10
Higher Education 33: 29–38, 1997. 29 c 1997 Kluwer Academic Publishers. Printed in the Netherlands. Is job sharing worthwhile? A cost-benefit analysis in UK universities GEOFF HARRIS Department of Economics, University of New England, Armidale, NSW 2351, Australia Abstract. Around 60 per cent of UK universities practice job sharing. This article is based on a survey of Personnel Directors in UK universities concerning their perceptions of job sharing. These responses were then used to carry out a cost benefit evaluation of job sharing from the universities’ perspective. If productivity of workers in the shared job rose by as little as 0.35 per cent – one third of one per cent – the estimated increase in Personnel Section costs would be covered. If there was a 5 per cent increase in productivity, the ratio of benefits to costs would be 14.3 to 1. Universities also save as a result of greater retention of staff, and there are also important benefits to society, particularly less overall stress and reduced unemployment levels. Introduction Job sharing may be defined as an arrangement where two or more people voluntarily share a full-time position and are entitled, on a pro-rata basis, to all the conditions of salary and other benefits of the position (Walton 1990, p. 7). Depending on the nature of the job and the preferences of the sharers, the job may be shared according to duties and/or time. A period of overlap is typically necessary for communication between the sharers. The distinction between job sharing and part-time work is an important one. In job sharing, the benefits of the full-time post are retained whereas part-time employees typically receive lower wage rates and non-wage benefits and have fewer opportunities for training and promotion. The motivation for job sharing originated from equal employment oppor- tunity considerations. At particular times in their working lives, women may find it more convenient to work less than full time. Job sharing is a way of allowing women to both continue in employment and to spend more time with young children. Another motive may occur for employees approaching retirement; they may wish to begin involvement in retirement pursuits whilst maintaining some employment. A motive for the provision of job sharing in the public sector, including universities, is related to recent increases in stress levels associated with public sector reform, which job sharing may reduce. A number of studies

Is job sharing worthwhile? A cost-benefit analysis in UK universities

Embed Size (px)

Citation preview

Higher Education 33: 29–38, 1997. 29c 1997 Kluwer Academic Publishers. Printed in the Netherlands.

Is job sharing worthwhile? A cost-benefit analysis in UKuniversities

GEOFF HARRISDepartment of Economics, University of New England, Armidale, NSW 2351, Australia

Abstract. Around 60 per cent of UK universities practice job sharing. This article is based ona survey of Personnel Directors in UK universities concerning their perceptions of job sharing.These responses were then used to carry out a cost benefit evaluation of job sharing from theuniversities’ perspective. If productivity of workers in the shared job rose by as little as 0.35per cent – one third of one per cent – the estimated increase in Personnel Section costs wouldbe covered. If there was a 5 per cent increase in productivity, the ratio of benefits to costswould be 14.3 to 1. Universities also save as a result of greater retention of staff, and there arealso important benefits to society, particularly less overall stress and reduced unemploymentlevels.

Introduction

Job sharing may be defined as an arrangement where two or more peoplevoluntarily share a full-time position and are entitled, on a pro-rata basis, toall the conditions of salary and other benefits of the position (Walton 1990,p. 7). Depending on the nature of the job and the preferences of the sharers,the job may be shared according to duties and/or time. A period of overlap istypically necessary for communication between the sharers. The distinctionbetween job sharing and part-time work is an important one. In job sharing,the benefits of the full-time post are retained whereas part-time employeestypically receive lower wage rates and non-wage benefits and have feweropportunities for training and promotion.

The motivation for job sharing originated from equal employment oppor-tunity considerations. At particular times in their working lives, women mayfind it more convenient to work less than full time. Job sharing is a way ofallowing women to both continue in employment and to spend more timewith young children. Another motive may occur for employees approachingretirement; they may wish to begin involvement in retirement pursuits whilstmaintaining some employment.

A motive for the provision of job sharing in the public sector, includinguniversities, is related to recent increases in stress levels associated withpublic sector reform, which job sharing may reduce. A number of studies

30

have documented increases in stress amongst university staff across a rangeof countries.1 One recent study (Daniels and Guppy 1994) tested for stressors(i.e. demands which cause stress) amongst 221 staff across all categories at oneUK university. They concluded that their sample’s psychological well-being‘may be very poor’ (1994, p. 138). Inadequate resources, having too muchwork to do and a feeling that the organization does not care for its staff werethe most commonly reported stressors. More detailed analysis found twounderlying dimensions to be crucial – work load/managerial stressors (thelatter largely associated with attending meetings) and role stressors (employ-ees’ uncertainty as to what was expected of them). The former stressor wasmost common amongst academics and the latter amongst secretarial/clericaland technical/engineering staff.

Such stress is costly. A study of six large Australian government organi-zations in 1990–91 (Toohey 1993) found that stress-related workers’ com-pensation claims made up 4.6% of all claims but 18% of all costs, averagingaround £15,000 per claim. This is, however, only a small part of the total costsof stress. The bulk of stress costs are less visible and occur in such forms asabsenteeism and sick leave, reduced work performance, service complaintsand staff turnover. The most common causes of occupational stress werefound to be workload (26%), trauma (10%), conflict with supervisors (9%)and forced relocation or deployment (9%). The annual cost to UK industry ofreplacing staff who underperform due to stress has been put at around £3000million (Cox et al. 1990, cited in Toohey 1993, p. 9). Job-sharing may be oneway of reducing stress (perhaps particularly that associated with workload)and its explicit and implicit costs to employers and to society.

There are reasons to expect an increase in productivity as a result of jobsharing. In section 2.2, the respondents mention greater flexibility, continuityand retention of skilled personnel. In addition, Walton (1988, pp. 76–77) citesevidence pointing to two other sources of productivity increase. First, there isan enhanced range of skills and experience brought to the job. Second, the totalenergy, motivation and commitment is greater; both job sharers and employersreport that workers are more energetic and creative during the hours they areat work, compared with full time workers. Walton also lists (pp. 78–79) somecosts of job sharing of which increased administration, greater training costs,and additional accommodation requirements during any time of overlap arethe most important. The evidence she cites, however, suggests that employersregard these as insignificant. The New Ways to Work (1987) survey found themost frequently mentioned problem areas were overtime payments, blockingof promotion for full-timers and public holidays. Walton comments (p. 79)that these are often the result of lack of specific guidance on how to deal withthe issues rather than being inherent problems. It should be recognised that

31

productivity might diminish if a high degree of cooperation and sharing ofinformation between the job sharers is not achieved.

Given this background, this article presents the results of a survey in UKuniversities with respect to attitudes towards and perceptions of job sharing.2

These data are then used in a cost-benefit evaluation of job sharing.

The survey

The data on which this article is based were collected by a questionnairesent to the Directors of Personnel in each of the 117 UK universities3 in July1994, with a follow up letter to non-respondents in October. A number ofthe questions were derived from a questionnaire employed in a study of localauthorities in 1986, the results of which are reported in New Ways to Work(1987). Questions were asked about whether the university had job sharing,and, irrespective of whether or not job sharing occurred, opinions were soughtconcerning its advantages and disadvantages and any additional costs which itmight impose. A total of 78 useable questionnaires were returned, a responserate of 66.1%. Of the 78, 51 were from old and 27 from new universities, andthe response rates from the two groups was very similar. Nineteen (24%) ofthe responding universities had a formally adopted policy on job sharing anda further 21 (27%) had present or future plans to put forward such a policy.With or without a policy, however, an impressive 61 (78%) of the respondinguniversities employed job sharers.4 For those with policies, the initiative hadmost often come from the Personnel Section (63% of cases) and typicallycovered all employees – academic, administrative and technical. In all butone case, the policy had received trade union support. The mean numberof jobs shared per university was 10.5; three universities had more than 20jobs shared, including one with 60. All but a handful of these jobs had twojob sharers, and the overwhelming proportion of sharers (96%) were female.Eighty six per cent of the posts were administrative, 10% were academic and4% were technical.

These data were supplemented by a follow-up questionnaire faxed to 30randomly selected universities in May 1996, which focussed on the possibleincrease in productivity as a result of job sharing and the rate of departure fromjob shared jobs as opposed to full time jobs. Twenty one useable responseswere received. Copies of both questionnaires are available on request to theauthor.

32

Table 1. Perceived advantages, problems and costs of job sharing

Number of responses

ProblemsLiaison between sharers 13Finding a replacement if one sharer left 11Payment for overtime 9Arranging public holidays 8Training costs 7

AdvantagesGreater flexibility 21Increased productivity 18Retention and recruitment of skilled staff 17

CostsSetting up costs 14Recruitment and administration of particular jobs 14

Problems

Respondents’ views as to potential or actual problems associated with jobsharing were sought. The two major potential or actual problems identified(see Table 1) were the necessity of close liaison and good communicationbetween the job sharers (13 respondents) and the potential difficulties offinding an appropriate replacement if one job sharer was to leave (11 respon-dents). A few respondents referred to the potential problems associated withpayments for overtime, public holidays (an issue when each person workedfixed days each week) and additional training costs. One respondent’s com-ment summed up the predominant view: ‘Sure there can be problems ofcommunication and so on, but most of our people are so happy to be sharingthat they take the time to do the appropriate talking with their partner. Maybeour Appointments people made some really good choices!’

Advantages

Questions were asked concerning the principal perceived advantages of jobsharing arrangements and these attracted more responses than those focussingon problems or costs. The main advantages were the greater flexibility andcontinuity which job sharing allows (e.g. at least part of the job is coveredwhen one person is away sick or on holidays, and both people may be availableduring peak periods), the possibility of retaining (and to a lesser extent attract-ing) skilled personnel who might otherwise be lost, with potential savings inrespect of recruiting and assimilating new staff, and greater productivity. In

33

some respects, increased productivity flows from the greater flexibility andretention, but it is also additional to these. Typical of the comments explain-ing this greater productivity were that job sharing doubles the range of skills,ideas, energy, and experience, that it allows for the ‘synergy of two committedindividuals’, and leads to ‘greater job commitment because people are doingwhat they want – both hours worked and the bits of the job they concentrateon’.

Costs

Two main costs were identified: the administrative and recruitment costsassociated with shared jobs and the setting up of any particular job, witharrangements to be made for two people. Few respondents attempted to esti-mate the additional input in person hours required by the Personnel Sectionto deal with a job shared post. For the 13 responses, however, the mean addi-tional input was a mere 3.2 person hours; a further 17 suggested a negligibleadditional input and three that it was considerable e.g. twice as much as for anon-shared post.

Cost benefit analysis

Theory

Armed with the forgoing information, we can now undertake a basic eval-uation of job sharing. Cost benefit analysis is a well established methodin economics of evaluating projects or programs. In brief, four steps areinvolved:1. Identifying all the relevant benefits and costs.2. Where possible, placing a value on each of these.3. Where the project or program operates over time, discounting the stream

of future net returns to bring them to a common time period (convention-ally, the present).

4. Calculating measures which the outcome of the project on program: netpresent value, internal rate of return or benefit cost ratio.

Evaluations can occur from a number of perspectives – the individualparticipants, the employing institution, the government – or society as awhole. Table 2 summarizes the principal benefits and costs from each ofthese perspectives. A further advantage of job sharing mentioned in the tableis that it ultimately increases the total numbers in employment. Resourcebenefits and resource costs represent additions to output (e.g. as a result ofgreater productivity) or additional costs (e.g. the use of additional resources

34

Table 2. The benefits and costs of one job sharing arrangement

Resource benefits Resource costs Transfers

Previouslyfull timeworkers

Time to do ‘other things’.Reduced stress fromwork

Gross income falls by£7500.Tax saved of £35001

Previouslyunemployedworkers

Reduced stress frombeing unemployed

Reduced time to doother things

Gross earned incomeincreases by £7500.Unemploymentbenefit falls by£3000.Tax paid of £1500.2

Employers Greater flexibility.Increased productivity.Retention andrecruitment of skilledstaff

Cost of additionalinputs inorganizingand administeringjob sharing

Increased tax as aresult of higher profits

Government Saved medical costs as aresult of reduced stress.Saved social costsof unemployment

Savedunemploymentbenefit payments of£3000.Tax revenue falls by£2000.3

Increases in taxrevenue from highercompany profits.

Society Reduced stress in society.The value, short term andlong term, of the ’otherthings’.Saved medical and socialcosts

Organizational andadministrative costs

1That is, tax on £15,000 at 33.3% minus £7500 taxed at 20%. The overall impact on the incomeof the previously full time worker is a reduction of £4000.2The net result is an increase of disposable income of £3000.3The government gains £1000 as a result of these transfers.

to produce a given cost). Transfers cancel out in that costs to one part ofsociety (e.g. taxpayers) are cancelled out by an identical gain to another(government). Transfers are not included in cost benefit calculations althoughthey are important from a distributional point of view. The figures are basedon the assumptions of one full time job being paid at £15,000 per annumbeing shared between two people, tax rates of 33.3% on £15,000 and 20% on£7,500, and an unemployment benefit rate of £3000 per annum which is nottaxed.

35

Table 2 lists the resource costs, resource benefits and transfers for eachrelevant group in society, and for society as a whole. It shows, given theprevious assumptions, that a previous full time worker will face a reductionin net income of £4,000 (down to £6,000 from £10,000); a previously unem-ployed person (whom we are assuming now shares the job) ends up witha net gain of £3,000; and the government, in a short term budgetary sense,gains £1,000.5 Two further points should be noted from this table. First, thekey issue is whether employers judge it to be in their interests to offer jobsharing arrangements. Second, there are important longer term gains to soci-ety in terms of the saved costs resulting from a less stressed society and asa result of lower unemployment rates. Another possible intangible and longterm benefit is that of better adjusted children as a consequence of more timewith the job sharing parent.

Some calculations

In this article, attention is focussed on the employing universities. Insofaras individuals wish to be involved in job sharing, they clearly perceive thepersonal benefits to exceed the costs. We now attempt to put some values on thecosts and benefits to universities identified earlier. It is, of course, extremelydifficult to estimate the impact on productivity occurring as a result of jobsharing. The approach used here was to estimate the breakeven productivityincrease i.e. the increase required to compensate for the increase in PersonnelSection costs. If we double the estimate of such costs to 6.4 hours, to be onthe safe side, this amounts, at £8.20 per hour (i.e. £15,000 � [52 weeks �35 hours per week]) to £52.50. The breakeven productivity increase requiredis therefore 0.35 or a third of one per cent (i.e. £52.50 � £15,000).6 In thesupplementary questionnaire, we asked whether such productivity increasewas considered likely to occur. Of the respondents, 14 regarded this as verylikely to be achieved and three as likely to be achieved; four did not know.Given that a productivity increase was discernible to a number of respondents,it seems possible that a productivity gain of five or even ten per cent might beachieved. If a productivity gain of five per cent was indeed achieved, the ratioof benefits to costs, still assuming 6.4 hours of additional Personnel Sectiontime, would be 14.3 (i.e. £15,000� 5%� £52.50). These, it should be noted,would be minimum benefit figures: the university would also gain from thegreater flexibility of its work force and would save as a result of retainingstaff.

There are two cost side considerations: the costs of setting up and admin-istering a job sharing policy and the costs related to the recruitment andemployment of sharers for a particular post. As to the cost of setting up a par-ticular job, we have one figure – 3.2 hours of additional input by the Personnel

36

Section. This is an important figure and deserves careful consideration. Theinstitutional input into job sharing can be broken up into two parts – setting upa job sharing policy, and appointing and dealing with the individuals involvedin a particular job sharing arrangement. The first of these, being motivatedby equal opportunity considerations and being a one-off event, cannot be rea-sonably counted as a cost of filling individual job sharing positions. Filling aparticular position involves the following five aspects to University PersonnelSection:

� involvement in the process of approval to fill a position

� advertising

� shortlisting of applicants

� interviewing/selection/offering appointment

� signing on new appointee(s)

We can identify two types of job shared positions, one fully vacant andone half vacant. In the former case, the only additional costs in filling a jobshared position are that two people will have to be selected (although around10% of universities with job sharers required applicants to apply as a team)and signing on two rather than one appointee. A half vacant post may occureither because a previous sharer leaves or because the incumbent full timerwishes to share. In the case of a previous sharer leaving, the remaining shareris usually offered the job on a full time basis (New Ways to Work 1987, p. 24).If they decline, all five steps would be involved in filling the half vacancy.The crucial question is whether half vacancies in shared jobs are more likelyto occur than full vacancies in full time jobs. Whilst there are two peopleand therefore more likelihood of departure, the attractions of job sharingmay mean lower departure rates. The supplementary questionnaire asked aquestion concerning rates of departure. Fifteen of the respondents reportedno difference between shared and full time jobs, and two respondents eachreported more frequently, less frequently and don’t know.

Taken as a whole, this discussion suggests that there will only be additionalPersonnel Section input in the case of half vacant posts where the vacant halfis not taken up by the other sharer, or if the rate of departure from sharedposts exceeds those from non-shared posts. To be on the conservative side,we shall double the estimate to 6.4 hours which amounts, at £8.20 per hour(i.e. £15,000 � [52 weeks � 35 hours per week] to £52.50. Assuming a 5%increase in productivity, these values result in a ratio of benefits to costs fromjob sharing of 14.3 (i.e. £750 � £52.50). Only if the additional PersonnelSection input amounted to 91 hours would the benefit of a 5% productivitygain be outweighed.

37

Conclusion

There are three variables of crucial importance in reaching a decision aboutthe worth of job sharing, viz., changes in productivity, additional PersonnelSection costs and the rate of departure of job sharers from their jobs comparedto full time workers. Given the difficulty of estimating the first, this articlehas estimated the increase in productivity required to break even, given thelikely increase in Personnel costs. The majority of follow up respondentsconsidered that such a productivity increase would be very easily attainable.Rates of departure of job sharers were reported to be little different to thoseof full timers. These data give grounds for confidence in believing that jobsharing is worthwhile for universities. One point which has not been examinedare the impacts on other stakeholders, such as work colleagues and students,and this could be a focus for further research.

The foregoing indicates that job sharing, as well as being beneficial toindividuals who choose to participate in it, is very likely to be advantageous toemployers, government and society. It is puzzling, then, why some employersoppose it. In the survey, a significant minority of respondents expressed theirdisapproval with the concept of job sharing – ‘it is not managerially sound’said one – although this was at times not consistent with their university’spolicy.

Such a view is also apparent in the study of Sidaway and Wareing (1992)who found that in most firms it is customary to consider that jobs should beproved suitable for part-time work, including job sharing, before this practiceis allowed. One exception is British Gas, where jobs have to be provedunsuitable for part-time work before a part-timer is refused them. Jobs whichwere task-related, technical and discrete or self-contained, were consideredmost suitable. On the other hand, managerial, supervisory and marketing jobswere frequently mentioned as unsuitable because they depend on personalcontacts, frequent communication, continuity and ad hoc problem solving.The study showed, however, that part-timers successfully managed thesetypes of ’unsuitable’ jobs, and the authors concluded that ‘too many jobs areregarded as unsuitable for these part timers. Without arguing that all jobs canbe done on a part-time basis, it does seem clear that many actually can’ (1992,p. 26), a point also supported by the experience of local government (NewWays to Work 1987, p. 25). As one university’s job sharing policy puts it ‘Allposts will be considered open for job sharing unless special circumstancesrequire otherwise’.

Two final voices in support of job sharing may be noted. The Associationof University Teachers is ‘strongly in favour of job sharing’ and has triedto encourage its more widespread take-up in universities (Personal commu-nication, May 2, 1994). The report of its Joint Working Party on Part-TimeUniversity Staff 1994 states that ‘job-sharing shall be encouraged, as the best

38

form of part-time work’ (para. 7). The CVCP’s Commission on UniversityCareer Opportunity (CUCO) surveyed UK universities in December 1993,and found 59% to have adopted job sharing schemes (CUCO 1994, p. 22).This followed an earlier report (CUCP 1991, p. 9) which recommended jobsharing schemes as a means of promoting equity and efficiency.

Notes

1 Recent examples include surveys of stress at individual universities e.g. Salford (The TimesHigher Education Supplement November 4, 1994, p. 3), London Guildhall (The Times HigherEducation Supplement November 4, 1994, p. 48), an unnamed British university (Daniels andGuppy 1994) and the University of New England, Australia (Dua 1994); and broader surveyse.g. the National Foundation for Educational Research survey commissioned by NATFHE (TheTimes Higher Education Supplement November 4, 1994, p. 5).2 Previous studies of job sharers include Solman et al (1986) (Australian teachers), and Chan-dler and Johnson (1989) (Canadian civil servants). The main study of job sharing in practiceis the New Ways to Work (1987) study of job sharing in local authorities in the UK.3 The seven colleges of the University of Wales and the 22 colleges which make up the Uni-versity of London were treated as separate institutions. The source was the Association ofCommonwealth Universities’ Commonwealth Universities Yearbook 1993.4 Another survey (CUCO 1994) found that 63 of the 106 responding universities (from a totalof 109) stated that they had job sharing. Since those universities responding to the presentsurvey were more likely to have job sharing, and the overall proportion of universities with jobsharing is probably around 60%.5 This reasoning is based on the assumption used by conventional economists that wages are ameasure of productivity.6 It is possible that the unemployed person was not in receipt of unemployment benefits, inwhich case their net income would rise by £6,000 and government tax revenue would rise by£1,500.

References

Chandler, F.J., Johnson, J.A. (1989). ‘Attitudes towards jobsharing: a case study of the Employ-ment and Immigration Commission’, Canadian Public Administration 32, 633–640.

Commission on University Career Opportunity (1994). A Report on Universities’ Policies andPractices on Equal Opportunities in Employment. London: CUCO.

CVCP (1991). Equal Opportunities in Employment in Universities. London: CVCP.Daniels, K. and Guppy, A. (1994). ‘An exploratory study of stress in a British University’,

Higher Education Quarterly 48, 135–44.Dua, J. (1994), ‘Job stressors and their effects on physical health, emotional health, and job

satisfaction, in a university’, Journal of Educational Administration 32, 59–78.New Ways to Work (1987). Job Sharing: Putting Policy in Practice. London: New Ways to

Work.Sidaway, J. and Wareing, A. (1992). ‘Part-timers with potential’,Employment Gazette, January,

19–26.Solman, R., Petty, M. and Laughlin, A. (1986). ‘Job sharing in teaching: attitudes of teachers

employed by the New South Wales Department of Education’, Australian Journal of SocialIssues 21, 126–142.

Toohey, J. (1993). Quality of Life Working Project. A Study of Occupational Stress inCommonwealth Government Agencies. Canberra: Australian Government Publishing Ser-vice/Comcare Australia.

Walton, P. (1990). Job-sharing: A Practical Guide. London: Kogan Page.