Retaining Employees in Small and Medium-Sized
Firms:
Examining
the Link with Human Resource Management
Terry H. Wagar, Saint Mary’s University
Kent V. Rondeau,
Executive
Summary
While the 1990s were a decade
characterized by downsizing and restructuring, there has been a recent focus on
attracting and retaining quality employees. Very little research has addressed
employee retention within small and medium-sized firms. This paper uses data from a national study of
manufacturing organizations in
Introduction
While researchers are now placing
considerable attention on the links between human resource management (HRM)
practices and the performance of the firm, most of this research has been
conducted in large organizations (Wood, 1999; Wright, Gardner, Moynihan &
Allen, 2005). Many of these studies have
examined the impact of human resource systems on the financial performance of
the organization. However, there is a
paucity of research investigating whether the adoption of certain human
resource management practices has the potential to affect employee retention
rates (Stovel & Bontis, 2002; Batt, Colvin & Keefe. 2002). The current study uses data from small and
medium-size manufacturing workplaces to investigate the factors associated with
the ability of an establishment to retain employees.
Why should small firms be concerned
about retaining employees? There is a
strong body of research indicating that employee turnover can be very costly
(Griffith & Hom, 2001). Potential
costs (Dess & Shaw, 2001) can include both "direct" costs (such
as separation, replacement, training and general administration costs) and
"indirect" costs (such as lower productivity and reduced customer
loyalty). While the loss of quality employees can be devastating for both big
and small firms, retaining key individuals may be particularly important for
small and medium-size businesses -- if a high-quality employee leaves the
organization, a smaller firm may be less likely to have a suitable internal
candidate or lack the resources to selectively recruit on the external
market.
Background
Although most of the research
addressing human resource management and firm performance has been conducted in
large firm settings, small and medium-size businesses are now adopting human
resource management practices with an expectation that these practices have the
potential to positively impact firm performance (Way, 2002). Following from Heneman, Tansky and Camp's
(2000) observation that there is a shortage of information on human resource
management practices in smaller organizations, it is not surprising that there
is very little research addressing the link between employee retention and
human resource management.
While there is an extensive body of
research on employee turnover, almost all of the work by social scientists has
been done at the individual (employee) level of analysis (Griffith & Hom,
2001). For instance, Kickul (2001) found
that employer breach of factors addressing autonomy and growth (such as
providing meaningful work, the opportunity to develop new skills, and
recognition of accomplishments) and rewards and opportunities (for instance,
providing opportunities for advancement and compensation tied to performance)
were associated with a greater intention to quit among small business
employees.
However, the importance of studying
employee turnover at the workplace or organizational level has been noted
(Delery, Gupta, Shaw, Jenkins & Gangster, 2000; Batt, 2002). Shaw, Gupta and Delery (2005) recently
investigated the relationship between voluntary turnover and organizational
performance using data from employers in the trucking industry and
organizations involved in the manufacturing of concrete pipe. They found that there was a negative
relationship between voluntary turnover and organizational performance but that
the relationship was attenuated as the level of voluntary turnover increased.
Marchington, Carroll and Boxall (2003)
examined the retention of drivers in small firms using data from the road
haulage industry in the
Two other studies examine the
relationship between employee turnover and HRM practices. Way (2002), using data from the National
Employer Survey (Phase II), found that a High Performance Work System (HPWS)
was strongly related to lower workforce turnover in his sample of small
businesses. Similarly, Batt (2002)
observed a very significant association between human resource management
practices and quit rates in her sample of call centers.
In addition to the lack of research
at the workplace or firm level, many of the previous studies have also failed
to differentiate between voluntary
and involuntary turnover (Shaw et
al., 1998; Shaw et al., 2005). Voluntary
turnover occurs when an employee makes a decision to leave the organization
(commonly referred to as a "quit").
Drawing from the individual level research on voluntary turnover, it can
be argued that the decision to quit a job may be affected by the attractiveness
of the current job, the relative fit that a person has to the job, and the
availability of alternative employment (Griffeth & Hom, 2001). While an organization can engage in practices
and activities to make the firm a more attractive place to work, the
availability of alternative employment is not directly controlled by the
organization. Consequently, this paper
focuses on the impact of human resource management practices in making the firm
a more attractive workplace.
According to Barney (1995), the
resource-based view of competition suggests that employers can create
competitive advantage by means of a highly productive workforce -- a valuable
source of competitive advantage because it cannot be copied quickly, is hard to
imitate, is a source of the firm’s competitive advantage, and is somewhat
specific to the climate or environment of the organization. Effective human resource management allows an
organization to attract, keep and develop higher quality employees that are a
relatively rare and valuable resource (Pfeffer, 2005).
It must be stated that there is yet no consensus from the
literature as to what specific HRM approach constitutes “high performance” work
practice. (Wright et al., 2005; Michie & Sheehan, 2005). As an
illustration, Pfeffer (1998), after an extensive review of the subject,
suggests that there are seven HRM practices of successful organizations (employment
security, selective hiring, decentralized work arrangements, performance-based
pay, extensive employee training, reduced status differentials, and information
sharing). More recently, Pfeffer
(2005:97) listed thirteen practices and observed that “there are interrelated
practices – I enumerate 13 but the exact number and how they are defined are
somewhat arbitrary – that seem to characterize companies that are effective in
achieving competitive success through how they manage people.”
The
strategic view of HRM implies that an organization’s human resources can be
“leveraged” to provide a source of competitive advantage. This perspective suggests that the
application and adoption of certain “bundles” of HRM work practices have the
ability to positively impact organizational performance by creating “powerful
connections”, or to detract from performance when “deadly combinations” of
practices are inadvertently placed in the mix.
This “configurational” (as opposed to contingent) approach suggests that
“bundles” of HRM practices, rather than discrete programs, act on
organizational performance. A firm’s
selection of which HRM practices to adopt is made so as to achieve “horizontal
fit” or internal consistency of HRM practices, as well as “vertical fit” or
congruence of the HRM system with other characteristics of the organization,
such as corporate culture and operative strategy (Huselid, Jackson and Schuler,
1997; Michie & Sheehan, 2005).
In light of the focus on human
resource management "systems" or "bundles" of human
resource management practices (Becker and Gerhart, 1996; Wood, 1999, Wright et
al., 2005) as a determinant of organizational performance, our major research
question is whether such systems are associated with voluntary employee turnover. It is anticipated that firms with a higher
use of HRM practices will have lower turnover.
In addition, we explore whether decision-making ideology, human resource
strategy and the presence of a human resource management professional or
department are related to lower quit rates.
Method
Data Collection
We collected our data by surveying
senior executives (typically the chief executive officer or the most senior
individual responsible for human resource management) in the manufacturing
sector. Employers selected to receive the questionnaire were identified and
randomly selected using a mailing list provided by Dun and Bradstreet
Dependent Variable
Our primary focus was on employee
retention, and we were particularly interested in the factors associated with
voluntary turnover (employee quits). Our
dependent variable, which we labeled the quit
rate, was measured by dividing the number of quits over the past year for
the workplace by the number of employees, adding one to remove zeroes, (quit
rates have a lower bound of zero), and then taking the natural logarithm. This procedure is similar to that used by
Delery, Gupta, Shaw, Jenkins and Gangster (2000).
Independent Variables
Our
initial objective was to examine whether employee retention was associated with
the human resource activities at the workplace.
In light of the focus on human resource management "systems"
or "bundles" of practices (Michie and Sheehan, 2005; Wright et al.,
2005), we developed an index of human resource management practices. For each of the twenty-four practices, the
establishment received one point (thus the minimum score was 0 and the maximum
score was 24). The twenty-four practices,
which address a variety of issues including communication, team-based programs,
training, compensation performance appraisal and employee selection, are
outlined in Figure 1. As Pfeffer (2005)
has pointed out, there is no magic list of practices that must be included in
the “bundle”. Our HRM practices index
contains many of the practices used in previous research (see, for example,
Way, 2002; Michie & Sheehan, 2005) but we acknowledge that our index does
not contain every possible practice.
Figure
1
HRM Practices Index
|
Specific
HRM Practice |
|
|
Communication: |
|
|
Employee
Attitude Survey |
|
|
Orientation
Program |
|
|
Employee
Career Counseling |
|
|
Information
Sharing Program |
|
|
Protection Programs: |
|
|
Employee
Assistance Program |
|
|
Sexual
Harassment Policy |
|
|
Recognition Programs: |
|
|
Employee
Recognition Program |
|
|
Internal
Promotion Policy |
|
|
Employee
Suggestions System |
|
|
Team Initiatives: |
|
|
Total
Quality Management Program |
|
|
Project
Teams |
|
|
Problem
Solving Groups |
|
|
Training: |
|
|
Formal
Training Program |
|
|
Employee
Involvement Training |
|
|
Personal
Computer Training |
|
|
Compensation: |
|
|
Profit
Sharing |
|
|
Productivity
Sharing |
|
|
Individual
Merit Pay |
|
|
Employee
Pension Plan |
|
|
Employee
Stock Ownership Plan |
|
|
Other Practices: |
|
|
Formal
Performance Appraisal System |
|
|
Flexible
Work Hours |
|
|
Employment
Selection Tests |
|
|
Written
Job Descriptions |
|
As a result of some
research addressing decision-making ideology (Goll, 1991) and human resource
strategy (Bae & Lawler, 2000), we also developed measures of these two
constructs. Decision-making ideology was
measured from a three-item scale adapted from Goll (1991) that assesses the
level of participative decision-making and open communications. Each item was measured using a six-point
scale (one = strongly disagree and six = strongly agree) and the scale's
Cronbach's coefficient alpha was 0.81.
A firm’s operational
human resource management strategy was assessed, based on the work of Huang
(1998; 2000). HRM strategy was measured
using an eight-item scale that addressed human resources and strategic
management decisions, the degree of devolution of human resource functions to
line management, human resources and long term planning, and human resources
and the future mission of the firm.
Again, each of the eight items was measured on a six-point scale (one =
strongly disagree and six = strongly agree), with the scale's Cronbach's
coefficient alpha at 0.87.
Although one may
question the use of a scale which was used in research on Taiwanese firms,
Huang’s (1998; 2000) articles were written in English and the questions making
up the scale were grounded in literature from mainstream Western human resource
management literature. Respondents to
the survey had no problems understanding the questions and the scale
reliability was very good. As Huang
(1998: 60) states, “The purpose of this research will, it is hoped, not only
promote a deeper understanding of HRM practices in Taiwan, but also help
determine the extent to which western HRM theories are applicable to oriental
social and cultural contexts.”
There has been some
research on the role of the human resource management department in achieving
organizational objectives (Ng & Maki, 1993). One may expect professional human resource
managers to be more aware of current issues in the field. Consequently, we examined whether the quit
rate was associated with the presence of a HRM professional or HRM
department. In the study, the presence
of an HRM professional or department was coded as one and absence as zero.
Control Variables
A number of variables
were entered into each of the regression equations to control for other factors
that were hypothesized to influence quit rates.
The control variables included the demand for the establishment's
primary product or service (one = substantial increase; six = substantial
decline), whether the business had experienced a merger over the past five
years (one = yes; zero = no), whether the business had permanently reduced its
workforce over the past two years (one = yes; zero = no), the union status of
the business (one = union; zero = nonunion), the age of the business (natural
logarithm of the number of years the business has been operating), and size of
the business (natural logarithm of the number of employees). Previous research (Shaw et al., 1998; Delery
et al., 2000; Batt et al., 2002) has demonstrated the need to control for such
variables.
As a check on
non-response bias, we grouped the survey responses depending on whether the
questionnaire was received on the first or second mail out. We then ran a logit model in which we
compared early and late respondents with reference to the control variables
(product demand, merger, permanent workforce reduction, union status, age of
the business, and size of the business) and each of the independent variables
(HRM practices, decision-making ideology, human resource strategy, and presence
of a human resources department). This
approach is based on the notion that late respondents are more likely to be
like non-respondents (Armstrong and Overton, 1977). The logit results showed no difference
between the early and late respondents.
Results
Respondent Characteristics
While respondent firms
varied in size, the average number of employees was 189. The mean number of years a business had been
operating was 38 years. About 23 percent
of firms reported engaging in a permanent workforce reduction of the workforce
over the past two years and respondents reported an increase in demand for
their primary product (mean score of 2.57 on the six-point scale). Although there was considerable variation in
the composition of the HRM practices index, the average number of practices per
workplace was about 16. The mean score
for the decision-making ideology scale was 4.43 (out of a maximum of 6),
suggesting a somewhat positive approach by the firms. However, the average score for the human
resource management strategy scale was somewhat lower (mean of 3.94 out of 6);
while some organizations are heavily involved in HRM strategy, other firms have
not placed much focus on this approach.
Quit Rate
Table
1 provides information on the quit rate experienced by the participant
organizations. As noted previously, the
rate is calculated as a percentage of the total workforce. By way of example, a score of 3 percent on
the quit rate index would indicate that for each 100 employees, a total of
three quit during the past year.
Table
1
Breakdown of Quit Rates
|
Quit
Rate |
Percentage |
Cumulative Percentage |
|
0 percent (No Quits) |
6.5
percent |
6.5
percent |
|
0.1 percent to 2.5
percent |
33.2
percent |
39.7
percent |
|
2.6 percent to 5.0
percent |
24.8
percent |
64.5
percent |
|
5.1 percent to 10.0
percent |
19.4
percent |
83.4
percent |
|
More than 10 percent |
16.1
percent |
100.0
percent |
As observed in Table 1, about 6.5
percent of the respondents reported that no employees quit the organization
over the past year. About one-third of
participants had a quit rate of between 0.1 percent and 2.5 percent and a
little over 16 percent had a quit rate exceeding 10 percent. In times when qualified job candidates are in
short supply, there are considerable costs in trying to replace good employees.
Quit Rates and Human Resource Management
We used OLS regression to
investigate whether a firm’s human resource management activities were
associated with quit behavior. These
results are provided in Table 2. The
first column of results (Model 1) contains the additive scale of 24 HRM
practices as well as the control variables.
The coefficient on the HRM practices variable was both negative and
significant at p <.01. In other words, employers with more HRM
practices (as measured by the HRM practices scale) had lower quit rates.
Table
2
OLS Regression Results:
Quit Rate and HRM
|
Variable |
Model 1 |
Model 2 |
Model 3 |
Model 4 |
|
HRM Variables |
|
|
|
|
|
HRM
Practices |
-0.032*** |
|
|
|
|
|
(0.009) |
|
|
|
|
Decision-Making
Ideology |
|
-0.121** |
|
|
|
|
|
(0.054) |
|
|
|
HRM
Strategy |
|
|
-0.182*** |
|
|
|
|
|
(0.057) |
|
|
HR
Professional/Department |
|
|
|
-0.149 |
|
|
|
|
|
(0.102) |
|
Control Variables: |
|
|
|
|
|
Product
Demand |
0.045 |
0.046 |
0.066 |
0.060 |
|
|
(0.047) |
(0.048) |
(0.050) |
(0.048) |
|
Merger |
-0.211** |
-0.253** |
-0.190* |
-0.244** |
|
|
(0.097) |
(0.098) |
(0.100) |
(0.098) |
|
Permanent
Reduction |
-0.043 |
-0.065 |
-0.065 |
-0.353 |
|
|
(0.108) |
(0.111) |
(0.112) |
(0.092) |
|
Union
Status |
-0.377*** |
-0.378*** |
-0.379*** |
-0.040*** |
|
|
(0.091) |
(0.092) |
(0.094) |
(0.110) |
|
Number
of Employees |
-0.238*** |
-0.284*** |
-0.263*** |
-0.252*** |
|
|
(0.075) |
(0.079) |
(0.078) |
(0.079) |
|
Years in
Business |
-0.073 |
-0.087 |
-0.070 |
-0.088 |
|
|
(0.055) |
(0.055) |
(0.056) |
(0.055) |
|
(Constant) |
3.657*** |
4.003*** |
3.950*** |
3.355*** |
|
|
(0.424) |
(0.504) |
(0.480) |
(0.428) |
|
F-Test |
10.877*** |