British Journal of Social Work (2022) 52, 1683–1702

doi: 10.1093/bjsw/bcab148
Advance Access Publication July 19, 2021

Downloaded from https://academic.oup.com/bjsw/article/52/3/1683/6327302 by City University of Hong Kong user on 23 June 2023
Social Worker Turnover under the Lump
Sum Grant Subvention System in Hong
Kong: Organisation-Level Analyses
Haijing Dai *, Niantao Jiang and Ruobing Li

Department of Social Work, The Chinese University of Hong Kong, Shatin N.T.,
Hong Kong

*Correspondence to Haijing Dai, Department of Social Work, The Chinese University of

Hong Kong, Shatin, N.T., Hong Kong. E-mail: [email protected]

Abstract
Increasing turnovers of social workers under the managerialist system of service sub-
vention in Hong Kong have aroused much concern, but systematic and evidence-
based investigation of the problem remains limited, particularly at the organisational
level. Using organisation-level data collected from the managers of 101 service organi-
sations, this study examines how organisational characteristics and management prac-
tices associate with the turnover rates of social workers. The results show that starting
salaries of social workers, income inequality between social workers and managers
and service types of organisations affect the odds of emerging social worker turnover.
Meanwhile, hiring practices that replace social worker positions with lower-paid jobs
and lack of peer support in larger service organisations increase the odds of both
emerging and severe turnovers of social workers. Differentiated employment terms
and income stratification amongst social workers in the same organisation, however,
do not have a significant relationship with social worker turnover. Our findings sug-
gest that social workers tend to demonstrate solidarity as a professional group under
the managerialist welfare and service administrations in Hong Kong. Organisational
justice in salary distribution and organisational respect for their professional values
and team spirits may retain them in service delivery for longer times.

Keywords: Hong Kong, managerialism, organisation-level analysis, social worker,

turnover, welfare reform

Accepted: June 2021

Frequent turnover of front line social workers is a shared problem in so-

cial service provision worldwide. To promote effective, consistent and

# The Author(s) 2021. Published by Oxford University Press on behalf

www.basw.co.uk of The British Association of Social Workers. All rights reserved.
1684 Haijing Dai et al.

high-quality service intervention, it is of vital importance to keep a healthy

and stable workforce (Barak et al., 2001), but the demanding job nature of
social work often leads to staff burnout and makes it difficult to retain so-
cial workers in service organisations (Wilke et al., 2017). Much ink has

Downloaded from https://academic.oup.com/bjsw/article/52/3/1683/6327302 by City University of Hong Kong user on 23 June 2023
been spilt to investigate how personal, work-related, professional and
organisational life aspects of social workers affect their intention to turn-
over (Shier et al., 2012), and organisational environment and climate factors
are often emphasised in recent studies (Middleton and Potter, 2015).
In Hong Kong, social service is organised and delivered in a wide
range of organisations, including local and international, non-profit and
for-profit and traditional and innovative organisations. Organisations en-
joy the freedom to define their service areas and goals and make their
own employment plans on human resources. Social workers could be the
primary service professionals in some organisations, but in other organi-
sations, service could be delivered mainly by care workers, medical
workers, counsellors, trainers and so on. Whilst there are no governmen-
tal or municipal agencies in direct service delivery, the government of
Hong Kong supports a number of local leading non-governmental organ-
isations through grants, which covers four major service areas: family
and child welfare services, rehabilitation and medical social services, ser-
vice for young people and others and service for the elderly.
In the organisations and service units that receive governmental grants
(subvented service units), the Social Welfare Department offers some
guidance on service delivery. Social service organisations that are not
funded by the government rely on charity organisations, foundations and
private donations for funding; meanwhile, organisations that receive gov-
ernmental grants are also allowed to seek other funding sources and es-
tablish non-subvented service units within the same organisations.
Before the year 2000, the government adopted an extensive subvention
system to fund and collaborate with the subvented service organisations,
and in particular, protect the salary, work benefits and promotion of social
workers. Under that system, the Department of Social Welfare, through fi-
nancial subvention, defined and regulated the number of posts and job
ranks (e.g. Social Work Assistant (SWA), Social Work Officer (SWA)) of
social workers in each organisation, and connected their salary, fringe ben-
efits and annual payment adjustment with the government’s payment scale
of civil service. The system built social work positions as desirable middle-
class professional jobs with security and achieved a very low rate of turn-
over amongst social workers, which was well below 10 percent (Lai, 2017).
But in 2000, the government decided to abandon this ‘inflexible, com-
plex and bureaucratic’ (Social Welfare Department of Hong Kong, 2000)
system and introduced the Lump Sum Grant Subvention System (LSGSS)
to social service organisations in Hong Kong. Under the new scheme, ser-
vice planning and management are shifted from the government to organ-
isations, as the Social Welfare Department no longer imposes rigid input
 Social Worker Turnover 1685

controls on Non-Governmental Organizations (NGOs’) staffing and salary

structure. Granting recurrent funding to service organisations in a lump-
sum nature, the new scheme hopes to give organisations greater auton-
omy and flexibility to deploy resources. However, as many local scholars

Downloaded from https://academic.oup.com/bjsw/article/52/3/1683/6327302 by City University of Hong Kong user on 23 June 2023
have pointed out, the reform, via de-regulation of service organisations,
results in severe cuts in salary and job security of social workers, creates
unfriendly workplace environment in organisations and therefore leads to
increasing turnover rates of social workers (Tam and Mong, 2005; Lam
and Tse, 2013).
According to the official data of Hong Kong, the total turnover rate
of social workers in 2018 was 14.7 percent (Social Welfare Department
of Hong Kong, 2019), comparable to the rate in England (15 percent)
(Department of Education of the UK, 2019) and a bit lower than child
welfare workers’ turnover rate in the USA (around 20 percent) (Social
Work Research Center of Colorado State University, 2019). Under the
LSGSS, Hong Kong’s welfare governance joins the global trends of state
withdrawal and promotion of managerialism (Macmillan, 2013), and so-
cial workers’ turnover rate has swiftly caught up with other regions in
the world. What contributes to social worker turnover in contemporary
Hong Kong? How does the LSGSS reform play an important role?
Despite the importance of these questions, systematic and evidence-
based research on them remains limited to date.
Globally, research on social worker turnover usually relies on the self-
reporting of intentions to turnover of social workers, and uses their
individual-level perceptions to measure workplace environment and cul-
ture. The limitations of the approach have been thoroughly discussed in
the previous studies, and organisation-level data collection and analysis
are often called to fully understand the problem (Shim, 2010). In this
study, in order to understand the impacts of the LSGSS on the practices
of service organisations in Hong Kong, we obtained data from the man-
agement of service organisations, instead of individual social workers, in
which they reported the actual turnover rates of social workers in their
organisations, as well as various aspects of their organisational opera-
tions. The data thus provide a unique opportunity to inquire into how
organisational variances in employment terms, salary distributions, staff-
ing strategies, service types and budgets associate with turnover deci-
sions of social workers, and to enrich the understandings of the
struggles, working conditions and career plans of social workers under
managerialist welfare governance, beyond individual-level differences.

Turnover of social workers

Working at the front line of welfare service delivery, social workers of-
ten experience great pressure in their everyday job routines, and their
1686 Haijing Dai et al.

intentions to turnover have aroused attention in different disciplines of

social sciences (Barak et al., 2001). Researchers have explored factors in
different aspects of the work and life of social workers to understand
their career decisions, and the factors could often be summarised as at

Downloaded from https://academic.oup.com/bjsw/article/52/3/1683/6327302 by City University of Hong Kong user on 23 June 2023
the different levels of personal characteristics, work-related life, profes-
sional values and organisational climate (Moynihan and Landuyt, 2008).
At the personal level, researchers identified that younger social work-
ers, women and members of minority groups are more likely to express
the intention to leave their jobs (Barak et al., 2006). However, scholars
believed that demographic characteristics of social workers at the per-
sonal level, instead of having direct effects on intentions to turnover, are
highly associated with work-related factors (Griffeth et al., 2000; Dysvik
and Kuvaas, 2013).
At the work-related level, Ertas (2015) demonstrated that job satisfac-
tion, pay satisfaction, creativity at work, professional development, pro-
motion based on merits and having a good work group substantially
reduce turnover intentions amongst social workers. Kim and Stoner
(2008) argued that job autonomy and social support from peers have a
negative direct effect on turnover intentions. Jung (2014) and Cho et al.
(2009) emphasised that quality relationships with supervisors and com-
munication efforts and routines in goal-setting can lead to decreasing
intentions to turnover of social workers. Demerouti et al. (2001) also in-
troduced the well-known model of burnout to understand stressors at
work and attitudinal outcomes, such as turnover of social workers.
When work-related factors have attracted much scholarly interest in the
literature on social worker turnover, some researchers believed that those
factors could apply to so many different work settings that they neglect
professional specificities in social service (Siefert et al., 1991). They instead
suggested that in a value-based profession, it is crucial for front line social
workers to perceive value compatibility in their work routines to reduce
their turnover intentions (Cho and Song, 2017). Siefert et al. (1991) found
when social workers sense that depersonalisation and emotional detach-
ment are expected at work, they are more likely to attempt to find a differ-
ent job. Empirical evidence also showed that social workers’ capacity to
stick to professional judgement would help them stay in their jobs for lon-
ger times (Middleton and Potter, 2015). Studies in both the USA (Shier
et al., 2012) and China (Jiang et al., 2019) discovered that social workers
with higher levels of professional commitment and higher scores in recogni-
tion of professional identity and values have lower intentions to turnover.
More recently, under the influence of organisational studies, many
researchers advocated for the importance to include organisational fac-
tors in the analysis on social worker turnover and shift the investigation
to a structural level (Wilke et al., 2017). Scholars found that social work-
ers’ identification with organisational culture and climate (Shim, 2010),
their sense of being included in an organisational citizenship (Shier
 Social Worker Turnover 1687

et al., 2012) and their perception of organisational justice (Barak et al.,

2006) can reduce their intentions to leave their current jobs in social ser-
vice organisations.
Although the necessity to examine organisation-level factors to under-

Downloaded from https://academic.oup.com/bjsw/article/52/3/1683/6327302 by City University of Hong Kong user on 23 June 2023
stand social worker turnover is recognised in the existing literature, the
organisational factors are oftentimes measured at the individual level as
personal perceptions of organisational variances. This approach has lim-
ited the validity and reliability of organisation-level analysis and hin-
dered the development of a multi-level understanding of the turnover of
front line social workers. Shim (2010), therefore, proposed collecting
and analysing data at the truly organisational level as a future direction
of research, which this study adopts and fulfils.

Impacts of the LSGSS on social worker turnover in

Hong Kong
Hong Kong government introduced the LSGSS to reform the
centralised budgeting system of social welfare and social service, and be-
lieved that under the new system, service organisations could have greater
autonomy and flexibility to deploy resources and re-engineer their serv-
ices to meet changing social needs (Social Welfare Department of Hong
Kong, 2008). In its vision, the reform aimed at encouraging service organi-
sations and their staff to take more active roles in managing their own
resources, which should benefit both parties in the long term.
However, in the past twenty years of implementation of the new sys-
tem, in contrast to the expectation of the government, dissatisfaction has
accumulated amongst staff members of social service organisations, espe-
cially social workers. The LSGSS compels subvented social service
organisations to become responsible for their own financial survival,
through a market-driven and cost-control approach in the planning and
development of social service (Lai, 2017). Under the LSGSS, organisa-
tions no longer need to follow the job ranking and payment scales of the
government. Some of them invent new job titles for social workers, such
as ‘Project Manager’ or ‘Service Director’, rather than using the original
‘Social Work Assistant (SWA)’ and ‘Social Work Officer (SWO)’ sys-
tem, as a measure to control costs (Lai and Chan, 2009). It is a common
practice that social workers employed under the new job titles are paid
30–40 percent less than the original scale, whilst their caseloads are un-
reasonably increased (Hong Kong Council of Social Service, 2004).
In the same organisation, when senior social workers follow the old pay-
ment scale yet junior social workers are employed under new job titles, in-
come stratification grows in a hierarchical organisational structure, which
can lead to distrust amongst social workers and tensions in the workplace
(Law, 2003; Tam and Mong, 2005). Meanwhile, as organisations now regard
1688 Haijing Dai et al.

value-for-money and financial efficiency as prime concerns, they are willing

to hire professional managers of human resources and finances with gener-
ous remuneration packages (Chang, 2002). Researchers believed that the
increasing payment gap between managers and social workers not only dis-

Downloaded from https://academic.oup.com/bjsw/article/52/3/1683/6327302 by City University of Hong Kong user on 23 June 2023
mays social workers, but also undermines the professional values of the
organisations (Tam and Mong, 2005).
Moreover, in order to achieve even higher cost–benefit effectiveness,
some organisations replace social worker positions, which offer stable sal-
aries for degree-holding professionals, with short-term or part-time posi-
tions of care workers and service workers that do not require degrees
(Lam and Tse, 2013). By reducing jobs in the category of social workers,
organisations manage to further control their operational costs but often
at the expense of service quality and well-being of their employees.
In such contexts, it is hardly surprising that the LSGSS reform has stim-
ulated strong feelings of anger and dissatisfaction amongst social workers
in Hong Kong. In the government’s review of the LSGSS in 2008, it was
recognised that having staff employed on different terms created work-
place conflicts, social workers demanded equal pay with their civil service
counterparts, and insufficient support for training and capacity enhance-
ment was provided in the whole welfare sector. The review believed that
such tensions had led to high turnover rates of social workers in that year
(Social Welfare Department of Hong Kong, 2008). Although the govern-
ment has adopted various measures to improve the implementation of the
LSGSS since the review, the fundamental problems in the system and
high social worker turnover rates have remained unchanged (Lai, 2017).
Despite the severe challenges social workers face under the LSGSS,
studies on social worker turnover are limited. Observations and qualita-
tive interviews scatter in pieces of criticism of the LSGSS (Law, 2003;
Kwok, 2004), and the analysis is conducted at the individual level. As
the LSGSS affects individual social workers mainly through reforming
the management practices of social service organisations, a systematic
organisation-level investigation is necessary to fully understand the asso-
ciation between organisational management under the LSGSS and social
worker turnover in Hong Kong.

Managerialism: an organisation-level framework of

analysis

Since the 1980s, governments over the globe have been searching for
flexible and cost–benefit efficient service models to replace the bureau-
cratic state in welfare governance (Nicholls, 2006), and managerialism
has become a prevalent choice. There is no consensus to date in a con-
clusive definition of the term, but managerialism in essence can be per-
ceived as a set of management ideas (Flynn, 2002). Forces external to
 Social Worker Turnover 1689

the daily operation of organisations, such as owners in business settings

and the state in the public sector, should restrain from intervention and
grant complete autonomy to the management for the best outcome in
organisational development (Klikauer, 2015). The LSGSS reform well

Downloaded from https://academic.oup.com/bjsw/article/52/3/1683/6327302 by City University of Hong Kong user on 23 June 2023
evinces state retrenchment and the managerialist turn in welfare gover-
nance in Hong Kong. Since the reform, the Social Welfare Department
has withdrawn from direct intervention in subvented service organisa-
tions, and good or efficient management practices in the organisations
have been emphasised.
Differentiated employment terms and stratified income levels are pop-
ular managerialist strategies to invite competition amongst employees
and increase labour value (Flynn, 2002). In social service organisations
under the LSGSS in Hong Kong, whether to abandon the original SWA/
SWO employment schemes and how to distribute income amongst junior
and senior social workers and agency managers are difficult management
decisions, which profoundly shape the work experiences of social work-
ers and their decisions to turnover.
Meanwhile, like their counterparts in corporates, managers in manageri-
alist service organisations allow different staffing quotas in different types
of positions to maximise cost–benefit efficiency and to establish an organi-
sational culture (Klikauer, 2015). For social service organisations under
the LSGSS, the number of staff positions they keep in the category of so-
cial workers is therefore an important indicator to explore, so as to under-
stand the management orientations and cultures of the organisations.
Adopting this organisation-level framework of analysis, this study sets
out to examine social worker turnover from a more macro perspective
and interrogates how organisational employment terms, salary stratifica-
tion and staffing strategies are associated with actual social workers’
turnover rates in the managerialist service organisations under the
LSGSS in Hong Kong, controlling service type and budget size of the
organisations. Our hypotheses include:
Hypothesis 1: Service organisations that adopt employment terms
other than SWA/SWO for social workers are more likely to have higher
levels of social worker turnover.
Hypothesis 2: Service organisations that offer lower minimum salary
and higher maximum salary to social workers are more likely to have
higher levels of social worker turnover.
Hypothesis 3: Service organisations that offer higher maximum salary
to management staff are more likely to have higher levels of social
worker turnover.
Hypothesis 4: Service organisations that keep fewer social
worker positions are more likely to have higher levels of social worker
turnover.
Hypothesis 5: Service type and budget size of service organisations are
associated with social worker turnover rates.
1690 Haijing Dai et al.

Research methods

The data

Downloaded from https://academic.oup.com/bjsw/article/52/3/1683/6327302 by City University of Hong Kong user on 23 June 2023
In 2019, sponsored by the Social Welfare Department of Hong Kong,
the research team collaborated with a consulting firm to collect organisa-
tion-level data from the management of social service organisations
funded by the LSGSS in Hong Kong. To make the data more relevant
for us to review the LSGSS, for organisations that had both subvented
and non-subvented service units under the LSGSS, they only needed to
report on the service units funded by the scheme. A questionnaire cov-
ering financial operation, human resource management, future blueprints
of development and managers’ views of the LSGSS in the fiscal year of
2017–2018 was emailed to each of the 164 service organisations sub-
vented by the LSGSS in April 2019. The Social Welfare Department
and the consulting firm also telephoned the managers of the organisa-
tions to encourage their participation and clarify the questions in the
questionnaire at the beginning of the project.
Although the study was solely funded by the Social Welfare
Department, the consulting firm and the research team were directly in-
volved in data collection and analysis, instead of government officials. We
followed scientific procedures in obtaining the full list of organisations and
service units, contacting respondents and following up with no-response
cases. All the information we collected from the organisations stayed anon-
ymous and confidential to prevent tracing the cases in the data-set to par-
ticular organisations. The data collection procedures were also approved
by the Research Ethics Committee of the home institution of the authors.
By the end of May 2019, 132 organisations replied and 112 of them
submitted valid responses, which gave a valid response rate of 68.3 per-
cent. Amongst the valid cases, eleven organisations did not hire any so-
cial workers at the beginning of the fiscal year of 2017–2018 and were
hence dropped from our analyses on social worker turnover. As a result,
101 service organisations funded by the LSGSS were included in our
data analyses in total.

Key measurements
Social worker turnover rates
In their response to the survey, the managers reported the total number
of social workers employed in their service organisations at the begin-
ning (1 April 2017) and by the end of the fiscal year (31 March 2018), as
well as the total number of social worker resignations during the fiscal
year. We then divided the number of resignations by the average
 Social Worker Turnover 1691

number of social workers in the organisation during the year and came
to the social worker turnover rate of that service organisation.
Amongst the 101 service organisations, 44.6 percent (n ¼ 45) reported
no social worker resignations during the fiscal year and thus had a turn-

Downloaded from https://academic.oup.com/bjsw/article/52/3/1683/6327302 by City University of Hong Kong user on 23 June 2023
over rate of zero. Compared with organisations that do not receive gov-
ernmental funding, service organisations and units under the LSGSS
usually provide jobs with better payment, more security and fringe bene-
fits that are more generous. It is therefore reasonable that the average
turnover rates of the organisations in the sample (mean ¼ 12.13 percent)
were lower than the official statistics (14.7 percent) in that fiscal year,
and many had no resignation cases of social workers at all.
In data analysis, we used zero and the official turnover rate (14.7 per-
cent) as the cutting points and categorised the service organisations in
the sample into three types: no social worker turnover (n ¼ 45, 44.6 per-
cent); emerging turnover (n ¼ 27, 26.7 percent); and severe turnover
(n ¼ 29, 28.7 percent).

Employment terms of social workers

In the survey, we asked whether the service organisations adopted em-
ployment schemes other than the SWA/SWO system to hire social work-
ers in the fiscal year. The respondents provided the answers as ‘Yes’ or
‘No’.

Minimum and maximum salary of social workers

The respondents to the survey reported the minimum and maximum sal-
ary they offered to social workers in their service organisations. The two
variables revealed not only the payment level of social workers, but also
the income stratification amongst social worker colleagues in the same
organisation under the LSGSS.

Maximum salary of management

Similar to the salary of social workers, we also asked the respondents to
provide the maximum salary of staff working in the management team.
We used this variable to understand the payment disparity between so-
cial workers and the management team in service organisations under
the LSGSS.

Total number of social workers employed

We used the total number of social workers employed at the beginning
of the fiscal year (1 April 2017) to indicate the number of social worker
positions in the service organisations. The number at the beginning of
1692 Haijing Dai et al.

the year could have influenced the working experiences of social work-
ers in the organisations and affected their decisions to turnover during
the fiscal year.

Downloaded from https://academic.oup.com/bjsw/article/52/3/1683/6327302 by City University of Hong Kong user on 23 June 2023
Service type
Using the official categorisation of subvented social services in Hong
Kong, we asked the managers to choose one service type from the fol-
lowing categories: Elderly Service, Family and Child Welfare Service,
Rehabilitation and Medical Social Service, Service for Young People
and Others and Multiple Services. Previous studies sometimes found
that social workers in Elderly Service had stronger intentions to leave
their jobs in Hong Kong (e.g. Hong Kong Council of Social Service,
2004); we hence used Elderly Service as the reference group in data
analyses.

Budget size
We retrieved the official statistics from the Social Welfare Department
on the annual budgets of the service organisations under the LSGSS in
the fiscal year of 2017–2018. We used this variable as an indicator of the
size of the service organisations.
The descriptive statistics of the key variables can be found in Table 1.

Data analyses

To evaluate the association between the key management practices of

service organisations and their social workers’ turnover rates, we built
a Multinomial Logistic Regression model, as the dependent variable
was categorical with three groups. Although the three categories of
turnover rates were ordinal in nature, the results of Brant test showed
that the three categories did not satisfy the proportional odds assump-
tion. We hence conducted a more conservative Multinomial Logistic
Regression, instead of an Ordinal Logistic Regression, for our data
analyses. We used Stata (version 14.2) to estimate the regression
models.

Results
A Multinomial Logistic Regression was performed to examine the
effects of NGO’s service types, budget size, employment terms of social
workers, social workers’ minimum and maximum salary, maximum sal-
ary of management team and total number of social workers employed
on social workers’ turnover rates.
 Social Worker Turnover 1693

Table 1 Descriptive statistics (N ¼ 101)

Variables Frequency Percentage

Mean SD

Downloaded from https://academic.oup.com/bjsw/article/52/3/1683/6327302 by City University of Hong Kong user on 23 June 2023
Social worker turnover rate
0 percent 45 44.55
0–14.7 percent 27 26.73
Above 14.7 percent 29 28.71
NGO service types
Family and child welfare services 6 5.94
Rehabilitation and medical social services 16 15.84
Service for young people and others 16 15.84
Service for the elderly 31 30.69
Multiple 32 31.68
Only SWO/SWA employment terms
Yes 81 80.20
No 20 19.80
Minimum social worker salary 20,599.37 5,774.82
Maximum social worker salary 59,184.90 28,514.03
Maximum manager salary 79,489.68 26,211.05
Budget size (ln) 17.06 1.59
Total social workers employed 42.20 41.96

The Multinomial Logistic Regression Model was statistically signifi-

cant (v2(20) ¼ 72.2, P < 0.0001), and its McFadden’s pseudo R2 was
0.4343. Results from the regression analyses are shown in Table 2. The
results demonstrated that service types of organisations, budget size, so-
cial workers’ minimum salary, maximum salary of management team
and total number of social workers employed had statistically significant
relationships with social workers’ turnover rates of the organisations.
Meanwhile, employment terms of social workers and maximum salary of
social workers were not significant statistically.

Emerging turnover versus no turnover

Compared with service organisations with no social worker turnover in

the fiscal year of 2017–2018, organisations providing family and child
welfare services or multiple services were more likely to have emerging
turnover rates (0–14.7 percent) than organisations providing elderly
services (the reference group): their multinomial log-odds were 10.52
(P < 0.01) and 6.20 (P < 0.05), respectively.
The budget size of service organisations also exhibited a positive rela-
tionship with social workers’ emerging turnover: one unit increase of
budget size in log form would increase the multinomial log-odds for hav-
ing emerging turnover by 4.12 (P < 0.001).
For employment terms of social workers in service organisations,
whether organisations adopted systems other than the SWA/SWO were
1694 Haijing Dai et al.

Table 2 Multinomial logistic regression on social worker turnover rate

(relative to no turnover)

Independent variables 0–14.7 percent (emerging Above 14.7 percent

turnover) (severe turnover)

Downloaded from https://academic.oup.com/bjsw/article/52/3/1683/6327302 by City University of Hong Kong user on 23 June 2023
Coefficient Coefficient
(standard error) (standard error)

NGO service types

Family and child welfare services 10.52 13.22
(3.77)** (1235.25)
Rehabilitation and medical social services 0 .48 0 .33
(2.11) (1.01)
Service for young people and others 4.88 1.86
(2.67) (1.42)
Multiple 6.20 0.70
(2.41)* (0.83)
Budget size (ln) 4.12 1.05
(1.27)*** (0.45)*
Only SWO/SWA employment terms 1.07 0.13
(1.18) (0.88)
Minimum social worker salary 0.0007 0.00007
(0.0004)* (0.0001)
Maximum social worker salary 0.00002 0.00002
(0.00002) (0.00002)
Maximum manager salary 0.00005 0.00001
(0.00003)* (0.00002)
Total social workers employed 0.04 0.02
(0.01)** (0.01)*
_cons 66.84 16.20
(19.85)*** (6.16)**

Note: *P <0.05, **
P <0.01, ***
P<0.001.

not significantly associated with the multinomial log-odds for having

emerging social worker turnover.
In terms of within-organisation salary stratification, social workers’
minimum salary displayed a negative relationship with emerging turn-
over: one Hong Kong Dollar increase in social workers’ minimum salary
would be expected to decrease the multinomial log-odds for emerging
turnover by 0.0007 (P < 0.05). At the same time, one Hong Kong
Dollar increase in the maximum salary of the management staff would
increase the multinomial log-odds for having emerging turnover of social
workers by 0.00005 (P < 0.05). Social workers’ maximum salary, how-
ever, did not significantly affect the multinomial log-odds for having
emerging turnover of social workers in service organisations.
Finally, regarding hiring strategies of service organisations, total num-
ber of social workers employed showed a negative relationship with the
emerging turnover of social workers: one staff increase in the total num-
ber of social workers employed would decrease the multinomial log-
odds of service organisations for having emerging turnover of social
workers by 0.04 (P < 0.01).
 Social Worker Turnover 1695

Severe turnover versus no turnover

We then explored the management practices of service organisations

with severe social worker turnover (over 14.7 percent) in the fiscal year

Downloaded from https://academic.oup.com/bjsw/article/52/3/1683/6327302 by City University of Hong Kong user on 23 June 2023
of 2017–2018, in comparison with service organisations that had no social
worker turnover in that year. Service types of the organisations did not
show a significant influence on the multinomial log-odds for having se-
vere turnovers of social workers in the organisations.
The budget size of service organisations, however, demonstrated a
positive relationship with severe social worker turnover, as it did with
emerging social worker turnover. One unit increase of the budget size of
service organisations in log form would increase the multinomial log-
odds for having severe turnovers of social workers by 1.05 (P < 0.05).
Employment terms of social workers and salary stratification in ser-
vice organisations (including minimum salary of social workers, maxi-
mum salary of social workers and maximum salary of management staff)
did not have statistically significant effects on the multinomial log-odds
for severe social worker turnover.
Hiring strategies of service organisations held an important role in se-
vere social worker turnover, as in emerging social worker turnover. The
total number of social workers employed in service organisations kept
showing a significant negative relationship with severe turnovers of so-
cial workers: one unit increase of it would decrease the multinomial log-
odds for having severe social worker turnover in the service organisa-
tions by 0.02 (P < 0.05).

Findings and discussions

Through Multinomial Logistic Regression analyses, we tested our five

hypotheses to understand social worker turnover under the LSGSS in
Hong Kong at the organisational level. Our Hypothesis 1 was not sup-
ported by the data, as whether organisations adopted employment terms
other than the SWA/SWO system for social workers did not significantly
associate with emerging or severe turnovers of social workers, compared
with no turnover. Previous studies argued that introducing non-SWA/
SWO schemes to hire social workers not only reduces the salary and
fringe benefits of social workers, but also creates tensions and distrust
amongst social worker colleagues in the same organisation (Lai and
Chan, 2009). It could therefore be an important explanatory factor for
increasing social worker turnover under the LSGSS in Hong Kong.
Evidence from our research, however, did not support this observation.
In Hypothesis 2, we hypothesised that stratification in salaries amongst
social workers tends to increase the odds of social worker turnover, and
it was only partially supported by the data. Lower minimum salary of
1696 Haijing Dai et al.

social workers was significantly associated with emerging turnovers of

social workers, but not severe turnovers, whilst maximum salary of social
workers had significant effects on neither. Meagre starting salary per se,
as expected, could push some social workers away from their jobs, but

Downloaded from https://academic.oup.com/bjsw/article/52/3/1683/6327302 by City University of Hong Kong user on 23 June 2023
not in a very large number. Contrary to some observations of local
scholars (e.g. Law, 2003), income disparity amongst social workers in the
same service organisation does not exert significant influence on the
turnovers of social workers through creating tensions and distrust
amongst social worker colleagues.
For Hypothesis 3, we tested the relationship between higher payments
to management staff and social worker turnover in service organisations
under the LSGSS, and the relationship was significant for the odds of
having emerging turnovers of social workers, but not for severe turn-
overs. Disgruntlement with resource allocation favouring managers of
service organisations and the accompanying income inequality between
social workers and management staff could lead some social workers to
leave their jobs, but still does not produce very severe turnovers of so-
cial workers under the LSGSS.
Hypothesis 4 explored the influence of staff hiring strategies of service
organisations under the LSGSS on social worker turnover. It was sup-
ported by our data that service organisations that kept fewer social
worker positions were more likely to have both emerging and severe
turnovers of social workers. When service organisations manoeuvre the
flexibility of the LSGSS to employ fewer staff in the social worker job
category to reduce costs, social workers tend to leave their jobs in both
small and large numbers.
The other factor that was significantly associated with the odds for
both emerging and severe social worker turnover was the budget size of
service organisations, which supported our Hypothesis 5. Ertas (2015) ar-
gued in the contexts of the USA that closeness amongst social worker
colleagues and good group work spirits are important factors to decrease
the intentions to turnover of social workers. In Hong Kong, it is possible
that in larger and usually more bureaucratic service organisations, social
workers feel distant from one another in the work environment. The
resulting sense of lack of peer support and loneliness eventually leads
social workers to leave their job positions, in both small and large
numbers.
As also hypothesised in Hypothesis 5, service types of organisations
were significantly associated with the odds for having emerging social
worker turnover, but not with severe social worker turnover. We used
Elderly Service as the reference group because it was often difficult for
this type of service organisations to retain social workers due to stressful
work environment, low payment and low professionalisation levels
(Hong Kong Council of Social Service, 2004). However, during the past
decade, faced with the growing challenge of population ageing, Hong
 Social Worker Turnover 1697

Kong government has substantially increased investment in elderly serv-

ices, which improved the employment conditions of social workers in
those organisations (Office of the Government Economist, 2019). Our
study also showed that in 2017–2018, organisations in Family and Child

Downloaded from https://academic.oup.com/bjsw/article/52/3/1683/6327302 by City University of Hong Kong user on 23 June 2023
Welfare Service and Multiple Service were more likely than Elderly
Service organisations to have emerging social worker turnover; and the
difference between Elderly Service organisations and service organisa-
tions of other types was insignificant to understand the odds for severe
turnovers of social workers. The findings provided some evidence for
the effectiveness of the investment of Hong Kong government in elderly
services in recent years.
Putting together our findings based on the five hypotheses, we discov-
ered that differentiated employment terms and salary stratification
amongst social workers in the same organisation were not significantly
associated with social worker turnover, although previous studies often
criticised these factors as negative consequences of the LSGSS.
Deteriorating minimum wage of social workers and salary inequality be-
tween social workers and management staff increased the odds for
emerging social worker turnover in service organisations. Nevertheless,
it was staffing strategies in social worker positions and budget size of
organisations that were related to both emerging and severe turnovers
of social workers. It is likely that recognition at the organisational level
of social workers’ professional skills and values and their needs of peer
support may help to retain social work talents under the LSGSS in
Hong Kong.

Conclusions and implications

Managerialism, which argues that problems of efficiency can have mana-

gerial solutions if state intervention in organisations remains limited
(Entemann, 1993), has gained much popularity and legitimacy in welfare
reforms and service planning in recent decades (Barberis, 2013). Hong
Kong’s reform of the LSGSS is often regarded as a textbook illustration
of the principles of managerialism, in which the state withdraws from
the bureaucratic administration of social services and grants flexibility
and autonomy to service organisations (Leung, 2002). This study system-
atically enquires into the actions of turnover of social workers under the
scheme, and reflects on the effects of the practices of managerialism
with organisation-level data.
Our findings suggest that in contrast to popular beliefs in the social
welfare sector of Hong Kong, differentiation of employment terms and
salary stratification amongst social workers, which have occurred in
many service organisations under the LSGSS, do not significantly associ-
ate with the turnovers of social workers. Social workers in Hong Kong
1698 Haijing Dai et al.

probably still see each other as a close peer group with solidarity, al-
though the managerialist LSGSS allows organisations to make distinc-
tions amongst them.
Whilst higher income of some fellow social workers does not bother

Downloaded from https://academic.oup.com/bjsw/article/52/3/1683/6327302 by City University of Hong Kong user on 23 June 2023
social workers in general, reduced starting salary of social workers and
increasing spending on the salaries of management staff—prevalent phe-
nomena under the LSGSS—can create discontents amongst social work-
ers and lead to their turnover decisions. Social workers seem to accept
employment stratification in their own professional group, but reject the
social worker/manager income stratification produced by the LSGSS. As
Barak et al. (2006) discovered from individual-level data, social workers’
perception of injustice in their service organisations strongly affects their
intentions to turnover. Our study, at the organisational level, provides
evidence to disentangle the complex concept of organisational injustice
in the contexts of Hong Kong—it probably means inequality between
social workers and managers under the LSGSS, but not stratification
amongst social workers.
Furthermore, management strategies to reduce costs in human resour-
ces by replacing social worker positions with lower-paid job titles, al-
though highly compatible with the principles of managerialism, fiercely
offend social workers and lead to their high levels of turnover. Lack of
peer bonding and team spirits amongst social workers in larger organisa-
tions could have similar effects. As mentioned in previous studies in
other contexts, organisational culture, employment climate, inclusion in
peer circles and an organisational citizenship are the important factors
to shape social workers’ turnover intentions (Shim, 2010; Shier et al.,
2012). Our findings echo these conclusions, yet have more specifically
identified the meaning of organisational culture and climate (as appreci-
ation and respect for social work professional values and skills) and peer
support (as closeness and team spirits amongst social worker colleagues)
in understanding social worker turnover under the LSGSS in Hong
Kong.
As the Hong Kong government claimed in their official reports (e.g.
Social Welfare Department of Hong Kong, 2019), they planned to stick
to the managerialist approach in social welfare administration and had
no intention to return to the inefficient and bureaucratic old system. In
order for the existing LSGSS to retain social worker talents and stabilise
social service quality, we would suggest the following interventions from
both service organisations and the government of Hong Kong, based on
the evidence discovered in the study.
In accordance with the spirits of managerialism, service organisations
could adopt varied employment terms to hire social workers, but there
needs to be an adequate bottom line for the minimum salary of social
workers and upper limits of salaries of the management staff. Service
organisations need to pay attention to the emerging income inequality
 Social Worker Turnover 1699

between social workers and managers under the LSGSS, whilst the gov-
ernment should carefully review the salary gaps in the annual reports of
organisations and sufficiently supervise the stratification.
To keep social worker talents, organisations under the LSGSS need to

Downloaded from https://academic.oup.com/bjsw/article/52/3/1683/6327302 by City University of Hong Kong user on 23 June 2023
restrain from replacing social worker positions with lower-paid job titles
in the name of cost control. Organisations should as well invest in team
building activities for social workers, especially in larger organisations,
to facilitate peer support and an intimate organisational culture. These
two practices should be encouraged, financially incentivized and closely
supervised in the implementation of the LSGSS at the governmental
level.
The impacts of managerialism, through growingly autonomous service
organisations, on service professionals and service quality in the public
sector have started to concern scholars and practitioners globally (Flynn,
2002; Shepherd, 2018). Managerialist principles, as evinced in the
LSGSS in Hong Kong, might bring efficiency and flexibility to the public
sector. Yet to control their harmful impacts on service professionals in
the welfare sector, the old ideals of income justice, professional skills
and values and team spirits need to be harboured in the new organisa-
tional environment and contexts.

Limitations
Our study suffers from many limitations. It is never easy to collect data
on human resources and financial management of service organisations,
especially in a place where principles of managerialism grant full auton-
omy and data confidentiality to them. In collaboration with our partners,
we only managed to reach the organisations and service units that re-
ceived funding from the Hong Kong government in 2017–2018, which
was a small portion of the NGOs in the region. Yet, because the LSGSS
and its managerialist principles affect these subvented NGOs and service
units most, they may still provide valuable data for our inquiries, but the
generalisability of our findings is limited to such organisations and ser-
vice units, instead of all the NGOs in Hong Kong.
Despite our repeated efforts to contact the management of the service
organisations, we reached a response rate of less than 70 percent and
the sample size for our analyses was small. With a small N, we were un-
able to include more factors in the regression models or test more com-
plicated relationships (e.g. pathway analysis, moderating effects)
amongst our key variables. In the explanation of statistical significance,
we need to rely on previous literature and observation, instead of con-
ducting further empirical analyses. The sponsorship of the Department
of Social Welfare of the research may also have influenced the validity
1700 Haijing Dai et al.

of the reporting of the service units. These factors could affect the valid-
ity of our interpretation and conclusion.
Nevertheless, the study contributes to our knowledge of social worker
turnover under welfare managerialism from the organisational perspec-

Downloaded from https://academic.oup.com/bjsw/article/52/3/1683/6327302 by City University of Hong Kong user on 23 June 2023
tive. It could be a starting point for future multilevel analyses or com-
parative research on social worker turnover in different social and
cultural contexts.

Funding
Research activities for this study were generously sponsored by a
Research Grant from the Social Welfare Department of Hong Kong.

References

Barak, M. E. M., Nissly, J. A. and Levin, A. (2001) ‘Antecedents to retention and

turnover among child welfare, social work, and other human services employees:
What can we learn from past research? A review and meta-analysis’, Social
Service Review, 75(4), pp. 625–61.
Barak, M. E. M., Levin, A., Nissly, J. A. and Lane, C. J. (2006) ‘Why do they leave?
Modeling child welfare workers’ turnover intentions’, Children and Youth Services
Review, 28(5), pp. 548–77.
Barberis, P. (2013) ‘The managerial imperative: Fifty years’ change in UK public ad-
ministration’, Public Policy and Administration, 28(4), pp. 327–45.
Chang, S.-W. (2002) ‘Managerialism and small NGOs in Hong Kong’, The Hong
Kong Journal of Social Work, 36(1–2), pp. 83–94.
Cho, Y. J. and Song, H. J. (2017) ‘Determinants of turnover intention of social work-
ers: Effects of emotional labor and organizational trust’, Public Personnel
Management, 46(1), pp. 41–65.
Cho, S., Johanson, M. M. and Guchait, P. (2009) ‘Employees intent to leave: A com-
parison of determinants of intent to leave versus intent to stay’, International
Journal of Hospitality Management, 28(3), pp. 374–81.
Demerouti, E., Bakker, A. B., Nachreiner, F. and Schaufeli, W. B. (2001) ‘The job
demands–resources model of burnout’, The Journal of Applied Psychology, 86(3),
pp. 499–512.
Department of Education of the UK. (2019) ‘Experimental statistics: Children and
family social work workforce in England, year ending 30 September 2018’, avail-
able online at: https://assets.publishing.service.gov.uk/government/uploads/system/
uploads/attachment_data/file/782154/Children_s_social_work_workforce_2018_text.
pdf (accessed December 10, 2020).
Dysvik, A. and Kuvaas, B. (2013) ‘Perceived job autonomy and turnover intention:
the moderating role of perceived supervisor support’, European Journal of Work
and Organizational Psychology, 22(5), pp. 563–73.
Entemann, W. F. (1993) Managerialism: The Emergence of a New Ideology, Madison,
University of Wisconsin Press.
 Social Worker Turnover 1701

Ertas, N. (2015) ‘Turnover intentions and work motivations of millennial employees

in federal service’, Public Personnel Management, 44(3), pp. 401–17.
Flynn, N. (2002) Public Sector Management, Harlow, Pearson.
Griffeth, R. W., Hom, P. W. and Gaertner, S. (2000) ‘A meta-analysis of antecedents

Downloaded from https://academic.oup.com/bjsw/article/52/3/1683/6327302 by City University of Hong Kong user on 23 June 2023
and correlates of employee turnover: Update, moderator tests, and research impli-
cations for the next millennium’, Journal of Management, 26(3), pp. 463–88.
Hong Kong Council of Social Service (2004) ‘Views toward lump sum grant’,
available online at: http://www.legco.gov.hk/yr04-05/chinese/panels/ws/papers/
ws0602cb2-1695-7c.pdf (accessed December 10, 2020).
Jiang, H., Wang, Y., Chui, E. and Xu, Y. (2019) ‘Professional identity and turnover
intentions of social workers in Beijing, China: The roles of job satisfaction and
agency type’, International Social Work, 62(1), pp. 146–60.
Jung, C. S. (2014) ‘Why are goals important in the public sector? Exploring the bene-
fits of goal clarity for reducing turnover intention’, Journal of Public
Administration Research and Theory, 24(1), pp. 209–34.
Klikauer, T. (2015) ‘What is managerialism?’, Critical Sociology, 41(7), pp. 1103–19.
Kim, H. and Stoner, M. (2008) ‘Burnout and turnover intention among social work-
ers: Effects of role stress, job autonomy and social support’, Administration in
Social Work, 32(3), pp. 5–25.
Kwok, C. W. (2004) A Study of the Lump Sum Grant Policy on the Provision of
Social Welfare Services in Hong Kong, PhD Dissertation, The University of Hong
Kong.
Lai, P. (2017) ‘Lump sum grant’, Cultural Research at Lingnan, 61(1), pp. 1–9.
Lai, W. H. F. and Chan, K. T. T. (2009) ‘Social work in Hong Kong: From profes-
sionalization to ‘re-professionalization’, China Journal of Social Work, 2(2),
pp. 95–108.
Lam, C. L. and Tse, H. K. (2013) ‘Adversity and resistance: Neoliberal social services
and social work in Hong Kong’, Critical and Radical Social Work, 1(2),
pp. 267–71.
Law, P. S. (2003) The Impact of the Lump Sum Grant System on Professional Social
Workers in Hong Kong, PhD Dissertation, The Chinese University of Hong Kong.
Leung, J. (2002) ‘The advent of managerialism in social welfare: The case of Hong
Kong’, The Hong Kong Journal of Social Work, 36(1–2), pp. 61–81.
Macmillan, R. (2013) ‘De-coupling the state and the third sector? The “Big Society”
as a spontaneous order’, Voluntary Sector Review, 4(2), pp. 185–203.
Middleton, J. S. and Potter, C. C. (2015) ‘Relationship between vicarious traumatiza-
tion and turnover among child welfare professionals’, Journal of Public Child
Welfare, 9(2), pp. 195–216.
Moynihan, D. P. and Landuyt, N. (2008) ‘Explaining turnover intention in the public
sector: Examining the roles of gender, life cycle and loyalty’, Review of Public
Personnel Administration, 28(2), pp. 120–43.
Nicholls, A. (2006) ‘Playing the field: A new approach to the meaning of social entre-
preneurship’, Social Enterprise Journal, 2(1), pp. 1–5.
Office of the Government Economist. (2019) ‘Population ageing trend of Hong
Kong’, available online at: https://www.hkeconomy.gov.hk/en/pdf/el/el-2019-02.pdf
(accessed December 10, 2020).
Shepherd, S. (2018) ‘Managerialism: An ideal type’, Studies in Higher Education,
43(9), pp. 1668–87.
1702 Haijing Dai et al.

Shier, M. L., Graham, J. R., Fukuda, E., Brownlee, K., Kline, T. J., Walji, S. and
Novik, N. (2012) ‘Social workers and satisfaction with child welfare work: Aspects
of work, profession, and personal life that contribute to turnover’, Child Welfare,
91(5), pp. 117–38.

Downloaded from https://academic.oup.com/bjsw/article/52/3/1683/6327302 by City University of Hong Kong user on 23 June 2023
Shim, M. (2010) ‘Factors influencing child welfare employee’s turnover: Focusing on
organizational culture and climate’, Children and Youth Services Review, 32(6),
pp. 847–56.
Siefert, K., Jayaratne, S. and Chess, W. A. (1991) ‘Job satisfaction, burnout, and turn-
over in health care social workers’, Health & Social Work, 16(3), pp. 193–202.
Social Welfare Department of Hong Kong. (2019) ‘Report on Human
Resources Needs in Social Work, 2019’, available online at: https://www.swd.gov.
hk/storage/asset/section/296/tc/SWMRS_Annual_Report_2018_Final-TC.pdf
(accessed December 10, 2020).
Social Welfare Department of Hong Kong. (2008) ‘Review report on the Lump Sum
Grant Subvention System’, available online at: https://www.swd.gov.hk/doc/ngo/
Report-eng.pdf (accessed December 10, 2020).
Social Welfare Department of Hong Kong. (2000) ‘Welfare sector subvention
reforms’, available online at: https://www.swd.gov.hk/doc/ngo/SectorReform_eng.
pdf (accessed December 10, 2020).
Social Work Research Center of Colorado State University. (2019) ‘Caseworker re-
tention survey report’, available online at: https://www.chhs.colostate.edu/ssw/wp-
content/uploads/sites/7/2018/11/arch-caseworker-retention-survey-final-report.pdf
(accessed December 10, 2020).
Tam, T. S. and Mong, L. P. (2005) ‘Job stress, perceived inequity and burnout among
school social workers in Hong Kong’, International Social Work, 48(4), pp. 467–83.
Wilke, D. J., Radey, M., King, E., Spinelli, C., Rakes, S. and Nolan, C. R. (2017) ‘A
multi-level conceptual model to examine child welfare worker turnover and reten-
tion decisions’, Journal of Public Child Welfare, 12(1), pp. 35–46.