...1 z.
alt“;
.1.
. 3.3.4 r
:0 h 4 $3.“!
r.) .
{r

,:l
W .
an?

x. "y...
‘K‘ .a I
sf. 1.3:. saw
I v. ‘

x
v ) ~
11“)

-
Lin,
. Juri 2 I r
I up i. ‘, .
:av:"xz:x

*u

5.." a)
3“ ! all.
rat: c112“...
.~..1,.I.l..v..

unaffiifi.
.: ..~.. .. c,

.14.... .
v It):
51
. Ah.-‘
:1»«

:uflufivfty.
mitt”.
. .

2 3n Lxlt...
211%.:

“n. It . .A It I
. {v (-0..
15.

5.....1. .5031. I 1
v3.— ... t I p.
.Av 151. .

1.13.! . ‘1'

uh. M'sirnL .10"

.1; I.

 

“‘Z JWWWIitiiiili‘lil'liiiill'ifillififla

3 1293 01417 2591

This is to certify that the

dissertation entitled

AN EMPIRICAL ANALYSIS OF
REA ACCOUNTING SYSTEMS, PRODUCTIVITY,
AND PERCEPTIONS OF COMPETITIVE ADVANTAGE

presented by
Julia Smith David

has been accepted towards fulfillment
of the requirements for

Ph . D . degree in Account ing

 

WW fflc

Major professor

Date July 14, 1995

 

MSUL: an Affirmatiw Action/Equal Opportunity Institution 0-1277‘

 

 

mflFfl _.~ _

" F.- -—‘.- *mv
- —~mv-- ~—.,..- 57"-

 

 

 

 

 

LlERARY
1 Michigan $iate
University

 

 

 

PLACE Ii RETURN 80X to mat-this checkoutilom your record.
TO AVOID FINES mum on or Moro date duo.

DATE DUE DATE DUE DATE DUE

        

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Vii—T“?

MSU ioAn Affirmative MM Opportunity IMW

 

 

 

AN EMPIRICAL ANALYSIS OF REA ACCOUNTING SYSTEMS,
PRODUCTIVITY, AND PERCEPTIONS OF COMPETITIVE ADVANTAGE

By

Julia Smith David

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Accounting

1995

ABSTRACT
AN EMPIRICAL ANALYSIS OF REA ACCOUNTING SYSTEMS,
PRODUCTIVITY, AND PERCEPTIONS OF COMPETITIVE ADVANTAGE
By

Julia Smith David

Weakness have been identified with traditional debit/credit accounting information
systems. In response to these, McCarthy (1982) developed the theory of REA accounting
systems that include detailed information about economic resources, events, and agents.
While several directed REA implementations have been described in the literature, no
empirical analysis has been performed to determine whether REA systems provide benefits
to the organizations using them. This dissertation, therefore, presents the first empirical
analysis of productivity and competitive advantage improvements from REA systems.

To analyze systems currently operating in Organizations, an Accounting System
Characteristics (ASC) metric was developed which is used to place organization’s systems
on a continuum between traditional and REA accounting information systems. In
addition, a questionnaire was developed to collect executive’s perceptions of Competitive
Advantage fiom their information system. This questionnaire relied heavily on the
theoretical model of Competitive Advantage proposed by Bakes and Treacy (1986).

These metrics were used during site visits to eight companies in the pulp and paper
industry. Executives of key functional areas completed the questionnaire of Competitive
Advantage during a preliminary meeting with the researcher. This meeting was followed

by two days of interviews with both executives and stafl‘ members throughout the

organizations. Each interview was structured using the ASC metric so that identical
information would be collected from each participant.

Hypothesis testing was performed after the completion of all eight site visits. The
results provide evidence that the more sophisticated systems are providing firms with
administrative efiiciencies, although they are not supporting interorganizational strategies
for competitive advantage. In addition, firms with more sophisticated systems are more
productive than those that use more traditional systems.

The major contribution of this work is that it provides a method for future
researchers to compare Operational systems. Key systems characteristics are identified,
and the ASC metric provides a method Of capturing information about these
characteristics. As a result, more detailed hypotheses may be posed, and evidence can be

gathered to evaluate the benefits that accounting systems provide to firms using them.

Bakos, J.Y. and M. E. Treacy. ”Information Technology and Corporate Strategy: A
Research Perspective.” MS Quarterly 10:2 (June 1986) pp. 107-119.

McCarthy, WE. ”The REA Accounting Model: A Generalized Framework for
Accounting Systems in a Shared Data Environment.” The Accounting Review (July
1982) pp. 554-77.

Copyright by
JULIA SMITH DAVID

1995

DEDICATION

to

Scott A David,

whose love and support made this possible

and

Jay Wilson Dorfl‘, IV,

who didn’t get the chance to finish his

ACKNOWLEDGMENTS

I would first like to thank the members of my committee for their time, insights,
and enthusiasm. Dr. William E. McCarthy has been a mentor since 1983 when I enjoyed
his undergraduate class and became fascinated with Accounting Information Systems.

Dr. Severin Grabski has been instrumental in my doctoral program, encouraging me to
participate in research fiom my first semester. Dr. Frank Boster has been a wonderfirl
statistical reference, working with me to understand the implications of the methods used
in this study. Dr. William Punch has provided a computer science perspective to this
study, challenging me to consider the complex nature of the problem being studied.

I would also like to recognize each of the companies that participated in this study,
and to thank all Of the contact people who scheduled my visits. However, to maintain
confidentiality, I am limited to saying that these people know who they are and that I
appreciate all of the time they gave to make this project a success. I am looking forward
to working with many in the firture. In addition, I am able to acknowledge the help of
Mr. Randy Johnson who was not only willing to listen to problems I experienced, but also
performed a “match maker” role, identifying potential companies and making
introductions. Without his influence, I would have been much less likely to have found

willing participants.

Elizabeth Connors and Tami Kuhn read of earlier drafts, and made insightful,
critical comments, challenging me to be more careful and specific in the text. In addition,
Greg Gerard was very helpful with the factor analysis and path analysis portions of this
document. Throughout my doctoral program the dinner group shared their insights into
how to best succeed in the doctoral program; I am sure that we will all continue to remain
true fiiends having shared so much while here. TO all, many heartfelt thanks.

The Department of Accounting has supported this research both financially and by
being flexible in my work schedule. Thank you for keeping this a four year program.

Finally, the road to this dissertation extends beyond the four years at Michigan
State University. My parents, Wayne K. Smith and Noreen B. Smith, “encouraged” me to
take education seriously since I started school. In addition, they have believed in me
throughout this most recent adventure. Dr. Raymond S. Schmidgall deserves credit for
being the first person to show me the challenges and rewards of the academic lifestyle. He
was also the first person to predict I would one day earn a Ph.D.

And most importantly, I must thank my husband, Scott, for his patience, love, and
support. When I questioned whether the doctoral program was the correct place for me,
he was able to stay focused on the goal at the end, and he reminded me that the
destination was worth the trip. Without him, I am sure I never would have made it this

far.

TABLE OF CONTENTS

LIST OF TABLES ............................................................................ xii
LIST OF FIGURES ........................................................................... xiv
Chapter 1 - Introduction .............................................................................. 1
Chapter 2 - Theoretical Foundations Of REA Accounting Information Systems ..... 4
2.1 Traditional Accounting ....................................................................... 4

2.2 REA Theory .............................................................................. 9
Chapter 3 - Construct and Hypothesis Development ........................................... 19
3.1 Competitive Advantage .................................................................... 19

3.2 Productivity ............................................................................ 30

3.3 Summary ............................................................................ 3 5
Chapter 4 - Methodology and Construct Development ........................................ 37
4.1 Introduction ............................................................................ 37

4.2 Construct Development .................................................................... 39

4.2.1 Accounting Systems Characteristics Questionnaire ............. 39

4.2.2 Competitive Advantage Survey .......................................... 42

4.2.3 Eficiency and Productivity Measures ................................. 44

4.3 Methodology ............................................................................ 45

4.3.1 Sample . ............................................................................ 45

4.3.2 Site Visit Procedures ......................................................... 46

Chapter 5 - Data Analysis and Results ................................................................ 50

5.1 Qualitative Results ........................................................................... 52
5.1.1 Key Systems Characteristics .............................................. 52
5.1.1.1 Support Critical Events ....................................... 53
5.1.1.2 Detailed History of Events .................................. 56
5.1.1.3 Integrated Data Processing .................................. 57
5.1.1.4 Data Availability .................................................. 60
5.1.1.5 Real Time Processing .......................................... 62
5.1.1.6 Directed REA Design and Implementation .......... 64
5.1.1.7 Elimination of Journal Entries .............................. 65
5.1.1.8 Summary ............................................................. 66
5.1.2 Key Organizational Characteristics ..................................... 67
5.1.2.1 Complexity .......................................................... 68
5.1.2.2 Size ..................................................................... 69
5.1.2.3 Organizational Structure ..................................... 70
5.1.2.4 Technical Sophistication ...................................... 71
5.1.2.5 Costs versus Benefits .......................................... 72
5.1.3 Summary of Qualitative Results ......................................... 73
5.2 Quantitative Results ......................................................................... 73
5.2.1 Preliminary Data Analysis .................................................. 74
5.2.1.1 Factor Analysis of Competitive Advantage
Metric ................................................................. 74
5.2.1.2 Factor Analysis of the User Satisfaction
Metric ................................................................. 84
5.2.2 Hypothesis Testing and Results .......................................... 86

5.2.2.1 Hypothesis One: Test of the Bakos and

Treacy Model ...................................................... 86
5.2.2.2 Hypothesis Two: REA Systems Provide
Competitive Advantage ....................................... 89
5.2.2.3 Research Question 1: REA Systems
Indirectly Provide Competitive Advantage ........... 91
5.2.2.4 Hypothesis Three: REA Systems Lead to
Operational Improvements .................................. 94
5.2.2.4.] Productivity Improvements ................... 95
5.2.2.4.2 Eficiency Improvements ....................... 98
5.3 Discussion of Results ..................................................................... 100
Chapter 6 - Contributions and Suggestions for Future Research ........................ 104
6.1 Contributions .......................................................................... 104
6.1.1 Accounting Systems Metric ............................................. 104
6.1.2 Operational Definition of REA Accounting Systems ........ 106
6.1.3 Competitive Advantage ................................................... 107
6.1.4 Efliciency and Productivity .............................................. 108
6.2 Future Research .......................................................................... 108
6.3 Conclusions .......................................................................... 110
LIST OF REFERENCES .......................................................................... 111
APPENDIX A: Accounting Systems Characteristics Questionnaire .................. 115

APPENDIX B: Competitive Advantage and User Satisfaction Questionnaire 123

APPENDIX C: User Satisfaction Questionnaire .............................................. 127

APPENDIX D: Systems Analysis Preliminary Information Form ...................... 128
APPENDIX E: Systems Analysis Tentative Schedule ...................................... 130

APPENDIX F: Company REA Diagrams and ASC Summary Sheets ................ 134

Table 1 -
Table 2 -

Table 3 -

Table 4 -

Table 5 -

Table 6 -

Table 7 -

Table 8 -

Table 9 -

Table 10 -

Table 11 -

Table 12 -

Table 13 -

Table 14 -

LIST OF TABLES

Weaknesses of Traditional Accounting Systems ............................. 8
Key Characteristics to Difi‘erentiate Traditional and

REA Accounting Information Systems ........................................ 53
Key Organizational Characteristics .............................................. 67
Executives Who Completed the Competitive Advantage

Survey ............................................................................ 76
Correlation Matrix for Questions in a Four Factor Model ............ 79

Correlations for Interorganizational Strategies in the
Modified Model of Competitive Advantage ................................. 81

Consistency Analysis for Administrative Strategies and
Competitive Advantage ............................................................... 82

Correlation Matrix for Development Characteristic
Questions ............................................................................ 84

Expected and Actual Correlations between User
Satisfaction Questions ................................................................. 85

Linear Regression Analysis of Factors Leading to

Competitive Advantage ............................................................... 87
Competitive Advantage Linear Regression with

Development Characteristics ....................................................... 89
Correlation Matrix Of Competitive Advantage Factors and

ASC Scores ............................................................................ 91
Demographic Statistics ................................................................ 95
Correlations between Productivity Measures and ASC Scores ...... 98

Table 15 - Correlations between Productivity Measures and
ASC Scores Removing Outlier ..................................................... 98

Table 16 - Correlations between ASC Scores and Measures of Efficiency... 100

xiii

Figure 1:
Figure 2:
Figure 3:

Figure 4:

Figure 5:
Figure 6:

Figure 7:

Figure 8:

Figure 9:

LIST OF FIGURES

Basic REA Representation ........................................................... 10
Revenue Cycle Representation in an REA Design ........................ 11
REA Model for a Manufacturing Firm ......................................... 12
Implementation Of Sales Transactions in an REA

Accounting System ...................................................................... 14
Causal Model Of Competitive Advantage ..................................... 23
Hypothesized Model Of Competitive Advantage .......................... 27

Hypothesized Model Of the Impact of REA Systems
and RQl ............................................................................ 36

Path Model Of an Indirect Effect of ASC on
Competitive Advantage ............................................................... 92

Path Model of a Direct Efi‘ect of ASC on
Competitive Advantage ............................................................... 93

xiv

Chapter 1 - Introduction

Computerized accounting systems have become standard features in modern
businesses, but it has been difficult to measure their impact on organizations. Much has
been written about the "productivity paradox" of increased computer spending coupled
with decreasing productivity in white collar workers (Roach 1991; Brynjolfsson 1993).
Other research has focused on whether investing in computers provides organizations with
competitive advantage (Porter 1985; Johnston and Carrico 1988; Johnston and Vitale
1988; Glazer 1993). Overall, results of these studies have been mixed.

Ifcomputer systems actually provide benefits to organizations, a possible
explanation for the negative results is that not all computer systems provide productivity
or competitive advantage improvements that can be measured empirically. Rather, only
computer systems with certain characteristics could provide measurable benefits. This
dissertation is the first empirical analysis to compare the efl‘ects Of traditionally designed
accounting systems and accounting systems that have been designed to reflect the business
processes, specifically systems that model resources, events, and agents (REA systems).
The major hypotheses are that REA systems can assist managers in improving both their
productivity and their organization's position in the market. Firms using more traditional
accounting systems that were designed to automate bookkeeping processes will report
significantly less, if any, improvements in productivity or perceptions of competitive
advantage compared to those using REA systems.

A field study approach incorporating personally-distributed surveys was used to
obtain in-depth information about eight organizations. Prior to visiting the organizations,

management provided financial and Operational information about their organizations that

could be used to measure productivity and market position. During the site visits,
executives and stafl‘ members from several functional areas completed surveys that
measured their perceptions of their firrn’s competitive advantage and how their computer
system helped provide any advantage. Each of these peOple was interviewed individually
to gather detailed information about what procedures they performed, where the
information they needed to perform these procedures was stored, and how their
information system provided productivity improvements. The firm's computer system
characteristics were subsequently scored along a continuum between a traditional
accounting information system and an REA accounting information system.

Two difi‘erent types of analysis were performed with the data gathered from the
site visits. First, correlations between system characteristics, productivity levels, and
perceptions of competitive advantage were identified to test the main hypotheses of this
study. Second, qualitative data fi'om the visits were used to provide a more detailed
operational definition of REA accounting information systems and to summarize the key
systems characteristics that provided benefits to the organizations participating in this
study.

This study makes several contributions to the REA accounting systems literature.
First, a metric has been developed that can be used to evaluate existing systems. It
identifies systems that are more similar to REA and those more like traditional systems. In
addition, it has been used to efl‘ectively communicate REA concepts to firm management,
and to provide them with a concise view of their organization and its system.

Second, the qualitative results provide an operational definition of REA systems

that can be used in future research. While most of the work in this area has focused on

 

Sn

theoretical characteristics of these systems, the key systems characteristics described in
Chapter 5 provide researchers a technique to differentiate firms between the traditional
and REA extremes. They can also be used to develop more specific hypotheses about the
benefits Of REA systems.

Third, the quantitative results are some of the first empirical data regarding the
benefits that may arise fi'om REA systems implementations. These results support a
modified version of the Bakos and Treacy (1986) model Of competitive advantage that
theorizes that information technology can provide competitive advantage if it either
improves administrative efficiencies or implements interorganizational strategies. In
addition, path analysis shows that the systems more like REA systems are perceived to
improve administrative eficiencies. However, the REA systems were not found to be
assisting with the interorganizational strategies.

Limited evidence is provided that REA systems are correlated with improved
productivity and eficiency. Specifically, there is a significant correlation between the
more advanced systems and productivity measured as sales per employee hour. In
addition, these systems are significantly correlated with improved eficiency in accounts
payables processing.

The remainder of this dissertation is organized as follows. Chapter two describes
REA accounting systems and highlights the difi‘erences between them and more traditional
accounting systems. Chapter three develops the hypotheses, while chapter four discusses
the methodology to be used and the development of the metrics used to test the
hypotheses. Chapter five summarizes the data analysis, and chapter six concludes with the

study’s contributions and suggestions for firture research.

Chapter 2 - Theoretical Foundations of REA Accounting Information Systems
2.1 Traditional Accounting

Traditional accounting has roots dating to approximately 1250-1400 (Kee 1993).
Luca Pacioli is credited with writing the first manuscript describing the double-entry
accounting system in 1494. He outlined concepts and processes that could be used to
manage small merchant ventures when the only tools available for record keeping were
paper and pens. He developed the ideas of the chart of accounts, ledgers and journals, and
double-sided entries for the journals. He also showed how the ledgers could be used to
help insure arithmetically correct figures, prevent fraud, and document the organization’s
transactions. His techniques were so successful that they have continued to be used
through the 20th century.

The first step in using these traditional accounting methods is to establish a
categorization structure called a chart of accounts. Each account represents an asset,
liability, equity, income, or expense amount and is assigned a unique account number.
These are listed in a general ledger that is used to record and report summary balances for
each account. Each transaction can then be described by a journal entry that changes the
levels Of two or more accounts. However, to limit the number of entries to the general
ledger, the accountant may first record a transaction in a subsidiary ledger that is a listing
of all transactions of one type. Periodically, these detailed listings are summarized, and
one journal entry is entered into the general ledger. Then the general ledger accounts and
balances can be used to prepare the financial statements. The income statement uses the

income and expense accounts to provide management with a financial report of how the

organization’s financial position changed over a period of time. The balance sheet shows
the firm’s assets, liabilities and equities at a specific point in time.

The following is an example of how a sales transaction would be recorded using
these traditional accounting techniques. First, the sale is recorded in a sales ledger. At the
end of a batch process called invoicing, summary sales totals are calculated from the sales
° ledger and are entered into the general ledger. If the chart of accounts has account
numbers for sales summarized by company, division, and warehouse, the following journal

entry may be recorded at the end of the invoice process:

9/22 1-10-00-01 Cash, main account 100,000
1-50-15-01 Sales, div 15, wh 1 60,000
1-50-20-02 Sales, div 20, wh 2 40,000

Record sales for 9/22 invoice processing

This account number structure uses a seven position account number to describe the
important components of each transaction. In this example, the first position specifies the
company that participated in the transaction (the organization operates several companies
although only company number one has had sales transactions during this period). The
next two positions identify the type of account that is being updated (10 stands for cash
and 50 stands for sales). The next two positions represent the division responsible for the
transaction (sales are subtotaled by division, but cash is maintained at the corporate level).
Finally, the last two positions are used to identify sub-accounts. For the cash account,
difi‘erent bank accounts are identified with these positions. Difl‘erent warehouse sales are

identified in the sales accounts.

While this method of recording business transactions is still being used in most
accounting systems today, the accountant’s tools have changed radically over the last 100
years. First, mechanical devices, such as peg boards and writing boards, were developed
in the late 1800’s (Key 1993). With the introduction Of such devices, clerical efficiency
was improved. In addition, adding and posting machines improved mathematical accuracy
and controls during the early 1900’s. Computers have improved emciency and accuracy
even more. Computers have also made difiicult and time consuming tasks (such as
generating financial statements from the general ledger) very simple and almost automatic.
They are able to store detailed information, to sort it in difi'erent sequences, and to provide
difi‘erent subtotals as needed. They have provided accountants the Opportunity to expand
their role in the organization beyond producing only financial statements and have enabled
improved accounting techniques and support.

As a result, new demands are being made on accounting departments. For
example, the chart of accounts may no longer be the only (or best) way for operational
managers to categorize their organization’s transactions. Rather than using financial
statements that report summarized financial information on a periodic basis, managers are
demanding non-financial information, different levels Of aggregation for different
managerial firnctions. They are also requesting reports in formats other than balance sheet
and income statements (Davenport 1993).

Because of weaknesses inherent in ledgers, journals, and financial statements,
accountants relying on them are unable to meet manager’s new requests. For example,
Table 1 identifies four major weaknesses that have been identified in the traditional

accounting systems (McCarthy 1982). When data are stored in joumai entry form, the

only information available consists of the general ledger accounts affected, the amount of
the effect, and the date of the transaction. No additional information, such as detailed
product, customer, or quality information is available. In addition, data are Often
summarized before a journal entry is entered into the system. As discussed, it is not
uncommon for sales information to be summarized so only one joumai entry updates the
general ledger each day. As a result, detailed information is not available to managers.
Similarly, the structure imposed by the chart of accounts limits the data categorization
available to managers. For example, sales managers are able to determine daily sales for
each division if their chart of accounts has separate sales accounts for each division.
However, information about sales by customer or sales by product would not be available
with this chart of accounts structure. Finally, because of the type of information stored in
traditional systems and the level of aggregation required, it is very difficult to combine
financial and non-financial information across the organization. As a result, organizations
with traditional accounting systems often operate additional systems to provide the
information necessary for other functional groups. These additional systems will contain

redundant data that are often inconsistent and not able to interface with other systems.

 

Table 1

Weakness of Traditional Accounting Systems

(McCarthy 1982)
1. Stored data have limited characteristics about transactions.
2. Data classification schemes are not always appropriate or supportive.

3. Aggregation level of stored data is too high.

4. Useful integration of financial and non-financial data across the organization is
dificult, if not impossible.

 

 

 

Andros, Cherrington and Denna (1992, p. 29) have identified other weaknesses of
traditional accounting systems including that they do not support organizations’ business
processes, but rather “institutionalize antiquated, inefficient, and inefi‘ective business
processes.” As a result, organizations Often accept current processing methods rather than
looking for more efficient procedures to “reengineering” their organization. One of the
first examples Of reengineering involves accounts payable processing at Ford Motor
Company (Hammer 1990). Before reengineering, Ford, like virtually all organizations,
performed three-way matching of purchase orders, receiving reports and vendor invoices
before cash disbursements were made. After reengineering, Ford’s management realized
that the disbursement was the important event (rather than the matching of the
documents), and that they had the information needed to prepare the disbursement upon

receipt of the goods. Management reengineered the process, and now distributions are

 

 

 

ac.

2.]

rec
din
idei
are
con
OCOl
dim
ever
an ex
Olga;

ACC

Ei'eni

made to vendors upon receipt of goods. Vendor invoices have been eliminated fiom the
process. After the process was modified, Ford’s accounts payable department has
achieved a 75 percent reduction in head count.
2.2 REA Theory

In response to the original weaknesses, McCarthy (1982) developed a new
accounting fiamework referred to as REA. With this approach, accountants model reality
directly by identifying the critical economic gesources, gvents, and agents (hence the REA
identification) in the organization’s operations, rather than focusing on a chart Of
accounts. A resource is defined as something that provides value, is under the firrn’s
control, and is scarce (Ijiri 1975). An event is an activity that increments or decrements an
economic resource and results “from production, exchange, consumption and
distribution” (Y u 1976, p. 256). Two agents, one internal and one external, participate in
every event. For example, CASH may be identified as a resource. It is increased through
an event called CASH RECEIPT in which a CUSTOMER (external agent) gives the
organization CASH (resource); the CASH RECEIPT event is performed by an
ACCOUNTS RECEIVABLE CLERK (internal agent).

In addition, the duality principle of REA recognizes the dual nature of exchanges.
Events are paired so the inflows of every resource are coupled with outflows of another
resource. For example, the CASH RECEIPT event increments CASH, and it is related to

an event called SALE that decrements the resource INVENTORY.

10

REA systems can be documented with entity-relationship diagrams. Each
resource, event and agent can be modeled as an “entity” (shown as a rectangle). The
relationships between the entities are represented by diamonds. Figure 1 shows the
“generic” REA diagram for an economic exchange. The diamond below the event entity
models the duality relationship between increment and decrement events. By showing
these duality relationships, this heuristic can be used to model all economic exchanges.
For example, the REA diagram for the revenue cycle discussed above is shown in Figure

2.

 

 

 

 

 

 

 

 

 

Inside
Stock flow
Agent
Resource . Event
* Outside
Agent

 

 

O

Duality

Figure 1: Basic REA Representation

 

 

 

 

11

 

 

 

 

 

 

 

Employee J
Finished . Sale . (Shipping Clerk)
f Goods ,

Customer
I l—O-l 9
Cash .
, Receipt ,

Figure 2: Revenue Cycle Representation in an REA Design

 

 

 

 

 

Employee
(Cashier)

 

 

 

 

If a complete analysis of the organization was performed, additional exchanges
would be added to the diagram. For instance, if the company being modeled is a
distribution company, then the INVENTORY would be incremented through a
PURCHASE, which is related to a decrement, CASH DISBURSEMENT, Of CASH.

Figure 3 is a model Of a manufacturing firm. In this firm, raw materials are
PURCHASEd and then converted into FINISHED GOODS. Other exchanges modeled
are FIXED ASSET ACQUISITION, SERVICE ACQUISITION, and EMPLOYEE

SERVICE.

12

ELE wet—58:52 a 3.. .032 a "n 0.59..—

 

 

 

13

Once the resources, events, and agents have been modeled, all of the necessary
data about these should be identified. For example, all of the data about CUSTOMERS
that would be required by sales, marketing, accounting, and any other organizational
members Should be identified. CUSTOMER attributes could include name, address,
credit limit, contact name, and phone number. Similar analysis would be performed for
each entity and relationship in the firm’s entity-relationship diagram.

After the important data are identified, the organization’s REA accounting
database can be created by implementing each entity, relationship, and attribute in a
centralized data repository. With application programs to enter, store, manipulate, and
report this data, an REA accounting information system will include detailed information
about the specific resources and people involved in the event as Opposed to a traditional
accounting system that includes only the transaction date, the account number, and the
financial amount. In REA systems, both financial and non-financial information about
events are accessible to all managers.

Continuing the earlier sales example, Figure 4 illustrates how an REA accounting
system would record the same transactions that were summarized in the day’s joumai
entry on page 5. The table shown is created for the SALE event. Attributes identified for
this entity were invoice number, invoice date, ship date, division, warehouse, total
amount, and total cost. In addition, the key attributes of the CUSTOMER and

EMPLOYEE table have been included to implement the relationship between SALE and

these entities.

l4

 

 

 

 

 

 

 

 

 

 

Invoice Inv. Ship Div- Ware- Total Inv. Total Sales- Cust
# Date Date ision _ house Amount Cost _ person

1 9/22 9/20 15 01 $ 60,000 8 30,000 1 C-05
2 9/22 9/22 20 02 $ 15,000 $ 10,000 2 C-10
3 9/22 9/22 20 02 $ 25,000 8 10,000 1 C-lO

 

 

 

 

 

 

Figure 4: Implementation of Sales Transactions in an REA Accounting System

 

 

 

The records in this table can be sorted and summarized to generate the same
joumai entry as in a traditional accounting system. However, additional “views” Of the
data are available for other system users with difi‘erent information needs. This illustrates
how REA accounting systems store much more data than traditional accounting systems.
To a large extent, the main difi‘erence between an REA accounting system and a more
traditional one is when data are “filtered.” In an REA accounting system, much less
transaction data are filtered before they are stored. The majority of the filtering occurs
when reports are generated, when only the information needed for the report is used.
Continuing the SALE example, data can be grouped and summarized by sales person for
the sales manager, or by customer for the marketing manager. Ifthe relationship between
SALE and INVENTORY is also implemented, the data may also be sorted by product for
the production manager. In addition, the data can be summarized to prepare the financial
statements when needed by management or to meet demands by outside institutions (such
as the SEC, etc.). On the other hand, traditional accounting systems place a filter on the
data before they are stored in the system. Therefore, the information that can be reported

is limited to the account information that is stored in such a system.

 

(”“3”

15

By implementing an REA accounting system, an organization can overcome the
traditional accounting system weaknesses identified in Table 1. First, weaknesses one and
four are the result of traditional accounting systems focusing on financial measures. REA
systems Store both financial and non-financial information about not only events but also
the related resources and agents. By storing this information in one system, it is available
to all system users, and REA accounting systems are able to summarize and report both
types of data.

REA systems are able to minimize weaknesses two and three because they store
detailed data about events and are able to produce difl‘erent “Views” of the data to meet
various management needs. This feature of REA systems minimizes the problems with
both data classification schemes and aggregation levels. REA data can be classified by
attributes other than the general ledger account numbers. The data can either be
aggregated in totals or studied as detailed transactions.

By focusing on REA modeling when designing a new system, organizations are
forced to evaluate their business processes, thus overcoming the fifth weakness of
traditional systems. By modeling each exchange in the organization, management has to
identify all resources and processes used to produce value to the customer. Activities that
do not provide value can be eliminated from the organization’s procedures (Geerts and
McCarthy 1995).

Using REA methods to guide a systems analysis process results in a “directed”
REA accounting system that meets the organization’s specific goals and is implemented
without a chart of accounts. Rather, all of the event data are collected, and a computer

procedure generates the financial statements.

16

Multiple directed REA proof of concept implementations have been reported in the
accounting literature (Gal and McCarthy 1986; Denna and McCarthy 1987). These
systems have been implemented in different processing environments, but both are able to
generate financial statements without a chart of accounts. In addition, each has provided
examples of how REA systems are able to meet other management needs. REA analysis is
also being used by IBM. It is one of the design philosophies they are using to reengineer
their accounting system (Andros, Cherrington, and Denna 1992). By focusing on their
economic exchanges and attempting to simplify these processes and the systems that
support them, IBM has been able to reduce the number of redundant systems used
worldwide fi'om 315 to 36.

In summary, research in REA accounting systems has been largely theoretical in
nature. Starting with McCarthy (1979 and 1982), papers have described the REA
fi'amework and have shown how these systems overcome the weaknesses inherent in
traditional accounting information systems. More recently, theoretical papers have applied
REA to more challenging domains such as manufacturing, showing that this model is very
robust (Grabski and Marsh 1995, Denna, Jasperson, Fong and Middleman 1995). In
addition, several papers describe REA systems that serve as proof of the concept
implementations such as Gal and McCarthy (1983). These research systems are able to
store detailed transaction data and produce traditional accounting reports. Some papers
also describe how these systems are able to meet additional management needs such as
manufacturing decision support systems (Denna and McCarthy 1987) and strategy support

systems (Revaz 1993).

 

011

0i“
5

Per:

17

Weber (1986) reports the only empirical analysis of commercially available
software. He studied several order entry systems to determine if these systems were
Similar to the REA systems recommended by McCarthy (1982). His results were that the
systems did, in fact, keep detailed information that was consistent with REA at a high
semantic level. However, at a lower semantic level, there were additional files that were
necessary, such as customer orders and ship-to-addresses. For example, the systems he
examined included CUSTOMER ORDERS. Although they have become important
components of these systems, Weber noted that these are events that are not included in
the basic REA template because no resource has been incremented or decremented. In
addition, they are not they captured by traditional accounting systems. While Weber
implies this is a weakness of REA systems, McCarthy (1982) had identified them as
contracts and as a possible addition to the REA constellation. Weber also identified files
that provided additional information about the resources, events and agents already
identified by REA, such as ship-to addresses for CUSTOMERS and SALES. These are
merely multi-valued attributes of the main entity to which they relate.

This work extends Weber (1986) in several ways. First, it examines the complete
information system, rather than focusing on the data fi'om one cycle within the system.
Second, it studies the organization’s systems at a high semantic level, rather than focusing
on the data level. Third, it looks at how the systems are used in organizations, rather than
studying vendor documentation. By doing this, the researcher is able to identify
organizational characteristics that influence the way the system is used. Therefore, user’s

perceptions of the system and their self-reported uses of the system are also gathered.

18

In addition, this study categorizes systems along a continuum between traditional
and REA accounting systems. The most traditional system, at one extreme of the
continuum, would consist of only journals and ledgers and would use journal entries to
produce financial statements. Directed REA systems, at the other extreme, would use
REA to guide the system development, and the resulting system would not use joumai
entries to a general ledger to produce the financial statements. Rather, the system would
produce them as a view from the detailed transaction data. To date, very few directed
REA implementations have been described. The exceptions are implementations by the
Price Waterhouse Geneva practice unit (Cherrington, McCarthy, Andros, Roth and Denna
1993) and IBM’s implementation of the REA ideas (Andros et al. 1992).

Although none of the firms in this study operate accounting systems at either
extreme of this continuum, a major goal of this study is to identify differences resulting
from systems being more similar to either the REA extreme or the traditional extreme. As
such, one of the contributions of this work is a more detailed, operational definition of the
continuum between the extreme system types. This definition, described in the qualitative
results sections of this dissertation, provides a fi'amework for additional empirical research

in REA systems and the potential benefits that they may provide.

19

Chapter 3 - Construct and Hypothesis Development

AS discussed, accounting information systems have evolved over time. Hopwood
(1987) described three common phases of information systems development. First,
businesses were owned and operated by a single person who could easily measure the
business’ success since he or she was involved in all of the transactions. Once
organizations grew too large to be monitored and controlled by an individual owner,
however, management relied on traditional accounting systems. More recently, another
shift has occurred. Information systems are broadening in scope to shift the focus from
solely financial measures of the organization’s performance to other measures such as
quality and flexibility.

He did not know why these shifts occurred, nor did he make hypotheses regarding
the environmental characteristics that would trigger these shifts. However, he realized
that the changes occur because the firms enjoy benefits from the type of system they use.
Many researchers have attempted to identify and quantify potential systems benefits. Two
benefits that have been important to both researchers and practitioners are improved
productivity and competitive position resulting fiom information technology. The
following sections of this chapter provide a summary of these research areas and develop

hypotheses for the current study.

3.1 Competitive Advantage
Organizations implement their strategic plans to improve their relative industry
position (i.e., to gain a competitive advantage). By using information technology (IT) as a

tool to support and reinforce an organization's strategy, firms can develop competitive

20

information systems (1qu and Beattie 1985). For example, the American Airlines
reservation system (Sabre) was provided to travel agents so they could make reservations
on-line (Wiseman and Mach 1984; Northrup 1991). This system enabled ticket
agents to reserve tickets for any airline, but American flights were listed first. Their
market share grew, largely due to ticket agent sales.

Porter (1985) identifies several strategies that can be used to highlight
opportunities for competitive information systems. He states that IT can be used to raise
entry barriers thus providing an advantage to the "first-mover" that adopts new
technology. For example, the first firm that implements an REA system in an industry
segment could store detailed information about customer preferences to improve customer
service levels. Once customers expect this level of service, potential entrants into the
market may have to provide similar services before customers would switch suppliers.

Second, IT can be used to increase negotiating power with suppliers. For
example, if a firm’s computer system is capable of storing price quotes fi'om several
vendors and vendor performance statistics, the firm will be able to negotiate more
successfirlly with its suppliers (Porter and Millar 1985). If a firrn’s IT enables an
organization to provide better service levels or improved flexibility, then the firm will
enjoy increased bargaining power over its customers.

IT can also create new dependencies for customers by increasing the costs they
would incur to switch to another vendor (”switching costs”). For example, American
Hospital Supply (AHS) developed a system for their customers so they could log directly
into the AHS system to check prices and product availability. Once the customers learned

the system, they would incur costs to switch to another vendor who ofi‘ered a difi‘erent

 

it

EX

ad

be

org

Dior

mi

21

proprietary system. As a result, AHS enjoyed significant customer loyalty as long as their
system continued to provide services similar to their competition (Porter and Millar 1985,
p.156)

Porter (1985) also states that IT can be a tool to develop new products or
substitute products ("unique product features"). For example, Federal Express used
technology to provide much faster and more reliable deliveries than its competition (Porter
1985, p. 171). IT can also create new business opportunities in “interrelated industries.”
These opportunities were very visible in the telecommunications and cable television
industries. Dramatic changes are taking place in these industries as the telephone
companies are broadening the types of services they ofl‘er beyond local or long distance
telephone services. They are also attempting to influence federal regulations so they can
enter the entertainment industry to provide cable television to customers using their
existing infiastructure. Entertainment products and services are being modified with
advanced computer technologies, and the distinction between these two industries is
becoming less clear. It is also possible for firms to expand into industries in which their
information system’s data become a product that is sold to information brokers. For
example, many information intensive companies such as grocery and retail chains have
sold their sales databases for large, undisclosed, amounts of money to (Lalete 1993).

In addition, by focusing on the production process during strategic planning, an
organization may be able to identify ways to use IT to minimize the time needed to
produce goods or to respond to changes in customer demands (”time eficiencies"). For
example, a Computer-Aided Design/Computer-Aided Manufacturing (CAD-CAM)

environment enables engineers to make modifications in product designs more easily.

22

Also, the system can automatically modify the production line to reflect these changes.
This results in reduced costs for product changes and improved response time to customer
requests.

While Porter’s recommendations appealed to organizations and were used by
researchers, they were not theoretically based. To fill this void, Bakos and Treacy (1986)
used well-established theories to build on these recommendations and develop a model of
how IT can lead-to competitive advantage (Figure 5). They categorize the factors fiom
Porter's (1985) fi’amework and hypothesize that firms gain competitive advantage either
by improving their bargaining power with external organizations or by improving their
efiiciency. The bargaining power hypotheses are developed using game theory in which
information can enable one player in a zero-sum game to earn increased profits at the
expense of the other player. They use transaction cost theory and bounded rationality to
support comparative efliciency as a method to gain competitive advantage. Because
managers have bounded rationality, an information system that decreases monitoring costs
or enables evaluation of more alternative situations will help managers overcome their

weaknesses and improve their decision-making process.

23

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Unique Product
Features Bargaining
Power
Switching >
Costs
Competitive
Advantage
Internal
Emcrency Comparative
Efficiency
Interorganizational
Efficiency

 

 

 

Source: Bakos and Treacy, 1986.

Figure 5: Causal Model of Competitive Advantage

 

 

To increase bargaining power, firms must either reduce the power their customers
enjoy or improve their power over their suppliers. Firms can increase their bargaining
power if they develop a product with unique features; then customers are not able to
substitute competitor’s goods. Similarly, if the firm can reduce its need for unique
products, it will have improved bargaining power over its suppliers. Bargaining power is
also improved by increasing customer switching costs or by reducing the costs of
switching suppliers. Bakos and Treacy (1986, p. 113) describe strategies to increase
customer switching costs as providing “unique and valuable information and services that
require idiosyncratic changes to the customer's organization.” As discussed earlier, the

American Hospital Supply system provided this type of benefit.

24

Firms can improve their comparative efficiency by improving either internal
emciency or interorganizational emciency. Internal eficiencies may be improved if a
process is redesigned to eliminate wastefirl steps and to automate the purely clerical ones.
An example of an interorganizational efficiency is Electronic Data Interchange (EDI)
through which electronic purchase orders are sent to vendors. This should reduce the
total paperwork and clerical time required to place orders, resulting in direct and indirect
benefits to both firms (Dearing 1990).

Bakos and Treacy (1986) also note that firms do not automatically enjoy
competitive advantage from information technology. Specifically, they state that the link
between technology must be strengthened with “strategy-literate information systems
planners and technology-literate strategic planners” (p. 116). This was operationalized by
David (1994) who defined a finn’s development characteristics as management’s
philosophy about using information systems for competitive advantage and the approach
they took when identifying systems enhancements to implement. This construct was
significantly related to management’s perceptions of the system’s ability to provide its firm
with competitive advantage.

Although much of the literature has focused on the positive factors that may lead
to competitive advantage, other articles have identified possible reasons why competitive
advantage fi'om IT may not be enjoyed. Several researchers (Porter 1985; Porter and
Millar 1985; and Senn 1992) have been concerned that any system developed may be
copied, so any competitive advantage created may not be sustainable. As a result,
organizations implementing new technology may only be raising the costs of doing

business for everyone in the industry. Once one firm implements a new system that

25

provides a temporary advantage, all other firms are forced to make similar investments. In
the end, no firm enjoys an advantage, and all incur increased costs. This has been shown
empirically in the hotel industry. Reid and Sandler (1992) conclude that hotel customers
quickly treat technological innovations as required amenities during their hotel stays.

Thus, total costs of operating hotels are increased, and none are provided with a
competitive advantage. Because of these concerns, the belief that competitive advantage
is not sustainable is hypothesized to be negatively related to managers' perceptions of
competitive advantage.

The Bakos and Treacy model has been extended here to include Development
Characteristics and Not-Sustainable as two additional factors that influence a firm’s
ability to enjoy a competitive advantage. In addition, firms may interact difl‘erently with
suppliers than with customers, so the components of Bargaining Power and
Interorganizational Efiiciency have been divided. Finally, production or administrative
stafi‘s could become more eficient, so Internal Eficiencies has also been expanded.

These changes result in the following hypothesis that is illustrated in Figure 6:

H1:

26

Manager's perceptions of competitive advantage will be positively related
to the ability of their systems to improve bargaining power, efficiency, and
their organization’s development characteristics. Manager perceptions will
be negatively afl‘ected by the fact that advantages are not sustainable.
(Competitive Advantage will be positively correlated with Development
Characteristics, Unique Product Features, Switching Costs, Bargaining
Power, Internal Efliciency, Interorganizational Efliciency, and
Comparative Efiiciency. Competitive Advantage will be negatively

correlated with Not-Sustainable.)

27

 

 

 

Unique Product
Features-Customer
Unique Product . ~
Features-Supplier Bgiagmg + Relationship
Switching

Costs-Customer - Relationship
Switching
Costs-Supplier

 

 

 

Competitive Not
Advantage Sustainable

 

Internal
Efliciency-Adminisu'ation

 

Internal + Relations '
Efficiency-Production

 

 

Comparative
Interorganizational Efficiency
Efficiency-Customer
Interorganizational
Efficiency-Supplier

 

Development
Characteristics

Source: Adapted fi'om Bakos and Treacy, 1986.

Figure 6: Hypothesized Model of Competitive Advantage

 

 

REA accounting information systems can assist managers in implementing each of
the Bakos and Treacy (1986) strategies for competitive advantage. Ifan information
system stores detailed event information, an organization using such a system could
provide that detail to their customers who do not have sophisticated information systems.
Ifcustomers values this information and can use it to improve their Operation, they will see
this service as enhancing the product they are purchasing. For example, if a customer’s
system does not maintain a detailed purchase history, the organization could produce
monthly inventory purchase reports for the customer. This service would difi‘erentiate the

organization’s product from those of competitors. Ifcustomers rely on the information

28

available from the organization’s proprietary information system, changing to another
vendor will be more costly for the customer.

An REA accounting information system can have positive efl‘ects on an
organization’s internal efficiency. First, storing detailed event data centrally should result
in productivity improvements by eliminating data redundancy. As a result, no effort is
required to insure that separate systems are updated or remain consistent. This can reduce
the amount of systems resources expended, and it may also reduce the control efforts
needed to maintain the systems. For example, by eliminating redundancies, Du Pont was
able to reduce the number of salary runs per year from 3,300 to 36 (Vincent 1993).

Second, new reports to meet specific information needs can be created more
quickly than they could be with a system that stored only sunrrnarized data or data in
different systems. For example, a sales manager may require a report that combines sales,
inventory, and financial information. If all data are stored locally, it is much easier to
produce this report than if the data are located on separate, independent systems.

Because management’s ability is linrited by bounded rationality, a system that
provides information in a format that enables more efficient data retrieval should improve
an organization’s internal strategy (Bakos and Treacy 1986). By design, REA accounting
systems should help management “plan, monitor and control” the key events of the
enterprise (Denna, Cherrington, Andros and Hollander 1993). Therefore, management
should perform their tasks more efi‘ectively, and outputs generated should increase or
required inputs should decrease. For example, if provided with better inventory and
production information, management should be able to identify problems in the production

process more quickly. Ifthey have the information necessary to correct the situation, they

29

may be able to significantly reduce the amount of scrap material used to produce the
finished goods. Overall, the production department would be more efiicient.

Denna et al. (1993) also recommend performing “reengineering” as a critical step
in the REA accounting system design process. Reengineering involves analyzing the
organization's processes to eliminate nonessential processes or events, to improve essential
process efficiency, and to enable new valuable processes. Throughout the reengineering
process, design participants are encouraged to think creatively about their business events
and attempt to make "radical" changes to the processes. For example, Davenport (1993,
p. 1) recommends organizational improvements through “efl‘orts to achieve 50%, 100% or
even higher improvement levels in a few key processes.” If process reengineering is
successful, significant eficiency improvements should occur because the overall result is
more streamlined processes that are only performed if they provide value to the customer.

Finally, an REA accounting system may improve interorganizational eficiencies
because the detailed transaction data are available for analysis. For example, Motorola
transformed its business to improve customer service levels, thus improving the
efliciencies between itself and its customers. Their system is able to process
“nontraditional information” including product quality data. When the organization was
implementing its quality programs, the former chairman asked for the quality reports to be
presented first in the monthly management meetings, and he left before the financial
information was presented. Because of this shift in focus toward quality, the number of

defective products has been reduced by a factor of 100 (Vincent 1993).

30

Because of the varied effects that REA systems should have on the organization,
and the weaknesses inherent in more traditional accounting systems, the following

hypothesis is tested:

H2: Executives of firms that have implemented REA systems will perceive that
their systems provide more competitive advantage than executives using
more traditional systems. (Accounting System Characteristics will be

positively correlated with Competitive Advantage.)

3.2 Productivity

Productivity is defined in the labor econonrics literature as “how much output can
be produced with as few inputs as possible” (Pka 1991, p. 192). Normally measured in
units, productivity measures focus on the quantity of inputs and outputs, rather than their
dollar value. Therefore, they can be used to monitor an organization’s success in
producing goods. While there have been different levels of productivity studied, the
definition of productivity used in this study is taken fi'om Davidson (1993). He defines
productivity measures at the company level, while eficiency is measured at a functional
level. For example, if fi'ont desk clerks are able to check in more guests after an
information system is implemented, then they are more emcient. On the other hand, if the
total number of labor hours per guest night is reduced, then the hotel firm’s productivity
has improved.

In the aggregate, productivity measures have been used to monitor and evaluate

the economy's ability to produce goods. For example, the US. Department of Labor’s

31

Bureau of Labor Statistics (BLS) measures productivity in the US. economy and in
segments of the economy. Researchers have used these measures to identify the
"productivity paradox” - white collar employee productivity seems to be decreasing
while, simultaneously, large computer expenditures have been made to increase their
productivity. Roach (1991), for example, found that although service industry firms have
purchased relatively more technology than manufacturing firms, their productivity has
been poorer.

This result has puzzled researchers. However, Panko ( 1991) identified several
problems with the BLS statistics that were used in the early productivity studies. First, the
BLS has difiiculties measuring units of outputs. Often constant-dollar sales are used as an
estimate for units produced, but both the sales figures and the deflation indexes may
introduce measurement errors to the productivity figures reported. For industries such as
not-for-profit and government, sales figures are not available, so the BLS estimates
outputs as equal to inputs. As a result, productivity for these segments never changes
from unity. Many segments of the service industry produce outputs that are dificult to
identify at all. For example, how would an accounting firm’s output be measured?

Should the number of audits be used? A weighted measure that considers the size of the
audit? Audit revenues? Industries that present significant measurement problems
(including finance, insurance, real estate) are omitted fi'om the BLS figures.

Another problem with aggregate productivity figures is that organizations may
substitute one input factor for another. For example, if a computer system is implemented
to reduce the number of people needed to package and ship merchandise, additional

personnel may be needed at the main ofiice to maintain the system. As a result, if the

32

number of orders shipped remains constant, the productivity for the ofiice workers will
appear to decrease. However, the productivity of the factory workers would
simultaneously increase.

In addition, Brynjolfsson (1993) identified four potential explanations for prior
productivity research results: (1) mismanagement of firm resources, (2) lags between
computer-related expenditures and productivity improvements, (3) measurement errors,
and (4) redistribution of resources within industries. He provided detailed examples of
each potential explanation and discussed the research approaches that could be used to
evaluate them.

Mismanagement of assets can occur if the managers approve computer
expenditures that improve their situation (status, slack time, etc.) while hurting the
company's financial position. If this is the cause of the productivity paradox, then there
are no benefits in productivity fi'om computers for firm owners.

Lags may occur because unproductive activities occur at the time of new computer
implementations. For example, training time may be necessary to prepare users for system
changes. When exanrining productivity levels during training, the number of labor hours
(including training) will increase without an increase in output. Also, the new users may
not be as productive at the time of implementation because they have not yet learned the
intricacies of the new system. Another cause in productivity lags is that the system
implementation may be phased in over time. Ifany of these lags occurs, one would not be
able to measure the productivity improvements until some time in the future. Kelley
(1994, p. 1420) provides evidence of lags. She compared the productivity levels for

several firms that had irnplrnented prograrnnrable control systems in manufacturing

33

processes and found that companies that had recently implemented such systems had not
experienced as much productivity improvement as earlier implementers. She asserts that
there is a lag between the time of implementation and productivity improvement; the
recent irnpiementers had yet to experience the improvement since they were still in the lag
phase.

Measurement errors may occur in two ways. First, the researcher may be
examining data that are too aggregated (such as BLS statistics) so the effects of the
computer are not measurable. Brynjolfsson and Hitt (1993) use firm-level computer
expenditures and sales figures to estimate the productivity provided fiom the computer
investments; they find a significant, positive relationship between the two. Second, the
researcher may be using inappropriate productivity measures. For example, some
researchers have used functional-level measures in their productivity studies. If researchers
used “checks cashed per hour” as a measure of bank teller productivity,‘ they would not
be able to identify productivity improvements arising fiom an ATM implementation. In
fact, they may detemrine that the tellers are less productive after the ATM is implemented
because many customers will use it for simple transactions, and the tellers will handle the
more dimcult ones. Therefore, the teller may process fewer checks after the
implementation than before.

It is possible that IT does not increase the overall size of an industry, but rather
shifts the market share among the firms in the industry. Brynjolfsson identifies this as the

redistribution efi‘ect of IT since fimrs that successfirlly implement technology may be able

 

‘Noficematmisisacnmllyanefliciencymeanueusingmedefimmnsmthissmdy.

34

to increase their sales by taking sales from other firms in their industry. There are many
researchers who have argued that information technology (IT) can provide competitive
advantage for individual firms (Porter 1985; Johnston and Carrico 1988; Johnston and
Vitale 1988). If this is so, firms with a competitive advantage may be the ones using IT
to improve their productivity. Firm-level measures would identify this type of
productivity improvements, whereas summary measures would show no productivity
improvements.

Different methodologies may be able to reduce the nrismeasurement and
redistribution difiiculties encountered in prior research. If only computer systems with
certain characteristics provide improvements, studies looking at more aggregated data
may not be able to identify any consistent impact of computers on productivity. By
focusing on the systems themselves, researchers may be able to identify the significant
characteristics that result in improved productivity.

As already discussed, having all data centrally located and available to all managers
can reduce both system processing costs and management efi‘ort. Therefore, both
managers and the information systems department can be more efficient. Ifthe managers
are more eficient, and none of the other workers become less efficient, then the
company’s productivity level will improve.

In addition, if a directed REA design is performed to develop the system, each
business process will be analyzed, unnecessary steps will be eliminated, and opportunities
for information technology to streamline the processes will be implemented. This will
result in internal eficiency improvements throughout the organization, and improvements

in the firm’s productivity.

35

The previous discussion leads to the third hypothesis:

H3: F inns with REA systems will be more productive than firms that use more

traditional accounting systems. (Accounting System Characteristics will be

positively correlated with Productivity.)

3.3 Summary

The three hypotheses are summarized in Figure 7. As shown, this model identifies
an additional research question: Do REA systems have a direct impact on competitive
advantage and productivity, or are they affecting bargaining power and comparative
eficiency directly and competitive advantage and productivity indirectly? Ifthere is a
direct connection between REA systems and competitive advantage, such a link would
provide evidence that the model proposed by Bakos and Treacy is missing critical
constructs that lead to competitive advantage. Therefore, in addition to testing the

individual hypotheses, this exploratory path model will be tested as Research Question 1

(RQl)-

36

 

 

 

 

 

 

  

 

 

 

 

 

 

 

 

Accounting
System \ 1
Classification Compgfive
\ Advantage
Productivity

 

 

 

Figure 7: Hypothesized Model of the Impact of REA Systems and RQl

 

 

37

Chapter 4 - Methodology and Construct Development
4.1 Introduction

While three hypotheses and a research question have been developed and are the
focus of the quantitative results reported, there are additional, qualitative goals of this
study. Most importantly, this study operationalizes the classification of current systems
along a continuum between traditional and REA accounting information systems. While
the extreme ends Of the continuum have been defined in previous literature (McCarthy
1982), the systems that fall in the middle have not been categorized. Therefore, a metric
was developed to capture the important information systems characteristics and to enable
the researcher to draw conclusions about those that provide organizations with benefits.
In addition, the results of the study are used to propose a more precise definition of the
characteristics that difl‘erentiate systems along the continuum.

These goals describe the cyclical nature of research. First, hypotheses have been
developed using existing theories and these are tested quantitatively. Second, the results
of the study are used to provide a more thorough model of information technology and to
identify the characteristics of REA accounting information systems. These characteristics
are used to develop a model that hypothesizes relationships between each of the
characteristics and potential benefits to the firm. This model can be used as the foundation
for firture research continuing this research stream.

These goals are best met using a field study methodology. As discussed in Trewin
(1988), field-based research allows the researcher to examine organizations and how
organizational characteristics afi’ect the success of their information system. This is

important when examining benefits derived from systems with specific characteristics

38

because it is likely that these systems only provide benefits to organizations with certain
characteristics. In addition, she cites Yin (1984) who says that field-based research can be
used not only for hypothesis testing, but also for categorizing phenomena and building
new hypotheses. In summary, field studies are the appropriate research methodology to
study “a research issue in transition, for which grounded theory or experience with
relationships among well-defined variables have not yet emerged” (Gosse 1993, p. 166).
This is certainly the case with REA accounting systems.

To meet the research goals, field research was performed to gather detailed
information from eight firms regarding their information systems characteristics and how
their systems afl‘ect both their productivity and their perceptions of competitive advantage.
To focus the field inquiry, four guidelines were followed (Gosse 1993). First, all of the
sites were selected from the pulp and paper industry, and all have approximately the same
level of sales. This minimizes confounding difi‘erences between the sample firms. Second,
to control the interview environment, measures of Competitive Advantage and Accounting
System Characteristics were developed and used to guide the discussions with firm
personnel. Third, to increase the “richness of the data” open ended questions were
included in the questionnaire. Subjects were encouraged to speak fieely and add insights
to the discussion. Finally, the data gathering procedures included interviews with people
across both firnctions and management levels within the organization. Individual
interviews were held with the goal of understanding system impacts on each person. The
group discussion at the end of the visit was used to confirm patterns that had emerged in

the data.

39

4.2 Construct Development
Two questionnaires were developed: one categorizes the firm’s information
systems and the other measures executives’ perceptions of competitive advantage. The

methods used to produce each will be discussed in the following paragraphs.

4.2.1 Accounting System Characteristics Questionnaire

As discussed in Chapter 2, most of the work in REA systems has been theoretical
in nature. Therefore, it was necessary to develop an instrument that would categorize ‘
information systems by placing them along the continuum between traditional accounting
information systems and REA systems. The survey was developed to incorporate the
following characteristics that were identified in the theoretical works as important in REA
systems:

0 No chart of accounts. Produce financial statements from detailed event and

resource data.

0 Support the organization’s critical events.

0 Store detailed data about the resources, events and agents.

0 Store non-financial information about resources, events and agents.

The survey uses a generic REA diagram for a manufacturing firm as its foundation.
It was developed in three phases. First, an REA model for a manufacturing firm was
developed. Next, this diagram was modified to incorporate contract events identified by
Weber (1986). The conversion cycle was then modified, by collapsing several events. For

example, the full REA diagram included two events for a raw material issue: an issue of

40

raw material fi'om the raw material storage area and an issue to work in process. These
two events were replaced with a single raw material issue event since they normally occur
simultaneously. Finally, full tracability was not enforced and resources such as
“advertising benefits” were eliminated since they are difficult to identify in practice.

The resulting diagram is used to identify business-critical events and it is modified
if the organization operated differently than the generic one. Once the organization’s
REA diagram is established, the information system is analyzed to determine if it supports
each entity on the diagram. A system would move toward the REA end of the continuum
with each event that it supports.

Questions were written to measure key characteristics of the accounting
department. For example, it is important to determine how the financial statements are
generated. Firms that do not have a general ledger would automatically be placed at the
REA end of the continuum. In addition, however, several accounting department
characteristics were identified that would provide evidence that the system was more
traditional or more REA. Ifthe company uses a very detailed chart of account structure
that incorporates many codes in each account, it is more likely relying on the traditional
accounting methods. Therefore, it would be classified as closer to the traditional end of
the continuum.

Finally, it was recognized that a system could be purchased from a vendor and that
it would have inherent characteristics that placed the software itself at a point along the
continuum. However, how the system is used depends on organizational characteristics
such as complexity, computer availability, speed of processing, flexibility of data retrieval,

etc. Therefore, the questionnaire includes an additional section to identify user’s

41

information needs and how the system is used. This section includes questions about the
processes the respondent performs. The next questions identify what information the
respondent needs to perform their daily functions and where it is available. In addition,
there are questions about whether the information system improves their productivity and
asks them to identify enhancements they would like to see in the system.

People fi'om several firnctional areas were asked these questions. Their
information system’s score was increased if it was able to provide the information needed
to perform daily firnctions. However, if the system did not provide information that would
have been available in an REA systenr, the information system’s score was reduced.

Questionnaire evaluation was performed in two phases. During the first phase, the
researcher completed the questionnaire for several firms that would not be included in the
main study. With each completion, the questionnaire was scored, and the results were
discussed with several researchers familiar with REA systems. The questionnaire was then
modified to gather additional information about the organization and their system, or the
questions and scoring were modified to improve the metric’s ability to rank systems along
the continuum.

The second phase of testing involved systems consultants from one of the Big 6
accounting firms. The consultants listened to a two hour seminar about REA accounting
systems. After the discussion, they were provided with a randomly-numbered, blank
questionnaire. They were asked to complete the functional questions fi'om an order entry
clerk’s point of view. One halfof the consultants were asked to complete the
questionnaire for the information system that they were currently replacing at one of their

clients. The other half completed it for the new system that was being recommended for

42

their clients. Additionally, they were asked to estimate how they would score the system
using a scale of one to seven where a one signified a traditional accounting information
system and a seven signified an REA accounting information system. They wrote these
scores on a separate sheet that also included their random number.

These questionnaires were scored, and the systems were ranked based on their
scores. Of the fourteen consultants that participated, the questionnaire scores ranked the
systems in the same order as the consultant’s estimates with only one exception. Upon
analyzing that consultant’s score and questionnaire, it was determined that the consultant
did not understand REA accounting systerrrs.

The final survey (Appendix A) includes ( 1) the generic diagram, (2) additional
questions about the system, accounting and MIS, and (3) the form that was used to
interview the functional employees (both stafi‘ and management level) within the
organization. The number of points added or subtracted to the organization’s total score
may seem arbitrary. However, based on the pilot test results, the scoring system

adequately difi‘erentiated between the information systems.

4.2.2 Competitive Advantage Survey

A survey was developed to measure executive’s perceptions of competitive
advantage (Appendix B). The Development Characteristics and eficiency questions are
those used by David (1994) to analyze how the executives of small, fast-growing firms
perceived comparative efliciencies and competitive advantage from information
technology. These questions were developed using the methods described by Moore and

Benbasat (1991) to improve internal reliability of the survey. First, five questions were

wrii
wer
first
que
imp
que

con

mi:

Uni
Cor
“iii
TeSU
Con

inter

43

written for each construct and were placed on individual note cards. These note cards
were distributed to students who placed them into categories for each related topic. The
first students to sort the cards read them and created category names in which to place the
questions. After their categorization was analyzed, the questions were re-written to
improve their understandability. Additional groups of students were provided with the
questions on note cards plus a list of the constructs and their definitions. This process was
continued until three questions consistently were placed into the correct category.

To complete the Competitive Advantage portion survey, additional questions were
written for Unique Product Features and Switching Costs. Since each firm may have
difl‘erent relationships with their customers and suppliers, five questions were written for
Unique Product Features-Customer, Unique Product F eatures-Supplier, Switching
Costs—Customer, and Switching Costs-Supplier. Thus, twenty potential questions were
written, and the same procedures were performed to verify the reliability of them. As a
result, twelve more questions were randomly added to the first portion of the survey.
Confirmatory factor analysis was performed on the survey responses to verify the metrics
internal reliability and validity. This procedure and its results are discussed in Chapter 5,
section 5.2.1.1, Factor Analysis of Competitive Advantage Metric.

Also included in this survey are questions that measure the User ’s Satisfaction
with the system. Seddon and Yip (1992) developed these questions to directly measure
User 's Satisfaction with general ledger accounting systems. Seddon and Kiew (1994, p.
10) recommend them as a “short, simple measure of IS success.” In this study, they are
used to confirm any positive perceptions of Competitive Advantage that are identified

during the interview process.

44

These four questions were also presented to all staff level employees who
participated in the study. To collect their perceptions, a one page questionnaire with a

brief instruction section was prepared. See Appendix C for a copy of this metric.

4.3 Efficiency and Productivity Measures

The demographic information provided by the organizations is used to measure
their eficiency and productivity. Emciency is defined as the number of output units fiom
a department or function divided by the number of units of input from the same function
or department. In this study, the efficiency of the order processing department is
measured as the number of orders divided by the number of order entry personnel. The
accounts payable department can be evaluated with the number of checks written divided
by the number of employees. Similar measures are available for accounts receivable and
production.

Productivity is a firm level concept, so overall input and output quantities are
needed to measure it. Firnrs participating in the study were asked to identify the unit of
measure they used to track output and the number of units produced over the last five
years. In addition, they provided sales figures which are often used as a proxy for units of
output. Input would be measured as total number of employees and the hours they
worked annually, the total number of administrative hours annually, and the total number
of factory employee hours. All of this information was collected with an introductory
questionnaire that was sent to the company before each visit began (Appendix D). The

goal was to have them complete this metric before the visit began.

45

4.3 Methodology
4.3.1 Sample

Potential firms in the paper industry were identified through several sources. First,
the SWd and Poor ’s Register of Corporations, Directors, and Executives (1993) was
used to identify firms in the Pulp and Paper SIC codes, located in the Midwest and
Northwest, with sales of $10 to $500million. Executives of these firms were sent letters
of introduction.2 After two weeks, the executives were called to schedule a site visit and
to gather preliminary data about the organization’s information systems. This
methodology generated three firms in the study. In addition, firms that had executives
who were board members of the Retail Packaging Manufacturer’s Association (RPMA)
were contacted. The RPMA is a trade association that provides resources to its members.
For example, it distributes a newsletter that discusses operations, management and
information technology, and trends in the industry. One RPMA firm is included in the
sample. Finally, firms were included through word of mouth. After the first few firms had
participated, they recommended additional fimrs and made calls of introduction.

This method of sample selection may bias the results. Each of the firms that
participated had a management team that was willing to commit considerable amount of
firm time to the project. They were all interested in receiving an evaluation of their system
and learning about other systems in the pulp and paper industry. Therefore, the

perceptions of these executives may be difi‘erent fi'om those of managers in the total

 

2Thisindustrywasselectedbecrrusetheresearcherhasexperiencehere.

46

population. However, the state of REA research requires detailed information about
operational systems and their uses. Therefore, this methodology and its inherent problems
are deemed appropriate. Future research promises to extend this work to confirm it in

larger samples or to identify characteristics that difi‘erentiate these sample firms from the

general population.

4.3.2 Site Visit Procedures

Each firm was scheduled for a two-and-a-half day site visit. Before the visits
began, the contact person fi'om the company was asked to complete both a tentative
schedule (Appendix E) and an initial questionnaire (Appendix D) that was used to collect
the detailed demographic information about the firm. These forms were to be returned to
the researcher before the visits. The schedule was required so the organization’s
personnel were aware of the time required for the analysis.

Each visit included a tour of the manufacturing facilities and an initial meeting to
introduce the researcher to the management-level participants in the evaluation. During
this meeting, the competitive advantage surveys were distributed. Depending on the time
available for the meeting, the executives either completed it during the meeting or before
they met individually with the researcher.

During the remainder of the first two days, the researcher met individually with
both managers and stafl‘ employees of the organizations. She carried a laptop computer to
each of these meetings and typed the interviewee’s responses as the system and
organization was discussed. It was determined that this typing was less obtrusive than

breaking periodically to enter or write the responses. In addition, this enabled the

47

researcher to keep very detailed notes of conversations, attempting to transcribe them
verbatim. These discussions were used to evaluate the organization’s system and classify
it along a continuum as an REA system or traditional system.

To'measure the system characteristics, she interviewed the executive responsible
for the information system and studied the system documentation.

To evaluate the organizational characteristics, she also interviewed managers and
stafl‘ in the key fimctional areas: sales, purchasing, production, and distribution. Since the
organizational structure differed between firms, the exact titles and levels of employees
varied. For example, one organization did not have a purchasing department. Instead,
purchasing was performed by several people. The goal of these meetings, therefore, was
to meet with as many people with the widest range of responsibilities as possible.

By interviewing several users of the system and by having executives fiom
different functional areas complete the competitive advantage survey, individual biases will
be reduced (Sethi and King 1994). For example, the executive responsible for information
systems may believe that the system is providing more competitive advantage than the
other executives who do not have any direct ownership in the system. The methodology
used in this study will reduce the impact of such biases.

The last halfday of the visit was used to hold a seminar for the firm’s management
team. This seminar accomplished four objectives. First, it provided an opportunity to
gather any missing data before leaving the firm. Second, it enabled ”dc-briefing” of the
executives, and the researcher was able learn more about their perceptions of their
information systems. Third, it allowed managers to correct any misconceptions the

researcher had about the firm. Finally, it gave the researcher an opportunity to provide

48

some service to the firms. This session was a selling tool that persuaded several of the
executives to participate in the study.

Because these sessions needed to meet each firm’s goals, the sessions were
tailored according to management’s instructions. As a result, difi‘erent firm members
attended the meetings. In some organizations only top management was involved; in
others the MIS staff was present; and in others all of the study participants were included.
During most of these sessions, the researcher provided a summary of the competitive
advantage and REA literature as it relates to information technology. In all cases, she
presented an initial analysis of the firm and its information system and highlighted potential
opportunities for improvement. She also was able to print a document of the Accounting
System Characteristics questionnaire that had been completed during the visit and to
provide it to management. A final version was sent at a later date after it had been edited
more thoroughly by the researcher. In addition, several firms were interested in selecting
new software, so discussions about beneficial selection procedures were often included.
Throughout, the researcher focused on the advantages of an REA-like system and on the
comments that had (been presented by the study participants.

In virtually all cases, the management was enthusiastic about this closing session.
They were intrigued by the concepts of REA, and they could easily understand their
systems representation in the generic REA diagram. In addition, they were impressed by
the documentation of their system, and the researcher’s ability to consolidate so much
organizational information after only two days at the firm. Therefore, the survey metric

that was developed provided benefits to the organizations participating in the study, and at

49

least five firms are interested in having the researcher perform ongoing, periodic analyses

over the next two years.

Chapter 5 - Data Analysis and Results

During the eight company visits, detailed information was gathered about each
company and its system. The company information included operational and financial
statistics, plus organizational structure, management philosophies, user perceptions, and
process flow analysis. Because both quantitative and qualitative data was collected, several
analysis techniques are used to evaluate the efl‘ects that the accounting information systems
had on the organizations. These results are reported in the three main sections of this chapter.
In the first section, the qualitative results are discussed and used to develop an operational
model of the characteristics that differentiate traditional and REA accounting information
systems. The second section discusses the statistical tests performed on the data to evaluate
the hypotheses developed in Chapter 3. The final section discusses the results, focusing on
how the quantitative results may be explained or reinforced using the qualitative insights
gained during the site visits.

Because each company visit was performed using the same questionnaire, the
similarities and difi‘erences between the firms were identifiable. In addition, each organization
had developed unique computer solutions to best meet the needs of their employees and
management. Section 5.1 presents the qualitative results of these company visits. The first
part of this section describes the systems characteristics that afl‘ected user satisfaction. The
important characteristics that differentiate systems are defined. Potential benefits from each
characteristic are discussed and current literature is examined for studies to evaluate these
characteristics and provide empirical evidence about their efi‘ect on organizations. This

information is used to develop a detailed model of the characteristics that differentiate systerrrs

50

51

along the traditional and REA accounting information systems continuum. Next,
organizational characteristics that affected the choice of accounting information system are
presented. These organizational characteristics provide logical reasons for the adoption of
systems that would otherwise be viewed as sub-optimal.

Section 5.2 of this chapter presents the quantitative results to determine if the systems
that are more similar to REA systems are providing firms with more benefits than the more
traditional systems. It describes the confirmatory factor analysis process used to evaluate
reliability of the Competitive Advantage and User Satijaction metrics. Regression analysis is
used to test hypothesis one: improved bargaining position and comparative emciencies lead
to competitive advantage from information systems. Path analysis is used to identify the efi‘ect
of REA versus traditional accounting information systems on competitive advantage. In
addition, correlations are presented to show the efi‘ect of accounting systems characteristics
on different efficiency and productivity measures.

The final section of this chapter discusses the implications of the qualitative and
quantitative results. While the data provides evidence that more advanced systems provide
improved administrative emciencies, they do not appear to aid in implementing
interorganizational strategies for competitive advantage. Organizational and systems
characteristics may affect both manager’s perceptions of their system and the benefits that

current systems actually are providing.

52

5.1 Qualitative Results

The Accounting Systems Characteristic questionnaire, developed prior to the site
visits, provided the structure for each of the visits. The researcher used it to first identify the
key business events, and then to gather information about how the firrn’s information system
processed these events, what data were stored, and how this information was translated into
financial statements. In addition, the questionnaire was used to gather user satisfaction
information from several systenrs users at different levels within the organization.

By using a consistent approach for each of the site visits, patterns about the
organizations and their systems emerged. This was especially true when the generic REA
diagram was modified to reflect the individual processing environment for each firm. These
diagrams enabled communication of detailed information about the organization and the
processing environment. They were fi'equently used as a focal point of the de-briefing
seminars at the end of each visit, and were used as the foundation for the qualitative analysis

that follows. (See Appendix F for each of the firm’s diagrams and summary scoring sheets.)

5.1.1 Key Systems Characteristics

When the site visits were completed, the researcher compiled and compared the
Accounting System Characteristics (ASC) questionnaires from all of the firms. Each firm’s
employees had discussed what information they used and where it was available. They also
identified weaknesses of their current systems, and systems enhancements they would
recommend. The researcher focused on the difi‘erences between the systems and the user’s

perceptions of them. She could then identify patterns of characteristics that difi‘erentiated the

53

firms with the more sophisticated systems from those with the more traditional systems. As a
result, seven key systems characteristics were identified (see Table 2). The following sub-
sections define each characteristic in terms of traditional versus REA accounting information
system, using examples fi'om the companies visited and evidence from other empirical studies

that have evaluated similar characteristics.

 

Table 2:
Key Characteristics to Differentiate
Traditional and REA Accounting Information Systems

1. Support all critical events.
2. Store a detailed history of events.

3. Store the data in an integrated data repository.

4. Have the ability to retrieve and manipulate the data to meet users needs.
5. Process the events as they occur.

6. Directed REA design and implementation.

7. Prepare the financial statements without joumai entries and a general ledger.

 

 

 

5.1.1.1 Support Critical Events

The theoretical definition of REA includes definitions of events and the duality
relationship between events which increment and decrement resources; a system’s ability to
proceSs information about these events is thus an important feature of REA accounting

systems. Using this as the foundation of REA accounting information systems, an important

54

component of the Accounting System Characteristics questionnaire was identifying events
that were and were not supported by the firm’s information system. Each event that was not
supported was crossed ofi‘ the generic E-R diagram for the firm. During the site visits, it
became apparent that this process identified significant system weaknesses that were the focus
of user dissatisfaction. For example, purchasing was performed manually by several
organizations. At Ann Arbor the purchasing director performed all inventory firnctions by
keeping a manual inventory book with a page for each raw material and finished good. This
process had a large, negative efl‘ect on several other departments in the organization. When a
customer placed an order, the customer service representative wrote it on an order form
which was passed to the purchasing director. He checked the availability of raw materials. If
available, he wrote the allocation in the inventory log and reduced the quantity available; if
not available, he placed a purchase order and updated the inventory sheet with the expected
purchase quantity. He returned the customer order form to the order clerk who examined any
notes fiom purchasing, and passed the form to an order entry clerk for entry into the system.
This process often took several days before the order was entered into the system. A similar
procedure was used to maintain the finished goods log.

These manual procedures caused many problems for the firm. First, the purchasing
director spent virtually all of his time maintaining the manual records and preparing manual
reports for accounting to calculate cost of goods sold each month. Second, the firm was not
able to process customer orders as quickly as they would have liked because there were
bottlenecks in the system. Third, because several processes were manual, the information was

not available to other employees who could have used it to better plan their activities. For

55

example, production did not have access to information about upcoming orders. Therefore, it
was common for customer materials needed for a production run to arrive at the plant before
the production manager was aware of the customer order. When this occurred, it took time
for the receiving clerk to discover the source of these materials and to determine storage and
scheduling requirements. Finally, it was difiicult to identify and measure inventory problems
because they were often hidden in the manual processes and logs.

It is also important that the system not only process the events, but that it processes
them adequately so the business can be supported. In the case of Columbus, the management
team had selected a software package and had implemented most of the functions. However,
because the system did not process customer orders or purchase orders sufficiently, there was
a considerable amount of manual processes that were performed in addition to the automated
process. In this company, the system was not able to reduce the amount of work needed to
process an order or to order raw materials. Management felt it either had not provided
eficiency improvements or had in fact reduced these functional area’s eficiencies.

Information systems that process key events have been hypothesized to lead to
efficiency improvements. For example, Kelley (1994) discusses the dificulty in linking
information systems to improvements in productivity because there are many confounding
factors that influence a finn’s inputs and outputs. However, she hypothesizes that firms
should enjoy emciency improvements, but only from systems that either automate a process or
improve the capabilities of existing machinery. Because a key feature of REA systems is that
they automate and support an organization’s key processes, her hypothesis can be restated

that REA systems will provide greater internal eficiencies than traditional systems. In

56

addition, if all of the key processes are included in the system, the firm’s productivity may also
improve.

Because of both the theoretical definition of REA accounting information systems and
the problems experienced at most of the firms that did not process all of their events on the
system, the first key characteristic of REA systems is that all of the major business events are
supported by the firm’s accounting information system. This is measured by adding points to
the systems Accounting Systems Characteristics score for each key event that is supported,
subtracting points fiom the system’s Accounting Systems Characteristics score for each key

event that is not supported.

5.1.1.2 Detailed History of Events

REA systems must not only process key business events, but must also store detailed
history records of the them for a significant period of time. The theoretical papers describing
REA systems rely on this feature for the financial statement generation; without detailed
history records, a system would not be able to generate a financial statement view without
some type of general ledger file. In addition, it is recognized that maintaining detailed records
will overcome many of the aggregation and classification weaknesses inherent in traditional
accounting systems.

This feature was also identified by the management of several of the companies
included in this study as a source of benefits fi'om the system. For example, one of Madison’s
recent systems improvements was to enhance the sales module to store two year's detailed

sales information. In addition, a complaint expressed by many of their employees was that

 

57

there were only three months of purchasing and inventory history. Because of this data
shortage, time was often spent sorting through manual files to identify inventory sources and
tracking inventory histories. Management felt that a system that maintained complete
inventory histories would provide significant efficiencies to the firm.

The company with one of the most advanced systems had four years of order history
available on-line to their order entry clerks. This information allowed the order entry clerks to
answer customer questions while the customer was on the phone, and re-orders could be
entered more quickly. In addition, detailed sales analysis could be performed to understand
sales trends and to identify key customers.

Keeping a detailed transaction history in the system is another theoretical tenet of REA
accounting systems. It would be this history that would be used to materialize the financial
statements. As a result, this has been identified as the second key characteristic of accounting
systems. To measure this feature, the length of time that transaction history is kept on the
system is written by each event. If any company-specific events are implemented in addition
to the generic events, more points are awarded for those events that have history stored for

over one year than for those with less than one year’s history.

5.1.1.3 Integrated Data Processing
Integrating data in a central storage location is an important characteristic of database
systems, which was the foundation for the irritiai REA research. In these systems, data

redundancy would be eliminated, and the data could be available to all systems users. This

58

feature was important to several organizations, and has been identified as the third key
characteristic of REA systems.

In this study, several companies operated more than one system to meet their needs. If
there were multiple systems, it was common to have redundant data and redundant processes
to store the data required in each system. In the extreme, Bloomington had a system that was
designed by the corporate office, and they were required to use it for order processing.
However, this system did not keep any quality information or results of tests that were run at
the end of each production run. Therefore, their management team hired a consultant who
developed a personal computer application to store and report this information. They had
another customized personal computer application to store sales analysis data. In this
company, similar data were entered into all three systems. Several people who were
interviewed commented on how they needed all of the data, but they also complained the
process to maintain the data was cumbersome and inefiicient.

Movement toward integrated systems has been documented in the pulp and paper
industry with the introduction of mill-wide control and information systems (MCIS). These
systems process information throughout the mill, and have been shown to improve quality,
flexibility, efiicierrcy, and on-time deliveries. In addition, firms using them have enjoyed cost
reductions and increased information distribution (Technology and Labor in Pulp,
Paperboard and Selected Converting Industries 1994). These systems are Similar to REA
systems as they are integrated systems that process events as they occur. While none of the
companies in this study have implemented this type of system, three have implemented

technology that performs real time monitoring and control of the production process, and one

59

of these has integrated this information into its main accounting information system to update
production records.

Integrated data is a system characteristic that has been hypothesized and shown to lead
to several benefits. Porter and Millar (1985) identified integration as a key to gaining
competitive advantage. They stated that information technology needed to aid organization-
wide communication if the firm was to receive benefits from its implementation (and that
decentralized systems would not provide as many benefits as integrated ones). In fact, they
state that “unless the numerous applications of information technology inside a company are
compatible with each other, many benefits may be lost” (p. 159).

Another benefit of integrated systems was documented at Conoco where an integrated
expert information system was implemented and used throughout the world. Belcher and
Watson (1993 p. 249) found that the integration of information from around the world was
“helping people feel more connected and informed,” and, as a result, it provided a “boost to
the morale of the group.” While it is difficult to translate this type of intangible benefit to
either competitive advantage or productivity improvement, the Conoco executives believed
that this was a true benefit of the system and that it could lead to other more tangible benefits.

Ifthe systems are not integrated, there are potential dificulties with insuring that
redundant data remains consistent across the systems. Fisher (1994) states that this can cause
inefficiencies for both the film’s computer personnel who have the “hassles of maintaining
multiple information systems” (p. 75) and their auditors who are faced with the problem of
reviewing both systems to determine which number is correct. As a result, integrated systems

can lead to improved control and less attestation risk.

60

Again, REA systems would, by definition, maintain all of the data in a central data
repository, and there would be applications that enabled data entry to the repository.
Therefore, this has been identified as the third characteristic of REA accounting information
systems. The Accounting Systems Characteristic questionnaire deducts points from firms that
operate multiple systems. First, 2 points are subtracted for each time (after the first) that
users would have to log into the systems. In addition, if users would have to physically log
into multiple terminals, five points are subtracted for every system after the first. This scoring
metric means that firms that have one integrated system will have higher ASC scores than
those that use multiple systems being run as a somewhat integrated system in which the users
use one terminal, but have to log in several times. Both of these types of firms will have

higher scores than firms that operate multiple distinct systems.

5.1.1.4 Data Availability

Storing the data is just one step in making it valuable to the organization. There must
also be a means to retrieve and manipulate that data to provide users with information.
Similar to data integration, data availability is an underlying assumption of well—designed
database systems. REA systems, therefore, would provide users with a method to retrieve the
detailed data about resources, events, and agents that is maintained in the system.

The firms in this study had difi‘erent levels of data accessibility, and unavailability often
resulted in user complaints. In one case (West Lafayette), all of the sales transactions were
entered into their accounting system, and over two years’ data were being stored. However,

the system the firm used was a personal computer program that was not able to be networked.

61

As a result, the accounting clerk used the machine throughout the day to enter all of the
transactions, to produce invoices and checks, and to produce the monthly financial statements.
No one else had access to the system, and people who needed sales information were required
to maintain spreadsheets for themselves. When orders were placed, the sales stafi‘ first entered
them into at least two spreadsheets, one for the total order that was sorted by customer, and
another that stored the sales detail for each item sold. Then the order was forwarded to the
accounting clerk for data entry.

East Lansing, although operating a system that integrated most of the critical events,
had similar concerns. Their system did not have a report writing facility, so much of the sales
analysis and production data was not available to executives. As a result, they were very
dissatisfied with their system, and they were considering purchasing a new one that would
provide access to the information they wanted to better manage their organization.

The spirit of REA accounting information systems stresses the ability to provide
financial and non-financial information to all users in the organization. Therefore, the fourth
important characteristic of an REA system is that there is a method of data retrieval. This
retrieval method could be pro-programmed reports or inquiry screens, or it could be a facility
that enables users to ask questions and structure the answers that they nwd. While the report-
writer capability enables more specific reports and may be able to provide more information, it
demands a high level of training to be effective. Columbus has this type of program, but only
the controller had the background necessary to use it. As a result, he infi'equently created new
reports, and once they were created, they were run at scheduled times. To the user’s point of

view, these reports were very similar to the programmed reports that came with the software.

62

The amount of data availability is measured in the Accounting Systems Characteristics
questionnaire by asking the users what information they use and where they locate it. Ifthe
information is retrieved fiom the system (either through a report or an inquiry screen),
positive points are awarded to the system, moving it closer to the REA end of the continuum.
If, however, the data is not available to the user, but it would be if the system were an REA

systenr, then points are subtracted fi'om the system’s score.

5.1.1.5 Real time processing

REA systems are designed following a model of the real world they represent. For the
information stored in these systems to reflect the real world, it is important that they be
updated as the real world changes, i.e., that they have real time updates to the data. This is
difi‘erent fi'om traditional accounting systems that embody batch processing to improve data
accuracy and to simplify the recording process. Therefore, the fifth key feature of REA
systems is that they process transactions as they occur.

The importance of this characteristic was evident at the companies with the most
sophisticated systems. These organizations had systems that had been customized to process
non-traditional business events such as customer quotes and ink formulation, but the users
werestilldissatisfiedwiththeir systems. Themainreasontheygavewasthatthedatainthe
system were not up-to-date. At Madison, inventory was updated by batch programs, rather
than with real time updates. As a result, the inventory quantity on hand in the system was
virtually always missing information on either purchase receipts or raw material issues. It was

not uncommon for four people (the production manager, the purchasing manger, and two stafi‘

63

people) to walk into the warehouse to check the availability of a key material needed for the
day’s production.

Even when the system processes most transactions in real time, there can be concerns.
At Champaign, the major complaint was that there were only daily updates to both the raw
materials issued to jobs and the jobs as they moved through production. Each roll of paper
was assigned a roll tag. When the roll was used in a process, the tag was removed and put in
a bin to be entered into the system. A new tag was created for the resulting product that
could become a raw material for another step in the process. All of the day’s tags were
entered during the aftemoon. Customer service people complained that they were not able to
answer questions with up—to-date information. Scheduling was concerned because it could
not see if the jobs scheduled were actually completed.

These are examples of complaints arising fi'om systems that process key events and
store the transaction detail for future inquiries. However, their weakness is that this
transaction information is not available as it occurs. This is another important characteristic
that differentiates REA systems from more traditional ones. An REA system is one in which
the database is generated fiom a model of the business, and the data should reflect reality to as
great an extent as possible. This feature is not measured directly in the Accounting Systems
Characteristics questionnaire. However, for future research it would be possible to add two
steps in its completion to capture this information. First, batch updates could be identified on
the generic REA diagram by drawing dashed lines between entities that have lags between
events and the updating of the resource and agent data. For example, if receiving information

is not entered and updated in real time, then a dashed line would be drawn between the

64

PURCHASE event and the INVENTORY resource. Second, a set number of points could be

subtracted for each event that is not updated in real time.

5.1.1.6 Directed REA Design and Implementation

The next important characteristic in the definition of REA systems is that they should
be designed following the REA template. Ifthis occurs, the economic events will be identified
along with their related agents and resources. In addition, the duality relationship between
events identifies the resources that are “given up” in order to “get” each resource. By
analyzing these exchanges, management can evaluate whether the exchange is economically
beneficial to the firm. When the system is implemented, the business process should be
designed to streamline each of the exchanges. Next a centralized database to store the
important information about the exchanges should be designed. Finally, the data would be up-
to-date and available to system users.

While this type of design had not been performed at any of the organizations, it was
the process used when making recommendations during the final seminar that presented the
results of the systems analysis. In several cases, the REA diagrams and philosophy sparked
ideas for systems improvements. For example, after discussing the revenue cycle at Ann
Arbor, they realized that they could enter the shipping information in the warehouse as the
goods were leaving the property. This would result in at least three benefits. First, the
invoices would be generated and mailed a day earlier, thus improving cash flow fi'om

receivables. Second, the system would be updated so customer service would know accurate

65

order status information. Third, management would have access to sales figures more
quickly.

If a company performed a directed REA design and implementation, the resulting
system should encompass the earlier key systems characteristics. However, implementation
compromises may occur if the organization determines that the costs of developing such a
system would outweigh the expected benefits. Regardless of the compromises, the
Accounting Systems Characteristics would award positive points to firms following the design
approach, which would be recognized as a reengineering exercise. AS such, systems resulting
fi'om this type of design would move more toward the REA end of the continuum, regardless

of the final comprorrrises that were necessary.

5.1.1.7 Elimination of Journal Entries

The final characteristic of REA accounting systems is that they do not generate journal
entries or maintain a general ledger. Rather, the financial statements are the result of View
materialization of the detailed transaction data stored in the system.

As expected, none of the organizations participating in this study had systems such as
these. However, the executives participating in the final sessions were interested in the
implications of such systems. Although their systems used joumai entries to update a general
ledger, they were able to identify areas in their current transaction processing systems where
the data was available to produce financial statement information. For example, the CEO at
East Lansing explained that their current system stored all of the detailed sales information

and could be used to prepare the sales portion of the income statement; he believed that their

66

transaction processing systems embodied many of the key features of REA systems. In fact,
he was correct as their system was scored at the top of those included in this study.

The research that looks at systems that do not use a chart of accounts or that have
been developed using a directed REA approach is limited. To date, the only published work is
that by Andros et al. (1992) who discussed the reduced number of systems that resulted and
the improved flexibility and efficiency enjoyed by IBM with the implementation of systems
that were developed following the REA approach. A major reason few researchers have
evaluated REA systems is that there have been very few directed REA implementations. As
they become more common, it will be important for researchers to evaluate their success,
identifying the costs and benefits of implementing and maintaining them.

To identify firms that have implemented a system that has adopted all of the REA
characteristics, the ASC questionnaire includes a question about how the system generates
financial statements. Ifthe system does not use joumai entries, 100 points are added to its
ASC score. This would insure that the system’s score would be significantly higher than any

other that had not been the result of a complete REA implementation.

5.1.1.8 Summary

These key characteristics (refer back to Table 2 for a summary) result in an operational
definition of REA systems. This definition can be used in future systems evaluations because
the accounting systems characteristics construct will no longer have two well-defined values:
traditional or REA. Rather, there are specific features of the systems that are along a

continuum.

67

5.1.2 Key Organizational Characteristics

One of the advantages of field-based research is that it provides Opportunities to study
the interaction between the organizations and their systems, rather than studying the systems
in isolation. When the site visits had been completed, it was possible to compare the
organizational characteristics that had influenced each organization’s systems decisions. It
became evident that the management teams were trying to make the best systems decisions for
their organizations, regardless of the system’s sophistication. In addition, organizational
characteristics influenced management’s perceptions of their systems. As such, examining the
system in isolation would not provide a complete view of the organization’s system. Instead,
the key organizational characteristics must be considered. Therefore, key organizational
characteristics (see Table 3) have been identified. The following sections discuss each of
these characteristics and their potential influence on the measurements and quantitative results

presented in this dissertation.

 

Table 3: Key Organizational Characteristics
1. Complexity.
2. Size.
3. Organizational structure.
4. Technical sophistication.

5. Costs versus benefits.

 

 

 

68

5.1.2.1 Complexity

Although the firms in this sample were selected from the same industry to minimize
confounding factors, it was evident that some of the firms had more complex processes than
others. Complexity in this Study is defined as the type of manufacturing process that is
performed, the number of products produced, and the number of outside agents that interact
with the organization. It became evident during the site visits that complexity influenced the
type of system that was needed to manage the organizations, as firms with more complex
environments implemented and supported more advanced systems.

For example, the system used by West Lafayette earned the lowest ASC score, but
management was satisfied with its performance. They used a single personal computer to
operate an integrated, ofi‘-the-shelf accounting program. All sales analysis reports that were
needed had to be generated fi'om spreadsheets. However, management was able to
successfully control their business using this system because the business was not very
complex. They sold only 15 products and did very little development of customer-specific
products. They produced twelve of their products on-site, and the production process had
only five steps. For their remaining products, they performed a limited amount of assembly,
relying on vendors for most of the production. Management believed that they were able to
monitor the organization’s progress with the information available. In addition, they had
estimated the costs of maintaining additional spreadsheets as considerably less than the costs

of implementing a new system.

69

It is possible that sophisticated systems cannot be cost justified in organizations that
do not have complex operations. For example, Porter and Millar (1985) claim that
opportunities for competitive advantage are most likely to arise in organizations that operate
with “information intensity.” Examples of information intensity include using a large number
of suppliers and/or customers, producing goods in many distinct product categories,
producing products that require many parts or many steps to complete, and producing

products that have a long cycle time.

5.1.2.2 Size

The size of an organization will also influence the type of system being used. Larger
firms in this sample had in-house MIS departments and were apt to modify their software to
meet new needs. They used technology to help monitor their processes, and to summarize the
large number of transactions. Their management believed that this type of technology was
able to improve eficiency and effectiveness, thus providing a positive return to the firm.
Managers of many of the smaller firms felt that investment in production technology was able
to provide a larger return than investment in information technology. Therefore, these smaller
firms were willing to accept pre-programmed software, and they modified their procedures to
work around any system irregularities.

In both cases, the management believed they were making a justified systems decision
for their firms. Therefore, the smaller organizations perceived their system to be “adequate”
for their present environment. The size of the organization, therefore, will influence

management’s decisions and their perceptions of their information system.

70

5.1.2.3 Organizational Structure

The organizational structure also imposed systems constraints on individual business
units. Ifa firm is composed of several divisions, the amount of centralization will have an
efi‘ect on the system and how closely it mirrors an REA system. By having additional business
units, the parent organization has to weigh the benefits of having a centralized, integrated data
base, with the costs of possibly not processing the individual business units transactions in
most efiicient manner.

For example, if the firm imposes a structure at the corporate level, then all of the
divisions will use that system, and the information can be consolidated and integrated for use
at the corporate omce. However, the systems may not be tailored to the individual business
units, and it will not be able to adequately process their key business events. This was evident
in this study, as Bloonrington was required to use their parent company’s order processing
system, even though it did not provide them with all of the sales information their
management and customers required. Similarly, Champaign included a small division that is a
distribution company for products used in conjunction with those produced by the main
division. Managers of this smaller division were constrained because they had to use the
system developed for the manufacturing firm although they themselves did no manufacturing.

On the other hand, allowing each business unit to establish their own information
system will result in two potential costs. First, the costs to develop the individual systems
could easily be greater than the cost to develop one system that would then be used at all

sites. Second, consolidating the divisional data from distinct, difi‘erent systems will be more

71

costly than transferring similar data from each of the divisions. For example, Madison had
purchased a company that was located out of state. It was allowed to continue to maintain its
old system, but this resulted in monthly manual processes to transfer the divisional information

to the centralized accounting information system.

5.1.2.4 Technical Sophistication

Management of the firms in this study exhibited a wide range in technical philosophy.
Some of the companies employed managers who were very technically aware, and these
companies were more likely to implement more advanced systems. In each case, there
appeared to be a technology “champion” who was able to evaluate technological innovations
and who could communicate expected benefits to the rest of the organization’s management.
They were able to convince the management team that new technology was possible to
implement in their organization, regardless of its size, and that the benefits would outweigh
the costs.

Most interestingly, managers in these firms were able to identify many more
weaknesses in their current systems, and they were not satisfied with how they performed.
Their Competitive Advantage scores were driven down by the few systems shortfalls they had
focused on, rather than the wide range of firnctions the system was adequately providing.
These managers were aware of new, more sophisticated technology, and they felt theirs was
already out of date. In several cases, they were pleased to hear fi'om the researcher that their
system was a leader in the industry; they had not realized that they were so far in fi'ont of

their competition.

72

5.1.2.5 Costs versus Benefits

Finally, it appeared that firms were willing to work with their current system until the
benefits of replacing it would clearly be greater than the costs. Before this occurred, they
would develop non-integrated personal computer applications that would support the
organization where the main system could not. For example, Evanston had developed a
personal computer application that would store Bill of Materials and Operations List
information about each product. Although they would have preferred to integrate this with
the order entry and inventory modules of their main system, they did not have the in-house
expertise to modify their main system, nor did they feel that purchasing a new system would
be cost effective. However, management believed that, if the organization continued to grow,
they would need a new system for better maintenance of specific information, along with more
detailed inventory and purchasing information.

These cost-benefit decisions will result in an organization’s ASC score changing over
time. When a new system is first implemented, it is likely that it will meet most of the firm’s
needs. Therefore, it will exhibit many of the characteristics of an REA system: it will process
key events, and store the detailed data in a centralized location. However, as the weaknesses
in the system become apparent, and stand alone applications are developed to overcome them,
the ASC score will decrease; more systems will be needed to satisfy users needs. When the
system becomes so inadequate that another is purchased, there will be another jump in the
ASC score. In all cases, the system in place may be the most cost efi‘ective for the firm at that

point in time.

73

5.1.3 Summary of Qualitative Results

By comparing and contrasting the firms and systems in this study, the author was able
to identify key characteristics that influenced successful systems implementations. First, a
more specific operational definition of REA systems was developed using the key systems
characteristics. In general, it is hypothesized that firms using systems incorporating more of
these characteristics will enjoy greater benefits than firms using systems that are more like
traditional systems. However, it is important to recognize firm-wide characteristics that may
influence an organization’s systems choice. Therefore, a list of organizational characteristics

that were important to this study’s firms was presented and discussed.

5.2 Quantitative Results

The goal of this section of the dissertation is to gain ‘a better understanding of benefits
that arise fi'om systems that are more similar to REA systems than those that are more
traditional. Specifically, it has been hypothesized that more advanced systems will enable
firms to enjoy competitive advantage and improvements in productivity. The following
sections describe statistical tests that were performed to provide evidence about these
hypotheses. In each section, the analysis relies on the quantitative data that was gathered

from the organizations during the site visits.

74

5.2.1 Preliminary Data Analysis

Several preliminary data analysis activities were required before the hypotheses could
be tested. First, the Competitive Advantage Survey was analyzed to determine if the
questions asked were representative of the different factors in the Bakos and Treacy model.
Factor analysis was performed to identify the model of Competitive Advantage that should be
used for the hypothesis analysis. Section 5.2.1.1 describes these procedures. In addition,
factor analysis was performed on the User Satisfaction responses received from the executive
and staff members who participated in the study. The results of this analysis are presented in

Section 5.2.1.2.

5.2.1.1 Factor Analysis of Competitive Advantage Metric

After the site visits were completed, factor analysis was used to verify the validity of
the questions in the Competitive Advantage survey. Confirmatory factor analysis was used
since the questions had been developed to measure specific, theoretically defined constructs.
As such, the questions could be grouped by factor in the Bakos and Treacy model.

Fifty-nine executives had completed competitive advantage surveys during the site
visits. Of these, four had missing responses for several questions, so they were orrritted fi'om
the analysis. The remaining fifty-five responses were used in the factor analysis. These
respondents were executives in the firms who had managerial responsibilities for various
functional areas. By having from executives across functional areas complete the
questionnaire, the summarized perceptions would be less biased. For example, one would

expect that the MIS Director would believe that the system was providing competitive

75

advantage. In addition, if the system supported customer service adequately, the executive
responsible for this area may have similar impressions. However, if the system was weak in
manufacturing processing, the executives responsible for this area may identify systems
weaknesses that the others would not. The author’s goal was to have each company
represented by the same number of executives from the same functional areas. Because
several firms had different organizational structures, this was not possible. For example, two
firms did not have purchasing departments established, but had purchasing performed by
various other executives. Similarly, several firms did not have a manager of distribution.
Therefore, participation was tailored to the finn’s organization with the goal of including
executives fi'om a wide range of functional areas. Several of the organizations had multiple
sales and marketing executives who were responsible for difi'erent product lines or customer
bases; therefore, the sample includes more sales and marketing executives than other

firnctional areas. Table 4 provides information regarding the survey respondents.

76

Table 4

Executives Who Completed the Competitive Advantage Survey

Number of Respondents/Company

Mean 6.88
Minimum 3
Maximum 9
Number of

Respondent’s Functional Areas: Exgutjves
CEO/Owner/Partner 4
Customer Service 3
Distribution 5
Finance/Accounting 7
Management Information Systems 3
Production 8
Purchasing 5
Sales/Marketing 15
Other _5_
Total 55

The methodology used to identify the specific questions to be included in the
competitive advantage analysis is similar to that of Sethi and King (1994). Correlations
between the question responses were calculated, and the resulting correlation matrix was used
to perform factor analysis of five constructs: Unique Products, Switching Costs, Internal
Efliciencies, Interorganizational Efliciencies, and Competitive Advantage. Six questions had
been written for each of the competitive advantage strategy constructs, and three were
included for Competitive Advantage. All of the questions in the survey were written to
determine if the film’s information system contributed to the firm’s competitive advantage or

individual strategy. Therefore, when executives agreed with statements in the survey, they

77

were not only saying that the firm enjoyed competitive advantage (or unique product features,
etc), but that their system helped them achieve this status.

Confirmatory factor analysis was performed to determine if there were five factors that
were measuring the five constructs in the Bakos and Treacy model. The questions fi'om the
survey were identified as indicators for one of the five factors in the model. The factor
analysis results were analyzed to identify questions that were inconsistent with the factor they
had been written to represent. The first level of analysis focused on the errors between
predicted correlations among the questions in each factor and the actual correlations from the
data. The predicted correlations were calculated using the Internal Consistency Theorem that
posits that the correlations between indicators within a factor should be equal to the product
of each indicator’s correlation with its factor (i.e. Iij = rm * If] ). Questions that had
significant errors were analyzed to identify ones that should be removed fiom the study. This
was an iterative process, removing one question from the model each time and performing the
analysis again. This isolated each question’s efi'ect on the constructs and overall model, and
the procedure continued until there were no significant errors between the actual and expected
correlations between the questions. Table 5 shows the correlation matrix of the independent
factors and the questions that were included at the end of this phase of the analysis.

The second level of reliability analysis was to calculate Spearman Brown’s Standard
Score Coeficient Alpha for each factor. Except for Switching Costs, the alphas for the
factors were over 0.75. Because of the low reliability of Switching Costs, this factor was

removed from the analysis.

78

The final level of reliability analysis was to perform additional tests to determine if
each factor was internally consistent, testing if the questions within each factor were
unidirnensional and parallel. Unidimensional means that each of the indicators within the
factor relate to the other indicators of the factor similarly. Parallelism means that all of a
factor’s indicators are Similarly related to the other factors in the model. This analysis showed
that all of the factors were internally consistent, except Internal Efficiencies. The indicators
of Administrative Eficiency were measuring a construct difi‘erent from those questions
written to measure Production Efliciency. Therefore, the construct was changed to measure
only Administrative Efiiciency and was linrited to questions 4 and 9 from the survey.

The test for parallelism failed for the Unique Product factor, and it was determined to
be similar to the Interorganizational Efl'iciency factor. These two factors were therefore
combined, and the analysis was performed again. At that time however, the resulting factor
failed to be internally consistent, i.e., the indicators of this factor were not measuring one
underlying construct. Bakos and Treacy had identified dificulties in differentiating the various
strategies that involve firms within the value chain. They recognized that the difi‘erent
strategies firms use to gain competitive advantage may often occur simultaneously and that
their efl‘ects may be hard to segregate. For example, they describe a situation in which a
product innovation could improve production eficiency, enhance product uniqueness, and
increase customer switching costs. Because firm strategies may encompass several of factors
in the original Bakos and Treacy model, the empirical model was modified to include only two

independent factors: Internal Strategies and Interorganizational Strategies.

79

an. ”LaeeecEm weaning—C m9 2e. ”A88 9.2258 m
E. ”Aaceocam 3.53 a 8n. e685 33:8 :
Hafiz 33880 28m eager. Peres 528%
38$ :08 .58 8 3.33:. 80903 05 @552 «can: 5 unease. .88..— ..32: .88.? 58 a .8 «mauso— 33 05 8e 28.. .58 .3. 05.
8682;
as seen 838 can League 2. see as: .2: .3 8.322.. 8822.8 Sean... 2.. 3% 3&5 2: 25% ones: 2:. .8.
.5 3:33.: E8885 05 e833 edema—oboe 188 05 28% 3.33% :09 05 323 0385 2:. gown—9:8 2a 8828 5.36:0 05 5 £8.52

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

we ea 8 R we an R an .2 7n 3 a. E a _- 3. was
we a. 8 8 _:. an en E _2 _n e... 2 .2 2 we a. .m—
2 we a 3. .3 a 2 a .2 _m~ 2 52 Lee 9. _o_ 2. _m
a 2 2- a _2 e. 2 a _e an em 2 .3 .3 We 3 S
3 a x. .5 an 3 em a a on 2 R a _8 2. en

a. :. a. .3 R an 2 3 an 8. R a a _n en- 2
n... me an .3 an 3 z _ 5 a X en. 2- .2. 3 en
a :. 3 .3 an 2 a n 3 2 :. a. 2 x“ no. a
_ 2. en en a 2 ”N n 2 ea 2 3.. 2

.8 3. en N- e- R 2 m- a- 2. an n

.3. an a a I. 2 3 z z a em. e

.2. 3 we 2 e. a e 8 3 R .3- e

a N. am ea .. L. an. L

L. m an :. 3. 7 _r a

e a. a a R _. _2 :

an em 2 n e z _N a

3. a E” .3

.9. 3. _8 A:

_2 a 7n 5

.3 Be a. _ _e

8 R «A a 2 n a e L a : a a” E 2 _e _
me. E m a . -

case: 8.5.5

.282 L80!— Eeh a a. 3.5850 Len 5.5.: gun—0.5.0 um 03:.

80

These factors were analyzed using the same procedures as described above. First,
the six questions relating to Internal Efliciencies were identified as those for Internal
Strategies, and all of the questions for Unique Products, Switching Costs, and
Interorganizational Efliciencies were combined for the Interorganizational Strategies
factor. The iterative analysis of errors between expected and actual correlations was
performed, and the constructs were tested for internal consistency and parallelism. After
several iterations, it became clear that the administrate efficiency questions identified one
factor, while the production efliciency factors were more similar to the
Interorganizational Strategies construct. Therefore, the Internal Strategies factor was
renamed Administrative Strategies. Table 6 shows the predicted and actual correlations
for the Interorganizational Strategies factor in the resulting model. Table 7 shows the
internal consistency analysis for both Administrative Strategies and Competitive
Advantage factors. Because there were fewer than four indicators for these factors, they
were analyzed by comparing predicted and actual correlations for questions outside of

their factor. The predicted correlations were calculated as m = rm * [3113* rpipj.

81

.22.... m. 8.33% 3823.8823. 3. 2...... 20.0880 .38 .583 9.8... gm
.32... .38 .58 a. 3.2.8... 2.058.... 0... gen... 8...... a. $.58. .92.... .89.. .3662... a .8 «95.3. .22... o... o... 28. 8.... .m... 9....

is... 8 E 8.....sz 2.. 238.88 .58.. 8.. 86.32.. 8223 882...... 38%.»... .5. .8. 3 8.3.8...
25.8.2.8 8.232.. 2.. so... 8.93... 5...... 2F .8. 3 3.8.8. 228...... 2.. :85... 8822.8 .38 o... .32.. :20 as... 035... .28. 2:

engage... 3:22.800 u <0
8.82.3 8.8.8283. u 9.
8.9.25 3023.59.88... u mo. ”9.2.3

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8 .3. a. .2 .8. S .e .3 .2. .3 .8 .2 an R an <0
.. .3 R .8 s. 2 2 .2 .2 _. n. 3 e. an 8 m...
8 _8 8 .3. 3 a. .8 .8 _: .2. 3 a. be 2 .3 mo.
.2 8 .2 N on 8 .R .8 2 a 2 a 2 .3 3

es .8. 3. an an _ce an e. h. n. 3. .n _on Na

.8. en a. .3 .on .e e. .n .. an s. .83 o.

z. 2 n- .2 e. . .. 3. 3. .n 3 a

3 2. .2 e. e. a 8 3. S e 3

e. .3 8 .8 2 R .N 2 a an

3 on R a 2 2 en 2 a. 3

a. .n a 2 2 3 2 3 e. .N

3 2.. 3 3. 8 on .n .8 a. 2

.2 .N 3 a. e. .3 3 .- n R

2 .N n. 2 e. n. 2 2 3 2

a 8 3n .3 2 a .3 R 3 .2

.m 3m .83 .3 e. b. an . e. .8. a.

3. h. _.m e. .8. .N a. .3. _.~ m.

2 3 .8 2 .2 .2 8 A... a _ n

3 2 e. .8 3 2 3 .3 2 .2 _~. .2 .2 R _...
<0 9. . , mo. .

2352... 02...»..an .8 .282 858:. 2.. a. 833...... 18.3.5983... .3. 38.3028 e 2......

 

82

833.8 83.8.33... 3888.38.89
32.8 3.33.883. 2888.52.82
.8282 2.0.0580 .583. 58.8%

35.3.2.8 .338 .2... 3.0.80... 803.2. .28 3855.8 8.85.89 3
8...... 3 .83.. 3 .3... u .8. an 3.3.3.8 28 88.3.2.8 v0.0.3...

 

.3 .3 3. 3 N

33 33 83 33 3.3

33 33 8 83 m...
.3 33 33 3. 83 8 m
.3 33 33 3 33 83 m
33 33 3 3. 3. 83 33 83 33 33 mm
.83 33 33 3. .3. 88 33 33 3- 3.- mm
33 33 33 3. 3. 33 88 83 3. 3. 43
.83 .3 .3 83 83 88 33 .8 8. 8. 3...
33 83 83 .. .. 33 8. 8. 3 8- N3
33 8 8 .. .. 33 3. .3 3- 3 3|.
33 .3 ..3 8. 8. 3. 33 .. 3. 83 N3-
33 .33 33 3. 3. 33 3. 33 3. 3. 33
83 33 33 3. 3. 33 .3 33 .3 33 .3
83 .33 33 .3. 3. 33 83 83 33 8 m
.3. mm 8|.- .. m m 3.3. 8|.. .3 m a
als8a.>..< 3.3.8.4....” 333.8388... . .3...
023.0:500 03.9.3.5:— < 029303—ch 02.9.3.5:— <

 

8.... .3 2... 8.2....

$3.533. 3.3.2.880 E... 8.3.35 o>..¢...»....:...< .8. 8.8.3.3.... 3:33.230 .3 2......

83

The three factors were also tested to determine if they were uniform and parallel.
In each case, they were both. As such, the data collected from the executives in this study
better fits the modified model of Competitive Advemtage in which firms select either
Interorganizational or Administrative Strategies to earn competitive advantage. Because
the data support this model, rather than the more detailed Bakos and Treacy model, it will
be used for the remaining analysis.

Four questions were included in the metric to collect information about the type of
information systems department each company operated, whether the organization
developed applications or purchased ofilthe-shelf software, and whether management
considered competitive advantage when they evaluated potential applications. These
questions were hypothesized to represent the factor called Development Characteristics.
Factor analysis was performed on these questions measure their validity. Table 8 shows

the predicted and actual correlations between these questions.

84

Table 8: Correlation Matrix for Development Characteristic Questions

1
45.1 39. 27.
59.04 41

32 42
72

 

Where DevChar = Development Characteristics

The lower triangle of this chart shows the actual correlations between the indicators
multiplied by 100. The upper triangles show the predicted correlations
multiplied by 100.

Predicted correlations are calculated as r; = rm "' rm- . There are no significant
differences between predicted and actual correlations.

The row is the factor loadings for the development characteristics factor.

Spearman Brown’s Standard Score Coeficient Alpha for Development Characteristics
is 0.739.

The factor questions were also tested for unidimensionality and parallelism. The
factor was determined to be both, so responses to these questions may be summed to

create a Development Characteristics measure.

5.2.1.2 Factor Analysis of the User Satisfaction Metric

To measure User Satisfaction (US), the four question metric from Seddon and Yip
(1992) was included in the Competitive Advantage survey that was completed by the
executives participating in this study. In addition, all of the stafi‘ employees that were
interviewed during the site visits completed a one-page survey that consisted only of these
questions (Appendix C). In total, US data were gathered from 86 people, 31 staff

members and 55 executives. The stafl‘ members performed various functions including

85

accounts receivable, accounts payable, customer service, order entry, billing, stafl'
purchasing, and production supervising.

Confirmatory factor analysis was used to determine if the four questions included
in the survey were a reliable measure of User Satisfaction. Table 9 shows the correlation
matrix for these factors. The lower portion of the matrix shows the actual correlations
between the questions, and the upper portion shows the predicted correlations. None of

the errors between predicted and actual correlations were significantly different fi'om zero.

Table 9: Expected and Actual Correlations between User Satisfaction Questions

1 3
83. 91 5
83 84. 48.
52.
55 5

Factor Loading: 95

 

The lower portion of the correlation matrix shows the actual correlation between question responses "
The uppelrogortion of this matrix shows the predicted correlation between question responses “ 100.
There are no significant errors.

There are two difl'erent tests for unidimensionality. The first one tests if the
factors not only are correlated with each other similarly, but that they represent the factor
with the same “quality” or strength. The second one relaxes the quality constraint to
allow the quality of the factors to vary. The combined test for a unidimensional factor that
had all items of equal quality was rejected for this factor. This failure was the result of

question four which was reversed coded and asked if the respondent was dissatisfied or

86

satisfied with the system. Afier the author received several surveys with inconsistent
responses to the last question, she asked a number of respondents why they had completed
the questionnaire as they had. They all admitted that they had not read the question
carefully and had made a mistake. If the parallelism test was modified to allow the quality
of the items to vary, then the data supported the model that there was one factor being
represented by these four questions. In addition, the Spearman Brown’s Standard Score
Coeflicient Alpha that measures the factor’s reliability was 0.90. This is similar to the
reliability factors published in earlier research using this survey. Therefore, the four

questions will be summed to create a User Satisfaction index.

5.2.2 Hypothesis Testing and Results
5.2.2.1 Hypothesis One: Test of the Bakos and Treacy Model

While the original Hypothesis One predicted that the factors of Bargaining Power,
Switching Costs, Internal Efliciency, and Interorganizational Efl'iciency would lead to
Competitive Advantage, confirmatory factor analysis revealed only two distinct factors
from the survey responses: Administrative Strategies and Interorganizational Strategies.
Therefore, the first hypothesis must be restated as Administrative Strategies and
Interorgtmizational Strategies will lead to Competitive Advantage.

To test this hypothesis, each individual’s survey response was used to calculate an
index for each of the constructs in the hypothesis. The responses from the ten questions
that were indicators of Interorganizational Strategies were summed to create an

Interorganizational Strategy index. Similarly, the two questions of Administrative

87

Strategies and the three questions of Competitive Advantage were summed to form their
index values.
These index values were used in a linear regression analysis of the following
equation:
CA = a+flJS+fi2AS+ e
where
CA == Competitive Advantage,
IS = Interorganizational Strategies, and
AS = Administrative Strategies

Both factors are hypothesized to be positively related to Competitive Advantage. Table

10 shows the results.

Table 10: Linear Regression Analysis of Factors Leading to Competitive Advantage

 

 

 

 

 

Regression Statistics
Multiple R 0.713637
R Square 0.509278
Adjusted R Square 0.490405
Standard Error 3.43196
Observations 55
ANOVA

W SS MS F Significance F

Regression 2 635.635 317.8175 26.9832 9.16E—09
Residual 52 612.474 11.77835
Total 54 1248.109

 

 

Coeflicients Standard t Stat P-value

 

 

Error
Intercept -7.52484 2.683061 -2.80457 0.007069
IOS 0.343696 0.051373 6.690231.54E-O8
AS 0.337281 0.147918 2.280184 0.026726
Where:
108 Interorganizational Strategies

AS = Administrative Strategies

88

As Table 10 shows, the coefficients for both of the strategies for Competitive
Advantage were positive and significant, which support the modified causal model for
Competitive Advantage. In addition, the overall model has an Adjusted R2 value of 0.49,
and the ANOVA results show an F score that is significant at less than 1%. All of these
measures support hypothesis one. They also provide some evidence that the strategies
proposed by Bakos and Treacy are seen by managers as methods to use computer systems
to gain competitive advantage.

If the Development Chwacteristics factor is added to the model, it is hypothesized
that firms with the higher development score, i.e. those that create their own applications
considering Competitive Advantage during application design, will have higher
perceptions of Competitive Advantage from their computer systems. Table 11 shows the
regression results when the measure of Development Characteristics is added.

Although Interorganizational Strategies is still a more important predictor of
Competitive Advantage than Administrative Strategies, Development Characteristics is
now the only significant cause of Competitive Advantage. However, the Adjusted R
Square statistic has decreased showing that the overall predictability of this model is lower
than the previous one.

These results provide some evidence that the firms that perceive they are enjoying
Competitive Advantage are those that manage their computer systems to achieve such
advantages. However, it is not possible to determine if manager’s perceptions are
influenced because they feel ownership for the information systems, or whether they are

actually enjoying a Competitive Advantage. Future research should be performed to

89

determine whether manager’s perceptions are accurate measures of Competitive

Advantage.

Table 11: Competitive Advantage Linear Regression with Development

 

 

 

 

 

 

 

 

 

Characteristics

Regression Statistics
Multiple R 0.639
R Square 0.409
Adjusted R Square 0.374
Standard Error 3 .803
Observations 55
ANOVA

3] MS F SigF
Regression 3 510.624 170.208 11.771 5.66E-06
Residual 51 737.485 14.460
Total 54 1248.109
Coeflicients Standard t Stat P-value
Error

Intercept -3.94 4.256 -0.927 0.358
External 0.096 0.066 1.448 0.154
Internal 0.126 0.108 1.167 0.249
DevChar 0.388 0.131 2.969 0.005
Where:

108 = Interorganimtional Strategies

AS = Administrative Strategies

DevChar -= Development Characteristies
5.2.2.2 Hypothesis Two: REA Systems Provide Competitive Advantage

Since the data fi'om the Competitive Advantage survey have been shown to be
reliable, they are used as the foundation for testing Hypothesis Two, a prediction that

Accounting System Classification will afi’ect Competitive Advantage. Specifically, the

90

hypothesis is that the firms using systems more like REA systems will enjoy more
competitive advantage than firms that have more traditional systems.

To test this hypothesis, average index values are calculated for each firm. The
Accounting System Classification (ASC) score is calculated fiom the ASC questionnaire
completed during the site visits. This score has two components. First, each firm’s
computer system is analyzed to determine if the key events are supported. Second, system
users are questioned about the information they use to perform their daily tasks and where
they get this information. These questions are scored to determine how the system meets
the user’s needs. All of the user scores are averaged to determine the average
organizational score for the firm. To weigh the organization and system scores similarly,
the average organization score is multiplied by ten; the result is summed with the system
score to calculate the firm’s ASC score.

Each firm had several executives complete the Competitive Advantage
questionnaire, and the responses to three of the question were summed to create an
individual Competitive Advantage index used to test Hypothesis One. For this analysis,
the responses fiom the executives of each firm were averaged to create a firm-level
Competitive Advantage index. These responses had to be averaged, rather than summed,
because there were not the same number of responses for each firm; one firm had only
three executives who participated while the maximum number of participants was nine.

To test Hypothesis Two, the firm-level Competitive Advantage index was
correlated with ASC. Lower scores on the Competitive Advantage index meant that the

executives believed more strongly that their system provided competitive advantage, and

91

higher ASC scores represented more advanced systems. Therefore, the correlation
between the two is hypothesized to be negative.

Table 12 shows the resulting correlation matrix that includes the three factors from
Hypothesis One and the ASC score. As shown, the correlation between ASC and
Competitive Advantage is negative, as predicted, but is not significant. However, ASC is
significantly negatively correlated with Administrative Strategies. Therefore, this provides

some evidence for the model posed at the end of Chapter 3 as Research Question 1.

Table 12

Correlation Matrix of Competitive Advantage Factors and ASC Scores

 

 

 

CA External Admin ASC
CA 1
External 0.874894 1
Admin 0.533935 0.253506 1
ASC -0.08044 0.250778 -0.54917 1
Where

CA - Competitive Advantage

External = External Strategies

Admin = Administrative Strategies

ASC = Accounting System Classification Score

5.2.2.3 Research Question 1: REA Systems Indirectly Provide Competitive
Advantage

To firrther evaluate the model presented as RQI, path analysis was performed on
the firm level data. The first model tested is shown in Figure 8. This model predicts that
ASC is an antecedent factor of the Acbninistrative Strategies (AS) and Interorganizational
Strategies (10S) constructs, which are both antecedent factors of Competitive Advantage

(CA). Higher ASC scores represent more advanced systems, while lower scores for other

facts
We

both

Fig

 

s(liar

data f

for the

92

factors represent stronger perceptions of each specific advantage from the computer
systems. As such, the link between ASC and AS, and the link between ASC and IS are

both predicted to be negative. The links to CA are both predicted to be positive.

 

 

 

 

 

 

0 55 Administrative
' ' Strat ’es
cg] 0.34
Accounting

System
Classification Competitive

Advantage

 

 

 

 

 

 

0.25

 

Interorganizational 0.78
Strategies

 

 

Figure 8: Path Model of an Indirect Effect of ASC on Competitive Advantage

 

 

Using the Spearman Brown’s Standard Score Coefiicient Alpha scores as the
reliabilities for the factors from the Competitive Advantage survey, this model was tested
using path analysis. The resulting path coefiicients are shown in Figure 8. As predicted,
the link between ASC and AS is negative; however, the link between ASC and IS is
positive. Both of the links to CA are positive, as predicted. In addition, the overall Chi
square value for the model is 0.41, which has a tail probability of 0.814. Therefore, the
data fails to reject the hypothesized path model.

These results provide evidence that more advanced systems do provide a means

for the firms to gain administrative efiiciencies. However, the main way that executives

 

93

perceive their systems as providing competitive advantage is through the implementation
of Interorganizational Strategies. Higher ASC scores were actually indicative of firms
that believed their system provided fewer benefits from these strategies.

A second model was tested to determine if ASC also had a direct effect on

manager’s perceptions of competitive advantage. This model is illustrated in Figure 9.

 

 

  
 

 

 

 

 

 

 

 

 

 

Accounting
System
Classification
0.25
Interorganizational
Strategies

 

Figure 9: Path Model of a Direct Effect of ASC on Competitive Advantage

 

 

The path coeflicients are also shown on this figure. The links that were in the
previous model have changed very little. The new link between ASC and Competitive
Advantage is negative, as predicted, but is relatively small. The Chi Square score for this

model is 0.39, which again shows that the data fail to reject this model. This provides

C0lll

Stra

the

pro

Illll'.

rati

acti

lllll

PFC

Prc

Col

94

limited evidence that ASC has a direct effect on CA, but it is not large. Again, the major
contribution from more advanced systems appears to be in improved Administrative

Strategies, rather than in Competitive Advantage.

5.2.2.4 Hypothesis Three: REA Systems Lead to Operational Improvements

Several measures are used to provide evidence about productivity and efficiency in
the organizations that participated in this study. As discussed in section 4.2.3,
productivity should be measured as the number of units of output produced divided by the
number of units of input. In this study, the definition is firrther restricted to represent a
ratio of outputszinputs at a corporate level. Segmented data, such as volume of a specific
activity, are defined as measures of emciency.

To collect emciency and productivity data, the contact person from each firm
completed an introductory questionnaire (Appendix D). This questionnaire asked for
information about each firm’s sales volumes, the number of invoices they sent, and the
number of checks written in recent years. In addition, the contact person was asked to
provide the number of units of product produced, and the unit of measure used by the
firm. The expectation was that these figures could be used to measure eficiency and
productivity, respectively. Table 13 shows summary demographic statistics that were

collected from the firms in this sample.

95

Table 13: Demographic Statistics

Demographics Standard
_Mean Min M M Count
Operational Statistics
Customer Orders 11,000 1,000 28,000 9,340 8
Invoices 19,800 1,800 56,900 17,620 8
Purchase Orders 3,400 500 6,000 2,070 6
Checks Written 7,300 2,200 17,200 5,630 8
Employee Statistics
Number of Employees 190 40 450 130 8
Number of White Collar 40 3 170 50 8
Employees
Number of A/P Employees 1 1 3 l 8
Number of MR Employees 1 1 3 1 8
Number of Customer Service 7 2 20 8
Employees
Number of Employee Hours 335,700 62,000 950,300 316,390 7
Number of White Collar 96,400 7,800 397,800 129,260 8
Hours
Accounting & Finance Wages $181,400 $40,000 $500,000 $154,880 7
Customer Service Wages $147,300 $43,300 $233,500 $86,430 6
Factory Wages $4,241,400 $200,000 $15,156,700 $5,212,410 7
Total A&G Wages $1,356,500 $120,000 $5,273,000 $1,944,170 6
M18 Expenses
Hardware/Software $83,800 $4,850 $382,000 $134,700 7
Payroll $41,300 $0 $194,000 $75,810 6
Supplies $72,600 $1,000 $417,400 $169,000 6
Other $35,700 $3,500 $160,000 $69,470 5
Total $206,900 $14,800 $1,153,400 $418,230 7
5.2.2.4.1 Productivity Improvements

Although all of the firms in this sample are in the paper industry, their production
measures varied considerably. Some firms were unable to identify a unit of measure or a

production quantity that they felt would provide any insight into their organization. These

96

companies produce very customized products; in effect, each production run is unique,
and the number of units produced on one run cannot be compared to those of any other
run.

In most other cases, the measure used by the firm was selected to help them
manage their operations, and to align incentives with corporate goals. For example, one
firm measured production by the number of hours that the equipment ran. While this may
seem to be a poor measure of productivity, their reasons behind it were rational. This
company also produces customized products, so comparing the physical units is not
meaningful. In addition, they have invested in several expensive, customized pieces of
equipment that are currently running below 75% of capacity. At this time, their main
strategy to improve profitability is to generate more customer orders and to use the
additional capacity available on the existing equipment. Because they have very accurate
machine usage statistics, this appears to be a good way for them to measure the
production volumes for their organization.

For this study, the number of units of output will have to be measured by the sales
dollars of the firm. Although this is frequently cited as a problem with productivity work,
there is no other common measure of output available to compare the eight firms. In
addition, this is an output measure used by at least one firm whose manager was unable to
identify a single unit of output for their firm. Instead, he calculated their units of
production by dividing sales by the average unit cost for the goods they produced. By
using this measure, they were able to provide a production figure to managers even

though they produced customized products in small batch sizes.

97

Several measures of inputs are used to calculate productivity. First, the number of
employee hours was provided by the firms. In addition, the number of white collar hours
was also available from the firms. Finally, the amount of wage expense is another measure
of the inputs into the production process.

To determine if firms that used systems that were more similar to REA systems
were more productive, correlations between the productivity figures and the ASC score
were calculated. As shown in Table 14, when the complete sample is included in the
analysis, none of the productivity measures are significantly correlated with ASC.
However, when a scatter plot of the data was analyzed for outliers, one organization
appeared to be driving the results. This company has the least complex operating
environment and has the lowest ASC score. It was determined that this company may be
significantly difl’erent from the others, so it was removed fi'om the population. The
correlations were recalculated (Table 15). In this case ASC is positively correlated with
the number of employee hours incurred in the previous year. However, there is a negative
correlation between the ASC score and the productivity figures that are calculated using
wage expense figures rather than hours. Therefore, if productivity is calculated using units
of inputs, the data is consistent with the hypothesis that systems more like REA systems
provide productivity benefits. However, if costs of the inputs are factored into the
equation, systems that are more similar to traditional systems appear to lead to
improvements in productivity. It appears that the companies that use more sophisticated

computer systems also pay their employees higher salaries.

98

Table 14: Correlations between Productivity Measures and ASC Score

 

 

ASC Sales/ Sales/ Sales/Total Sales/
Emp Hour WC Hour Wages A&G$
ASC 1
Sales/Emp Hour 0.074 1
Sales/WC Hour 0.042 0.609 "‘ l
Sales/Total Wages -O.299 0.1 16 -0.155 1
Sales/A&G$ -0.099 -0. 152 -0. 186 0.932 1

 

* Using t-test of Pearson Product Correlation, p < 0.05

Table 15: Correlations between Productivity Measures and

 

 

ASC Score Removing Outlier
ASC Sales/limp Sales/WC Sales/Total Sales/
Hour Hour Wages A&G$

ASC 1.000
Sales/Emp Hour 0.556 "‘ 1.000
Sales/WC Hour 0.046 0.577 " 1.000
Sales/'1' otal Wages -0.520 0.239 -0. 159 1.000
Sales/MOS -0.443 0.056 -0.208 0.948 " 1.000

 

" Using t-test of Pearson Product Correlation, p < 0.10

** Using t-test of Pearson Product Correlation, p < 0.05
5.2.2.4.2 Efficiency Improvements

Because it is dificult to isolate a computer system’s influence on overall measures
such as productivity statistics, it is often important to analyze more disaggregate statistics
such as eficiency measures. Using this analysis, it may be possible to identify specific
efi‘ects of the system, and to explain how the system afi‘ects the organization (Sethi and
King 1994). There are several measures of emciency that can be used to compare the
firms in this study. First, operational statistics provided by the firms can be used to

calculate the following eficiency measures:

is t]
res1

efli

are

rep
soc

waj

prc

of

in»

of

fro

En

99

Customer Orders/ii of Customer Service Employees
Customer Invoices/# of Customer Service Employees
Customer Invoices/# of Accounts Receivable Employees
Checks Written/# of Accounts Payable Employees

Another measure that may proxy for efiiciencies arising fiom the computer system
is the User Satisfaction (US) index. As discussed, this index is compiled from user
responses to questions about the computer system’s efi‘ect on user efficiency and
effectiveness. If this measure is correlated with the ASC, then more sophisticated systems
are correlated at least with user’s perceptions of their efficiencies.

Correlations between these measures and the finn’s ASC score are calculated and
reported in Table 16. This table shows that the only significant correlation between ASC
score and an efiiciency measure is with Checks per A/P Clerk. It is possible that the three-
way match process has been automated (or reengineered) in the firms with more
sophisticated systems. Therefore, the firm’s Accounts Payable department is able to
process more vendor invoices with fewer clerks.

The correlation between Customer Orders per Customer Service Employee is
insignificant when all of the firms are included in the analysis. However, upon inspection
of the data, it appears that one firm may have had an error in the number of orders and
invoices they reported. They wrote that they received fewer than 2,000 customer orders
in 1994, but prepared over 50,000 customer invoices. It seems unlikely, given the nature
of their business, that they produce 25 invoices fiom each order. If this firm is removed
from the analysis, the correlation between Customer Orders per Customer Service

Employee and ASC score is 0.58. While insignificant, this correlation is in the positive

dir

all

It

p12

an;

Vis

100

direction and provides some evidence that the other firms appear to be enjoying an

efiiciency improvement with the introduction of more sophisticated systems.

Table 16: Correlations between ASC Scores and Measures of Efficiency

 

 

 

ASC C 0/ Inv/ Inv/ Chk/ UserSat
#CSEmp #CSEmp AR AP
ASC l
CO/#CSEmp -0.018 1
Inv/#CSEmp -0.061 0.709 * l
Inv/AR 0.251 -0.108 0.296 1
Chk/AP 0.512 "”" 0.241 -0.051 0.188 1
UserSat -0.249 -0.547 " -0.256 0.227 -0.489 1
Where:

ASC = Accounting Systems Characteristics score

COMCSEmp = Customer Orders per Customer Service Employee
luv/#CSEmp = Customer Invoices per Customer Service Employee
Inv/AR = Customer Invoices per AIR Clerk

Chit/AP - Number of Checks Written per A/P Clerk

UserSat = Summary measure of user satisfaction

‘ Usinga t-test ofthe Pearson Productcorrelation, p <0.05
“ Usingat-testofthePearsonProductcorrclation,p<0.lO

5.3 Discussion of Results

This chapter provides several measures of the benefits provided from information

systems, focusing on incremental benefits fi'om systems more like REA accounting

information systems than more traditional systems. By examining eight firms in detail, it

was possible to become familiar with their operational environments and the demands

placed upon their systems. In addition, statistical analysis was performed to provide

analysis of three hypotheses that had been developed earlier in this dissertation.

The qualitative analysis relied heavily on information gathered during the site

visits. In all cases, the users throughout the organization were interviewed. They were

It i
chi
we:
53’5'
efii
CV11

Spe

whi

ber

to r
Strc
that
thei

do]

Sl’St

101

asked to evaluate their current system and to make recommendations for enhancements.
It was through these discussions that the researcher was able to identify key systems
characteristics that would provide solutions for the problems that the firms in this study
were experiencing. As such, the key characteristics were often the enhancements that
systems users recommended, believing that these changes would improve either their
efficiency or efi‘ectiveness. This analysis and the list of key characteristics provides
evidence that not all systems provide benefits to their users; rather systems that have
specific characteristics will be more beneficial to organizations.

The field study methodology also allowed the researcher to identify situations in
which systems characteristics and organizational characteristics would interact and
influence the amount of benefits that the system could provide. Again, these insights can
be used to identify organizations that may benefit fi'om more advanced systems and those
that will not require the expenditures.

These findings can be used to explain some of the interesting quantitative results.
The Bakos and Treacy model was modified so only two difl‘erent strategies were predicted
to result in competitive advantage: Administrative Strategies and Interorganizational
Strategies. The data from the CA survey supported the hypothesis that executives of firms
that used their systems for either of these strategies believed that their systems provided
their firm with competitive advantage. Therefore, there is evidence that these executives
do perceive their system as providing competitive advantage.

However, when the model was expanded to include the accounting information

system, the results showed that systems more like REA were being used to improve the

102

firrn’s administrative efiiciencies, but they were not being used to implement
interorganizational strategies. This is an important finding as it was the interorganizational
strategies that were the main cause of competitive advantage from information

technology.

These puzzling results may be explained through insights provided during the site
visits and the interviews with the executives. One explanation is that the executives who
had more sophisticated systems were also more sophisticated in their knowledge of
information systems. It was common for these executives to understand the current
technology available, and to be critical of their systems that they believed to be out of
date. Therefore, these executives may have perceived their systems as providing less
competitive advantage than they desired, and their CA scores would be lower.

These results may also arise because the smaller firms are the ones that have been
forced to implement systems to meet their customers needs. For example, one firm had a
personal computer devoted to EDI processing of large customer orders. While they
believed that this system provided them some benefits, it would have had a negative
impact on their ASC score since this personal computer is not integrated with the firm’s
main system. As a result, several redundant processes were necessary to record the orders
and to distribute the electronic customer invoices. In addition, several spreadsheets were
maintained to capture the information in the orders.

A final explanation is that the bulk of the energy in information systems
advancement has been to automate the administrative firnctions such as accounts payable

and order processing (improvements in both were correlated with higher ASC scores).

pro
the
5UP
the

yet

53'5
trai
POE
neg

ins?

lee
inn
org
adc

tee

103

Practitioners have not yet devoted the time necessary to produce systems that meet the
other user’s needs. In this sample, most of the major systems complaints centered on the
production process. In many cases, the users were unhappy with systems that could only
perform batch processing of production data or with systems that did not support the
production environment at all. \Vrthout meeting production needs, it would be difficult for
the firms to respond quickly to customer needs or to improve their power over their
suppliers. As a result, the path coefiicients would accurately reflect the situation within
the firms in this sample: their systems are solving some management problems, but have
yet to tackle the interorganizational links within the value chain.

The quantitative productivity and efficiency provide some additional evidence that
systems more like REA accounting information systems provide more benefits than more
traditional systems. The amount of sales per employee and checks per A/P clerk are .
positively correlated with the firm’s ASC score. However, sales per labor dollar is
negatively correlated with the ASC score, and the remaining efiiciency variances are
insignificantly related to ASC score.

Overall, the research approach used in this study has provided detailed insights into
companies in the pulp and paper industries. The qualitative and quantitative analysis
techniques have provided evidence of benefits fiom systems more like REA systems. The
interviews with a wide variety of employees provide a more complete view of each
organization and its system than a survey or archival study would have gathered. In
addition, the quantitative results support the hypotheses that firms use information
technology to gain competitive advantage and improvements in productivity and

efficiency.

Chapter 6 - Contributions and Suggestions for Future Research

6.1 Contributions

This dissertation makes four major contributions to the accounting information
systems literature. First, it presents a questionnaire that can be used to evaluate systems
along the continuum between traditional and REA accounting information systems.
Second, it uses qualitative information collected from eight organizations to develop a
more precise operational definition of REA accounting information systems. Third, it
presents a survey that captures manager’s perceptions of competitive advantage and
results that show accounting information systems as providing such an advantage. This is
especially true if they are able to assist management in implementing interoroganizational
strategies. Finally, evidence is provided that the type of accounting information system
used will influence an organization’s productivity and eficiency. The following sections
discuss each of these in more detail.
6.1.1 Accounting Systems Characteristics Metric

Except for Weber (1986) and Andros et al. (1992), the research in REA
accounting information systems has been theoretical and normative. This study is,
therefore, one of the first to evaluate information systems and provide evidence about the
benefits of more sophisticated systems. One of the major dimculties in attempting this
study was that there was no operational definition of REA accounting information
systems. Weber (1986) had focused only on the revenue cycle and he had studied vendor
documentation to gain an in-depth understanding of the data storage structure of these

firms. The goal of this study was to understand how the systems were being used in live

104

105

firms, and how organizational characteristics influenced the manner in which the systems
were used. Therefore, it was not possible to focus on only the data.

The first step in this project was to develop the Accounting System Characteristics
questionnaire. This questionnaire was developed to gather high level information about
each system’s data, how the data were processed, and how the users were able to firnction
with the system. The questions included in the survey were designed to capture qualities
of REA and traditional systems, so that the overall score an organization received would
place them along the continuum between the two poles. Through pilot testing and use in
this study, it appears this metric successfirlly ranks the systems evaluated. Therefore, it
could be used in firture research as a method of gathering specific information about
organizations and their accounting information systems.

This metric also may provide benefits to practitioners. The accounting firm that
participated in the pilot study is very interested in using the metric to gain a thorough
understanding of potential client’s systems quickly. In addition, the companies that
participated in this study perceived their completed study materials (often 25-30 pages
long) as valuable. They believed that the list of potential enhancements could be used
management to prioritize new systems requirements. They felt that the descriptions of the
process employees performed were valuable if consultants were to come to the
organizations. Management planned on having the consultants read the ASC questionnaire
before they performed any interviews with the hope of minimizing the consultant’s time

and, therefore, fees.

106

6.1.2 Operational Definition of REA Accounting Systems

Although the ASC questionnaire was developed using the theoretical tenets of
REA, the accounting literature has not included a detailed definition of how to
operationalize an REA system. However, by using a consistent questionnaire at each of
the organizations, this study presents results of comparing and contrasting the systems and
their efi‘ects on the organizations. Using this information, it was possible to identify the
key characteristics that move a system along the continuum between traditional and REA
systems.

This definition could be used to categorize prior research of systems benefits by
identifying which systems characteristics have provided benefits to firms. For example,
Mia and Chenall (1994) studied the relationship between manager performance and the
use of broad management information systems (MIS). Their definition of a broad MIS is
one that supports several sites and provides non-financial information about several key
business processes, and the data are not aggregated. These systems also provided timely
information and supported management decision models. These systems would be closer
to REA systems than to traditional accounting because they support the key processes,
they maintain detailed histories, and managers are able to retrieve information to meet
their needs. To measure manager performance, they asked the participants’ supervisors to
rate their performance. Analysis of five firms in difi‘erent industries that used broad MIS
demonstrated that system use was correlated with performance and that the relationship
was much stronger for the executives in marketing roles than for those in production.

Assuming that the supervisors gave higher ratings to managers that performed their

107

firnctions either more efficiently or more effectively, this study provides evidence that
systems more similar to REA systems than traditional ones were able to provide
improvements in internal efficiencies, especially to the administrative staff. These
emciencies may result in competitive advantage for a firm or increased productivity.
Therefore, the results of the Mia and Chenall study are very similar to those fiom this

dissertation.

6.1.3 Competitive Advantage

This study reports the development of a survey that can be used to measure how
management’s perceptions about their systems and the efl‘ect these systems have on the
firm’s competitive position. Specifically, the Bakos and Treacy (1986) theoretical model
of competitive advantage was operationalized as a survey, and the survey was distributed
to several executives in each organization to collect a well-rounded view of the systems.

Through confirmatory factor analysis, it was determined that the managers in this
sample were using two distinct strategies to gain benefits fiom their systems: they used
their systems to improve operational efiiciency and they used them to improve
interorganizational relationships with customers and suppliers. Using these insights, it was
determined that the interorganizational strategies that were supported by the information
systems were more strongly linked with competitive advantage from systems. In addition,
the systems that were more like REA systems were used to improve administrative

eficiencies but not interorganizational eficiencies.

108

6.1.4 Efficiency and Productivity

The quantitative analysis provided evidence that the more advanced systems did
provide benefits to firms using them. Overall productivity improved if it is measured in
terms of the sales generated per labor hour. However, other productivity measures, such
as the number of units of output by the number of units of input, were not possible to
calculate. Even though the firms in this study were in the same industry, they produced a
wide range of products, and often produced a different custom good with each production
run. Therefore, summing the number of units produced would not provide a meaningful
statistic.

Because of dificulties in measuring productivity and identifying the link between
the information systems and productivity, efficiency measures were also calculated. By
comparing them with the type of system the organization used, it was determined that
eficiency improved for accounts payable clerks and customer service clerks in
organizations that used systems more like REA systems. Other measures of efficiency,

however, were not found to be significantly affected by the type of system used.

6.2 Future Research

There are many opportunities to continue this line of research. This study is one of
the first empirical analyses of REA systems, and, although it provides some evidence of
the efi‘ects these systems have on organizations, it has several weaknesses that must be
addressed in future research. The following section identifies some of these weaknesses

and makes recommendations for how they may be overcome in the firture.

109

First, it would be beneficial to perform a longitudinal study to determine how a
new REA system will affect an organization. This research design would control for the
organizational characteristics that make comparing the benefits fiom systems across
organizations. In a longitudinal study, many of these characteristics would remain
constant. Therefore, productivity and efiiciency figures could be calculated for operations
under the old and new systems, and the results compared. Then the changes could be
attributed to the new system.

Second, firture research should involve directed REA designs and implementations.
Unfortunately, none of the firms in this sample have participated in such a design.
Therefore, it is dificult to make claims about benefits that would accrue to those who
implemented systems with the most advanced REA characteristics (directed
implementation and no journal entries). If these development approaches were included in
future research, the costs and benefits of such implementations could be quantified and
verified.

Third, organizational characteristics must be included in firture studies. For
example, complexity should be considered in firture studies that evaluate the benefits of
REA accounting systems versus traditional accounting systems. Unless the organizations
in the study have complex environments, the benefits arising fiom sophisticated REA
systems may not ofl‘set the costs necessary to develop and maintain them.

Finally, firture research should be performed on larger samples. While it was
important to gather detailed insights fiom organizations for this study, this approach
limited the number of companies that could participate. However, now that the key

systems and organizational characteristics have been identified, it may be possible to

110

develop a questionnaire that could be sent to a large number of companies. Results of

such a study may be more generalizable to the economy as a whole.

6.3 Conclusions

This dissertation describes a field study experiment that studied the interaction of
organizations and their accounting information systems. It presents the first empirical
analysis of REA accounting information systems, and includes results that show both
qualitative and quantitative analyses of the benefits that arise from them. By nature of the
research methodology selected to address these questions, the study provides a detailed
operational definition of REA systems, and it challenges future researchers to use this

definition in firture systems development projects and large sample studies.

LIST OF REFERENCES

LIST OF REFERENCES

Andros, D., J. Cherrington, and E. Denna. “Reengineer Your Accounting, the IBM Way”
Ihe Financial Executive (July/August 1992) pp. 28-31.

Bakos, J.Y. and M. E. Treacy. ”Information Technology and Corporate Strategy: A
Research Perspective.” MS Quarterly 10:2 (June 1986) pp. 107-119.

Belcher, L. and H. Watson. “Assessing the Value of Conoco’s Expert Information
System.” MIS Quarterly (September 1993) pp. 239-253.

Brynjolfsson, E. “The Productivity Paradox of Information Technology: Review and
Assessment.” Communications of the ACID! (December 1993) pp. 66-77.

Brynjolfsson, E. and L. Hitt. “New Evidence on the Returns to Information Systems.”
Working paper, MIT Sloan School (1993).

Cherrington, 1.0., W.E. McCarthy, D.P. Andros, R Roth, and EL. Denna. 1993. Event-
Driven Business Solutions: Implementation Experiences and Issues. Proceedings
of the Fourteenth International Conference on Information Systems. Orlando.

p. 394.

Davidson, W.H. “Beyond Re-engineering: The Three Phases of Business
Transformation.” IBM Systems Journal (1993) pp. 65-79.

Davenport, T. Process Innovation: Reengineering Work through Information
Technology. Harvard Business School Press, Boston, MA, 1993.

David, J ., “Management’s Perceptions of Competitive Information Systems: A Survey of
Small Business Executives.” Working paper, Michigan State University, 1994.

Dearing, B. “The Strategic Benefits of EDI.” The Journal of Business Strategy
(January/February 1990).

Denna, E., J. Cherrington, D. Andros, and A Hollander. Events-Driven Business
Solutions: T oday's Revolution in Technology, Business One Irwin, IL, 1993.

Denna, E. and WE. McCarthy. “An Events Accounting Foundation for DSS.” in C.W.
Holsapple and AB. Whinston (eds) Decision Support Systems: Theory and
Applications (Proceedings of the NATO Advanced Study Institute on Decision
Support Systems, Maratea, Italy), Springer-Verlag Publishing Company (1987)
pp. 23 9-263.

, J. Jasperson, K. Fong, and D. Middleman. “Modeling Business Processes.”
Journal of Information Systems (forthcoming).

Fisher, J. “The New Finance.” Journal of Accountancy (August 1994) pp. 73-80.

111

112

Gal, G. and WE. McCarthy. (1983) “Declarative and Procedural Features of a
CODASYL Accounting System,” in Entity-Relationship Approach to Information
Modeling andAnalysis, P. Chen (ed.), North-Holland, pp. 197-213.

and . “Operation of a Relational Accounting System.” Advances in
Accounting 3 (1986) pp. 83-112.

 

Geerts, G. and WE. McCarthy. “The Economic and Strategic Structure of REA
Accounting Systems.” Working paper, Michigan State University, 1995.

Glazer, R. “Measuring the Value of Information: The Inforrnation-Intensive
Organization.” IBM Systems Journal (1993) pp. 99-110.

Gosse, D.I. “Cost Accounting’s Role in Computer Integrated Manufacturing An Empirical
Field Study.” Journal of Management Accounting Research (Fall 1993) pp. 159-
1 79.

Grabski, S.V. and RJ. Marsh. 1995. “Integrating Accounting and Advanced
Manufacturing Information Systems: An ABC and REA Approach.” Journal of
Information Systems (forthcoming).

Hammer, M. “Reengineering Work: Don’t Automate, Obliterate.” Harvard Business
Review. (July-August 1990) pp. 104-112.

Hopwood, A “The Archaeology of Accounting Systems.” Accounting, Organizations and
Society (1987) pp. 207-234.

Hufl; S.L. and ES. Beattie. "Strategic Versus Competitive Information Systems,”
Business Quarterly. (Winter 1985) pp. 97-102.

Ijiri, Y. Theory of Accounting Measurement, American Accounting Association (1975).

Johnston, HR and S. Carrico. “Developing Capabilities to Use Information
Strategically.” MS Quarterly (March 1988) pp. 37-48.

Johnston, HR. and M. Vitale. “Creating Competitive Advantage with Interorganizational
Information Systems.” MS Quarterly (June 1988) pp. 153-165.

Kee, R “Data Processing Technology and Accounting: A Historical Perspective.” The
Accounting Historians Journal (December 1993) pp. 187-216.

Kelly, M. “Productivity and Information Technology: The Elusive Connection.”
Management Science 40:11 (November 1994) pp. 1406-1425.

LaPlante, A “Turning corporate data into profitability.” Info World (October 18, 1993):
p. 63.

113

Mia, L. and RH. Chenall. “The Usefirlness of Management Accounting Systems,
Functional Difi’erentiation and Managerial Effectiveness.” Accounting,
Organizations and Society (1994) pp. 1-13.

McCarthy, W.E. “An Entity-Relationship View of Accounting Models.” The Accounting
Review (October 1979) pp. 667-686.

. ”The REA Accounting Model: A Generalized Framework for Accounting
Systems in a Shared Data Environment.” The Accounting Review (July 1982) pp.
554-77.

Moore, G. and I. Benbasat. “Development of an Instrument to Measure the Perceptions of
Adopting an Information Technology Innovation.” Information Systems Research
2:3 (September 1991) pp. 192-222.

Northnrp, G. ”MIS and Competition. " 3X/400 Information Management (January 1991)
pp. 23 -24.

Panko, R “Is Ofice Productivity Stagnant?” MS Quarterly (June 1991) pp. 190-203.
Porter, M.E. Competitive Advantage, The Free Press, New York NY, 1985.

Porter, ME. and V.E. Millar “How information gives you competitive advantage.”
Harvard Business Review (July-August 1985) pp. 149-160.

Raymond, L. “Organizational Characteristics and MIS Success in the Context of Small
Business.” MS Quarterly (March 1985) pp. 37-52.

Reid, RD. and M. Sandler. “The Use of Technology to Improve Service Quality.” The
Cornelle Quarterly (June 1992) pp. 68-72.

Revaz, E. “Corporate and Environment Transformations: An Event Based Approach of
Financial and Accounting Information Management.” Workshop on Accounting in
Europe, Edinburgh UK (1993) pp. 1-39.

Roach, S. “Services Under Siege-The Restructuring Imperative.” The Harvard Business
Review (September-October 1991) pp. 82-92.

Seddon, P. and M.Y. Kiew. “A Partial Test and Development of DeLone and McLean’s
Model of IS Success.” working paper, The University of Melbourne, June 1994.

Seddon, RB. and KS. Yip. “An Empirical Evaluation of User Information Satisfaction
(UIS) Measures for Use with General Ledger Accounting Software.” Journal of
Information Systems (1992) pp. 75-92.

Senn, J. "The Myths of Strategic Systems: What Defines True Competitive Advantage?"
Information Systems Management (Summer 1992) pp. 7-12.

114

Sethi, V. and W.R. King. “Development of Measures to Assess the Extent to Which an
Information Technology Application Provides Competitive Advantage.”
Management Science 40: 12 (December 1994) pp. 1601-1627.

Standard and Poor ’s Register of Corporations, Directors, and Executives, Standard and
Poor’s, NewYork, NY (1993).

Technology and Labor in Pulp, Paperboard and Selected Converting Industries. US
Department of Labor, Bureau of Labor Statistics (June 1994): Bulletin 2443.

Trewin, J. “The Need and Opportunity for F ield-Based Research in Accounting
Information Systems.” Journal of Information Systems (P all 1998) pp. 104-118.

Vincent, D. “How Eight Firms Transformed—With Technology” Financial Executive
(March/April 1993) pp. 52-58.

\Vrseman, C. and I. MacMillan. "Creating Competitive Weapons from Information
Systems." Journal of Business Strategy (Fall 1984) pp. 42-29.

Yin, RK., Case Study Research: Design and Methods (Sage Publications, 1984).
Yu, SC. (1976). The Structure of Accounting Theory. The University Press of Florida.

APPENDICES

APPENDIX A

APPENDIX A: Accounting Systems Characteristics Questionnaire

 

 

 

 

 

 

l 16
REA Evaluation Form

System Overview

1. Use the appropriate, traditional E-R diagram to begin evaluating the system.

a.

For each event (entities with names written in bold, italic print):
1. If not supported by the system, draw an 'X' through the event.
ii. If it is supported by the system,
* Write the length of time that the records are stored, such as EOD
(end of day), EO2Y (end of 2 years), Until Complete Payment.
"‘ Write "GL" after the time if the results of the event are
automatically posted to the general ledger.
Identify any business-specific events that are supported by the system
i. Draw them on the diagram
ii. Identify characteristics as in a.
For each resource and agent
i. Identify any missing from the diagram, and add them
ii. Draw an 'X‘ through any of the traditional ones not supported by the

system.
Relationships
i. Draw an ”X” through any relationships that are neither eggph'gifly or
Mimplemented in the system.

ii. Draw a circle around the relationship if it is implicit 1y implemented (i.e. it
could be 'reconstructed‘ procedurally).

AS. Scoring:

A.
B.

C.

Add 5 points for each traditional event supported by the system.

Add 5 points for each new event whose data is stored until EOD; 7 points if
stored until EOM, and 10 points if stored longer.

Subtract 5 points for each traditional event with more than 1% of the
business ' transactions or more than 500 transactions not supported by the
system.

117

Additional System Questions:

1.

How did the organization decide how long to store the events? (+10 if decision
included in Events-Driven design)

If you have purchased applications programs, list them here:

 

 

What programming language(s) is/are used to write the applications?

 

How are the explicit relationships implemented?

Pointers

Data Base

Repeating groups
Physically contiguous

If someone wanted information fi'om every cycle, how many terminals would be
needed? (-5 for each over 1)

How many times would they have to log in? (-2 for each over 1)

Describe your hardware environment (including any client/server configurations).

Page Points

1.

1 18
Accounting/General Ledger

How are your financial statements generated? (Check appropriate)

Chart of Accounts used to prepare General Ledger and Statements (-5 points)

No Chart of Accounts. Separate view of the data (+100 points. This system

2.

automatically counts as an Events-Driven system, but continue with survey to
record characteristics and implementation decisions made.)

List the manual journal entries entered into the system.

 

 

 

 

(-10 points for each event)

Is information fiom the Chart of Accounts down-loaded to personal computers where it is
manipulated to generate additional reports used by management?

Yes (.10 points) No (0 points)

Who has requested the downloaded data?

For what purpose?

Are G/L account balances transferred from one system to another General Ledger system
in the organization? (either manually or automatically)

Yes (~10 points) No (0 points)

How many days after the last day of the months usually pass before the final financial
statements can be prepared?

Days (-1 point per day)
How many codes are embedded in each g/l account number? For example, the account
number 01-573-01 that represents division 01, sales manager 5, salesperson 73, sales (01)
has 4 codes.

codes (-2 points per code)
Page Points

119

M18 Characteristics
Is the current system the result of a "re-engineering” exercise?
__ Yes (+10) _ No (-10 points)
If so, describe 2 major changes that occurred as a result:

8.

b.
(Score 10 points for each change that shows evaluation of business events)

Are users outside of MS able to generate ad hoc queries of data in the system (excluding
311 queries)?

__ Yes (+10) _ No (0 points)
If so, how are these queries supported?
__ Query program within the system
_ Add-on query support (such as SQL)
_ Download data and manipulate on personal computers
__ Other:
How many users use this facility at least once a month?
(+2 for each with a maximum of 10 points)
Is there a steering committee to guide the MIS department?
Yes No

If so, list each non-MIS and non-Accounting active member of the committee:

 

 

 

 

 

 

(Score 2 points for each functional area represented on the steering committee)

Page Points

“hat

(Sco
rddi
for 1

Are

Give

4.

120

What are the 5 most recent systems projects that have been completed?

 

 

 

 

 

(Score +10 points for each that supports an already supported event; +20 for
adding support for a new event: -10 for supporting the Chart of Accounts; 0 points
for fixing program bugs)

Are customers able to get information from your system (either on-line or in reports)?
Yes (+10) No

Give examples:

Page Points

Interviewed: Date:

121
Functional Department: XXX Staff or Manager

 

Prmess:

1.

What are the three most important sources of information used to guide this department?
Are they available fiom your system?

 

(Score +5 for each that is generated by the system but is not a result of the Chart of
Accounts; -5 for each not generated by the system that could have been included in
an Events Based system)

Does the system help your department be more productive?

_ Yes __ No

If so, describe 3 ways it does:

a.

b.

c.

(Score +5 for each productivity tool that is a result of an events-driven design)
What are the two major enhancements you'd like to see made to the system?

a.

b.

How often do you use an on-line inquiry to get information about your department?

Never (-5) Daily (+5)

Weeldy (+2) > Once per Day (+10)

Page Points

122

Summary Sheet

Implemented applications (Chart)
Other system characteristics (p. 1)
System Total
Organization characteristics - Accounting
Organization characteristics - MIS
Purchasing characteristics
Production characteristics
Sales characteristics
Distribution characteristics
Organization Total
Grand Total

 

 

 

 

 

 

 

 

 

 

 

APPENDIX B

APPENDIX B: Competitive Advantage and User Satisfaction Questionnaire

While computer systems can be used for processing daily transactions, they can play other
roles in organizations. This survey attempts to assess the ways in which your organization uses
your information system as a competitive tool.

Before you begin to answer these questions, please take a minute to think about how your

computer systems have significantly altered either your organization's business, or that of your
customers or suppliers.

123

124
1. Please indicate you much you agree with each of the statements on the following page using the
following 1 to 7 scale:

1 = Strongly Agree 2 = Agree 3 = Agree Somewhat
4 = Neutral
5 = Disagree Somewhat 6 = Disagree 7 = Strongly Disagree

Because of our computer system, customers would incur a cost to switch to another supplier.

When we develop or purchase new computer applications, we think about how they will help us
compete.

Our computer system helps us control our production costs.
Our computer system has reduced the time it takes our administrative stafi‘ to do their jobs.

Our computer system enables us to modify our products so customers value them differently than
our competitors’ products.

Because of our computer system, we have been able to sell different products than we would be
able to without it.

We use our computer system to communicate with our suppliers.

Because our customers rely on unique information fiom our system, they would have to increase
their costs if they were to purchase products fiom another supplier.

Our computer system helps us control our administrative costs.
We develop new applications for our computer.

Our computer system has made it possible to work with more vendors without incurring extra
costs.

Our customers use a unique, proprietary interface to log into our computer to place orders or get
information.

We are able to produce products more quickly because of our computer system.
The computer system has helped us produce our products at a lower cost.

Most of the applications we have developed or purchased that make us more efficient are used by
everyone in the industry.

Our computer system has helped us earn a favorable competitive position in our industry.
Our customers use information from our computer to improve their profitability.

We rely on our supplier’s computer system so we are not able to switch to another suppliers
without increasing some of our costs.

Our computer system has helped us shift our raw material components toward more generic,
commodity products.

Our computer system has not led to any efiiciency improvements for our administrative
departments.

Our computer provides our customers with information that makes them more emcient.

125
1. (Cont) Please indicate you much you agree with each of the statements on the following page using
the following 1 to 7 scale:

1 = Strongly Agree 2 = Agree 3 = Agree Somewhat
4 = Neutral
5 = Disagree Somewhat 6 = Disagree 7 = Strongly Disagree

Our computer system provides no competitive advantage.

Because of our computer system, our product is perceived to be difi‘erent than our competitors’
products.

Our computer system has no impact on our relative position in our industry.
We use our computer system to communicate with our customers.
We make very few changes to our computer system or its applications.

Because of our computer system, we do not rely on our suppliers for information about raw
materials.

We try to manage our information systems with the goal of being the first company to develop or
purchase new computer applications that provide competitive advantage.

Because of our computer system, we no longer have to purchase custom raw materials, but can
now use more generic materials.

Because our competitors can copy our system, we are not able to sustain any competitive
advantage from emciencies that our computer provides.

Applications we have purchased or developed to improve our eficiencies will probably be adopted
by other firms in our industry.

Because of our computer system, we have more flexibility in the vendors we can purchase
inventory and supplies from.

Our computer system has helped us become less reliant on any individual vendor.
Our computer system has no direct impact on our customers' businesses.

Because of our computer system, we are able to use less expensive raw materials in our production
process.

Because of our computer system, we can purchase inventory and supplies more efiiciently.

126

Please CIRCLE the number that best answers the following four questions.

How adequately do you feel your accounting information system (AIS) meets the information
processing needs of your area of responsibility?
adequate l 2 3 4 5 6 7 inadequate
How efiicient is your AIS?
eflicient 1 2 3 4 5 6 7 ineflicient
How efl'ective is your AIS?
efi'ective 1 2 3 4 5 6 7 ineflective

Overall, are you satisfied with your AIS?

dissatisfied 1 2 3 4 5 6 7 satisfied

List the three most important features of your organization that difi‘erentiate it from your
competitors and provide you with a competitive advantage:

A.

 

B.

 

C.

 

Estimate what percent of your competitive advantage arises fiom your information system.

%

 

Please provide your position in your organization:

 

APPENDIX C

127

APPENDIX C: User Satisfaction Questionnaire

This brief survey is designed to measure your overall impressions of your

Accounting Information System (AIS). After you complete it, we will discuss the
system and how it supports your activities in much more detail. Your results will be
kept confidential, so please answer these questions honestly.

Please CIRCLE the number that best answers the following four questions.

How adequately do you feel your accounting information system (AIS) meets the
information processing needs of your area of responsibility?

adequate 1 2 3 4 5 6 7 inadequate
How eficient is your AIS?

efi‘icient 1 2 3 4 5 6 7 inefl‘icient
How efi‘ective is your AIS?

eflective 1 2 3 4 5 6 7 ineflective
Overall, are you satisfied with your AIS?

dissatisfied 1 2 3 4 5 6 7 satisfied

APPENDIX D

SLAVE— I-AVuunwErhAbs‘Ih- bun-hmEm-sdrhk Um!~fi..finfi< UEUHUsfim can x-A-ZBLA.<

128

 

muse: 3:00-823 mo cox—Esz

 

 

9:5: 8.3—qu mo cons—=2

 

 

macho—men.“
83.5w 88380 mo 59:52

 

macho—mam M2 me .3832

 

$05.qu an}. ac con—=32

 

 

m8>o_aEm $86.23 .«0 cope—:2

 

macho—92m mo con—.52

 

meander—£5 8.3—qu

 

 

 

coat? 3.00:0 mo Lon—=32

 

 

3090 8223 m0 con—=52

 

 

8202: we con—=32

 

coo—ah 2090 38830 «c aonEsz

 

$53009.—

 

 

oooooea ea:

 

9:582 .uO «a:

 

 

cae—

 

a:

 

nae—

 

nae—

 

vmxe— GH>

 

E28305
558:0

 

“8:28: 38:82? wage—.8 2: 0235 .N

\ \ “Em 30> 33E A

me. « BEEF .553:qu .59» mean? .5 9.30.:— 03 32— 8 2.2.8.5 mite—.8 a... .3323 83..—

.Eom 52.5.5...— beséeam 35.5. 25.26 ”a amaze—a3.

 

129

as 8:-31:
3o: $8-23:

as: E .583 seem
E30225 88m Sumac:
38.9.80 gunman 562 «32

259 §Em 0:3.

2...; 5.3 «5.8... o. Eats ":28. scream. 9.5:; .8283: e. E... a... 5 as...

 

33:06 28 03853834 :33.
beacon

 

83am LoEofiau

 

 

005:5 v5 :c:=c8<

 

“”285qu =ciam

 

 

aoeo

 

3:85

 

=cH~am

 

o§3¢0mb§3w83

 

685qu mg

 

 

0:52: 82

 

£05 305

 

saw

 

 

558:0

 

 

ema—

 

 

aa—

 

nea—

 

mam—

 

a}: a?

 

”momentum 305cm wage—.8 05 0353 883 .n

APPENDIX E

130

as... 5.... 2.... 538.5
22.82%... .59. .u 2.9.26 5.3 9.82.. o. .53.... 96.8. 2:3. .5. . .8. a 8.5....

.272123
.m 2: o. 256.. 2.. 2.... 2.3... .5... “.an ...o .8 89%.. 2.9026 .2... 0.2.2.8 2.. 3.0383 26.. so» 02.0

... .ow . .2... 0...... 8.2.. .28 0938:. a 8.2 ...B 02...... 2.. $52.88 2.3.... .a :38... .3 Eu . .. .omgrmmm A23
:8 o. m. 2.. :28. o. .33 .2... 2. ... 2.. =8 e. 8... .8. 88... 2...... 82.. 3.8.9.8 .6 22.2.8 2.. Scan 2.0.32... >5 2:... :9. ..

3E. . 2&2. as... 5.... 3...... o. as . o. 2...... 92.998 322.. as... ..
6.3.2.2..

..2.. .6 0.2.8 0..» 2522...... 2.. .a... 8 8:... 5.? 2.83.8 :22... wfivem... .23 :25... magma 2.. 0.0.9.5... 88... .N
seas. .... .o .. 38......

oz... . >2. 2.. co 8.23.2. 2.. 88. c. 3... 2.53 . .5533. .36.. .098 2.. e. .32. o. 26.. .2. 0.. 8.23.2. .82 ..

H2..... a. .8... 2.8.6. o... 5.3 .. 82.58 o... 6.2.2.3 2.. oases one...
.23.... .8: 2.. a. w...=..o..§. 2. EB 25 2.... 83.2.2. 29.3.... 2:2»... 2.. 358.. «.02.» 2.52.0. 2....

«nu—632:3:—
aoeef? 22......7. 2.2...
58:232.: 5:32.23: 393.50 3.5950: .uheuaEeU

2.62.8 3.3.3... 22...... 3.3% 2.. 2.5.2....»

131

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

....8
.880 .2928... 8.8....
>889... 6.2.8.. 3.583.. E22583. ....... 98».
....8
22.2.. 85....
o_£>.os......=o8< 8.88.. ....... 21..
22.1.3. 85......
02.8.... =.....8< B22583. 28.. .
8.888..
83% 8.8.8.... 862m .82. v
8.3.355.
8.22.5.5
83m
8.888..
8.35,.
8.888..
88m 8.8.8.... 8.888.... .8 we
.52 .5. .9. .. 88.... 8mm 22...... 2......»2.

 

 

 

88...? 22......2 08....
88.2.8... 8.82.3.2: 8589

2..—88¢ 92.88... ”29......< 28.28

8.58.5: .. Nan. .

832......va "88.80

 

132

 

 

£ng

 

 

 

 

 

 

 

 

 

 

 

 

 

“.3822 .mSm aonznfima act—.355 _o>o_..m8m .38 on
max—«9%
o>u=ooxm :oaBEma .8955me o>u=ooxm .52— _
mama—«5‘
”“3808 baa 55395 noun—62m _o>o_.b3m 58 on
max—SE
o>u=ooxm nowosvoi gag—55 o>usooxm So: _
max—£2
cacao:— bflm 335m 38880 outflow 88830 _o>o_..m5m SE om
max—£2.
o>u=8xm mofiwugoiaz mofimhgoxa—Z o>u=ooxm Bo: g
”.Beuaasm ”53.2.5 wig wsfioégm .aa on
332$
998$ 9.6225 mfimaaobm 2533m— So: _
25h 2:3
3252 cam 25 3533::— £>§< 8:25 mania magnum

 

5:23: 221.52 2.23
58:22:43 .AoEazuflmh: ”38.80

o_=_§_om gaseoh u£m>_a=< magnum

Aethouaev "N .95

£59.80: ”.33ch

 

133

 

90.76.. .5 353053.. 337.3% 3.
mascaooo$oo§£m

.353:qu

83m

3:023...

3.3.32

€322.

9.36% 333.8..—

mmmbfla 2.36%
.3 333% E... marooswfiom
£33533. .3

 

 

 

 

 

 

«83 9:39.30 .3 333% 3.3.. m
93:. 0.5a.
83.—3:3.— bm>_3< 3.3.59 ”.53”.— ”5.330:

 

3.3.5: 233:2 23.:
5332.34: zeaazefimm: .3330

3.323m 958.3h Ema—3.2. magnum

.60....ho3n: “n .95

893.330: ".93....30

 

APPENDIX F

134

 

 

 

 

.....................

 

 

 

 

3.8.}. x. 3538: mm. H.
at. is D
Ali-9
839.... iv.
:3 as
[.053
.83»... .
as.
so
.10
33.8....
68°
i6
lllllllll I _
n .
35.3. 2 «828
:5

 

 

38.5 D3335 0.3x .23 2.33.5 <5— ».31330 n..— £2.33.

Ala—e

 

 

135

Summary Sheet - Ann Arbor

Implemented applications (Chart)
Other system characteristics (p. 1)
System Total

Organization characteristics - Accounting
Organization characteristics - MIS

Department Scores
Purchasing Executive
Estimating Stafl‘
Production Executive
Production Stafl’

Sales Executive

Marketing Executive
Customer Service - Custom
Customer Service - Gift wrap
Accounts Receivable
Accounts Payable

Average department characteristics

Organization Total
Grand Total

-14
-9

-65

10

-10
-20
-10
10
20
35
25

6.7

67
-7

136

 

Evin .280 8:83... $2 324,
$0.39.... 3.. 326 305.0 VA

8:39.. x. 3.9.5.82 _,I l.

 

3.5. 2.35 D

 

 

 

 

 

 

 

 

 

cits-e
iv. 83.—In
in 815.
81>
RIG .3
:5. It}...
i
1.0 .lo
582%
no— i
a; , .16 it...
lieu
.. I! 35.2
In:
_
.5. 53>
g

 

 

137

Summary Sheet - Bloomington

Implemented applications (Chart)
Other system characteristics (p. 1)
System Total

Organization characteristics - Accounting
Organization characteristics - MIS

Purchasing

Purchasing Staff

Accounts Receivable

Accounts Payable

Production

Production Stafl‘

Production: Engineering

Quality Assurance

Scheduling

Sales

Sales Stafl‘ (Order processing)

Distribution

Distribution Stafi‘
Organization Total

Grand Total

10
-14

-55
12

15
10

-15
15
3O
10
-5
20
35

77
73

 

 

 

 

 

 

 

 

 

 

 

 

139

Summary Sheet: Champaign

Implemented applications (Chart)
Other system characteristics (p. 1)
System Total
Organization characteristics - Accounting
Organization characteristics - MIS

Accounts Receivable

Accounts Payable

Purchasing Executive

Purchasing Stafl‘

Sales/Marketing: Eaton

Customer Service

Distribution

Production

Average department characteristics

Organization Total
Grand Total

75
-7
68
-57

25
20
15
25
—5
35
10
-10

14.38
151
219

E<l=5501§8m “Iii-..

v..:..........

EpisassueSalsaaaaz ..............
£523
"32850 98593850 0
5.2.82»
8395
833.
83....» will-.. _
_ E .
al.; " _ _ . 3195..—
uuuuuuuuu _
3.19 Agency
8395 i 2 _ m i
833. 33.:
as;

 

 

 

 

 

 

 

 

 

141

Summary Sheet: Columbus

Implemented applications (Chart)
Other system characteristics (p. 1)
System Total
Organization characteristics - Accounting
Organization characteristics - MIS

Estimating - boxes
Estimating - specialty
Accounts Receivable
Accounts Payable
Purchasing

Partner - Production
Partner - Sales
Distribution
Production
Production - Specialty
Customer Service
Customer Service

Average department characteristics
Organization Total
Grand Total

40

36
-12
-8

15
-15

10

-3
-IO
-10
-20
-10

-5

-3.58
-35.8
-20

 

 

 

 

143

Summary Sheet: East Lansing

Implemented applications (Chart)
Other system characteristics (p. 1)
System Total
Organization characteristics - Accounting
Organization characteristics - MIS

Accounting Executive

Billing

Aooormts Receivable

NP and Production Labor Tracking
Operations and MIS Executive
Sales and Marketing Executive
Purchasing, non roll materials
Customer Service
Production/Scheduling Manager
Distribution Stafi'

Continuous Improvement Manager

Receiving (EDI) Roll Stock

Average department characteristics
Organimtion Total
Grand Total

75
-2 1
54
-44
5O

25
30
25
20
20
10
-20
20

30
15
25

17.08
171
231

 

SE 5%? 5 £5

111.111.

llllll

 

IIIIIIL

 

 

Ill-all—

.lllllllllll. Ill.

 

 

 

 

145

Summary Sheet: Evanston

Implemented applications (Chart)
Other system characteristics (p. 1)
System Total

Organization characteristics - Accounting
Organization characteristics - MIS

Sales/Marketing Management (Div2)

Sales/Marketing Management (Divl)

Sales/Marketing Staff (Order Entry)

Production Manager

Production Supervisor (Div2)

Production Supervisor (Divl)

Distribution Manager (Shipping and
Receiving)

Accounts Payable

Invoicing

Div3 Management

Average department characteristics

Organization Total
Grand Total

15

15

-41
32

25

15
25
-5
15

10
25

11.5
81
96

 

 

 

 

 

 

 

 

 

147

Summary Sheet - Madison

Implemented applications (Chart) 30
Other system characteristics (p. 1) -7

System Total 23
Organization characteristics - Accounting -48
Organization characteristics - MIS 50
Purchasing 25
Purchasing Staff 5
Production 15
Production Staff 25
Sales 20
Sales Stafl‘ 5
Distribution 0
Distribution Staff 15

Organization Total 1 12

Grand Total m

 

 

PeaehTree. Excel
, EDI

 

 

 

 

 

 

149

Summary Sheet: West Lafayette

Implemented applications (Chart)
Other system characteristics (p. 1)
System Total
Organization characteristics - Accounting
Organization characteristics ~ MIS

Purchasing Executive
Production Manager
Sales Executive
Marketing Manager: Radio Marketing
Marketing Executive
Marketing

Specialty Sales
Order Processing
Customer Service
AIR and NP

Design Department

Average department characteristics
Organization Total
Grand Total

~21
~17
~32

- l O
~1 5
~1 5
~20

~5
~20
~20
~20
~20

~12.73
~127
~176

HICHIGRN STRTE UNIV. LIBRRRIES
lllll Will 1 ll 1| “1| |N| llll 1| 1 Ill ||||l 1|? WI UNI
31293014172591