The deteriorating experience of university workers is central to understanding the pay element of the current 2019 UCU dispute. The ability of people’s take-home pay to cover increasing expenses is critical, but so too are working in conditions experienced in the context of accelerating casualisation, rapid staff turnover, and threats of redundancy. This aspect of the industrial dispute is harder to encapsulate in a claim, as issues are less easily quantified, statistics are patchy, and institutional practices differ.
In this USSbrief, we explore the statistics used by the employers’ body, the Universities and Colleges Employers Association (UCEA), which negotiates on our pay and conditions. We focus on just one aspect of their work here: their reporting of staff turnover, which, counter to experience, suggests that casualisation rates are declining. Their use of strategic ignorance, statistical measures, and selective benchmarking suggest their data are not to be trusted.
UCEA produces a variety of reports on the Higher Education (HE) workforce, which includes evaluation of staff turnover. These appear to show a fairly promising picture of relative stability across the sector, with lower turnover figures than wider industry. Their 2019 workforce report suggests the overall turnover was 11.8% for professional service staff and 8% for academics.
However, on closer inspection, UCEA’s figures on staff turnover are deliberately engineered to portray the HE sector in a good light. We suggest they operate through an organised logic of strategic ignorance (to use a term that sociologist Linsey McGoey operationalises to good effect). Or, to use another technical phrase, they are based on bollocks. Strategic ignorance is achieved by the misleading ways in which staff turnover figures are collected, represented, analysed, and benchmarked. What UCEA has produced cannot simply be carelessness. It is arguably constructed to mislead.
These figures matter, as reviewing data on staff working conditions, and on issues including equality and precarity, often form a key part of UCEA’s offer in pay negotiations. UCEA recently made much of its commitment to ‘work on gender and ethnicity pay gaps, casual employment arrangements and workload’ in its press release on pay negotiations (26 November 2019). UCEA also regularly produces infographics and reports that stress the positive and improving picture.
Yet, the figures on staff turnover reported by UCEA are a great deal lower than we might expect from experience. In our institutions, we see many people moving through on fixed contracts and hourly paid workers coming and going. It turns out these do not count in the figures on staff turnover constructed by UCEA. Even once these increasingly typical, and so-called ‘atypical’, staff are excluded, the figures diverge from what we see every day in the increasing incidence of redundancies, which are hitting some institutions particularly hard. We should be looking carefully at staff turnover data.
2. Why focus on staff turnover data?
UCEA depends on data collected by the Higher Education Statistics Agency (HESA). HESA is the body responsible for collecting and disseminating HE data for its ‘statutory customers’, including the Office for Students (OfS), Department of Education (DoE), Higher Education Funding Council for Wales (HEFCW), and so on. It is also the body from which UCEA draws its statistics about staffing trends. There is a fair amount written about the problems of HESA staff data. UCU have raised concerns about their representation of key sector indicators, such as staff–student ratios (Court 2012).
The way UCEA reports, shapes, and black-boxes data on staffing trends and turnover mean it is difficult to create an accurate picture of the security of employment in the sector. Many of us are aware that the positive stories on casualisation rates that UCEA and HE employers spin in press releases conflict sharply with what we know and see, and the negative reporting that does sometimes make it into the media.
We focus here on staff turnover data. This measure does not address all the difficulties of UCEA data, but it can provide us with a snapshot of the landscape. It illustrates the extent to which precarity, disruption, casualisation, and restructuring are — as is widely assumed by HE workers — the new ‘normal’. It also tells us something significant about the manner in which our employers and UCEA are operating in how they represent the state of the sector, and the environment in which we work, to governments, funding bodies — and, crucially, to us.
There is no magic figure for what the ‘right’ percentage of staff turnover in any organisation ought to be and there are differences between sectors based on a variety of factors (Booz 2018). There is a balance to be achieved, which suggests a goldilocks approach to turnover — not too low, nor too high, but just right. As Booz notes, ‘no one wants their business to be a revolving door’.
Higher retention rates indicate an organisation is providing an attractive workplace where employees wish to stay, an effective environment for employees to undertake their jobs, a good staff community, and employee wellbeing. Staff turnover, as well as employees’ desire to leave or their willingness to recommend that others join the organisation, are key measures that human resources (HR) and organisational managers use to assess the ‘health’ of their organisations.
The literature is clear that a high turnover can be a significant issue for an organisation in terms of cost, disruption, and impact on the workforce and productivity. These costs consist of ‘not only the accumulating recruitment and selection costs, but also the negative impact on current staff’ (Murphy 2019). There are good reasons for organisations to keep a firm check on staff turnover so as to avoid the continuous and wide-scale disruption associated with a high turnover of staff.
But UK HE has become a revolving door. There are not only gaps in staff, but also gaps in data. Much staff turnover is not showing up in UCEA figures.
We explore below some key issues with UCEA’s methodology on staff turnover figures. We show how key categories of staff are excluded from data and how this distorts the full extent of staff turnover in UK HE. UCEA’s exclusion of a significant staff population also undermines its attempt to benchmark staff turnover rates in other sectors, which report in a fundamentally different way. Combined, these missteps mean that UCEA fails to deliver informative data on staff turnover.
UCEA would appear to have opportunistically chosen to exclude ‘uncomfortable knowledge’ (Rayner 2012). The result is the presentation of a rather different picture of UK HE, achieved through deploying the power of ‘strategic ignorance’ (McGoey 2019). We make this argument by documenting the specific population excluded, the deviation from normal staff turnover reporting practice, and the opacity of data collection and reporting practices in HE.
3. Missteps in accounting for staff turnover in UK HE
Step One: select out precarity
The story of how UCEA selectively represents staff turnover goes back at least a decade. In Recruitment and Retention of Staff in Higher Education 2008, UCEA highlighted turnover rates for 2006–2007 as follows: academic staff turnover rate was 6%, technical staff 7%, and administrative/professional staff 8%. The report notes the rates are almost identical to the 2005 survey, points towards a consistently low level of labour turnover in the HE sector, and celebrates ‘a positive picture of the HE labour market overall’.
While boasting of the attractiveness of the sector evident in these low staff turnover figures, UCEA does concede that staff turnover figures are difficult to compare to other sectors. It suggests ‘a meaningful comparison with turnover rates in the wider economy is notoriously difficult to achieve, given different methods of calculations and varied definitions of leaver categories’ (pp. 8–9).
Yet, these pesky methodological considerations are quickly forgotten, as UCEA proceeds to draw data from the CIPD Annual Survey Report of 2008, suggesting ‘[T]he overall turnover rate in the UK was 17.3%. The average was highest in the private sector at 20.4%, with a public sector average of 13.5% [and that given] the highest turnover rate within the HE sector was 10% — for manual and clerical staff — it appears that higher education compares relatively well with the wider economy’ (p.9).
The report then flashes up a quote from a pre-1992 university exclaiming there is ‘hardly any turnover with academic staff; the recruitment that is carried out is as a result of either a fixed term replacement when an academic has been successful in obtaining research fellowship or retirement’. In 2008, UCEA thinks staff turnover rates are superb in UK HE. But even at this point, there are major omissions in terms of which staff count and which are excluded from turnover figures.
UCEA has, remarkably, excluded all staff on fixed-term contracts in UK HE and is calculating staff turnover on the basis of open-ended contracts only. Two critical issues for staff turnover figures flow from this.
The first is the sheer magnitude of this excluded population. A significant proportion of the entire staff body employed by HE providers in the UK are employed on fixed-term contracts, amounting to one third of all academic staff and over one-tenth of non-academic staff.
The second concern is the pressure of insecurity experienced by this excluded population. Fixed-term contract workers are one of the most transient and precarious contract types. By focusing on open-ended contracts only, institutional ‘staff turnover’ figures are pre-selected to present highly stable turnover figures and a far lower overall turnover figure.
The decision to exclude fixed-term staff from staff turnover figures, and to present what remains as the overall staff turnover in HE, is methodologically flawed and extremely misleading. Reported total staff turnover rates are more normally calculated as
a ‘crude-wastage’ rate and does not distinguish the reasons why individuals leave an organisation. It covers all types of employee departures including voluntary resignations, redundancies, dismissals and retirements. It is calculated by taking the total number of leavers in a specified period (usually 12 months) and expressing the number as a percentage of the number of people employed during that period’ [our emphasis] (XpertHR 2019).
While it is common practice for organisations to undertake separate evaluation of different function groups and job levels, this is not what UCEA is presenting. The significance of this is clear. UCEA has not included all types of employee departures nor the total number of leavers in the figures it is using to represent staff turnover in HE. Its headline turnover rate only includes a select population.
Those on fixed-term contracts are not the only HE workers excluded from the staff figures. Others critical for the functioning of the HE sector are excluded as a matter of course, including those categorised as ‘atypical staff’ and other HE ‘workers’. 2008 heralds the start of UCEA using techniques to present a rosy picture of staff turnover. We already see why it is so problematic as a measure of the extent of casualisation and precarity in UK HE.
Step Two: trivialise total turnover
Fast forward to 2017–18. UCEA is still publishing its summaries, the exclusion of staff on fixed term contracts endures, and UCEA has switched its benchmarking group from CIPD to the (largely paywalled) XpertHR Survey. The XpertHR survey report figures for 349 private and public sector organisations in the UK and reports an average overall turnover rate for 2018 of 20.9% (with a median of 18.8%).
In contrast, in its Higher Education Workforce Report 2019, UCEA reports that the median total employee turnover rates for academic staff on open-ended contracts in 2017–18 was 8.0%. It notes that rates at pre-1992s are slightly lower, and post-1992s slightly higher. We will come to the difference between voluntary and total turnover in a minute. UCEA celebrates again, stating that these median figures are ‘well below wider UK averages for staff in similar organisations’ (p.24).
Here is UCEA’s presentation of the non-academic staff turnover. It states the median total turnover for all professional service staff employed on open-ended contracts was 11.1%. It also notes turnover figures for pre-1992 HEIs were slightly lower.
There are three points we want to note here. First is the continuing exclusion of staff on fixed term contracts, other university workers, including so-called atypical staff. ‘Atypical’ staff are defined by HESA as those people whose working arrangements are ‘not permanent, involve complex employment relationships and/or involve work away from the supervision of the normal provider’. HESA gives institutions discretion on how and whether to report staff under this category. Some institutions choose not to report any ‘atypical’ academic staff at all. The exclusion of precarious staff from the data is perpetuated by UCEA, by HESA, and by institutions.
Second, is that using the median has a smoothing effect, presenting trends in a particular light. As the XpertHR Survey 2019 explains, the median falls at the ‘exact midpoint in the range of resignation rates and reduces the impact of exceptionally high or low figures in the overall calculation’. XpertHR consistently provides both the mean and median. While UCEA routinely chooses to report lower median figures without comment, the XpertHR report provides both mean and median, and notes the significance of the difference between them.
A further issue is the difference between voluntary and total turnover in the UCEA figures. UCEA accounts for the difference between resignation rate and total turnover by pointing to voluntary and compulsory redundancies (0.3%), retirement (1%), and other reasons such as death in service. UCEA’s figures are used to suggest academics generally hang around until they retire or die. UCEA trivialises the total labour turnover rate by suggesting the difference between the two is largely external to the operation of the sector.
In contrast, the XpertHR Survey 2019 takes the gap between resignation rate and total turnover more seriously. They illustrate these figures with comments from survey respondents which ‘demonstrate that restructuring and resulting redundancies are now a part of day-to-day business’. The full report is behind a paywall, and so we do not know which sector is speaking here, but we do know that resignations, voluntary redundancy, normal retirement, compulsory redundancy, and fixed term contracts are all part of the day-to-day business of universities too. And they are being taken more seriously by their chosen benchmarkers than by UCEA itself.
These steps, critically — and perniciously — exclude a sizeable proportion of the staff body from staff turnover and play down total staff turnover figures. They point to techniques that produce politically significant absences when partial UCEA data are compared with data from the wider economy. Benchmarking is used to spin a narrative of relative stability for UK HE. This leads us to the third misstep.
Step Three: benchmark apples with beetroot
We conclude this part of the argument by returning to what UCEA identifies as the ‘notoriously difficult to achieve’ comparisons with staff turnover in broader industry. Badly put, UCEA claims that the HE sector benefits from good staff turnover and employment conditions by comparing data that are compiled in different ways, making systematic errors in UCEA’s favour. This is how its positive story is constructed.
Let us return to our earlier analysis of the 2019 UCEA report. This states:
Employee turnover in the sector remains low relative to external benchmarks and has fallen since the 2017 survey. Across all UK HEIs, HESA data show a median resignation rate of 7.6% in 2017–18, based on all staff employed on open-ended contracts and total turnover was 11.8%. Rates of turnover at Pre-1992 HEIs are slightly lower. The median resignation rate in the wider economy was 12.9% in 2018 according to an XpertHR survey of 349 private and public sector organisations with median total turnover standing at 18.8%. (p. 18)
Under the section for academic staff turnover, the UCEA report states:
Employee turnover rates for academic staff are lower than those for professional services staff and therefore well below wider UK averages for staff in similar organisations. As shown in Table 4, the median institutional resignation rate across the whole sector was 5.0% for academic staff on open-ended contracts in 2017–18, while total turnover is 8.0%. (p. 24)
It is clear how the exclusions embodied in UCEA data become a tool of strategic ignorance, once we know the exclusions of staff type and types of turnover data. These exclusions enable UCEA to reconstruct the narrative around many aspects of staff turnover and precarity in UK HE, and continue to proclaim a positive picture of relative employment stability in the sector. The failure to include all employees in an overall turnover figure, and then to draw comparisons with XpertHR reports which operate on a very different basis (i.e. total staff turnover includes all staff), is deeply problematic.
Comparisons are being made by UCEA that appear to be manipulated to present UK HE in a highly favourable light and to attempt to suppress the rate of turnover. To compare apples with beetroots in the examples above, without any information on how data are calculated, is a significant step in presenting the sector in a stable and favourable light.
Robust comparisons are actually not so difficult if data are sampled, presented, and interpreted in a similar way — even if we may want to question what and when benchmarking is useful. We show how in the next section.
4. Recalculating the hidden workforce of UK HE
Step One: add meaning to average figures
Let us start again and calculate staff turnover in good faith and in a similar way to XpertHR. There are three steps to our argument for reconstructing a more accurate picture of precarity: redoing averages, situating comparisons, and questioning everything.
First, in line with standard ways of reporting staff turnover, let us present all FTE staff reported to HESA, rather than missing out the third of academic staff on fixed term contracts. This calculation still critically excludes so-called atypical staff, as HESA does not include these in leavers’ data. We cannot replicate non-academic staff leavers figures as they are not available to us on HESA. We will also exclude the ‘Voluntary Resignation Rate’ as this data is not easy to extract from public-facing HESA data. But we have to start somewhere.
The total labour turnover figure for academics is then easy to generate from HESA data. This represents a population of 211,980 academic staff across UK HE and shows academic staff starters and leavers by HE provider (HESA 2019). The mean and median figures are drawn up using the methods XpertHR use to calculate their staff turnover figures. It returns an average for UK HE that is over double the median reported by UCEA.
These recalculated figures show higher, fluctuating, rates of staff turnover across the sector over the last ten years. The average 2011/12 staff turnover rate across the sector was 17.23% (15.81% median), in 2014/15 it was 16.78% (14.85% median), in 2015/16 it was 16.78% (15.47% median), and in 2016/17 it was 16.67% (15.51% median).
These average turnover rates hide some pockets of incredibly high staff throughput. Below are some examples of the mean turnover figures in the 2017/18 HESA data returned by specific higher education providers (HEPs). Note that these figures still exclude atypical staff.
This institutional data (see Appendix) show some enormous shifts in staffing obscured by averages. In some cases, such as Ulster University, the majority of the academic staff during the 2017/18 academic year appear to have been changed. It is not clear what is going on behind this remarkable statistic and it needs to be looked at more carefully. But it is not the only one. In the period from 2014–2018 several HEPs experienced academic staff turnovers in excess of 25% in one or more academic year, sometimes year after year. These are likely to have significant and deleterious effects on staff across these organisations.
Step Two: situate comparisons
Now we have some more robust figures on staff turnover, we can redo the comparisons used to maintain claims of good and stable working conditions in UK HE. We do so acknowledging our figures are still partial in their focus on academic staff and exclusion of other UK HE workers. But what we have produced gets us closer to a truthful comparison with the XpertHR data used as a benchmark by UCEA. The picture is far less rosy than that presented by UCEA, who report just 8% academic staff turnover in UK HE.
This table raises further questions of what we need universities to be and do, and the organisations we want to compare them to. We find universities sitting on a par with manufacturing and production. Our rates of staff turnover have overtaken those in the public sector by some way.
The figures still show less turnover than the rapid rates in the private sector. But that is likely to be expected. And it starts to raise pressing questions over how people in universities are able to sustain the capacity for teaching and research on which many other industries depend.
In a global context, based on a sample of half a billion professionals, the LinkedIn Insights report provides an equally sobering set of comparisons. It also suggests UK HE has become an increasingly ‘liquid sector’. The LinkedIn methodology merits critical scrutiny, but it has potentially interesting headlines on which to base further research and reflection on how UK HE is being reimagined and redefined through its labour practices.
The LinkedIn study reports a worldwide professional staff turnover rate of 10.9%. This is calculated on the basis of the percentage of LinkedIn members who indicate they left an organisation in 2017, meaning the data have absences and its presentation disguises variations. The reported global turnover by sector ranges from 9.4% in health care to 13.2% in software. In this report, the combined category of ‘global government/edu/non-profit’ sit in the middle of the table at 11.2%.
In this global comparison, our recalculated academic staff turnover for UK HE appears high. It now exceeds global tech companies and is nearer to that experienced in restaurants (17.2%) and retail (16.2%). These comparisons might more accurately represent the experience of working in the software-defined and consumer-focused environment of UK HE than the UCEA figures with which we started.
Step Three: question everything
We are not the first to question how far UCEA figures mask the extent of casualised and precarious staff in UK HE. This was also the focus of the 2018 UCU report on Precarious Education. Kernohan’s (2018) piece ‘Who Works in a University?’ similarly asks questions around which workers (1) count as staff, and which do not; (2) which staff are counted by the HESA, and which are not, and (3) how much, or little, data needs to be provided around different staff categories. Answering these involves descent into a definitional hellscape of indeterminate data.
This is the final stage of our argument. To prevent the reproduction of strategic ignorance of this increasingly liquid and invisible workforce, we need to question everything. The more one delves into HESA definitions the more it becomes apparent that HESA figures provide an extremely partial view of the UK HE workforce.
Our analysis here is just the start. The diagram below attempts to show quite what a narrow population of the UK HE workforce is counted in UCEA’s headline figures. The failure of UCEA to reflect on the limitations of its official figures and discuss those workers they systematically exclude should raise serious concerns about UCEA as a representative body.
HESA figures exclude agency workers. There will be a veritable army of temp staff employed through recruitment agencies, such as Hayes Specialist Recruitment Ltd, Reed, Unitemps and so on, who move through the Higher Education sector without being counted. FOI data highlights that these figures are likely to be very significant in practice (e.g. Scott 2017; Michaels 2019).
Also excluded are self-employed and honorary staff who often play key roles in delivering projects by providing additional professional services without university contracts and creating vibrant intellectual environments.
PhD students are increasingly involved in the delivery of seminars, tutorials and other aspects of higher education. These individuals do vital work in departments, in often the most precarious contexts, but they do not count as staff as they are rarely employed on an ‘academic contract’. As, Kernohan (2018) states, ‘[w]ithout an academic contract you don’t get identified as academic staff in institutional data returns to HESA’.
The fixed-term staff, with whom we started this piece, are recorded by HESA but not reported at all by UCEA.
We can close our argument by illustrating the implications of this by focusing on the atypical staff — both academic and non-academic — who have such an uncertain status in UK HE staff turnover figures. Data on atypical non-academic workers is no longer collected by HESA (Kernohan 2018). Reporting of atypical academic workers is discretionary for institutions; some use this category to report and some do not. Given these ambiguities, perhaps we can conclude UCEA have sensibly decided not to use this dataset at all.
Yet we find the UCEA’s 2019 report turns precisely to this data to celebrate how, against ‘the direction of travel cited in the “casualisation” campaigning’ the ‘number of atypical (casual) staff has fallen by 16.1% since 2011–12’ (p.3). It does not mention here that academic atypical staff is a discretionary reporting category for HESA. It does not discuss how many members of casualised staff are returned within the category of fixed term staff by institutions instead, where they will not be counted by UCEA. These figures cannot be robust. They certainly do not counter the direction of travel cited in the ‘casualisation’ campaigning in UK HE. Step Three: question everything.
This USSbrief has looked at how UCEA trivialises a range of methodological concerns around the production and representation of staff turnover figures that are critical to any serious account of the UK HE workforce. UCEA appears to invest considerable energy in presenting staff turnover in a good light; better than is actually the case. Given the position of UCEA, the data and intelligence available to it, and the specific techniques it deploys to construct its crude staff turnover figures, our assumption is that UCEA is deploying strategic ignorance.
This analysis aligns with that of UCU (2018), which noted that, ‘[f]aced with press interest in precarious employment in their sector, and presumably keen to keep their subscribers happy, UCEA have attempted to cloud the issue by publishing a calculation that they claim shows there is no real issue in higher education’ (p.7).
This clouding of the state of the workforce in UK HE goes well beyond the active exclusion of fixed-term workers, though this constitutes a substantial proportion of staff in HE. UCEA uses statistics and benchmarking to obscure further patterns of staff precarity. We suggest these slips in reporting staff turnover from 2008 to 2019 are unlikely to be the result of clumsiness. They seem designed to provide UK HE with convenient figures to evidence fights against casualisation, support stories of stability, and protect the individual reputation of HEPs.
Overall, the HESA data from 2009 onwards demonstrates an overall growth of staff, which UCEA does not hesitate to tell us. However, what we do not see in UCEA figures is the extent to which the workforce is precarious, the extent to which layers of workers are shed each year, and the extent to which staff are left in tentative employment — waiting to learn whether their contracts will be renewed in the following academic year (or not).
While UCEA tells us UK HE is moving against the tide of casualisation, fuller analysis of the HESA figures tells a different story. We see that close to a fifth of the workforce of FTE staff (not including atypical staff) are lost each year as a matter of course. This is just the tip of the iceberg. It does not include atypical staff, agency staff, self-employed staff, or PhD students who are increasingly a critical part of the HE workforce, and many more. UCEA cannot argue these are issues with HESA data. Their decision to exclude further populations of the UK HE workforce in their presentation of staff turnover figures exacerbates the issue.
When we include fixed-term workers in the staff turnover figures, despite knowing that this fails to include a far larger number of UK HE workers, the overall results are catastrophic. Looking in detail at the staff turnover figures of individual HEPs reveals some institutions where the entire workforce are experiencing extraordinary uncertainty about their employment and this is happening year after year. These are critical issues that any future settlements need to take into account, as are measures to improve the accuracy and trustworthiness of the data on which they will be evaluated.
Appendix: Academic Staff Turnover from HESA data
This paper represents the views of the authors only. The authors believe all information to be reliable and accurate; if any errors are found please contact us so that we can correct them. We welcome discussion of the points raised and suggest that discussants use Twitter with the hashtag #USSbriefs88; the authors will try to respond as appropriate. This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.