied by 300, 583 hours.
5 - Harter, J K, Schmidt, F L, Kilham, E A & Asplund, J. W, Q12 Meta-analysis report, Gallup Consulting, 2008
6 - US Merit Systems Protection Board, The Power of Federal Employee Engagement, MSPB, Washington, D.C., 2008, p. 29.
7 - Australian Public Service Commission, State of the Service Report 2015-16, Commonwealth of Australia, Canberra, 2016.
8 - In all correlations presented between PMES and WFP data at agency level, agencies with less than 100 responses to the PMES or with an FTE of less than 100 in the WFP were excluded to ensure adequate sample size.
9 - Government Sector was chosen for a number of reasons: External to Government was too small to allow for sub-group comparisons, changes in PUA rates for SOCS and External to Government were positive, and the contribution of these agencies to the overall PUA rate was small (they represented less than 5% of the total FTE entitled to paid sick leave or carers leave).