Organisational Change and Employee Turnover
Kevin M. Morrell, John Loan-Clarke and Adrian J. Wilkinson

Total Word Count: 4515 K v Mor ls f ayaD c r R sa h t et iP Ditl ‘ dln ei r liai l er ot a eer Su n h h sie Moei n e n ol c d ,s td lg E p ye unvr k.m.morrell@lboro.ac.uk m l e T roe , o ’

John Loan-Clarke is a Lecturer in Organisational Behaviour, j.loan-clarke@lboro.ac.uk

Professor Adrian Wilkinson is Professor of Human Resource Management, a.j.wilkinson@lboro.ac.uk

All authors are based at: Loughborough University Business School, Loughborough LE11 3TU, UK.

Key Words: Employee Turnover, Organisational Change, NHS

1

Organisational Change and Employee Turnover

Abstract This paper investigates employee turnover, using data from a recent study of 352 NHS nurse leavers. We explore individual decisions to quit in a context characterised by change, and suggest a mechanism whereby organisational and contextual change can prompt individual leaving decisions. We develop and test hypotheses designed to shed light on the links between organisational change and individual decisions to quit. We then develop a theoretical, heuristic model describing the relationship between organisational change and turnover. The managerial implications of this model are outlined and the paper concludes with an agenda for future research.

Introduction Employee turnover is a much studied phenomenon (Shaw, Delery, Jenkins and Gupta 1998: 511). Indeed, one recent meta-analysis (Hom and Griffeth: 1995) alone reviewed over 800 such studies (Iverson 1999). However, there is no universally accepted account for why people choose to leave organisations (Lee and Mitchell 1994), even though it is predominantly in instances where the employee makes the decision (rather than in cases of involuntary turnover) that organisations and organisational theorists have an interest. Voluntary turnover is of interest because in most cases, this represents the bulk of turnover within an organisation. Such instances of turnover also represent a significant cost, both in terms of direct costs (replacement, recruitment and selection, temporary staff, management time) but also, and perhaps more significantly, in terms of indirect costs (morale, pressure on remaining staff, costs of learning, product/service quality, organisational memory) and the loss of social capital (Dess and Shaw 2001).

Although there is currently no accepted framework for understanding the turnover process as a whole, a wide range of factors have been found useful when it comes to interpreting employee turnover, and these have been employed to model turnover in a range of different organisational and occupational settings. These include: job satisfaction (Hom and Kinicki 2001); labour market variables (Kirschenbaum and Mano-Negrin 1999); various forms of commitment (see Meyer 2001 for a review); equity (Aquino, Griffeth, Allen and Hom 1997);

2

psychological contract (Morrison and Robinson 1997) and many others (see Morrell, LoanClarke and Wilkinson 2001 for a review). However, there is little research specifically exploring the link between organisational change and turnover and we suggest that this is a gap in the literature. No-one would seriously challenge the idea that mismanaging organisational change can result in people choosing to leave. Indeed, it may even result in the highest performing (and therefore the more employable) employees leaving (Jackofsky, Ferris and Breckenridge 1986). However, explaining the mechanisms underlying how and why such change can result in differential rates of turnover is more open to question.

The Unfolding Model Of Voluntary Employee Turnover We develop and apply a recent, influential account of the turnover process (Lee, Mitchell, Holtom, McDaniel and Hill 1999) to the problem of how organisational change can influence i i dadc i so u . e ea s19)ufl n m dlo vl t yunvr n v uleio tqi Le tl (99 ‘no i oe f o n r t oe di sn t ’ dg ’ ua r represents a divergence from traditional thinking (Hom and Griffeth 1995), by focusing more on the decisional aspect of the phenomenon, and indeed it is based on a theory of decision making (Beach 1990). The underlying premise of the model is that people leave organisations in different ways, and it outlines five pathways describing various decision processes any one of which a leaver may go through before finally quitting. This represents a i poe etvr n a sacut w i a r tc d o ne t d g lpol s nm rvm noeuirtcon , h h r e r t t udra i a ep ’ ti s c e sie sn n l e decisions as influenced by the same factors and considerations, and this approach is therefore a significant departure from the founding fathers of modern research into turnover (March and Simon 1958) as well as from other influential thinkers (Porter and Steers 1973; Price 1977). L e tl t oy l e pai sh rla i l j r g vn cn l ism pol s e ea sh r a o m hs e t o s g , rn eeta p yn o e ep ’ ’ e s s e e n e ai a e decisions to quit, and it is this – rather than the model as a whole – which forms the theoretical focus for this paper. They refer to this event as a shock, though it is important to e pai t t sok ed ob uepc d Asok s ec bd s eesr t ‘ ae m hs eh a hc ne nte nxet . hc idsr e a ncs yo s k s a e i a ,h e p ye f mt ilhry (e ad t e 19: 1)r l t gh i ahtuk m l esr h re a ’L e n Mihl 91 18,e e i t d t sn o o e t g c l f cn e e a costs (Becker 1960), inertia (Mercer 1979) and a wish to remain in employment (Sheridan and Abelson 1983) have each been found to be important factors influencing turnover. We suggest that a better understanding of the role shocks play in precipitating decisions to quit is key to understanding the relationship between organisational change and employee turnover.

3

This is because the notion of shock offers a way to understand how there can be linkages between change at the level of structures and social settings on the one hand, and individual agency on the other. In other words, to understand how organisational change influences employee turnover.

This approach is helpful to our understanding of organisational change management because the evaluation and measurement of change initiatives involves balancing two considerations.

Firstly, in terms of the external context, assessing change initiatives involves determining the necessity for the imposition of change at an organisational level.

Secondly, in terms of internal resources, for most organisations it will be important to determine the likely impact such change has for individual employees, and to manage the consequences of this change at various stages during the process.

The second point is particularly important to consider if such change leads to increased turnover and a loss of social capital, which may be critical to organisational success (Dess and Shaw 2001). Understanding the role that shocks typically play in employee turnover can improve evaluation of the impact of change on individual employees. In the light of recent research, and in light of the findings introduced in this study, we argue that managers of change can benefit from employing an understanding of shocks in two ways. Firstly, seeing shock as the first stage in many leaving decisions gives managers a useful heuristic device to think about intervention, in other words, to stop people leaving. Secondly, assessing the incidence and type of turnover prompted by these shocks can enhance the ability of managers and organisations to monitor change.

Avoidable Turnover One means of diagnosing the amount of influence organisations have over turnover, is to look at et to h h eiosoev a dsr e a ‘ o al b l vr(a p n th x nt w i dc i tl e r ec bd sa i b ’ ye e C m i e e c sn a e i vd e a s o 1991; Morrell et al 2001). In other words, is it a case of employee instigated turnover that could have been prevented? We suggest it is particularly important to assess and understand this during a period of change, because employee turnover can be used as an index of organisational health. Supplementing a crude measure of turnover, such as the base rate -

4

(number of leavers in a year / average number of employees in a year) * 100 - with a measure of avoidability, can inform more effective management of employee resourcing.

For example, if a firm can identify that the bulk of their voluntary turnover is unavoidable, they may profit better from initiatives that seek to manage turnover post hoc, such as by streamlining recruitment processes, rather than spend on theorised preventative measures, such as increasing pay. We might call this a control model. On the other hand, if the bulk of turnover is avoidable this offers the potential for directed intervention - a prevention model.

If organisations introduce change and experience a resultant increase in turnover, it is important for them to be able to identify whether this change is typically avoidable, or unavoidable in order to manage it effectively. Determining this will enable them to manage better the trade off between attending to the competitive context on the one hand, and maintaining internal capability on the other. This can be illustrated with three hypothetical scenarios:

First, if turnover has increased as a result of the implementation of change, and this turnover is mainly unavoidable (i.e. the organisation could not influence it because the change has happened), then - bearing in mind that turnover results in substantial indirect costs - an organisation can quickly calculate some measures of the cost of the change, setting these against the supposed benefits. In a sense, this represents an ideal scenario, one where the internal impact of change is easy to identify and to understand, and where it is simple to calculate cost-benefit. These leavers represent the proverbial eggs in the omelette, or to use a m la aa g, e cn e ec bd sncs r csai ’ O cus im ngri it y nl yt y a b dsr e a ‘ees y aule . f or f aae n ir o h i a ts e s the organisation are unaware that these casualties are unavoidable, they may try spurious initiatives designdo e i t m r r et g w s o r or so ‘ ai sao s e tr a h ,e e n n a at fe uc ,rc s g hdw ’ tn e p s i e s e h n .

Second, if turnover has increased as a result of the implementation of change, and the levels of avoidable and unavoidable turnover are approximately equal, then it will be beneficial to look more closely at the phenomenon and uncover those areas where intervention will result in lower levels of avoidable turnover. This represents a mid-point, where the internal impact of change is difficult to understand, but signalling substantial room for improvement. C n ni t m la aa g, t snt c cag cu r u i ‘ncet ll ss ot u gh it y nl yi h i a e hne ol e l n uacp b o e’ i n e ir o n i sn d st ae s if managers pursued either a pure control or prevention paradigm. On the other hand, where

5

it is possible to correctly identify patterns of turnover and control the costs of some unavoidable turnover, while minimising some instances of avoidable turnover, this is aa gu t m ng g f cvl a e i t ‘ g f a . nl oso aai e et e ,l i n h f o w r o n f i y bt e o ’

Third, if turnover has increased as a result of the implementation of change, and it is predominantly avoidable, then this implies that the process is being mismanaged, and that an organisation is passing up on the chance to retain its staff. In this instance change could r u i t ‘ a e fh Lgt r ae a futile and needless loss of valuable employees. e l nh c r o t i Bi d’ st e h g e h g , However, if change managers can identify and successfully intervene in those areas that would otherwise lead to avoidable decisions to quit, that would represent a notable victory, and the effects of chane ol b m t a d gi t f cvl ‘r g g h t os ak g cu e igt aa s e et e bi i t r p bc d i e n, f i y n n e o hm ’ oe .

To recap, if turnover is generally avoidable, this offers the potential for directed intervention, and thereby prevention. If it is unavoidable, it will be better to concentrate on managing the phenomenon by reducing its cost, and thereby control turnover after the event. As there is the potential for this process to be disastrously mismanaged, we suggest that there is a need for organisations to assess patterns of avoidability in the overall profile of employee turnover. This level of measurement is needed in order not to incur unnecessary losses, or wrongly try to prevent something when resources would be better spent managing the consequences. This is illustrated below:

Take In Figure 1

Having outlined the theoretical background to the research, we can now move on to discuss the empirical elements of this study.

The Study We tested the unfolding model by studying the leaving decisions of 352 NHS nurses, using a slightly mod i vro o L e tl (99 qet na e L e tl a s d d 2 ie e i f e ea s19) usoni . e eahd t i 29 fd s n ’ i r ue accountant leavers in the US and so some changes in the questionnaire were necessary to reflect differences in national and organisational context. Other changes were informed by a short pilot of the questionnaire with 15 nurses and midwives, and we made some additional

6

improvements based on a theoretical critique of the model. Our original sample frame comprised voluntary leavers in the financial year 2000-2001, at eight NHS Trusts. The Trusts were drawn from three regions and three of the larger Trusts had recently undergone mergers, with one other facing the prospect of merger in the near future. Another of the Trusts was undergoing a substantial programme of development, including the building of a new multi-storey wing. T ee rs w rntn n w y hsn sr r eti ’ fh N Sa a hl Iw u hs Tut e oi ay a coe a ‘ pe n t eo t H s w o . t ol s e e s av e e d not be possible to do this with just eight Trusts in any case, moreover, as we are exploring the processes involved in turnover decisions, the unit of analysis is the individual leaver. Nonetheless, taken together the Trusts represent a diverse range in terms of location, size and t e T e a fu m d m s e Tut w i a ec ‘ r ’ in the sense they are not y . hr r or ei i d rs , h h r ah r a – p e e u z s c e ul based exclusively in a large city, and four large acute Trusts, each of which comprises a teaching hospital or hospitals. A total of 1,190 surveys were sent out via the Trusts, of these, 368 were returned during the period from the last week in April 2001, to the first week in September 2001. Sixteen surveys were excluded from the analysis because the respondent was not a nurse (two cases), or because the turnover was involuntary (ten cases), or because there was too much missing data to be able to analyse the responses (four cases). The final sample size is thus 352. Taking into account those surveys that were wrongly addressed and returned, this represents an overall response rate of 31%, which is significantly higher (p