Professional Documents
Culture Documents
Volume Twelve
2000
INTRODUCTION
In recent years, scholars and practitioners have expressed concerns with
traditional performance measures that focus solely on financial metrics such as
return on investment or net earnings (Atkinson et al. 1997; Ittner et al. 1997;
Kaplan and Norton 1996; Lynch and Cross 1991; Shields 1997). The balanced
scorecard (BSC) approach to management (Kaplan and Norton 1992, 1993,
1996) has gained prominence in management accounting research as a way of
integrating financial and nonfinancial performance measures (for reviews, see
Atkinson et al. 1997; Ruhl 1997; Shields 1997; Simons 2000). The BSC views
organizational performance from four dimensions: financial (or shareholders),
customers, internal business processes, and learning and growth. Atkinson et
al. (1997, 93–94) suggest that BSC has the potential to provide planners with a
way of expressing and testing a sophisticated model of cause-and-effect in the
organization—a model that provides managers with a basis on which to manage
the drivers of desired outcomes.
While work of Kaplan and Norton has added to our knowledge of how a BSC
can provide managers with an integrative framework to manage organizational
activities, little evidence is available outside the U.S. on current practice in the
area (Hoque et al. 1997; Creelman 1998).1 Atkinson et al. (1997, 94) note that
“the balanced scorecard is among the most significant developments in
management accounting and thus, deserves intense research attention.” This
paper is an attempt to contribute to the limited body of knowledge in this area.
1
The purpose of this research is to search for a relationship between BSC
usage and (1) organization size, (2) product life-cycle stage, and (3) strength of
market position. It also looks for a contingent relationship between
organizational performance (the outcome variable) and the match between BSC
usage and the three contextual variables described above. The framework for the
research is illustrated in Figure 1.
FIGURE 1
Framework for the Study
Organization Size
Strength of
Market Position
The next section of the paper briefly reviews the relevant literature and
develops the research hypotheses. The research method applied is described in
section three. Section four presents our results. The final section concludes the
paper.
2
from sub optimizing by improving one measure at the expense of others. 3 To
achieve a balance among the four dimensions of the BSC, a company should pay
attention to all of them.
Organization Size
Contingency theories of organizations developed by Burns and Stalker
(1961), Lawrence and Lorsch (1967), and Woodward (1965) suggest that size
may affect the way organizations design and use management systems.
Numerous accounting studies have drawn on this theoretical framework.
Merchant (1981, 1984) claims that organizational growth poses increased
communication and control problems. Bruns and Waterhouse (1975), Ezzamel
(1990), and Libby and Waterhouse (1996) suggest that as firm size increases,
accounting and control processes tend to become more specialized and
sophisticated.
Others in the organizational literature (e.g., Burns and Stalker 1961;
Chandler 1962; Pugh et al. 1969) claim that size is related to greater
decentralization and structuring of activities because of information processing
constraints upon senior management. Furthermore, the need to stimulate
effective communication flows becomes more apparent in larger organizations
where the behavioral orientation characterizing management controls in small
organizations become unworkable. As a consequence, in large business
enterprises, a broader set of information and measurement issues arises (Kaplan
and Atkinson 1998). Small companies frequently do not require elaborate
performance evaluation techniques, as the strategy setters, usually the owners,
are close to the “action.” Based on this a priori reasoning, it is proposed here
that larger organizations are likely to place a greater reliance on a BSC approach
to management than are smaller organizations.
3
measures of whether the firm is “doing the right thing.” In addition, financial
outcomes from important decisions may be too far in the future when the firm is
in an early stage, and nonfinancial indicators like new product development and
customer-response measures give earlier indications of whether appropriate
decisions have been made.
Market Position
Market position in this study refers to a company’s revenue share in relation
to its competitors in a particular market. Following the Galbraith (1977)
argument, Merchant (1984) suggests that for firms in a weak market position,
the demand for adaptability and creativity would be greater than the demand for
internal communication, while for firms in a strong market position, the demand
for internal communication would be greater. He argues that when an
organization has a strong market position, the use of budgets for internal
controls would be more prominent than for a firm in a weak market position.
A BSC can satisfy companies’ greater internal communication needs, as it
facilitates decisions and actions that support strategies based on the needs of
stakeholders, internal and external customers, regulatory bodies, managers, and
employees and requires involvement by all levels of the organization (Kaplan and
Norton 1996; Atkinson et al. 1997). We propose that organizations with a strong
market position have a greater demand for internal communication, and thus
are likely to place greater emphasis on the use of a BSC. Conversely, a weak
position in a particular market may create a lesser demand for internal
communication. A lesser demand for internal communication suggests a lower
deployment of sophisticated management systems such as the BSC.
The preceding discussion can be summarized with the following hypothesis:
H1: Balanced scorecard usage is positively associated with: (a) larger
organization size; (b) businesses with products at the early/growth
stage; and (c) businesses with a strong market position.
RESEARCH METHOD
A questionnaire4 with a cover letter and a postage-paid, self-addressed
envelope was mailed to the chief financial controllers of 188 Australian
manufacturing firms in July 1997. These companies were randomly chosen from
the Business Who’s Who of Australia (Dun & Bradstreet 1997). Forty-three of the
188 questionnaires sent out in the first mailing were returned. A second mailing
resulted in a further 27 returned questionnaires. Four of the 70 respondents
failed to complete the questionnaire, citing reasons such as contravening
company policy and staffing constraints. Consequently, the adjusted usable
response rate is 35.1 percent.
A sample of non respondents was contacted by telephone to investigate
reasons for nonresponse. The reasons given for nonresponse were consistent
with those who had returned the questionnaire without completing them, that is,
either due to staffing constraints or contravening company policy. To test for the
existence of possible response bias, t-tests for two independent samples were
undertaken by testing first and second mailing returns as suggested by
Oppenheim (1966, 34). No statistically significant differences in the mean scores
on the firm size, industry, or performance indices between the early and late
responses were noted, suggesting the absence of response bias. The average
experience of the participants in the company was 6.5 years, with a range of four
months to 25 years. Table 1 provides a profile of the responding firms. In this
study the responding firms were independent business units, not the head
offices.
TABLE 1
Profile of Responding Companies (n =
66)
Number of Employees n Organization Type
4
Metal Product
Machinery & Equipment
300–449 5 Textile, Clothing, Footwear & Leather
Printing, Publishing & Recorded Media
Metal Product
Machinery & Equipment
450–999 7 Textile, Clothing, Footwear & Leather
Printing, Publishing & Recorded Media
Petroleum, Coal, Chemical & Associated Product
Metal Product
Machinery & Equipment
1,000 or greater 18 Food, Beverage & Tobacco
Textile, Clothing, Footwear & Leather
Metal Product
Machinery & Equipment
Measurement of Variables
Organization Size
Organization size was measured using three measures: sales turnover, total
assets, and number of employees. These three measures were highly correlated
(p < .001). We conducted our analyses using each of these measures as a proxy
for size, one at a time, but the results were similar throughout. Consistent with
previous studies (Bruns and Waterhouse 1975; Merchant 1981, 1984; Ezzamel
1990; Libby and Waterhouse 1996), only those results obtained when the
number of employees was used as a proxy for size are reported. Due to the
nonnormality of “size,”5 it was transformed using the square root of the variable
for use in regression analysis. Table 2 presents the descriptive statistics for this
variable in its original and transformed form.
TABLE 2
Descriptive Statistics
(n = 66)
Standard Theoretical Actual
* Transformed statistics.
5
a
Used in additional regression analysis reported in Appendix B.
NA = not available.
BSC Usage
BSC usage was measured using a 20-item scale similar to that developed by
Hoque et al. (1997). The instrument comprised items that incorporate Kaplan
and Norton’s (1992) four dimensions of the BSC. It asked respondents to
indicate the extent to which each item was used to assess their organization’s
performance on a fully anchored, five-point Likert scale ranging from 1 (not at
all) to 5 (to a great extent). A reliability check for the measure in this study
produced a Cronbach alpha (Cronbach 1951) of 0.81, which is considered to be
well above the lower limits of normal acceptability (Nunnally 1967). It should be
noted that our BSC measure might not pick up the strategic linkages of a real
BSC usage; it does pick up firms’ tendency to use quantitative measures
(frequency and extent of reporting) of several kinds in assessing performance.
This construct shares with the BSC construct the idea that financial measures
alone are insufficient.
Appendix A presents the descriptive statistics for each item of the scale. A
principal components analysis (PCA) with varimax rotation was performed to
determine whether the measures used in the survey can be grouped according to
6
the BSC’s four perspectives. The results of the factor analysis, which also appear
in Appendix A, are roughly consistent with the four perspectives identified by
Kaplan and Norton (1992).
A mean score was calculated for each of the four perspectives. The primary
hypothesis tests were performed using an average of these four perspective
means to represent overall BSC usage. Additional tests were performed using
each of the perspective means separately. Table 2 presents descriptive statistics
for the measure.
TABLE 3
Correlation Matrix
(Pearson Coefficients) (n
= 66)
Organization Product Life- Market Overall BSC Organizational
Variable Size Cycle Stage Position Usage Performance
*, **, *** Significant at the 10 percent, 5 percent, and 1 percent level, respectively.
Organizational Performance
Organizational performance was measured by appraising five dimensions of
performance: return on investment, margin on sales, capacity utilization,
customer satisfaction, and product quality. The instrument is conceptually
consistent with Kaplan and Norton’s (1992) BSC theorizing. Following the
procedure used by others (e.g., Merchant 1984; Abernethy and Lillis 1995),
respondents were asked to indicate their organization’s performance compared
to their competitors along the above five dimensions on a scale from 1 = below
average to 5 = above average. Although not presented here, it is noted that the
five dimensions in the performance instrument were positively and significantly
related to each other (p<0.05). A PCA of the five items revealed that all items
loaded on to a single factor with eigenvalue 2.52. The factor score was saved for
use in the regression analyses. The Cronbach alpha for this scale was 0.75,
indicating satisfactory internal reliability of the scale. Table 2 presents
descriptive statistics for the measure.
RESULTS
Table 2 provides descriptive statistics for all variables. Pearson correlation
coefficients appear in Table 3. As expected, BSC usage is positively and
significantly correlated with organization size (r = 0.25, p < 0.10), product life-
cycle stage (r = 0.26, p < 0.05), and organizational performance (r = 0.46, p <
0.01). Although strength of market position is positively correlated with BSC
usage (r = 0.18), it is generally not statistically significant (p-value > 0.10). The
contextual variables, organization size, product life-cycle stage, and strength of
market position are not significantly related to each other, suggesting that
multicollinearity is unlikely. Tests of multicolinearity via tolerance and variance
inflation factor (VIF) presented in Table 4 revealed that multicollinearity does not
pose a problem in interpreting these results. Furthermore, tests of nonlinearity
and heteroskedasticity of the data indicated no major problem for regression
analysis.
TABLE 4
Regression Results (H1)
X4 (BSC Usage) = α0 + β1X1(Size) + β2X2 (Product Life Cycle Stage) + β3X3 (Market Position) + e (n = 66)
Coefficient (predicted
sign
Variable in brackets) Estimate t-value p Tolerance VIF
TABLE 5
Analysis of Variance (ANOVA): Two-Way Interaction Effects (H2–H4)
Sum of Mean
Source of Variation Squares DF Square F Sig. of F
Closing Remarks
We close this paper with a discussion of what we can infer from this study.
The findings reported in this paper suggest that large firms make more use of
the measures in the questionnaire used (Appendix A) than do small firms. They
do not suggest whether that is because large firms get (or expect to get) more
benefit from these measures, or because they can spread the fixed cost of
information systems over larger output and therefore find additional measures
more affordable, or because they are more likely to pursue performance-
measurement fads (or are more likely targets for consultants, more concerned
about their image as up-todate management because public image matters more
to them, etc.).
The results also suggest that firms with more new products make more use
of new-product-development measures; however, we do not know which way
causation runs here. Emphasizing these measures might lead to more new
product development, either in reality or because people are gaming the
measure. Or conversely, a firm’s proportion of new products might be
exogenously given; if it is large, then the firm sees a greater need to measure
performance in new-product development and launch.
14
Descriptive Statistics and Principal Components Analysis of the BSC Items 7 (n = 66)
7 The boldface data represent the factor loadings that were greater than 0.40. b As noted in the last section of this paper,
firms use these measures because they are an important part of their portfolio.
APPENDIX B
Additional Regression Analysis with Independent Variables (Organization Size, Product LifeCycle
Stage, and Market Position) and Four Perspectives of the BSC (H1)
Variable Coefficient Estimate t-value p Tolerance VIF