DEVELOPMENT or A DINAIIIIC SIMULATION MODEL _ 5 -‘ ‘
FOR PLANNING PHYSICAL DISTRIBUTION. SYSTEMS! ‘ ’ I
THE FINANCIAL IMPLICATIONS OF WAREHOUSING ;

DECISIONS A ’

Ihesis for the Cegreeiof PILL). , f »
MICHIGAN STATE UNIVERSITY
MICHAEL LANCE LAWRENCE

   
     

 
 

LIBRARY

Ihdhjfiganlstau:
University

Uhhy’

 

This is to certify that the

thesis entitled _
Development of a Dynamic Simulation Model 'for Planning '
Physical Distribution Systems: The Financial Implica—
tions of Warehousing Decisions

presented by

Michael Lance Lawrence

has been accepted towards fulfillment
of the requirements for

Ph.D . degree in Business Administration

MW L Queue/Z
mam,

 

 

 

Date May 19, 1972

0-7639

 

 

 

 

 

 

 

ABSTRACT

DEVELOPMENT OF A DYNAMIC SIMULATION
MODEL FOR PLANNING PHYSICAL DISTRIBUTION
SYSTEMS: THE FINANCIAL IMPLICATIONS
OF WAREHOUSING DECISIONS

BY

Michael Lance Lawrence

This thesis is based on a dynamic computer
simulation of a large scale physical distribution system
develOped by a faculty and doctoral student research team
under a Michigan State University industrial research
grant. The project and model name is Long-Range Environ-
mental Planning Simulator (LREPS) and the model was built
to experimentally study the behavior of physical distribu-
tion systems.

The specific purpose of this thesis was to use the
LREPS model to study the effects of distribution warehousing
decisions on financial variables.

To accomplish this purpose, four steps were taken:

1. The researcher participated on the LREPS research
team to build the computer simulation model upon
which experimentation and sensitivity analysis
could be performed to study the total impact of

five alternative warehousing decisions on

Michael Lance Lawrence
financial variables in modeled companies from the
health cares products and appliances wholesale
industries.

2. The experimentation and sensitivity analysis were
performed, with only one variable exogenously
changed per experiment and the resultant changes
in financial variables were observed.

3. Those financial variables which were "significantly"
affected under varying economic conditions in each
of the two industries by the alternative ware-
housing decisions were identified from the results
of the experimentation.

4. The results were studied for general relationships
which would help explain the interaction between
warehousing decisions and changes in financial
variables.

There were several results which should help financial
management to better anticipate the effects of warehousing
decisions and aid distribution managers in making correct
warehousing decisions. Further, experimental analysis
yielded several by-products of interest to finance. The
findings of primary interest are the effects on financial
variables of the addition of a warehouse and of the shift
from private to public warehousing.

Every experiment which involved adding a second ware-
house resulted in a drop in the average level of accounts

payable of 13 to 21 percent. The drop in accounts payable

Michael Lance Lawrence
in every case resulted in a decline in the average cash
balance; an increase in short term debt; and a drop in the
debt service coverage ratio and net earnings, which re-
flected the fact that interest bearing debt replaced a
cost free source of financing. The major implication of
this finding is that managers contemplating an additional
warehouse should consider the adverse effects which such
a decision will have on the financial structure and
liquidity position of the firm and include these con-
siderations in the decision-making framework.

The addition of a second warehouse caused a sub-
stantial rise in inventories in the health cares company
but only a small increase in the appliances company.
However, the change in accounts payable was relatively the
same between the two industries. Resultantly, the strain on
the financial structure and liquidity position caused by the
drop in accounts payable is aggravated in the health cares
company by the substantial rise in the level of inventories.
That this rise in inventories was not financed by a rise in
accounts payable led to the following analysis of the re-
lationship between warehousing additions, inventories, and
accounts payable:
1. Increases in safety inventories are financed from
cash or some other source, not from accounts payable.
2. Increases in re-order inventories which occur
because of the addition of a second warehouse result
in an increase in accounts payable at the beginning

of each re-order cycle and an increase in the

Michael Lance Lawrence
length of the re-order cycle. The net result of
the two effects is that the average level of
accounts payable is left unchanged, meaning that
the increased inventories are not financed through
accounts payable.

3. The average cost of carrying inventories shifts as
the result of the addition of a warehouse if the
addition causes re-order inventories to increase.

Each experiment involving a change from private to
public warehousing resulted in a drop in the level of sales
required to break even on net income, a drop in the
variability of earnings and cash flows, and an improvement
in the liquidity and debt service coverage ratios. These
changes individually and in total reflect a much improved
defensive posture against market and economic reversals
if the firm uses public warehousing.

Furthermore, replacing a private with a public ware-
house resulted in an increase in accounts payable because
the expenses incurred through the use of a public warehouse
all give rise to payables. Many private warehousing expenses
do not. These increased payables are a small but permanent
source of financing and are another reason that public ware-
housing puts less strain on the liquidity position of the
firm than private warehousing.

The research by-products of interest to finance center
on the experimental verification that short term loans have

a strong tendency to decline when sales turn downward and

Michael Lance Lawrence

to build up when sales rise. The Specific observations are:

Debt

1. The inclusion of short term loans in the ———r——
Equity

ratio will cause the ratio to give false danger
signals during upturns.

2. The debt service coverage ratio may cause undue
alarm if interest or re-payments on short term
loans are included in its construction.

3. The acid test ratio is a signal of a poor liquid
position if sales turn up, not a danger signal of
the succeptibility of the firms liquidity position
to sales reversals.

A major limitation of this research is that the model
firms are small, based on two industries, and limited to one
or two warehouse systems. Another is that the research is
based on wholesale companies and generalization of some of
the findings to manufacturing firms would be dangerous. The
need for more sophisticated output monitoring - model ad-
justment feedback mechanisms is a third major limitation.

Of necessity, the range of possible future economic and market
conditions under which the decision alternatives were tested
was also quite limited. Future research to rectify each of
these limitations and test the consistency of the findings

is in order.

DEVELOPMENT OF A DYNAMIC SIMULATION MODEL FOR
PLANNING PHYSICAL DISTRIBUTION SYSTEMS:
THE FINANCIAL IMPLICATIONS OF

WAREHOUSING DECISIONS

BY

Michael Lance Lawrence

A THESIS

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

DOCTOR OF PHILOSOPHY

Department of Accounting and Finance

@Copyright by
MICHAEL LANCE LAWRENCE

1972

To my loving wife, Janis

ACKNOWLEDGEMENTS

The basic computer model used in this research, LREPS,
was developed by a Michigan State University research
team. The members of the research team, Dr. D. J. Bowersox,
the faculty adviser, Dr. 0. K. Helfrich, Dr. E. J. Marien,
Dr. V. K. Prasad, Dr. P. Gilmour, F. w. Morgen, Jr., and
R. T. Rogers are gratefully acknowledged. The Specific
research performed for this thesis might not have been
possible without the technical and programming help of
John Dean, an aide to the research team.

The industrial firm which provided financial and
informational support to the LREPS project is acknowledged,
as is the American Warehousemen's Association which pro-
vided financial and informational support for this
specific research. The personal encouragement of Mr.
Donald Horton of the American Warehousemen's Association
is especially appreciated.

The doctoral committee for this thesis consisted of
three members of the Michigan State University faculty:

Dr. Donald J. Bowersox, Professor of Marketing and
Transportation; Dr. A. E. Grunewald, Professor of
Finance; and Dr. Harold Sollenberger, Associate Professor
of Accounting. Dr. Bowersox, who was co-chairman, was

especially helpful in the technical and conceptual

iv

aspects of model building. The assistance of Dr.
Grunewald, also co-chairman, in keeping the research
headed in a meaningful direction and in interpreting the
enormous volume of experimental output was invaluable.
Dr. Sollenberger's help in financial modeling and his
tolerance of being the third member of a co—chaired
committee is sincerely appreciated.
The typing assistance and patience of Mrs. Chris
Metcalf, Mrs. Kathy Craig and Mrs. Bonnie Keithley
is acknowledged and appreciated. Mrs. Metcalf is also
thanked for typing and preparing the thesis in final form.
My wife, Janis, is greatly appreciated for her help,

understanding, and encouragement.

TABLE OF CONTENTS

DEDICATION . . . . . . . . .
ACKNOWLEDGEMENTS . . . . . . .
LIST OF TABLES . . . . . . .

LIST OF FIGURES . . . . . . .

Chapter

I.

II.

III.

INTRODUCTION . . . . .
The Basic LREPS Model . .
Detailed Problem Statement
Organization of the Thesis

BACKGROUND TO THE RESEARCH .

The Concept of Financial Position

Measuring Financial Position

Anticipating the Effect of Decision
Alternatives on Financial Position

Concluding Remarks . . .
LREPS: FINANCE VERSION . .

Introduction . . . .
General Features of LREPS- F
The Supporting Data System
Operations System . . .
The Operations System:

Measurement Sub-System .
The Operations System:

Monitor and Control Sub-System

The Report Generator System
Concluding Remarks . . .

vi

Page
iii
iv

viii

15
22

25

27
3O

38
40

46

46
47
50
51

79
100

102
108

Chapter

IV.

V.

VI.

EXPERIMENTAL DESIGN . . . . . .

Techniques and Problems in the
Design of Simulation Experiments
Design of Experiments on LREPS—F .

EXPERIMENTAL RESULTS AND ANALYSIS .

Experimentation on the Modeled
Health Cares Company . .

Model Changes Required for the
Appliances Company . . . . .

Comparison of Results Between
Industries . . . . . . .

Warehousing Decisions, Inventories,

and Accounts Payable . . . .

SUMMARY AND CONCLUDING REMARKS . .

Addition of a Second Warehouse .
Public vs Private Warehouse
Decisions . . . . . . . .
Research By-Product to Financial
Analysis . . . . . . . .
Limitations and Implications for
Future Research . . . . . .

BIBLIOGMPIiY O O O O O O O O O 0

vii

Page
110
111
115

124

125
146
149
158
177
178
183
186

187

194

LIST OF TABLES

Table Page

3.1 Financial Statement Items and the

Associated Operations System Routines . . 57
3.2 A Summary Overview of the Finance

Associated Routines and Their Associated

variables 0 O O O O O O O O O O 59
3.3 Financial Variable and Routine Code

Glossary . . . . . . . . . . . 62
5.1 Finance Variables Significantly Affected

by the Addition of a Second Private

Warehouse in the Health Cares Company . . 127
5.2 Distribution Variables Significantly

Affected by the Addition of a Second
Private Warehouse in the Health Cares
Company 0 O O O O C I O O O O O 128

5.3 Comparison of Financial Variables Signifi-
cantly Affected by the Addition of a
Second Private Warehouse Under Varying
Growth Rate Assumptions for the Health
Cares Company . . . . . . . . . . 130

5.4 Comparison of Distribution Variables
Significantly Affected by the Addition
of a Second Private Warehouse Under
Varying Growth Rate Assumptions for the
Health Cares Company . . . . . . . 131

5.5 Financial and Distribution Variables
Significantly Affected by Using One
Public Instead of One Private Warehouse
in the Health Cares Company . . . . . 136

5.6 Comparison of Certain Key Variables to
the Public vs Private Analysis Under
Alternative Assumptions Concerning
Source of Financing . . . . . . . . 140

viii

Table Page

5.7 Comparison of Financial Variables
Significantly Affected by Using One
Public Instead of One Private Ware-
house Under Varying Growth Rate
Assumptions for the Health Cares
Company . . . . . . . . . . . . 142

5.8 Comparison of Financial Variables
Significantly Affected by the Addition
of a Second Private Warehouse Under
Varying Growth Rate Assumptions for
the Appliances Company . . . . . . . 150

5.9 Comparison of Distribution Variables
Significantly Affected by the Addition
of a Second Private Warehouse Under
Varying Growth Rate Assumptions for
the Appliances Company . . . . . . . 152

5.10 Comparison of Variables Significantly
Affected by Using One Public Instead
of One Private Warehouse Under Varying
Growth Rate Assumptions for the

Appliances Company . . . . . . . . 156
5.11 A Simplified Illustration of the

Relationships Between Changes in Sales

and Working Capital . . . . . . . . 161

ix

LIST OF FIGURES

General Description of Firm-Distribution

Audit 0 O C O O O O O O O 0

Stages of the Physical Distribution
Network . . . . . . . . . .

LREPS Systems Model Concept . . . .

LREPS Systems Design Procedure . . .

Flowchart: Processing, Preparation, and

Shipment Routine . . . . . . .

Flowchart: Inventory Depletion-—Cost of

Goods Sold Routine . . . . . .

Flowchart: Inventory Management Routine

Flowchart: Inventory Replenishment Routine
Flowchart: Inventory--Accounts Payable
Update Routine . . . . . . . .

Flowchart: Accounts Receivable Routine

Quarterly Sales and the Earnings Change
from Experiment 1.1 to 1.4 . . . .

Quarterly Accounts Payable, Experiment
1.01 vs 1.02 C O O O O 0 O 0

Inventory Re-Order Quantities and the
Weighted Average Cost Per Unit of
Carrying Re-Order Inventories . . .

Page

12

13

67

69
71

72

75

77

138

167

171

CHAPTER I
INTRODUCTION

Correct business decisions require explicit consider-
ation of the effect which the decision will have on all of
the activity centers (functions) of the organization.
Traditionally, this has not been possible because of the
poor quality of information concerning the interaction of
functional areas and because of the limitations of quanti-
tative decision techniques. Thus, firms typically have
been divided into "manageable" functions with the implicit
understanding that each functional manager should sub-
jectively consider the effects of his decisions on the other
functional areas of the firm.

The advent of electronic data processing, improvements
in computer technology, and advances in the management
sciences have substantially increased the number of vari-
ables which can be included in quantitative decision
analysis.; Recently, interest in the interaction of the
functional areas within the firm has grown rapidly. The
application of systems analysis as a tool of decision
.making is a logical outgrowth of these recent deve10pments.

Systems analysis can be considered an extension of

. . . 2
Ithea role of management dec1s1on mak1ng. In a general sense,

2
the systems concept involves viewing a united system of
objects as a hierarchy of ranked sub-systems integrated
into a single system.3 Johnson, Katz, and Rosenzweig de-
fine the concept of the business firm as a system in this
way:

A business firm is an integrated whole where each

system, sub-system and supporting sub-system is

associated with the total Operation. Its structure
therefore is created by hundreds of systems arranged
in hierarchical order. The output of the smallest
system becomes input for the next larger system
which in turn furnishes input for a higher level.
The systems concept contends that optimum decisions cannot
be made on the basis of individual functions because of the
complex inter-relationships between the functions.
Decisions in the firm should be concerned with the final
outcome, not with individual phenomena along the way.5
Optimization of the objective of the firm is often frus-
trated by optimizing individual functions.

Computer simulation is an operations research tool
which facilitates systems analysis. The word simulation
has been used in many different contexts to mean many
different things. For the purposes of this research, it
is defined as follows:

Simulation is an iterative operations research
tool which involves building a model to imitate the
Operation of a business system and then performing
experiments with that model in order to generate
answers to specific questions, provide information,
and study the behavior of the system. The model
itself is usually written in mathematical form and
may contain either stochastic or deterministic vari-

ables, or both.

1?1€= ultimate tool for management decision making from the

3
systems perspective would be a simulator of the total firm
and simulating the major sub-systems of the firm is a logi-
cal first step toward reaching this ultimate goal.

A major sub-system of the firm which lends itself most
readily to systems analysis is physical distribution, which
has been defined by the National Council of Physical Dis—
tribution Management as follows:6

A term employed in manufacturing and commerce to
describe the broad range of activities concerned with
efficient movement of finished products from the end
of the production line to the consumer, and in some
cases includes the movement of raw materials from the
source of supply to the beginning of the production
line. These activities include freight tranSportation,
warehousing, material handling, protective packaging,
inventory control, plant and warehouse site selection,
order processing, market forecasting and customer ser-
vice.

The performance of the physical distribution system
(and its management) is measured by two standards: (1)
level of customer service and (2) the total cost required
to attain that level.7 Typically, service and cost move
directly but non-proportionately with each other. A firm
might find that the cost of achieving a service level of
95% of the "optimal" level is double the cost of a 90%
service level. The design and management of physical dis-
tribution systems involves striking the best overall
balance between service and total cost. The best overall
system will seldom if ever be service maximizing or cost
Iminimizing. Rather, it will attain reasonable service
8

levels at a realistic total cost.

Determining this optimum balance is complicated by

4
inverse cost relationships between the activity centers
of the system (transportation, inventory, warehousing,
communications, and unitization.) Cost savings adjustments
in one of these centers often causes cost increases in one
or more of the other activity centers. Thus, finding the
total cost to compare to the potential service level of a
system design alternative is an involved process. The de—
sign process is further complicated by ever-changing en-
vironmental conditions. The Optimum system for any given
time period will likely not be Optimum in successive time
periods. Thus, the flexibility Of alternative systems
to adjust to future change is a third objective which
complicates the problem Of designing physical distribution
systems.

In its prOper perspective, then, physical distribution
system design must primarily be concerned with the cen-
tralized and integrated management of the movements system
in such a way that decisions are based not on individual
objectives and functions but rather on the total performance
of the system over a prescribed planning period. Prior to
the development of systems simulation, planning from such
a viewPoint was largely a discussional art rather than a
quantitative decision science.

A dynamic computer simulation of a large scale
physical distribution system which will be used for planning
and designing purposes from a systems perspective has been

develOped by a team of faculty and doctoral candidates under

5
a Michigan State University industrial research grant.
The project and model name is Long-Range Environmental
Planning Simulator (LREPS) and the research has two major
objectives:
1. To conceptualize, construct, and
computerize a dynamic simulation of
a large scale physical distribution
system.
2. To use the computerized model in
experimentation to examine questions
about and study the behavior Of
physical distribution systems.
The first Objective has been achieved, as LREPS is

operational and performing according to specification. The

research monograph Dynamic Simulation of Physical Distri—

 

 

bution Systems explains the conceptualization Of LREPS and
offers a general overview Of the project.9 The development
of the mathematical model is explained in a doctoral

dissertation entitled Development of a Dynamic Simulation

 

 

Model for Planning Physical Distribution Systems:

10

Formulation 2E the Mathematical Model. The computeriza-

 

 

tion of the math model is reported in a doctoral disserta-

tion entitled Development of a Dynamic Simulation Model

 

 

for Planning Physical Distribution Systems: Formulation
ll

 

 

of the Computer Model.

 

The second goal of the project has been partially
achieved. A recently completed doctoral dissertation

entitled Development 9f 3 Dynamic Simulation Model for

 

 

Planning Physical Distribution Systems: Validation of

 

the Operational Model has validated the design and output

 

6
of the model for scientific experimentation.12 A disserta-
tion currently in progress considers the statistical design

13 This

ramifications Of experimentation with the model.
dissertation involves modifying the basic model in order
to study the effect which distribution warehousing desicions

have on another major sub-system Of the firm: the finance

sub-system.

The Basic LREPS Model

 

The development and conceptualization of the basic
LREPS model are presented in detail in the above mentioned
monograph14 and in the first two dissertations from the

project.15’l6

The following summary comments concerning
the basic model are presented to lend continuity to this

volume.

General Framework

 

The basic LREPS model imitates the Operations of the
modeled physical distribution system from the end Of the
manufacturing activity to the transfer of product to
customers. The five major components of the physical
distribution system which are included are as follows:

1. The fixed facility system, which is concerned
with when, where, and in what size and form
warehouses should be included in the total
distribution system.

2. Inventory, which is concerned with where,

when, and in what volumes finished goods
should be held in the system.

.7

3. Tran5portation, which is concerned with the
movement, and the form and timing of move-
ment, into and out Of the fixed facilities.

4. Throughput (or unitization), which is
concerned with movement of goods within
the fixed facility and with the physical
picking and preparation of customer orders.
As such, it is also concerned with the in~
ternal management of fixed facilities.

5. Communication, which is concerned with the
flow of orders and other information between
the firm and its customers and between various
stages of the firm.

There are three major stages of the modeled system

at which activities occur, originate, and/or terminate.

The three stages are presented in graphical form in

Figure 1.1. The type Of distribution system they represent
is summarized in Figure 1.2. These stages are:

l. The Manufacturing Control Center (MCC) and
its associated Replenishment Center (RC) at
which products are manufactured and stored.

Each MCC produces only a partial line of
products and each product is manufactured
at no more than two MCC's.

2. The Distribution Center (DC) which is an
intermediate fixed facility between the RC
and the marketplace.

3. The demand unit (DU) which can by design be
either an individual customer or a group Of
geographically agglomerated customers.

The second (DC) stage requires further discussion,

as there are four different possible forms Of distribution
centers. A primary distribution center (PDC) is one which
handles all products and possesses a design capability Of
serving all DU's within a defined region Of the total

market area. It differs from the second type, the Remote

Distribution Center - Full Line (RDC-F), in that the RDC—F

8

PHYSICAL DISTRIBUTION SYSTEM

 

MANUFACTURING CONTROL CENTERS (MCC)
MULTI-LOCATION
EACH PRODUCES LESS THAN FULL LINE
EACH PRODUCT IS PRODUCED AT MORE THAN ONE MCC

REPLENISHMENT CENTERS(RC)
MULTI-LOCATION
EACH STOCKS ALL PRODUCTS MANUFACTURED AT MCC

DISTRIBUTION CENTERS (PDC) (RDC)
MULTI-LOCATION
FULL LINE - PRIMARY DC (PDC)
FULL OR PARTIAL LINE - REMOTE DC (RDC)
CONSOLIDATED SHIPPING POINT (csp)

TRANSPORTATION
COMMON CARRIER - TRUCK, RAIL, AIR

INVENTORY
STOCKS AT RC, PDC, RDC

COMMUNICATIONS
COMPUTER, TELETYPE, MAIL, TELEPHONE

UNITIZATION
AUTOMATED OR MANUAL

PRODUCT PROFILE

 

MULTI-PRODUCT LINE
KEY PRODUCT GROUPS FOR EACH CUSTOMER CLASS OF TRADE

MARKET PROFILE

 

MULTI-CUSTOMER CLASSES OF TRADE
TOTAL U.S. MARKET

COMPETITIVE PROFILE

 

MULTI-COMPETITORS

Figure l.l--General Description of Firm-Distribution Auditl

1D. J. Bowersox, et al., Dynamic Simulation of Physical

 

Distribution Systems, Monograph (East Lansing, Michigan:
Division of Research, Michigan State University, Forth-
coming) .

 

 

 

STAGE 1:
MANUFAC-
TUBING

CONTROL q'N
CENTERS

\
AND RE- ‘ '\ \\\
PLENISH- I \ ~
MENT ’ ‘ ‘\

CENTERS ,/ ‘ ‘ “

 

 

 

STAGE 2: RDC RDC
DISTRIBU- PDC FULL PDC PARTIAL
TION LINE LINE

CENTERS

 

 

 

 

 

 

 

 

 

 

 

 

~u._47
.17
.J'
,v
N ———v

\.
STAGE 3: / I ‘\ ’
”M” Q @
UNITS

PD REGION

 

 

 

 

PD REGION ]

————— INFORMATION FLOW -————-———PRODUCT FLOW

 

REGION...THE REGION IS DEFINED BY THE ASSIGNMENT OF RDCS AND DUS TO A
PDC.

MCC......EACH MANUFACTURING CENTER PRODUCES A PARTIAL LINE.

RC.......REPLENISHMENT CENTERS STOCK ONLY PRODUCTS MANUFACTURED AT
COINCIDENT MCC.

RDC......REMOTE DISTRIBUTION CENTER, FULL OR PARTIAL LINE.

PDC......PRIMARY DISTRIBUTION CENTER, EACH PDC IS FULL LINE AND
SUPPLIES ALL PRODUCTS TO DUS ASSIGNED TO THE PDC REGION:
PRODUCT CATEGORIES NOT STOCKED AT THE PARTIAL LINE RDCS IN
THE REGION ARE ALSO SHIPPED BY THE PDC.

DU.......THE DEMEND UNIT CONSISTS OF ZIP SECTIONAL CENTER(S).

CSP......CONSOLIDATED SHIPPING POINT.

 

 

 

Figure l.2--Stages of the Physical Distribution Network1

1D. J. Bowersox, et al., Dynamic Simulation of Physical
Distribution Systems, Monograph (East Lansing, Michigan:
Division of Research, Michigan State University, Forthcoming).

 

10
is designed with a full line of products but with the
re3ponsibility to serve only part Of the DU's in a pre-
scribed region. The third type, a Remote Distribution
Center-Partial Line (RDC-P) supplies only a partial line
of products to its assigned DU's with the other products
(usually the slower moving products) being supplied from
the PDC to which the RDC-P is linked. The fourth type of
DC, the consolidated shipping point (CSP) is merely a
break-bulk point to which the aggregate demand for several

DU's can be shipped.

Model Design Criteria

 

The general framework discussed above described the
physical distribution system of most of the larger sized
manufacturing firms. Based on this general framework, the
LREPS conceptual model was formed under the following de-
sign criteria:

1. The construction should be in modular form
and the model should be universally applicable
to industrial and consumer products firms
after only minor changes in design.

2. The model should enable testing of trade-offs
between cost and service and among the various
cost functions.

3. The model should embody the capacity to measure
the extent to which the desired physical distri-
bution system change as the environmental con-
ditions changes; that is, it should embody a
sequential decision process.

4. The constraints Of computer resources and
reasonable real world validity should be met.

11

Conceptual Design Of LREPS

 

The major systems and sub-systems through which LREPS
was modularly constructed are summarized in Figure 1.3.
The Supporting Data System (the input system) is run off-
line and exists to facilitate design analysis and the pre-
paration and reduction of data for input into the Operations
System. The Operations System is the model system which
imitates the real world physical distribution system de-
scribed in Figures 1.1 and 1.2. The third major system,
the Report Generator System is designed to print the
simulation output in several Optional management reports.

The LREPS model is discussed in more detail in Chapter
III through description Of these three systems and their

sub-systems.

System Identification

The design procedure used in the development of the
LREPS simulation model is summarized in Figure 1.4. The
first step, "Problem Definition and Feasibility Study,"
was based on a collection and analysis of data to determine
if the objectives Of the research were attainable. The
outputs of this step were a detailed problem statement,
the specifications for the mathematical model, and the
design criteria for the Operational model. This served as
input to the conceptualization of the mathematical model,
which consisted of the specification of (1) system bound-

aries and assumptions, (2) the inputs and outputs, (3) the

12

_

.huwmuo>wco mumum :mmfl20fiz .noumommm mo cowmw>wa II II
.mEmummw :oflusnfluumfla Hmowmunm mo coaumHsEHm anocmo ..Hm um .xomumzom .n .0

.Amcweoonuuom

”savanna: .mcflmcoq ummmv cmmHOOGOS

H

unmocoo Hmooz msmumsm mamquIm.H Gunmen

H

 

 

 

xuamwm ho wu<az=om

 

 

 

 

NBHAHmeMAh

moH>mmm
mmA<m
Hwou

mo mMMDmdmx

 

._
_
_
_

L
_
_

 

zuamrm
moa<mmzwu
Hmomma

_
_
_
.t
_

_ lbIMmlL

 

 

 

machodm
mHnoz
mm4<m

mmuz<zo
mrm am

nomwzoo
>zH

o a z

 

 

Mo<mnmmh

 

 

 

 

 

C

 

xmgm
mUH>Mmm
mm4<m

Hmoo

 

 

 

mmz<me
HHZD
>2H
2200
04m

who

 

 

 

 

 

fl

 

mmmnmo
mo
UOAA<

mma<m
A<5H0<

 

FIFE

 

 

 

 

zuamwm GZHH<Mmmo

 

4 J_

 

mmoau<m Hmoo

mmnbm onmHomQ
mmz<MH

tz Huaaomm

— tz mmzoamao
mmmnmo

Hm<ummom mqum
mmmamz<m<m HQ:

 

<H¢a

— zmemwm ‘
— quHmommbm _

 

l3

 

 

 

    

PROB DEFN &

  
  

 

 

 

  
  
  

 

 

 

 

FEAS STUDY
4A
MATH MODEL
DEVELOPMENT
SCOPE
OF
:4). RESEARCH

 

 

 

 

 

(TERMINAL)

PROCESS
‘fi BLOCK

I/O
BLOCK

 

  
 

VALIDATION
0' OIEL

    

 

 

   
  

 

 

 

 
 

EXP DESIGN &
MODEL USAGE

 

  

 

 

(END) 1

Figure l.4-—LREPS System Design Procedure

1D. J. Bowersox, et 31., Dynamic Simulation Of Physical
Distribution Systems, MOnograph (East Lansing, Michigan:
Division of Research, Michigan State University,
Forthcoming).

14
constraints of the inputs, outputs and system configuration
measurement, (4) criteria for comparing alternative systems,
(5) the design parameters to be used for experimentation,
(6) the estimation of parameters via further data collection
and analysis, and (7) comparison to real world data as a
measure of model validity.15 The next step was the computer-
ization of the mathematical model which involved the
following:16

1. Specification of the computerized data base.

2. Selection of the programming language.

3. Identification of each Of the computer model
activities with their respective data bases
and programming them.

4. Identification of activities and information
flows, flowcharting and coding, and testing
and debugging; for each individual sub-program
associated with the LREPS procedure.

5. Combination of the subprograms and their re-
spective inputs, outputs, and data base require-
ments.

6. Testing, debugging, and Operationalizing the
combined subsystem models and calibration to
the actual system.

The following step in the system design was Model Validation,
which involved demonstrating that the model is a reasonable
representation of reality. The validated model was the

input into the final process in the system design, Experi-
mental Design and Model Usage. The focus of this thesis

is on the last step. The model is used to experimentally

study the effect of warehousing decisions on variables Of

the finance sub-system of the firm.

15

Detailed Problem Statement

 

Financial position of the firm refers to balance
between liquid and productive assets in the asset
structure, the split between long term and short term
debt and equity in the financial structure, and the extent
to which the asset and financial structures complement
one another. In the broadest sense it encompasses all of
the components (variables) of the finance sub-system of
the firm. Financial position is typically managed by
exception. That is, various financial ratios are com-
pared to prescribed standards and changes are made only
when the ratios deviate substantially from their standards.
There has recently been much criticism of the waste in-
herent in financial management by exceptions and manage-
ment from a preventive orientation has been encouraged.
Preventive management suggests anticipating adverse changes
in financial position and preventing their occurrence.
Changes can only be anticipated, however, if their causes
are understood and identification of the causes of change
in financial position is a necessary condition for
preventive management.

There are two basic premises to this research:

(1) Much of the change in financial position Of the
firm over time is the result of overt management
decisions in other functional areas of the firm.

(2) Most operational decisions within the firm
trigger changes, either directly or indirectly,

in many of the variables of financial position.

It follows from these premises that a major step toward

16
more enlightened financial management is to identify the
effects which various classes of management decisions have
on the components of financial position. A better under-
standing Of the total financial effects of specific manage-
ment decisions would also increase the likelihood Of correct

decisions.

Objectives of the Research

The objective of this research is to study the nature
of the effects of decisions in the warehousing function
on variables of the finance sub-system. This is accom—
plished by:

1. As a member of a faculty and student research
team, building a dynamic computer simulation
model upon which experimentation and sensitivity
analysis can be performed to study the impact of
alternative warehousing decisions on financial
variables of modeled firms.

2. Performing the experimentation and sensitivity
analysis, with only one variable exogenously
changed per experiment and recording the re-
sultant changes in financial variables.

3. Identifying those financial variables which are
"significantly" affected by each warehouse de—
cision alternative under varying sales growth
rate assumptions and in different industries.

4. Studying these results for general relationships
which would help describe the nature Of cross
functional interaction between warehousing de-
cisions and financial variables.

Financial position (or financial health) is an area

of the firm which has traditionally not been considered in
distribution system decisions. A better understanding of

the interaction between warehousing decisions and financial

17
variables would enable managers to more accurately antici-
pate the total financial effect of decisions made in the
warehousing area. Better informed decisions could be made
in the warehousing area and changes resulting from ware-
housing decisions could be anticipated in the finance area.
Management of financial position from a preventive rather

than a corrective approach would be closer to reality.

Warehousing Decision Alternatives

The warehousing decision alternatives selected for
study are the various possible combinations Of public and
private warehouses for one and two warehouse systems at
two possible locations in a study region. The study region
is limited to the twelve state area which includes:
Nfissouri, Nebraska, Iowa, North Dakota, South Dakota,
Minnesota, Wisconsin, Illinois, Indiana, Michigan, Ohio,
and Kentucky.

Chicago is the single location which minimizes the
total ton-mile distribution requirements for the twelve
state area, assuming that sales are perfectly correlated
with population. As such, Chicago will always be included
as either a public or private warehouse. Given that
Chicago is in solution, Detroit is the location for the
second warehouse which Offers the biggest reduction in
total ton miles. However, problems with the directional
flow of traffic significantly reduce service capability

from Detroit and the next best location, Columbus, is

18

selected as the second location for the two warehouse
systems. Since the possible combination of a private
warehouse at Columbus and a public warehouse at Chicago is
eliminated as being unrealistic (Chicago will always handle
the largest volume), the possible one and two warehousing
schemes remaining to be tested experimentally are:

l. A private warehouse at Chicago.

2. A public warehouse at Chicago.

3. Private warehouses at Chicago and at
Columbus.

4. Public warehouses at Chicago and at
Columbus.

5. A private warehouse at Chicago and a
public warehouse at Columbus.

Sales Growth Rates

 

The experimentation is performed assuming 5%, 15%,
and -5% annual growth rates in sales to test the sensitivity
Of the relationship between warehousing decisions and
financial variables to changes in economic and market
conditions. Daily sales vary randomly within each
experiment but annual sales are designed to average the
assumed growth rate bias over the five year horizon Of

the model.

Types of Industries Involved

 

The LREPS model is based on a data survey of a major
company in the health cares industry. Modeling the finan-

cial system of such a company would require expanding the

19
model to include the manufacturing system. To avoid such
a monumental task, it is assumed that the model represents
a company in the health cares products wholesale industry.
The required financial information not obtained in the
basic data survey were obtained by taking the median average
data for this industry from the Robert Morriss Associates
and Dun and Bradstreet industrial financial statement
studies. The model has been pre-tested and various para-
meters have been adjusted to assure that the model output,
before experimentation, is a reasonable approximation of
the reported financial and Operating statistics for the
average company in this industry with annual sales in the
ten million dollar range.

To test the consistency of findings across industry
lines, the experimentation is repeated for the home appli-
ance products wholesale industry. Since the basic model has
been built with universal application a prime consideration,
extensive basic model changes are not required for modeling
the second industry. Changes in the data input are neces-
sary, however. The necessary data for changing distribution
system cost and structure characteristics and product and
market characteristics were provided by two major firms
in the appliance industry. The required financial informa-
tion was obtained by taking the median average data for
the industry from the Robert Morriss Associates financial

statement study.

20

Financial Variables

 

The financial variables included in this research

consist

l.

2.

3.

of the following:

Each item in the balance Sheet and income state-
ment of the modeled firms.

Twelve selected financial ratios.

Several proxy measures of risk which are based
on the regression of quarterly sales and
adjusted quarterly net income. Adjusted net
income is defined by net earnings, after taxes
plus depreciation less annual principle re-
payments on long term debt. The specific
statistics from this regression analysis are:
a
a. The formula X = E , formed by trans-
forming the linear regression equation,
which is a measure of the level Of
quarterly sales below which adjusted
net income would become negative.

b. Total variance, which is the total
dispersion of adjusted net income over
the twenty quarter period and is a
measure of the stability of earnings.

c. Residual variance, which is the amount
of total variance not explained by the
relationship between sales and net
income. It is a measure of the un-
certainty Of earnings beyond the in-
stability which occurs because of
sales variability.

The financial statements and the twelve ratios are

included as traditional measures of financial position and

performance. The special statistics are used as measures

of the more academic concept of risk.

Distribution System Measurement Variables

 

Although the objective of the research is to study

the interaction between warehousing decisions and changes

in financial variables, certain financial changes may

21
occur indirectly from warehousing decisions through the
direct effects which the decisions have on distribution
system variables. Therefore, measures of distribution
variables are included as possible tools for explaining
the observed changes in financial variables.

The two standards by which distribution systems are
most commonly measured are service and costs. For this
research, the service levels will be measured from two
approaches: stock-out frequencies and order cycle time.
The measures of stock-out frequencies used for this
research are:

1. Number of stockouts.

2. Average days per stock-out.

3. Standard deviation days per stock-out.

The measures of order cycle time employed are:
1. Percent of total sales dollars delivered to
customers within 5, 6, 7, and 8 days from
the time the order is placed.
2. Percent of total orders delivered to cus—
tomers within 5, 6, 7, and 8 days from the
time the order is placed.

Costs are measured for the total physical distribution
system and for each of its components which are separated
for costing purposes as follows:

1. Inbound transportation

2. Outbound transportation

3. Warehouse operations expense (throughput
cost)

4. Warehouse facilities expense

22

5. Communications expense

6. Inventory expense

Ogganization of the Thesis

 

This thesis is divided into six chapters. The first
chapter has served as an introduction to the research.
Chapter II explains where this research fits in the main—
stream of research in the finance area and reviews the
pertinent finance literature. Chapter II also identifies
and defines certain significant finance variables and
describes in general terms why a multitude of variables
and reports are required to measure financial position of
the firm.

Chapter III describes the LREPS model as it has been
adjusted to perform this research and describes the addi-
tional data gathering and assimilation required for model
adjustment and experimentation. Chapter IV briefly reviews
pertinent general concepts of experimental design, discusses
the design problems encountered in this research, and
describes the experimental design adopted. Chapter V
reports the results of experimentation, the analysis of
those results, and the conclusions drawn from the analysis.
Chapter VI summarizes the objectives and findings of the
research, describes several limitations, and discusses

implications for future research.

CHAPTER I--FOOTNOTE REFERENCES

1Donald J. Bowersox, Edward W. Smykay, and Bernard J.

LaLonde, Physical Distribution Management (New York: The
MacMillan Company, 1968), p. 8.

2Stanford Optner, Systems Analysis for Business and
Industrial Problem Solvifig (Englewood Cliffs, N. J.:
Prentice-Hall, 1965), p. 8.

3Howard L. Timms, Introduction to Operations Managgment
(Homewood, Illinois: Richard D. Irvin, 1967). p. 82.

4Richard A. Johnson, Fremont E. Katz, and James E.
Rozenzeweig, The Theory of Management of Systems (New York:
McGraw-Hill, 1963). P. 91.

 

 

 

 

 

5Optner, op ci .1
6Bowersox, op. ci ., p. 2.
71bid., p. 323.

81bid., p. 116.

9Donald J. Bowersox, et al., Dygamic Computer Simula-
tion of Physical Distribution Systems, Monograph (East
Lansing, Michigan: DivisiOn of Research, Michigan State
University, Forthcoming).

10O. K. Helferich, Development of a Dynamic Simulation
Model for Planning Physical Distribution Systems: Formula-
tiOn of the Mathematical Model (unpublished Doctoral Disser-
tation, Michigan State University, 1970).

11Edward J. Marien, Development of a Dynamic Simulation
for Planning Physical Distribution Systems: Formulation of
the Computer Model (unpublished Doctoral DiSsertation,
Michigan State University, 1970).

12Peter Gilmour, Development of a Dynamic Simulation
Model for Planning PhySical Distribution Systems: Valida-
tion of the Operational Model (unpublished Doctoral Disser-
tation, Michigan State University, 1971).

 

 

 

 

 

 

 

 

23

24

13Richard T. Rogers, Development of a Dynamic
Simulation Model for Planning Physical Distribution Systems:
Design Of Experiments (unpubliShed Doctoral Dissertation,
Michigan State University, 1971).

14

 

 

 

Bowersox, et. a1; Op. cit.

15Helferich, op. cit.

16 . .
Mar1en, Op. c1t.

l71bid.

18Ibid.

CHAPTER II
BACKGROUND TO THE RESEARCH

The financial variables considered in business de-
cisions traditionally have been expected profits and return
on investment. In recent years, the concept Of risk is
receiving considerable attention from the academic community
as an additional financial decision variable. The total
risk complexion of the firm is popularly treated in two
parts: business risk and financial risk. Business risk
is the variability of net Operating income.1 It is a
function of the uncertainty of sales and other environ-
mental variables; and of the Operations structure through
which sales and other exogenous inputs are transformed
into Operating cash flows.

Financial risk is Often defined as the extent to which
business risk is magnified or accentuated by the financial
structure Of the firm.2 The addition of debt to the
financial structure adds interest expense and debt prin-
cipal repayment to the total fixed expense structure,
which increases the probability that the gross Operating
margin from sales will not cover the fixed Obligations Of
the firm under adverse economic conditions. The de—

finition Of financial risk is often broadened to include

25

26

the degree Of complementarity between the level of liquid
assets and the maturity of the debt within the firm.3'4
Since both business risk and financial risk are related to
the nature of the firm's assets, the two are overlapping.
This is one explanation of why risk has been measured in
total rather than in its component parts.

Measures of risk in the finance literature are
typically based on the variability of expected return.5'6
No explicit recognition is given in these measures to the
importance of the shifts which the decision in question
would cause in the asset and financial structure Of the
firm, and these shifts are integral determinants Of
business and financial risk.

The most SOphisticated approach used generally by
practitioners to explicitly consider risk as a decision
variable is the addition of a risk premium to the discount
rate in calculating net present value.7 The inadequacy of
information concerning the variability Of returns on
projects Often prohibits use of techniques Of measurement
suggested in the academic literature. Nevertheless,
practioners do recognize the importance Of risk as a
decision variable, as reflected in the following statement
by Tucker:9

Profit planning must plan for continuous profits

which are made consistent with the financial state

of health Of the company. The maximization of

profit is no guarantee of financial health, and

in fact, when profits are maximized to the exclu-

sion of other considerations, the company can get
into serious difficulty.

27
That risk is important as a second parameter is not
disputed by practitioners; how to measure it is. In
measuring the risk nature of the firm as of a given point
in time, practical analysts have long depended on the

concept of financial position.10

The Concept of Financial Position

 

Although authors writing on the subject of financial
analysis and control agree on the importance of financial
position, there is wide disagreement concerning exactly
what it is. Arthur Dewing, one of the earliest of the
finance authors, felt that financial position is the
asset and liability structure of the firm as stated in
the balance sheet.11 Many writers feel that financial
position is primarily determined by the solvency of the
firm, as measured by liquid assets and the coverage those

12’13 It is also

assets give to immediate obligations.
often defined as the split between equities and liabilities
and the maturity structure of those liabilities.l4 Others
feel that financial position is the measure Of a firm's
ability to withstand adverse market and economic con-
ditions.15
Why is there such a divergence of Opinion among
respected authorities on such an important element Of
financial management? Felix Kollaritch has succinctly

characterized this problem.16

28

Financial position is many different things to
many different peOple. . . . .A creditor with

no substantial interest in the borrower's

business (a trade creditor, for example) will

base his judgment on the enterprises liquidating
value.....(Creditors who have a substantial
interest in the debtor's business) will use two
yardsticks to measure the security behind the

loans extended. First, they will decide whether

or not the assets in their relationship to pro-
duction are sufficient to assure solvency and
continuation Of the enterprise. (The second
yardstick) is whether or not the assets of the
enterprise can be maintained.....(The owner's)
primary interest lies in the financial worth Of

his investment and in the profits to be derived
from it. As such he is interested in the company's
maintaining the vitality in asset structure required
to grow in the future.....TO financial management,
the management group primarily responsible for

and interested in the financial position of a
business, the term 'financial position' is synony-
mous with the collequial use Of the term in business,
namely its debt paying ability as a going concern
(including asset acquisition, labor, creditors,
etc.) Immediate paying ability of the firm will

be judged on the basis of a continuing enterprise.

The orientation of this research is the manager's
vieWpOint. However, Kollaritch's description of financial
position from the vieWpOint of financial management is
incomplete and inaccurate for at least three reasons.
First, it is widely accepted among finance authorities that
the ultimate financial Objective of the firm is maximiza-

tion Of the wealth of the shareholder.l7'18

Managing
financial position in a manner which investors find un-
desirable will penalize the market value Of the company.
Second, a continuing enterprise must always be concerned

with obtaining future debt financing and therefore must

be concerned with the appearance of the firm to outsiders.

29

Third, the financial performance Of the firm under
future conditions depends largely on the financial position
of the firm today. If future sales increase significantly
but the firm has a disproportionately high level Of current
and liquid assets, the firm will not realize the full pro-
fit potential from the upswing. If future sales turn down
and the firm has a disprOportionately high level of fixed
assets and little borrowing capacity with suppliers of
outside financing, the solvency Of the firm may be en-
dangered. In summary, the degree of flexibility with
which the firm can adjust to uncertain future conditions
is a function of the vitality of the asset structure, the
vitality Of the liability structure, and the extent to
which these structures complement one another.

Wright warns that maintaining this character of
vitality and complementarity should be a prime objective
Of financial management and refers to it as "financial

balance."19

He suggests arranging financial position so
that the firm can benefit substantially from favorable
future conditions while providing the required protection
against future adverse conditions.

Borrowing from Wright's concept Of financial balance,
financial position for the purposes of this research is the
vitality of the firm as a going concern to beneficially

adapt to ever changing future conditions and refers to the

balance within and among the asset and financial structures.

30

Measuring Financial Position

 

Just as there is no accepted single definition for
"financial position," there is also wide disagreement on

how to measure it.

The Balance Sheet

 

For the early finance writers, the balance sheet
served as a measure of financial position. Dewing, for
example, defines the balance sheet as "a statement of the

"20 At

financial position Of a firm as of a point in time.
even this early date, however, many Of the shortcomings Of
the balance sheet as a measure Of financial position were
recognized. Graham and Dodd, for example, warn that the
balance sheet is an extremely static measure and individual
items mean little unless considered in relation to other

21 Guthmann

items and compared to some objective standard.
and Dougall are especially critical of the static nature
of the balance sheet and insist that whether individual
items are in an improving or decaying trend is far more
important than their absolute values.22 Today, many
accountants and financial analysts are displeased that
balance sheet information is on a historical rather than
a reproduction basis and that wide latitude in accounting
practices makes it difficult to know what the absolute

values actually represent.23

31

The Sources and Uses Of Funds Statement

The general purpose of this statement is to exemplify
the changes in financial position which occur as the result
of operations in any given time period. The most common
technique of calculation is to subtract each balance sheet
item in time period t—l from the corresponding item in
time period t. Guthmann and Dougall, who were among the
earliest to recommend the use of the funds statement,
feel it rectifies the static nature of the balance sheet.24
Their contention that it is primarily a tool for under-
standing from where cash has come and to whence it has

25’26 Bierman

gone is shared by many prominent authors.
explains that the funds statement is preferred to the
balance sheet by many analysts because accounting conven-
tions(historica1 rather than reproduction cost, for
example) have produced a balance sheet which is a residual
of accounting procedures rather than a meaningful statement
of financial position.27
As a measure of total financial position, however,
the funds statement lacks concrete value. Financial items
must be considered in relation to each other and compared

against some Objective standard Of measure to have

economic meaning.

Egnancial Ratios

 

As early as the 1890's financial analysts recognized

‘that individual financial statement items only carry

32
significance in relation to other items.28 At a much later
date, Tucker explains the rationale for considering finan-
cial data in the form of ratios as follows:29

Primary data are defined as those data which can-

not in themselves be used to predict or appraise

objectively performance, progress, and profit-
ability. Primary data are used as the ingredients
for various types of ratios which do give economic
meaning to events and permit objective diagnosis,
decision making, and assessment of all areas of

a company's economy and activity.

Early writers, notably Dewing, warn that hard and
fast rules for ratios are dangerous because acceptable
financial policies in one industry are not necessarily
acceptable in others. In the period 1900-1920 a signifi-
cant development in the art of financial analysis was the
establishment of industry averages as rules Of thumb
against which the ratios of individual companies could

be compared.30

Alexander Wall developed average ratios
for 981 companies, stratified by industry and geographical
location. His measures were incomplete and were based on
a very insufficient sample, but were noteworthy as a
beginning.31 Ten years later, Roy Foulke began keeping
averages of twelve ratios from a wide sampling of
companies which was the first form of the fourteen key
ratio averages published currently by Dun & Bradstreet.32
The use of financial ratios in conjunction with industry
averages satisfies the two most commonly heard criticisms
of financial statements. Through ratio analysis items

are considered in relation to other items and Objective

standards Of measure are available in the form of

33
industry averages.

Evaluating financial position through ratio analysis
has been Often criticized. As early as the 1930's, Gillman
expressed concern over ratio analysis for the following
reasons:33

1. Changes in ratios can result from movement

in either the numerator or denominator and

the cause of ratio change is not always clear.

2. Ratios divert attention from the comprehensive
view of the firm.

3. Even within the same industry, a particular
ratio with a certain value might indicate
safety for one firm and trouble for another
because of varying management attitudes and
philosophies, market characteristics, and
so on.

Gillman's criticisms, however, were largely ignored for
many years, probably because Of several studies in the
early 1930's which demonstrated the highly predictive
power of certain financial ratios in anticipating bank-
ruptcies. The results of the studies are of significant
interest. Winakor and Smith conclude from their studies
Of firms which experienced financial difficulties from
1923—1931 that deterioration in the net working capital
to total assets ratio preceded financial difficulties and
often began to deteriorate ten years in advance of bank-
ruptcy.34 Fithatrick added successful companies to his
sample to serve as a control group against which the
unsuccessful firms could be compared, a precaution which

Winakor and Smith failed to take. He concludes that the

deterioration in any one Of a number Of ratios can

34

predict bankruptcy, but that the most powerful are net
profit to net worth, net worth to debt, and net worth to

fixed assets.35

Merwin analyzed financial reports for
six prior years for a larger number of "continuing" and
"discontinuing" firms. Among his findings are that net
working capital to total assets; net worth to debt; and
the current ratio are the most sensitive predictors of
discontinuance.36
A noteworthy Observation is that in total the three
studies indentify seven ratios which are significant ex-ante
predictors of financial difficulties. Of the seven, only
one is a measure of profitability and return. Two ratios
are concerned with balance within the capital structure,
two with the balance within the asset structure, and two
with the complementary relationship between the asset
structure and the capital structure. This evidence strong-
ly indicates that there are other financial variables as
important to the firm as profitability.
Wide use of the traditional financial ratios to
measure financial position continued into the 1950's. In
their 1951 edition Graham and Dodd state:37
The tangible factors affecting the quality of a
company may be measured through the use Of certain
key ratios. .....(The security analyst) should
then develOp a number of key ratios which will
throw light on the company's over-all performance,
on the safety of its senior securities, and on the
attractiveness of its common stock for investment.

In the late 50's and early 60's displeasure develOped

with the static nature of the traditional ratios, which

35

are calculated from items in the balance sheet and income

38,39

statement. The following are examples Of criticisms

aimed at the traditional ratios:

1. Long term debt totaling one million dollars
would appear the same in traditional ratios
for a ten year, 8 percent loan as for a
thirty year, 4 percent loan, even though the
former requires over two times more outflow
of cash each year.

2. Lease payments, which constitute a major
fixed outflow Of cash each year for many
companies do not appear in the traditional
ratios.41
To alleviate these and other shortcomings, much attention
is given in recent years to "coverage ratios" such as
the number of times working capital covers average daily
cash flows and the number Of times annual earnings plus
depreciation covers annual interest expense and principle
re-payment. Although the more traditional ratios
(liquidity, profitability, and leverage ratios) continue
in wide use, the coverage ratios are significant additions
to the measurement Of financial position.
Despite the widesPread use of the traditional and
coverage ratios, many finance academicians question the
value of ratio analysis. Their disenchantment is largely

with the use Of industry averages as objective measures.42

Donaldson summarizes this disenchantment as follows:43

Even assuming strict comparability, which is hard
to establish, there is no proof that the companies
concerned have arrived at their current debt pro—

portions in-a deliberate and rational manner. In
view of the wide variations in debt policy within
any industry group, there can be little real mean—
ing in an industry average. And what happens if

36

every other member of the group looks to the other

for guidance? The most that can be said for this

approach to debt policy is that the company con-

cerned can avoid the appearance of being atypical

in the investment market so far as its capital

structure is concerned.

Unlike most critics of ratio analysis, Donaldson does
prOpose an alternate measure Of financial position.

However, in his judgment financial position is synonymous
with defensive position. His technique centers on pre-
paring the firm for any "reasonably" possible future ex-
tended economic downturn and is not concerned with priming
the company to take Optimum advantage of economic and mar-
ket prosperity.

Although ratio analysis is strongly criticized, no
viable alternatives for measuring total financial position
have been found. Resultantly, academicians are again
testing empirically the utility of ratio analysis. Baruch
Lev used the financial statements on COMPUSTAT and the
changes in these statements over time to test the hypothesis
that firms manage their financial position by periodically
adjusting their financial ratios to industry averages.44
He concludes that they do. The implication of his work
is that, as many finance people suspected, firms use
ratio analysis to manage their financial position from a
corrective orientation.

Beaver analyzed the ability Of individual ratios to
accurately classify firms as "failed" and "non-failed",

45

using a classification test. The five ratios he found,

in order of predictive strength are: Cash flow/ total debt,

37

net income/total assets, total debt/total assets, working
capital/total assets, and the current ratio. Beaver notes,
however, that since several of these ratios are highly
correlated, the predictive ability of certain ratios could
be their correlation to other "strong" predictors.

Altman built a multiple discriminant analysis model
based on ratios which he found to be much more accurate in
predicting failure than are the individual ratios standing

alone.46

The five ratios in his most predictive model are:
working capital/total assets; retained earnings/total
assets; earnings, before interest and taxes/total assets;
market value of equity/book value Of total debt; and sales/
total assets. However, while these five ratios in com-
bination constitute the most accurate predicting model,
they are not individually the five best predicting ratios.
Therefore, Altman's conclusions are not necessarily in-
consistent with those Of Beaver. Both authors conclude that
financial ratios are significant tools for analyzing finan-
cial position. Several variables in combination should
Offer a better prediction than any single variable.

Beaver and Altman considered financial position only
as the defensive posture of the firm. Carlson recognized
this shortcoming and, borrowing from Altman's conclusion
that several variables are better predictors than any
single variable, developed a model to predict "financial

47

efficiency" based on several financial ratios. The

individual ratios in his model, which he explains are not

38
all superior predictors in univariate analysis, are:
the current ratio, the cash turnover, the inventory turn-
over, the debt ratio, and the dividend payout ratio. He
concludes that financial ratios used in a proper research
framework are valuable tools both for management planning
and control and for financial analysis by outsiders such
as bankers, credit managers, and investors.

In conclusion, financial ratios are not academically
the ideal technique for measuring financial position.
However, empirical research evidence indicates that in-
dividual ratios can do a reasonable job of evaluation and
that jointly, they are extremely accurate. It is also
demonstrated empirically that ratio analysis is used in
managerial control and analysis and no technique is shown

to do a better job in practical analysis.

Anticipating the Effect of Decision
Alternatives on Financial Position

 

Financial statements and ratios are based on past per-
formance and as such lack a crucial dimension as informa-
tion tools for effective financial decisions. Thomas Neff
alludes to this problem:48

He (the financial executive) should use history

as a benchmark but concentrate on the future.

His most valuable contribution is not in telling

management where they have been, but in telling

them where they are going.

Neff is referring specifically to the use of pro-forma

financial statements for planning purposes. Pro-forma

financial statements are projections of how the financial

39

statements will look in the future as the company Operates
over time. The importance Of pro-forma statements for
planning purposes is emphasized by Raymond Kent:49
A projected balance sheet makes it possible to
judge to what extent an enterprise's financial
position will be strengthened or weakened as a
result of its planned activities up to the date
for which it is planned. .....Often, projected
balance sheets reveal errors in planning or
expectations not readily discoverable elsewhere.
This does not mean, however, that pro—forma statements
are incorporated into the decision making process. They
are used for planning purposes only. This literature
review substantiates that financial planning, asset
management decisions, and management Of financial position
are performed within the firm as separate functions.
Financial position is typically managed on an
exceptions basis. That is, individual items are adjusted
when the ratios deviate from prescribed standards.
Financial management by exception is concerned with de-
tecting areas of over - and under - investment, lack of
capital, trends toward insolvency: measuring what has been
and comparing it to what Should be. The research by
Baruch Lev reported earlier demonstrates that management
by exception (corrective management) is the technique used
today to manage financial position.50
In the past such techniques were appropriate because
so much potential existed in "skimming the cream" that

detailed attention to subtle, hidden relationships could

hardly be justified. However, the texture Of industrial

40
life has changed. As Tucker explains:

Industrial competition has steadily moved into

the subtler areas of manufacturing structure to
where the little things are important. Profit
making now comes from unmasking hidden costs, from
policing against insiduous creeping change, from
lighting up a hitherto unknown profit area.5l

Support is building for financial management from a pre-
ventive rather than from a corrective vieWpoint. In
discussing the importance Of financial balance, Wright
says:52
Financial managers too Often await results of
operations before seeking to adjust financial
position. This is too late in time to begin
efforts to control financial position.
In writing on the future direction of financial management,
Jaedicke states that in the future knowing what causes
items to change and financial position to decay and
taking steps to prevent such changes will become a major

responsibility of financial management.53

Concluding Remarks

 

This research is concerned with understanding the
causes of change in financial position. By understanding
the causes, the changes can be anticipated, considered in
decision analysis, and prevented if so desired. Many of
the causes are exogenous. Such environmental variables as
economic conditions, industry sales, and cost structures
are largely out of control of the manager. However, a
premise of this research is that much Of the change in fi-

nancial position is the result of overt decisions in

41

functional areas of the company. Another underlying
premise is that most Operational decisions within a
company trigger changes in many of the variables which
make up the financial position of the company. A major
step toward more enlightened financial management, then,
is to provide management with a capability of predicting
the total financial impact of various classes of decisions.

Tucker made a significant first step toward
scientific specification of total financial impact.54
Through the use of a multitude of ratios, he developed a
large number of general relationships among ratios and
developed composite indices by which management can pre-
dict the total financial effect of decisions made in the
production and sales areas.55

HOwever, he omitted a major area of the Operations
system of the firm - the distribution sub-system. This
research studies the effect Of a specific class of
decisions in the distribution sector, warehousing de-

cisions, on the financial position Of modeled firms in

two industries.

CHAPTER II--FOOTNOTE REFERENCES

1James Van Horne, Financial Management and Policy

(Englewood Cliffs, N. J.: Prentice Hall, Inc. 1968), p. 198.

2Allan E. Grunewald and Erwin Nemmers, Basic

Managerial Finance (New York: Holt, Rinehart, and Winston,
1971), p. 67.

3Van Horne, 9p. cit.

4Eli Schwartz, "Theory of the Capital Structure
of the Firm," Journal of Finance (March, 1959), p. 18.
5J. Fred Weston and Eugene F. Brigham, Managerial
Finance (New York: Holt, Rinehart, and Winston), p. 181.

6Van Horne, 9p. cit.

 

 

7Weston, 9p. cit.

81bid.

9James Tucker, Managerial Control Through Ratio
Analysis. (New York: McGraw-Hill, 1963). P. 211.

loHarold Bierman, Jr., "Measuring Financial Liquidity,"
Accounting Review (March, 1960), p. 628.

llArthur Dewing, Corporation Finance (New York: The
Ronald Press, 1931).

12

 

 

 

Bierman, pp. cit.
l3James Walters, "Determination of Technical Solvency,"
Journal of Business (January, 1957), p. 30.

14Van Horne, pp. cit.

' 15Gordon Donaldson, Cogporate Debt Capacity
(Cambridge, Mass.: The Harvard Press, 1964).

16Felix Kollaritch, "Can the Balance Sheet Reveal
Financial Position?", Accounting Review (July, 1960),
p. 482.

 

 

42

43

17Ezra Solomon, The Theory Of Finance Management

(New York: Columbia University Press, 1963).

18Eugene Lerner and Willard T. Carlton, "The
Integration Of Capital Budgeting and Stock Valuation,"
American Economic Review (September, 1964).

19Albert W. Wright, "Maintaining Balance in
Financial Position," Management Accounting (September,
1969), p0 150

 

 

20Dewing, pp. cit.

21Benjamin Graham and David Dodd, Security
Analysis: Principles and Techniques (New York: McGraw
Hill, 1951).

 

 

 

22Harry Guthmann and Herbert Dougall, Corporate
Financial Poliqy (Englewood Cliffs, N. J.: Prentice-Hall,
Inc., 1955).

23

Robert Sprouse, "The Measurement of Financial
Position and Income: Purpose and Procedure," Research
in Accounting Measurement (Jaedicke, ed. American
Accounting Association, 1966), p. 43.

24

 

Guthmann, pp. cit.

5Weston, pp. cit.

26Van Horne, pp. cit.

27Bierman, pp. cit.

28Roy Foulke, Practical Financial Statement Analysis
(New York: McGraw-Hill, 1926).

29

 

Tucker, pp. cit.

30James O. Horrigan, "A Short History of Financial
Ratio Analysis," The Accounting Review (April, 1968), p. 285.

31Alexander Wall, How To Evaluate Financial Statements
(New York: Harper and Brothers, 1936), P. 68.

32

 

 

Foulke, pp. cit.

33Stephen Gillman, Analyzing Financial Statements
(New York: The Ronald Press, 1925), pp. 111-112.

34Horrigan, pp. cit.

351bid.

44

36Charles L. Merwin, Financing Small Corporations

(National Bureau Of EconomiE Research, 1938).

 

 

37Graham, pp. pip.
38Walters, pp. cit.
39Bierman, pp. pip.
4ODonaldson, pp. pip.
41.1.2101-

42

Edward I. Altman, "Financial Ratios, Discriminant
Analysis, and the Prediction Of Corporation Bankruptcy,"
Journal of Finance (September, 1968).

43Donaldson, pp. cit.

44Baruch Lev, "Industry Averages as Targets for
Financial Ratios," Journal Of Accounting Research
(Autumn, 1969).

45William Beaver, "Alternative Accounting Measures
as Predictors of Failure," Journal of Accounting Review
(January, 1968), p. 113.

46Altman, pp. cit.

47C. Robert Carlson, "A Financial Efficiency Model,"

(unpublished dissertation, Michigan State University, 1971).
48

Action,

49Raymond Kent, Corporate Financial Management
(Homewood, Illinois: Richard D. Irwin, Inc. 1969), p. 274.

50Lev, pp. cit.

 

 

 

Thomas Neff, "Financial Reports Aimed at Management
" Financial Executive (January, 1967), p. 26.

 

 

SlTucker, pp. cit.

 

52Wright, pp. cit.
53Robert J. Jaedicke, Charles P. Bonini, and Harvey
M. Wagner, Management Controls: New Directions in Basic
Research (New York: McGraw—Hill, 1964), p. x.
.54

Tucker, pp. cit.

 

55

Ibid.

45

CHAPTER III

LREPS: FINANCE VERSION

Introduction

 

The purpose of this chapter is to describe the version
of the LREPS model capable of modeling the finance sub-
system. ‘Although the basic LREPS model was built based
on an extensive data audit of a sponsor company, LREPS -
Finance Version (LREPS-F) is not intended to imitate the
Operation of a "real world" company. Rather, information
was drawn from several different sources and the model was
pretested with the Objective Of building a reasonable ap-
proximation of a "typical" firm in each industry. In
instances where an input to the model is not important to
the results Of test questions, the value Of that variable
was estimated.

The validity of the model was evaluated by comparing
the output of test runs to industry average Operations and
financial statistics reported in several data sources.

This chapter describes LREPS-F with the data input and model
parameters for simulating the hypothetical "typical firm"
within the health care products wholesale industry. The
changes which are necessary to simulate a typical firm in

the appliance wholesale industry are explained in Chapter V.

46

47

General Features Of LREPS-F

 

The model is programmed in GASP II-A, an event oriented
simulation language. It is divided into three overlapping
sub-systems. A sub-system is a combination of model com-
ponents and activities linked in a logical arrangement.
Components are the functional entities Of the firm (e.g.
demand component and transportation component) and are
formed in the firm by combining activities. An activity,
the most basic element in the model, is the process of
action in the firm. Activities involve the interplay among
the most elementary inputs and outputs. The values of the
material, monetary, or informational flows which the inputs
and outputs involve are termed variables.

There are three general types of model variables:
exogenous, status, and endogenous. An exogenous variable
is an input variable, the value of which is not changed by
the Operation of the model. A status variable describes
the state Of a system component (or one Of its attributes)
for any given point in time. The values Of these variables
change over time as Operations proceed. For example, a
balance sheet item is a status variable. The balance
sheet status variable "cash" describes the value of a
system component, cash, as of a given point in time. An
endogenous variable is a model variable whose value is
dependent on the operation Of the model and is considered
an output of one phase of the model and an input to another

phase. A given variable is Often a status variable for

48
certain purposes and an endogenous variable for other pur-
poses.

The relationships among variables and components are
modeled through algorithms. An algorithm is the quantita-
tive or logical description of the interaction among vari-
ables, activities and/or components; and it is the func-
tional form by which an activity processes inputs into out-
puts. A routine is a computer sub-program which links
activities into components and/or activities, components,
and variables into a sub-system. A routine is usually a
logical arrangement Of several algorithms. The sequencing
of routines is the process by which the model is made to
operate a meaningful order.

Events are the timing mechanisms which activate
routines within the model. Each event usually activates
several routines, which are sequentially attached in some
orderly framework. The occurrence of events is controlled
by the Executive Routine Of the Monitor and Control Sub-
system. The events which are to occur on a particular day
are stored in an event file and are picked by the executive
on the following basis:

1. All variable events are picked on a first in
first out basis.

2. All fixed time events are picked in the order
in which they are "fixed" to occur.

Fixed time events are events which occur with some pre—
scribed regularity, such as daily, weekly, and end—of—quar-

ter. Variable events are events which occur Sporadically

49

as the value of certain status variables reach prescribed

levels.

LREPS-F includes eight fixed time events and four

variable time events, which are the following:

1.

10.

11.

The event that initializes the LREPS-F
Operating System for start of LREPS-F
simulation planning horizon cycle,
Beginning-Of—Cycle, BOCYC.

The event to initiate the normal sales
processing activities for the day, DAILY.

The event to initiate the processing Of
the end and beginning of week activities,
WEEKLY.

The event to initiate the processing Of
beginning-of—month activities, MONTHLY.

The event to initiate the processing of

the end and beginning-of—quarter activities
required for Operating System information
input/output and control of the feedback
responses, QUARTERLY.

An event similar to the Quarterly event,
but with some additional half-year
activities, HALF-YEAR.

An event similar to the Half-Year event,
but with some additional yearly activities,
YEARLY.

The event that completes the LREPS Operating
System activities of a simulation cycle and
generates the required information to
terminate the execution, End-Of—Cycle,
(EOCYC).

A variable event that initiates the Oper-
ations Sub-system processing of a MCC order
arrival from the DC, MCC-Order-Arrival,
(MCORA).

A variable event that initiates the Oper—
ations Sub-system processing of a DC shipment
arrival, DC-Shipment Arrival, (DCSHPAR).

A variable event that initiates the process

50
of paying trade accounts generated by the
receipt of replenishment inventories,
ACCPYDS.

12. An event to imitate the payment of principle
on long term debt at the end of the seventh
week and every thirteenth week thereafter,
QTRPRP .

The clock (CLOCK) is the driving mechanism which
identifies the day in process and by which the model moves
from one day to the next. When all events listed in the
event file for a particular day (T-NOW), have occurred,
CLOCK advances to the next time and the events filed for
occurrence at T-NOW + l are processed.

Events do not activate all of the routines in a sub-
system at the same time and Often routines from more than
one sub-system are called by the same event. The remainder
of this chapter describes the routines within each system

and sub-system. The event by which each routine is acti-

vated is identified in the associated discussion.

The Supporting Data System

 

The Supporting Data System is the avenue by which all
exogenous information is inputted to the model at the
beginning of each experimental run. The Specific facilities
which exist in the distribution system scheme and the number
and location of suppliers are identified exogenously in
LREPS-F. Changing the value of exogenously inputted data
and the location of fixed facilities in the system through
the Supporting Data System is the manner in which experi-

mentation is performed on the model.

51

To simplify the discussion, the data listed in the
Supporting Data System, and a brief comment on the analysis
required to prepare it, are discussed in conjunction with
the particular sub-systems of the Operations system which

the data supports.

Qperations System

 

The Operations System is the system of the model in
which the imitation of the firm occurs. It traces the flow
of materials, money, and information (demand, Operational,
and financial) through the structure of the distribution
system. For the health cares company, the distribution
system consists of 125 customer demand units in a twelve
state region; supplier's (MCC'S) at Chicago, Philadelphia,
and Dallas; a distribution center (DC) at Chicago; and, in
eighteen experiments, a DC at Columbus, Ohio. The Operations
System is described according to the sub-systems and routines

'of which it consists.

Operating SyStem: Demand and Environment Sub-System

The primary purpose of this sub-system is to imitate.
the volume characteristics and product breakdown of orders
flowing from the customers to the DC'S Of the modeled firm.
The sub-system Operates in four separate but over-lapping
routines, which are:

1. Generation of daily sales for the total
company.

52

2. Assignment of daily sales to individual
DC's.

3. Determination Of the volume and geographic
destination Of each other and the volume Of
each individual order sold.

4. Summaries at the end of each day of the
total volume and product characteristics
of sales at each DC.

Generation Of DailypSales. Daily sales (DLSLS) are

 

generated in the model by randomly selecting a normal

deviate (Z score) and plugging it into the formula

+ *.._O_

DLSLS = 21

ppm

M and. O are the mean and standard deviation respectively
of the probability distribution from which monthly sales
are being sampled for the month in process. Twenty-one

is the number of working days in each month. The twenty-
one daily sales points selected in a month will total to

M, actual monthly sales, which will be the same as M only
by chance. M is a random variable Of the joint probability
distribution which is Obtained by combining the twenty-

One daily probability distributions, which are defined by

O o

:flZI

The value Of M for any given month is determined by

the formula .
pp_1

Ml(l+12

 

W.
1

where

3|
N

the mean of the probability distribution
from which monthly sales will be generated

53

for the i th month.

3|
N

1 the mean monthly sales for the first month
in any run.

r = annual growth bias in sales.

1 = the month for which sales are being generated.

Wi = a weighting factor between .1 and 2.0 used
for building seasonality patterns into the
experimentation.

The value of o for any given month is determined by the

 

r i
formu1a 01(l+ 1—2—)
Oi =
W.
1
0i = the standard deviation for the probability

distribution from.which monthly sales will
be generated for the i th month.

01 = the standard deviation for the first month
in any run.

r,i,Wi = as defined above.

0 is allowed to grow at the same rate as M to assure a

constant value for the co-efficient of variation.
The input parameters for this routine for the health

cares company were determined by studying the monthly
sales patterns of the sponsor company, calculating the
standard deviation of monthly sales after the growth bias
had been removed, and determining the standard deviation Of
daily sales which would be required to attain the identified
monthly standard deviation. NO significant seasonality

characteristics were found in the monthly demand patterns.

54

Sales Assignments to DC'S. Daily sales assignments

 

are made to an individual distribution center based on the

following formula:

 

n
.:l WTIDX(DCiDU.)
DLSLS (DCi) = DLSLS (TOT) * 3m 3
p p z WTIDX(DUj)
j=1
where
DLSLSp(DCi) = sales for day p at DCi.
DLSLSp(TOT) = sales for day p for total company.
n
2 WTIDX(DCiDU.) = the sum of the weighted indices for
j=1 3 all DU's assigned to DCi.
m
X WTIDX(DU ) = the sum of weighted indices for all
j=1 J DU's in the company.

The weighted index for each DU is based on the popula-
tion of the DU and the relationship found in a cross
section analysis between population patterns and sales
for the sponsor company. Several independent variables
were examined in the cross section analysis and population,
with a correlation co-efficient of .95, was found to be
the best predictor of sales.

Demand Unit Sales Assignment. Total DCi sales are

 

assigned to individual customers by randomly picking orders
from an order file and randomly assigning them to a DU. The
relative probability that any DUj will have any given order

assigned to it is
WTIDX(DUi)

 

M23

WTIDX(DC.DU.)
j=1 1 3

55

A basic assumption is that all customer orders emanate
from the center of DUj. Each postal zip code in the twelve
state study region is assigned to a DU. The center of each
DU is a Zip Code Section Center, (ZSC) although each ZSC
is not necessarily the center of a demand unit. The demand
unit is identified by the number of the ZSC which serves
as its center and by the longitudinal - latitudinal bearings
Of that ZSC. The 1970 population of each DU is listed
in the data base and the running model updates the popula-
tion each running year based on Rand McNally geographical
area population growth rates.

Order Picking. The order file (Supporting Data System)

 

from which orders are picked consists Of ten-order blocks
of sample invoices. The sample Of invoices was systemati-
cally selected from the files of the Sponsor company and
the sample size was determined with Type I and Type II
errors set at .01. Orders are randomly selected ten at a
time and 10 percent of the total sales dollars and weight
in an order block are randomly assigned to DU's ten
different times.

Specific sales information is summarized by the DCi
level and is kept only for fifty representative or
"tracked" products. Because of the cost and impracticality
of keeping track of all of the products sold by the
company, these fifty products were selected to be tracked
from day to day for inventory management, shipment and

costing purposes. Each tracked product represents a group

56
(of products and extrapolation factors are used for deter-
mining total product sales, inventory, shipment, and costing
patterns based on the patterns of the tracked products.
The characteristics of each tracked product (sales value,
pounds, cube, and cost of goods sold per care unit) are
listed in the order file Of the data base.

Summary Of Output. The output of the Demand and En-
vironment sub-system serves as the input to the Operating
sub-system. The information is summarized as follows:

1. Total daily sales for the company, in

dollars, weight, cube, orders, and
line-items.

2. Total daily sales for each DC in

dollars, weight, cube, orders, and

line-items.

3. The total sales for each DU in dollars
and weight.

4. The sales volume for each tracked
product at each DC.

Qperating System: Finance Routines

Finance is not a model sub-system in LREPS-F. Rather,
routines associated with cash and financial information
flows are located in each Of the sub-systems of the
Operations system. To provide an overview of the total
finance function in LREPS-F, a general outline of the
routines associated with cash and financial information
flows are presented in Tables 3.1, 3.2, and 3.3.
Detailed description of the routines is deferred to the

particular sub-system in which each routine is located.

TABLE 3.1.--Financial Statement Items and the
Associated Operations System Routines

 

 

 

Item Code Associated Routines
Cash CASH PAYAC, ARADJ, LAMGT, CSMGT,
FTXPY, CPBUD, DIVPY, INTRS,
MKSCY, PRRPY, TOPRX

Marketable Securities MKTSC CSMGT
Accounts Receivable ACCRC DLSLS, ARADJ
Inventory INVTY IDCGS, INRPL
Fixed Assets PLNEQ CPBUD, TOPRX
Total Assets TOTAS (RESIDUAL)
Accounts Payable ACCPY PAYAC, INRPL
Taxes Payable TXPAY FTXRT, FTXPY
Deferred Taxes DEFTX FTXRT
Short Term Loans STDBT LAMGT
Current Maturities -

Long Term Loans LTDTC PRRPY
Long Term Loans LTDBT PRRPY
Equity SHEQT EAFIT, DIVPY
Sales SALES DLSLS
Cost of Goods Sold CSTGS IDCGS
Gross Margin GMARG (RESIDUAL)
Administrative and

Selling Expense ADSLX TOPRX
Outbound Transporta-

t1on Expense OTBDX TOPRX

TABLE 3.1.--Continued

 

 

 

Item Code Associated Routines
Inbound Transporta-

tion Expense INBTX TOPRX
Throughput Expense THPTX TOPRX
Communications Expense COMMX TOPRX
Straight Line

Depreciation SLDPX TOPRX, FTXRT
Interest on Debt CSDBT INTRS
Net Operating Income NOINC (RESIDUAL)
Marketable Securities

Income MSINC MKSCY
Marketable Securities

Expense MKSCX CSMGT
Net Income EBFIT (RESIDUAL)
Federal Income Taxes FINCT FTXRT
Net Income, after

Taxes EAFIT NETIN

 

59

.moaom>floomu Ou mmamm

 

 

 

 

 

msmeou can Imumamo Omooa omooa San
ION Oumumcmm Sasoocmm ammmo mmqam mamas Imonazm enema mucmSumsflna means
I>eoomm monsooo<
NEZOU
memB
xomzH
meBO
.mmaou Damssmm Hoe «ammo xqmoa s<a
Omaspmnom mucsooom smm swoop we>zH qmmzH ImOIozm eased monsooum
mo ucmsamm
.anmwmm monsooom
Ommouocfl .muoucm>sfl wavud
mo Hm>auum choomm >e>zH qmzmo mammmoa qmmzH Damsnmflcmam
Imm muoucm>oH
.>Houco>ca Eoum oomuu
Loom one oHOm wows: we>zH
mo umoo wumasoamo moemo mmqam mqmoo quao mOOOH vaom mpooo
mo omoo .coflu
Imammp muoucm>cH
.m.oo
pom m.Uo on cmflmmm one
mmamm Saame mumumcmo mmqam qmzmu mOOOH quao woman mmamm Seamo
ooflumomamxm osmooo oode mwawuoom pdm>m oooo memz
mmabmflum> poumwoommd Ocfluoom
ooumwoommd
mOHomHHm>

poumfioommd Memes pom mmcfluoom omumfloomma wooedwm Ono mo 3mfl>um>o mumssdm «II.N.m mqm<a

60

.mmmo HOHHOO CCOH

 

 

 

omumaosooom we mmumu mmmo oqood mezoz
Dmmumuafl Refine Samauasz emomo omega am>>¢ Imonozm mmezH DmmumucH
smoom
mmao
xmaqm
.maommmm monsooom pom x2200
ammo on muomeumonom xamme
museumoummm oxmz meZH
.xmoz How mmmcmmxm xomeo omzam mums mmcfluoom omomm
mcflumummo Ham mumasoamo xqmoa qmzmo Imouozm Smaoe Ixm meaumumdo
.mOmomuom
mdflumoo pom OEOOGH
How maomflum> mofluma
IoEooom mumflumoummm Op amend emcee
pasooum amoa .suflusomm omega smasm was Guava: mannaum>
comm ca ucsosm pea manna Owes: Imouozm ama>< masnmasssooa
.emnflsamu ma Owes: noes xeomz
umsflwm .mufieaa mane Owes: wan
Lummoom «mmao mummsoo «mmmo «mmao Leonozm eozmo usmgwmmcmz ammo
.OOHHDOOH we mcmoH
sum» uuoam sues Dmsflca
.muflsfla manmummoom ammao sea Demsmmmamz
OD mmOHq mummsoo emoem maoqu ILOIozm Hosea mummmm cflswaq
doflumomamxm pseudo MDQSH mmdfluoom ucm>m mooo mEmz
mmaomflum> pmumfloommm Odfluoom
pmumHOOmmm

 

 

pmocHDGOUII.N.m momma

61

 

 

 

 

.DGOEQHDOO Omzqm mm44m mmmw
om: + ucmsmomammm mmao xmoqm Imonozm cameo mcflummasm Hmuaamo
.mmxmu umumm eommm mmemaoo Damssmm
.mmcflcumm Dmc mo won mmao eHmam Imouozm wm>Ho newnfl>ao
.Oeoeq
ou emoeq some mamm
ummmcmne .mmflufiuoums mmemmbo pomswmmmm
ucmuuso CO «\H and emoeq oeoeq ImOIozm semen SHSROCHHN
.manmsmm mmxmu cam ammo ammo mmemaao
soup DOMHDQSm Immxmu sea samxa samxe Imouozm waxes Damssmm xme
hmmxmu omnnmwoo ou
Axmoqumoowmv mo N\H pea
.moomm co memn mane xaemo mezoz waxes
used mmxmu mamasoamo samxe moowm ILOIozm emxem meoucH Hmumamm
.GOADOHO
IOHQOO mafia uemflmuom
co comma .mmxmu sebum mzmeH ezmzmeaem mezoz
.msoucfl um: mamasoamo mzoozH age Imouozm zHemz msoocH umz
mmmomnom
xmu How coflomfiooummo mezoz doHumHomummo
muflmfle mummsIeoISSm xaosm omzqm Imouazm moasm emumHsSsoom
.mmmp HMHHOO
.w.2 amenasssoom an mmmo mezoz OsoocH mmauau
assume Seams Samfluasz oszz mange om>>a Imouozm some: Isomm mabmumxumz
Soflumcmamxm pseudo oomcH mmdwusom ucm>m OUOU mamz
mmabmflum> omumfloommm oofluoom
omumfloommm

 

 

poocflucoolI.N.m mqm<e

62

TABLE 3.3.--Financial Variable and

Routine Code Glossary

 

 

 

Code Description

ACCPY Accounts Payable

ACCRC Accounts Receivable

ADSLX Administration and
Selling Expense

ARADJ Accounts Receivables
Adjustment Routine

AVUPD Accumulating Variable
Update Routine

CASH Cash

COMMX Communications Expense

CPBUD Capital Budgeting Routine

CSDBT Interest on Debt

CSMGT Cash Management Routine

CSTGS Cost Of Goods Sold

DEFTX Deferred Taxes

DIVPY Dividend Payment Routine

DLSLS Daily Sales

EAFIT Net Income, after taxes

EBFIT Net Income

FINCT Federal Income Taxes

FTXPY Tax Payment Routine

FTXRT Federal Income TAXES Routine

GMARG Gross Margin

IDCGS Inventory Depletion, Cost

of Goods Sold Routine

63

 

 

 

TABLE 3.3. --Continued

Code Description

INBTX Inbound Transportation
Expense

INRPL Inventory Replenishment

INTRS Interest Routine

INVTY Inventory

LAMGT Liquid Management Assets
Rout1ne

LTDBT Long Term Loans

LTDTC Current Maturities -
Long Term Loan

MKSCY Marketable Securities
Income Routine

MKSCX Marketable Securities
Expense

MKTSC Marketable Securities

MSINC Marketable Securities Income

NETIN Net Income Routine

NOINC Net Operating Income

OTBDX Outbound Transportation
Expense Routine

PAYAC Payment Of Accounts
Routine

PLNEQ Fixed Assets

PRRPY Principle Repayment

SALES Sales

SHEQT Equity

SLDPX Straight Line Depreciation

64

TABLE 3.3.--Continued

 

 

 

Code Description

STDBT Short Term Loan

SYDDP Accumulated Depreciation
Routine

TDPRX Operating Expense
Routine

THPTX Throughput Expense

TOTAS Total Assets

TXPAY Taxes Payable

 

65

A summary of the financial flows is presented in
Tatiles 3G1. and 3.2. Table 3.1 lists the items in the
balance sheet, and income statement, their code names, and
thee code name of routines in the model which directly cause
change in that item.

Table 3.2 ;presents a schematic of the finance -
amssociated routines, approximately in the order in which
tfliey occur in the model. The information presented in
Table 3.2 includes: the code name of the routines; the
event which activates the routine; the name of the routine;
the code name of the finance variables and finance related
‘variables which link to it in the model, both input and
output; and a brief explanation of the routine. In Table

3.3, each code name used in this section is listed in
alphabetical order along with the variable or routine which
it represents. Discussion of the financial ratios and
Special statistics, which are calculations based on data in
the financial statements, is deferred to the Report

Generator System.

Opprating System: Operations Sub-System

The purpose of this sub-system is to imitate the flow
of orders, material, cash and information through the
distribution system. The first activity is the arrival of
batches of orders at the distribution center as an output
of the D&E sub-system. This section described the

Operations Sub-System through a discussion Of the routines

66

trixggered by the arrival of orders, which are:

l. The Order Processing, Preparation, and
Shipment Routine.

2. The Inventory Depletion - Cost Of Goods
Sold Routine.

3. The Inventory Management Routine.
4. The Inventory Replenishment Routine.

5. The Inventory Arrival - Accounts Payable
Adjustment Routine.

6. The Accounts Receivable Adjustment Routine.

7. The Liquid Asset Management Routine.

8. The Cash Management Routine.

9. The Accumulating Variable Update Routine.

The Order Processing, Preparation and Shipment
Routine. This routine is activated by the DAILY event.
Figure 3.1 traces through the physical flow of a batch
of orders from a DU from the time it arrives at a DC. The
available inventory for each tracked product on the order-
is checked to determine if the line item can be filled. If
sufficient inventory is not available for one or more of the
products, stock outs are reported (see the Measurement sub-
system for stock - out level reporting). Order processing
and preparation are assigned a time lag Of two days and
shipment occurs at the end of this lag. All outbound
shipments are less - than - truck - load shipments by
common carriers. The elapsed transit time before the
order is received by the customers is generated in the

following steps:

67

 

Cm» mo
I

 

CALCULATE

CT1

l

 

 

 

 

CT

 

 

 

 

RECORD NO
‘ STOCKOUTS '

 

 

 

YES

RECORD
SALES
VOLUME

1

WEIGHT
DS-DU
LINK

I
c 3

 

 

 

 

 

 

 

 

 

Figure 3.l.--Flowchart: Processing Preparation,

Routine

Order Arrival.

Calculate - order
processing and
preparation.

Assign transit time
lag.

Is Inventory available
for all products?

Record sales volume in
appropriate order cycle
time accumulating
variable for the DC.

Record the weight of
this order in the

DC - DU. accumulating
weight vdriable.

Process another
order.

and Shipment

68
l. The distance from the DC to the destination

DU is calculated by calling the distance

routine, which reads longitude and latitude

of the DU and DC from a list in the data

base. The straight line distance between

the two points is calculated based on the

difference between their navigational

bearings and this distance is increased

by 17 percent to allow for highway circuity.

2. For each DC, three geographical radii are

listed within which the expected transit

times are l, 2, and 3 days respectively.

The appropriate expected transit time is

selected based on the particular radii

within which the DC-DU distance falls.
The transit, order processing and order preparation times
are summed to Obtain normal order cycle time and the total
weight for these orders are added to the weight accumulating
variable for this length of order cycle.

The real world times required for processing and prepar-
ing orders were determined by tracing an order through the
sponsor company and Obtaining estimates from the appropriate
manager of the time required for each activity. The transit
times over each MCC-DC link, and the maximum distances from
each DC which can be reached by outbound shipments in 24,
48, and 72 hours were gathered by polling the motor common
carriers serving the DC'S. The survey was conducted through

the Sponsor company.

The Inventory Depletion, Cost of Goods Sold Routine.

 

This routine (IDCGS) is activated by the event End-of—Day.
It is flow charted in Figure 3.2. At the end of each

day, the total sales units for each tracked product, DLSLS
(IPT), are subtracted from the inventory on hand (IOH) for

that product. The cost of goods sold for each product,

 

GLSLS (ITP) >

IOH

 

 

 

 

 

CALC
COGS

ACCOUNTING
ADJUSTMENTS

1
Cm D

 

 

 

 

 

 

 

 

69

Daily sales units individual
tracked product.

Subtract DLSLS (ITP) from inventory
on hand for previous day.

Multiply cost per unit of ITP and
extrapolate up for all products
represented by ITP.

Add cost of goods sold to
accumulating variable COGS,
MTD and to status variable,
inventory dollars.

Repeat this process for all
ITP.

Figure 3.2.--Flowchart: Inventory Depletion - Cost of
, Goods Sold Routine

70

CSTGS (ITP), is calculated by multiplying the number of case
units times the unit cost, which is listed in the data
base. CSTGS (ITP) is extrapolated up to reflect the total
cost of goods sold for the category Of products which it
represents, CSTGS (CATi). The sum Of CSTGS (CATi) for all
categories is added to the income statement variable "cost
Of goods sold" (CSTGS) and subtracted from the balance
sheet variable "inventory" (INVTY).

Daily Inventory Management Routine. Subsequent to
the adjustment of unit inventories by the amount of daily
customer shipments, the level of inventory on hand for
each tracked product, IOH (ITP) is checked to determine if
replenishment is necessary. This routine is flow-charted
in Figure 3.3. IOH (ITP) is compared to the re-order point
ROP (ITP) to determine if a replenishment for that product
should be ordered. ROP (ITP) is a variable parameter and
is calculated in the Monitor and Control Sub—system
Controller function.

If the answer is that IOH ROP, the replenishment
que is checked to see if replenishment has already been
ordered. If not, a replenishment order is placed for the
amount M - IOH (ITP), which is a variable also provided
from the M&C Controller function.

The Inventory Replenishment Routine. Figure 3.4
depicts the inventory replenishment shipment routine,
which is associated with the event MCORAR (MCC order arriv-

al). The total volume in an inventory replenishment order

71

 

 

 

IOH ROP

 

 

 

 

 

ORDER
EOQ

 

 

 

 

I
C 3

Inventory on hand,
individual tracked
product.

18 the level of
inventory units less
than the re-order
point?

Is this tracked
product already
re-ordered?

Place a replenishment
order for EDQ-T.

Select another
ITP.

Figure 3.3.--Flowchart: Inventory Management Routine

72

 

( RE—ORDER >

 

 

 

ADD TO
R—QUE

 

 

  
 

 

< RETURN

 

 

GENERATE
SHIPMENT

 

 

 

 

 

RECORD
WEIGHT

 

 

 

l
C m 3

Figure 3 . 4 .--Flowchart:

Inventory

Re—Order for EOQ-T.

Add EOQ—T to replenishment
que, MCC:k—DC:i.

Have ten working days
elapsed since the arrival
of the first order. If
not, go on to next routine.

Generate time lag to
ascertain day Shipment
will arrive at DC.

Record the weight of the
shipment in the appropriate
weight category for the

MCC - DC. link.
k 1

Replenishment Routine

73

is determined by extrapolating the amount ordered Of the
tracked product. However, the re-order quantity for most
of the products is sufficiently small in that even after
extrapolation a Single replenishment order will not normally
justify a truck-load inbound shipment. Thus, when a re-
plenishment order arrives at the Supplier (MCC) the total
ordered volume (EOQ-T) is placed in a que, which accumulates
the total weight Of outstanding orders for the particular
MCC-DCi link. Shipment of the orders in the que is
generated when the total weight reaches 24,000 pounds or
after ten working days from the day the first order is
placed in the que, whichever comes first. 24,000 pounds
is the minimum weight on which the truck-load rates are
applicable and ten is the number of days beyond which the
cost of additional buffer stock required to protect against
stockouts exceeds the expected gain from truck—load freight
rate savings.

When the replenishment shipment is dispatched from
the supplier (MCC) to the DC, a transit time lag is
generated. Each MCC-DC replenishment link is assigned a
transit time in the data base. An event is scheduled for
the arrival of the replenishment Shipment at Shipment date
plus elapsed transit time. The total weight Shipped is
also recorded and served as input to the measurement system.

The Inventory-Accounts Payable Adjustment Routine.

 

The purchase Of inventories creates an account payable.

Thus, the arrival of an inventory replenishment Shipment

74
triggers a series of adjustments in IOH (ITP), for each
tracked product which is replenished; the inventory account
(INVTY); and accounts payable (ACCPY). The Inventory -
Accounts Payable Adjustment Routine is summarized in Figure
3.5.

Upon arrival of the replenishment shipment, a time
lag for handling and processing is generated. Subsequently,
the replenished units for each tracked product are added
to IOH (ITP). The purchase cost for the tracked products
is calculated and extrapolated up to represent purchase
cost for the group of products which the tracked products
represent, PRCST (CATi). PRCST (CATi) is then added to
INVTY and to ACCPY. Payment of PRCST (CATi) to the
supplier is scheduled for the 25th following working day,
at which time this prescribed amount is subtracted from
CASH and from ACCPY.

The twenty-five day lag in Accounts Payable is
determined by dividing twenty-one days, the average number
of days sales in accounts payable for the industry on the
average, by .85. .85 is the quotient Of expenses which
generate accounts payable divided by sales. It is based
on the estimate that all of inventory purchases and trans-
portation expenses give rise to trade accounts; and that
50 percent Of the administrative and selling expenses,
throughput expenses, and communications expenses create
trade accounts. In the running model, accounts payable

turnover closely approximates the industry average.

75

 

(REP TWO

 

CALCULATE
TIME LAG

1

UPDATE
IOH

I

CALCULATE
PURCHASE COST

1

UPDATE
INVENTORY
DOLLARS

1

UPDATE
ACCOUNTS
PAYABLE

1

SCHEDULE
A/P

I
C 3

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Arrival of replenishment shipment
@ DC.

A time lag is generated for
handling and processing the
delivery.

Inventory on hand is updated by
the units of each ITP delivered.

The purchase cost of each ITP
is calculated, extrapolated
for product classes, and
summed.

The total purchase cost of the
recorder is added to inventory
dollars and to accounts payable.

A time is schedules for paying
the associated A/P.

Figure 3.5 --Flowchart: Inventory - Accounts Payable Update Routine

76

Account Receivables Adjustment Routine (ARADJ).

 

Accounts receivable (ACCRC) are created by daily sales.
They are depleted when payment is received. The amount Of
ACCRC paid Off each day is a stochastic variate and a
function of the total amount Of ACCRC outstanding at the
beginning Of the day. Since the health cares industry
average turnover of accounts receivable is thirty three
days, an average of three percent Of outstanding A/R is
paid Off each day. At the end Of each day, the percentage
of outstanding receivables paid-off is randomly selected
from an approximately normal distribution with mean .03 and
standard deviation .005. The standard deviation is esti-
mated. Receipts for the day are subtracted from ACCRC and
added to CASH. The dollar amount Of sales for the day
(DLSLS) are added to ACCRC. This routine is flow-charted
in Figure 3.6.

The Liquid Asset Management Routine. Total liquid

 

assets (TLQAS) are defined as cash plus marketable
securities. The preservation Of a prescribed amount Of
liquidity among the assets of the firm is necessary to
support Operations. In LREPS—F liquid assets fluctuate
daily within a prescribed band. If the level of liquid
assets at the end of a day is less than the lower boundary,
Tl’ liquidity is added through Short term borrowings, in
an amount sufficient to return the level of liquid assets
to a prescribed replenishment level, M. If the level

of liquid assets at the end Of the day exceeds an upper

77

 

3
l

A/R
TNOW~1

I

Level Of accounts receivable from
previous day.

“~aqppppp

 

 

 

 

 

SELECT Randomly generates outstanding
A/R T0 A/R for which payment will be
BE REC'D received.

UPDATE The amount received 13 added

to cash and A/R.

 

 

 

 

DAILY Reads in daily sales dollars,
SALES total firm.

 

Adds daily sales to sales, MTD,

UPDATE and A/R.

 

 

 

 

1
C mm)

Figure 3.6.--Flowchart: Accounts Receivable Routine

78

boundary, Tu’ a sufficient amount Of short term loans is
repaid to bring the level of liquid assets down to M.

The parameters for this routine are calculated Offline
and exogenously inputted. The value of M is the average
level Of liquid assets for a firm in the health cares
industry with annual sales of $10MM. The upper and lower
boundaries are calculated using a standard replenishment
inventory control formula.1 The mean and standard deviation
of the distribution of daily cash flows are calculated from
test runs of the model and a normal distribution is assumed.
A lag of two days to transact short term borrowings and re-
payments is assumed. Interest charges on outstanding Short
term loans are assumed to be 6 percent and per transaction
costs are assumed to be $100. An assumed management para-
meter is that the probability Of running out of cash in
the two day lag required to transact a loan should be less
than .01.

Cash Management Routine. The level of cash also fluc-

 

tuates daily within a prescribed band. The upper and lower
boundaries and the target level of cash are prescribed

with the same replenishment inventory control model used
for managing liquid assets. The level Of cash is ad-
justed by buying and selling marketable securities. Be-
cause of the absence Of marketable securities from the
asset structure Of the industries modeled, this routine

is inactive for the research currently performed.

79

The Accumulatinngariable Update Routine. The pur-

 

pose of this routine is to record the end-of—day value of
several status variables in accumulating variables for
costing and income reporting purposes. At the end Of each
day the amount in Marketable Securities is added to
accumulating dollar days of marketable securities (ADDMS).
The amount in short term loans and long term loans at

the end of each day are also closed into accumulating
dollar day variables, accumulating dollar days of Short
term loans (ADDSL) and accumulating dollar days of long
term loans (ADDLL). In the Measurement sub-system, these
accumulating variables are all multiplied by a daily
interest rate or a daily rate of return to Obtain weekly

revenue and cost totals for the associated expense items.

The Operations System: Measurement Sub-SyStem

 

The purpose of this sub-system is to measure the
service levels attained by the distribution system, to
measure the costs Of Operating the firm, and to prepare
all financial items for inclusion into financial state-

ments and reports.

Measures of Service.

 

The general levels Of service which are measured by
this research are stockout frequencies and total elapsed

order cycle times.

80

The Stock-Out Measurement Routine. If inventory is

 

insufficient to cover a line item on an order, a stock-out
is reported to the Stock-Out Measurement Routine. The
amount reported is the case units ordered on the line item,
and it is added to a status variable, case units back—
ordered (CUBO). At the end of each day, CUBO is added to
an accumulating variable, accumulated stock out days delays
(ACSODL). When a replenishment Shipment is received contain-
ing a back-ordered product, all back orders for that product
are removed from CUBO. At the end of each week, the follow-
ing service statistics are calculated for each DC and for
the total company: total number of stockouts, average
delayed days per stockout, and standard deviation delayed
days per stockout.

Average delayed days per stockout and standard devia—
tion days delays per stockout are calculated based on the

following formulas:

AVVDL == 3%?
SDDDL = WM%%SQ - AVVDL2
where

AVVOL = average days delay per
stockout

SDDDL = standard deviation days delay

ACSODL = accumulated stockout days
delay

WKSLS = weekly sales, in case units

MSADD = mean of the square Of accumu-

lated stockout days delays

81
These service statistics are calculated for each indivi-
dual DC and for the total company.

Total Elapsed Order Cycle Times. The generation Of

 

elapsed time per order for order processing and preparation
and for shipment transit times are explained in the
Operating Sub-system. The total sales dollars for each
order are added to the appropriate accumulating variable
for sales volume which is delivered to the customer in k
days from order arrival date, where k = 4, 5, 6, 7 or 8 or
more. The total order cycle time actually includes an
allocation for the average days delayed per order because
Of back-orders. However, this allocation is a constant
and does not affect the relative distribution. Thus, at
the end of the month, the percentage Of sales delivered
with order cycle times Of 4, 5, 6, 7, and 8 or more days
is calculated for a given distribution center by simply

dividing monthly DC sales into each Of the appropriate

accumulating variables.

Measures Of Cost

 

The process of costing operations in the model are
explained according to the function they serve. There are
two cost accounts which are associated with general
management. These are:

1. Administrative and Selling Expenses.

2. Cost of goods sold (purchase cost Of
inventories).

82
There are six components Of total cost to the physical dis—
tribution system:
1. The outbound transportation component.
2. The inbound transportation component.

3. The throughput component.

4. The communications component.
5. The facilities component.
6. The inventory component.

Finally, there are two expense areas which are common to
both functions. These are:

1. Interest expense.

2. Federal income taxes.
Cost-of—goods sold calculations are described in the
Operations sub-system inventory depletion routine. Each
of the other nine costing sub-routines are explained here
along with the financial activities associated with each
of the expenses.

Administrative and Sellipg Expense Routine. This

 

routine is activated by the Weekly event. Administrative
and selling expenses include all expenses incurred by the
firm except those expenses associated with the actual
physical distribution system and the purchase price Of
inventories or securities. Weekly administrative and
selling expenses (ADSEX) are calculated using an equation

of the functional form:

ADSEX = FADEX (1+9)t + VADEX * WKSLS

83

where
FADEX = fixed weekly expenses.
VADEX = variable cost per sales dollar.
9 = an annual growth rate in
fixed expenses.
t = the year of operation (l,2,....5)

ADSEX is added to the income statement endogenous
variable ADSLX. The amount .5 ADSEX is added to the
balance sheet status variable accounts payable (ACCPY) and
subtracted from cash. An event is scheduled for the twenty-
fifth following day to subtract the amount .5 ADSEX from
cash and from ACCPY to reflect payment of the accounts
payable generated.

The values of FADEX, VADEX, and g are used in the
model as residuals to aid in designing the model for the
base experiment as a reasonable approximation of the
modeled industry.

Outbound Transportation Expense Routine. A com-
prehensive survey of freight rates from each potential
distribution center to a sample of destination points in
the study was undertaken through the sponsor company. The
rates for the various weight breaks for the class of
products shipped were summarized into a weighted average
freight rate for each sample DC — DU linkage and these
weighted averages were regressed against the apprOpriate
distances. The resulting equations had correlation co-

efficients significant at the .01 test level. Based on

84
this survey, outbound freight rates are generated in the
model from equations which are of the general form:

FRTRAT (DC * B (1+R4)* DIST

l

t
= * *

where

the estimated weighted
average freight rate for
shipment from DCi to DUj.

FRTRAT (DCiDUj)

R1,R2 = adjustment factors used
for reflecting negotiated
rates.

R3,R4 = adjustment factors used

to reflect general shifts
in the transportation rate
structure.

t = the year (1,2,....5)
A,B = the regression co-efficients
from the sample rate

analysis.

DIST = the distance from DCi tO DUj.

The adjustment factors (R1, R and R4) are all set at

2' R3!
1.0 for this experimentation. The total outbound freight
expense at a particular DCi for a given week WKOTBTC(DCi)

is calculated from the formula

n *
WKOTBTC (DCi) = Z [FRTRAT (DCiDUj) VOLWTj]
j=1
where
VOLWT. = the total sales volume in pounds
3 shipped from DCi to each DUj.
n =

number Of DU's served by DCi'

The total weekly outbound freight expense for the company

(WKOTBEX) is

IIM a

WKOTBTX (DCi)

j l

85
where
m = the number of DC's in the system.
WKOTBEX is added to the income statement variable OTBDX and
to accounts payable (ACCPY). Payment of the appropriate
trade account is scheduled for the twenty-fifth day after
the expense is occurred.

Inbound Tranpportation Expense Routine. Each DC
receives replenishment shipments from three supplier's
(MCC'S) which are located in Chicago, Philadelphia, and
Dallas. Motor common carrier freight rates were Obtained
through the Sponsor company for each Of the Six MCC - DC
supply links (three MCC's to two DC'S) for three different
weight categories, which are: under 5,000 pounds, 5,000
to 24,000 pounds; and over 24,000 pounds.

Replenishment shipments are recorded in the Operations
sub-system according to the weight category in which the
shipment falls. The total weekly inbound freight expense
from a particular DCi [ITCST (DCi)] is calculated from
the formula
3 3
ITCST (DCi) = z 2

[FRTRAT (MCC -DCi,CATe * VOLWT...

k
k=1 l=l

...(MCCk-DCi.CATe)]

where

FRTRAT (MCCk-DCi,CATe) = the freight rate for shipments to
DCi from MCCk for weight category
e.

86

VOLWT (MCCk-DCi,CATe) = the total replenishment volume over

the linkage which moved in ship-
ments Of weight category e.

The total weekly inbound transportation expense for the
company (ITCST) is the sum of ITCST (DCi) for all DCi.
ITCST is added to the income statement variable INBTX

and to accounts payable. Payment of the associated trade
credit is scheduled for the twenty-fifth day after the
expense is incurred.

Throughput Expense Routine. Throughput (or facilities
operation) expense is the cost Of the physical activity
of loading, unloading, handling and storing inventories.
For privately owned and Operated warehouses it includes
the following cost activities:

1. Unloading inbound shipments.

2. Handling and moving the inventories within
the warehouse, with related costs such as:

a. Lift truck rentals, maintenance,
and operating costs.

b. Packing materials.
c. Automated picking units.
3. Taxes and utilities.
4. Warehouse management and clerical salaries.
5. Loading outbound shipments.
The sponsor company owns and Operates five warehouses
of sizes ranging from three million dollars volume annually
to one hundred million dollars annually and ranging in

degree of automation from manual picking to 80 percent

87

automated picking and loading. The private warehouse
throughput expenses used in modeling this industry are

based on a comprehensive analysis Of the elements Of the
standard costing reports from each Of these warehouses

over a three year period. The analysis involved examination
of changes in cost characteristics between different types
and sizes of warehouses; Of the upward shifts in the ex-
pense structure of the warehouse over the three year period;
and Of the changes in operating costs resulting from

sales volume increases.

For the volume through public warehouses, throughput
costs are the charges against the company by the public
warehouseman. These charges consist primarily of three
elements:

1. Handling charges.

2. Storage charges.

3. Accessorial charges for such things as preparing
products for shipment.

The public warehouse throughput expenses are based on
rates quoted by public warehousemen of the cities included
in the model. The rates were gathered through the
auspices of an independent trade association and were
approved as being representative by the association.

The throughput cost parameters are stored in the
Supporting Data System for warehouses by type and Size.
The two types Of feasible warehouses in LREPS-F are
private distribution centers and public distribution

centers. The possible sizes (sales volume for the

88

modeled company moving through the DC) are:

Sipp' Capacity
1 Under $5MM
2 $6MM - $11.99MM
3 $12MM - $23.99MM
4 $24MM — $35.99MM
5 Over $36MM

For the private distribution centers, a fixed cost per
week plus a variable cost per pound of throughput is
listed for each Of the five sizes in the data base. For
the public distribution centers, a variable cost per pound
Of throughput is listed for each of the five Sizes.

The weekly throughput cost for a particular DCi

[TPCST (DCi)] is calculated from the formula

TPCST (DCi) = FTPEX (O,p) + VTPEX (O,p) * VOLWT (DCi)

where
FTPEX = the fixed weekly expense for a
DC of size and type p.
VTPEX = the variable cost per pound for

a DC of size and type p.

VOLWT (DCi) the total pounds shipped out

of DCi during this week.

For public facilities, FTPEX = 0. The weekly total
throughput cost for the company (TPCST) is the sum of
TPCST (DCi) for all DCi. TPCST is added to the income
statement variable THPTX. For private DC's, .5 TPCST (DCi)
is subtracted from cash and the same amount is added to

accounts payable. For public DC's, TPCST (DCi) is added

89

to accounts payable. The trade accounts are paid by
scheduling a deletion of both cash and accounts payable
for twenty-five days after the expense in incurred for
the amount .5 TPCST (DCi) for all private DC's and for
the amount TPCST (DCi) for all public DC'S.

Communications Expense. Communications expense is

 

essentially the cost Of processing orders within the
company. The cost parameters are listed in the Supporting
Data System and are computed in two parts. There is a
fixed communications expense and a variable expense per
order and line item for the total company. There is also
a fixed communications expense and a variable expense

per order and line item listed for each size and type

of distribution center. The types and sizes are the

same as for throughput costing.

Weekly orders and line items for the total company
and for each DC are inputted from the Operations Sub-
System. The weekly communications expense for each DC
is calculated by multiplying the per order variable ex-
pense and per line item variable expense for the size and
type of DC times the weekly DC orders and line items, and
adding the results to the weekly fixed communications
expense for that size and type of DC. Total weekly company
communications expense (CMCST) is calculated in three
steps:

1. Weekly company total orders and line items

are multiplied times the company level
variable expense per order and per line

items and these products are added to
company-wide fixed communications expense.

90

2. The total communications expense for each
DC are summed.

3. The outputs Of steps 1 and 2 are summed.

CMCSID is added to the income statement variable COMMX and
.5 CHHCST is subtracted from cash and added to accounts
payadole. Payment Of the appropriate trade account is
scheduled for the twenty-fifth subsequent day.

Fixed Facilities Expense Routine. The fixed
facilities component exists only for private distribution
centers and serves the purpose of describing the expense
associated with the fixed investment in facilities.

Thereare two components to fixed facilities expense.
The first, the cost on the financing required for the
:hrvestment, is an actual cash flow. The second, deprecia-
tion, is an accounting allocation. To further complicate
this routine, the cost Of capital has an economic meaning
for measuring the cost of the physical distribution system
and takes on a different value for purposes Of composing
an income statement. Further, depreciation is based on the
streight line method for financial reporting purposes and
on the sumeof-the years-digits method for tax purposes.
This gives rise to deferred taxes.

As input into the physical distribution system cost
report, fixed facilities expense includes (1) straight
line depreciation, and (2) a financing charge equal to the
weighted average cost of capital times the undepreciated

book value of the facility. As input to the income state-

ment, fixed facilities expense includes only straight line

91
deprenciation. The treatment of interest for financial
reporting purposes, depreciation of facilities for tax
purpxoses, and deferred taxes are explained in later
expeuise routines. The following discussion describes
thezzfixed facility expense routine for the purpose Of
costing the distribution system.

The initial investment in fixed facilities for the
five different sizes of private distribution centers are
listed.in the Supporting Data System. The investment
figures listed in the data base are the estimated cost
Of building the warehouse, laying the rail-spur tracks,
and purchasing the necessary automated materials handling
Systmme. For the health cares company, investment in
materials handling equipment is 20 percent Of total
investment for DC sizes 1, 2, and 3 and 35 percent for
Sizes 4 and 5.

Materials handling equipment is depreciated over
fifteen years and the remaining investment over fifty
years. The information for this routine is based on a
Study Of the private facilities Operated by the sponsor
company, which range in annual dollar sales volume from
three million to sixty million dollars. At the end of
each week, the initial investment, S, for a given DC Of

Size e is picked from the Supporting Data System and

Pingged into the formula

s
*_ 'k
2600 + a 780

 

WKDEP (DCi) = (l-a)

92

where

the expense Of depreciation for

WKDEP (DCi)
the week in process for DCi

a = the percent of the total in-
vestment for this Size of
facility in materials handling

equipment
2600 = 52 weeks * 50 years
780 = 52 weeks * 15 years

The total weekly depreciation charges for the company

(WKDEP) are calculated from the formula

2 .
WKDEP = I: WKDEP (DC.) + W
._ 1 52
1-1
where
SDNEQj = the straight line depreciation

charges for year j on all
equipment added during years
(1, 2,...j-l).

See page for an explanation
of this value.

WKDEP is added to the income statement variable, straight
line depreciation expense; and to the total physical
distribution system management report variable, depreciation
expense (DEPEX).
The weekly fixed facilities total cost of financing
WFTCC (DCi) is calculated for a given facility through the
w

equation WFTCF (DCi) = [s - ACSLD (Dci)] * E?

where

ACSLD (DCi) accumulated straight line

depreciation for DCi’

w = the annual weighted average
cost of capital.

93

The annual weighted average cost of capital is calculated
off-line and is exogenously inputted to the data base.
WFTCF (DCi) is added to a physical distribution system
report variable, fixed facilities financing expense (FFFNX).

Inventory Cost Routine. The inventory component
generates expenses against the system in two forms: the
cost of ordering and the cost of carrying inventory. The
cost of ordering inventory is included in communications
costing and is not listed separately.

Although the carrying cost of inventory is not listed
separately in the financial statements, it exists in an
economic sense and consists Of two elements:

1. Obsolescence and deterioration Of products,
taxes, insurance, and other miscellaneous items.

2. The cost of the capital invested in inventory.
The first element consists largely of general items and is
included in Selling and Administrative Expense in the
income statement. A true measure of the cost of capital
does not appear in the financial statements. While an
interest expense is subtracted from net operating income,
it certainly does not represent the cost Of all capital
used by the company. The limited measure Offered by
interest expense does not identify the functions responsible
for the expense. Thus, the cost for carrying inventory
listed in the physical distribution costing report cannot
be identified in the financial reports.

The weekly cost of carrying inventory (WCCIN), which

is an input to the physical distribution system report

94

1&5 calculated with the following equation:

WCCIN = INVTY * —%
where
INVTY = the level Of the balance

sheet variable inventory
at the end of each week.

Z = w + c
c = the annual per unit cost of
taxes, obsolescence, and
other miscellaneous expenses
of carrying inventories.
WCCIN is added to the physical distribution system report
variable, inventory carrying cost (INCST).

Cost of Debt Routine. The routine for calculating

 

and reporting interest expense in the model is associated
with the event DAILY and with the event WEEKLY. Debt
exists in the modeled firm in two essential forms: short
term debt and long term debt. Short term debt is

modeled assuming a revolving credit agreement under which
the firm may borrow up to two million dollars. Interest
charges are 6 percent per annum for the outstanding balance
and .25 percent per annum for the unused balance.

At the end of each day, the amount of outstanding
Short term loans is added to an accumulating variable,
short term lean dollar days (STLDD), which totals the
number of dollar days on which outstanding interest must
be paid. At the end of each week, STLDD is multiplied by
.06

'52- to determine the weekly interest charge on outstanding

short term debt (WISTD). The weekly charge for the credit

95
agreement (WCRCA) is calculated in the following steps:

1. STLDD is subtracted from $10MM, the amount of
unused dollar days which would accumulate
under the credit agreement if no borrowings
occurred during the week.

.0025

52 to

2. The difference is multiplied times
Obtain MCRCA.

 

The total weekly Short term debt expense (TWSTD) is
WISTD + WCRCA.

At the end Of each day, the amount of outstanding
long term loans, including the current position of long
term debt, is added to an accumulating variable, long
term loan dollar days (LTLDD). At the end Of each week
LTLDD is multiplied by 4%; to determine the weekly interest
charges on outstanding long term debt WISLD. Total weekly
interest charges (TWICH) are TWSTD + WISLD. TWICH is
subtracted from cash and is added to the income statement

variable, cost of debt (CSDBT).

Federal Income Tax Expense Routine. Federal income

 

taxes each quarter take on two different values, the tax
expense for the income statement and the actual taxes to
be paid. The weekly federal income tax expense (WFTXX) is

calculated using the following formula:

WFTXX = 172—3—0— + .52 (WKOPY)
where
$7500 = annual taxes on the first
$25,000 of earnings.
WKOPY = weekly Operating income,

before taxes.

WFTXX is added to the income statement variable, federal

96

income taxes (FINCT). Actual taxes due from each week's

Operations (WFTXP) are calculated using the following

formula:
WFTXP = ilggg + .52 (ADJNI)
where
ADJNI = Net income, before deprecia—

tion and taxes, less accelera-
ted depreciation.

WFTXP is added to the balance sheet variable, taxes
payable (TXPAY). The difference WFTXP - WFTXX is added
to the balance sheet variable, deferred taxes (DEFTX).
Accelerated depreciation for the five-year period is
actually calculated at the beginning Of each run and the
amount of depreciation for each year is stored in the
Supporting Data System. The accelerated depreciation for
any given year for a particular distribution system is

calculated by the following formula:

(l-a) * s + (a) (15-j)S
so 120

 

 

ACDCHj (DCi)

where

accelerated depreciation
charges for year j for

ACDCH. (DC.)
:1 l

(DCi)
S = the original investment
in DC..
1
a = the percent Of the total

investment in the facility
which is in materials
handling equipment

j = the year for which
depreciation is being
calculated.

li—l = the formula for calculating
SOYD depreciation for year
j for an investment being
depreciated over fifteen
years.

Only the investment in materials handling equipment is
eligible for accelerated depreciation.

For each year, ACDCHj (DCi) is calculated and the
n
total ACDCHj (DCi) is stored in the Supporting data
i=1

system. Accelerated depreciation for each week is

n
ACDCH.
i=1 3 .
52

 

Other Adjustments and Preparations Of Financial Variables
The five special routines associated with the finance
system are described in this section. They are:
1. Income on marketable securities.
2. Marketable security transactions expenses.

3. Quarterly Principle Re-payments on Long-
Term Loans.

4. An annual Capital Budgeting Routine for
replacement and additional equipment.

5. The Dividend Payout Routine.

Income on Marketable Securities Routine. The income
from marketable securities is calculated each week by
multiplying the accumulated dollar days Of marketable
securities (ADDMS), an input from the Operations Sub-system,
by the daily return per dollar invested in marketable

securities. The result, yield on marketable securities,

98
is added to cash and to the income statement variable,
marketable securities income (MKSCX). This routine is
inactive for all experimentation performed in this re-
search because of the relative unimportance of marketable
securities for the industries modeled.

Transactions Expense Routine. When each marketable

 

securities transaction occurs, the dollar amoung of the
transaction is plugged into the formula

MKSTE

$50 + $.000025A
where

A = the dollar amount of the
transaction

and the result, marketable securities transaction expense
(MKSTE) is subtracted from cash and added to the income
statement variable marketable securities transaction
expense (MKSCX). This routine is inactive for all experi-
mentation performed in this research.

Quarterly Principle Repayments Routine. The initial

 

level of long term loans is repaid at the rate Of 5 percent
per year (1.25 percent per quarter) of the original loan.
The even QTRPRP occurs to pay this amount in week seven

and every thirteenth week after week seven for the re-
mainder of the model (week 7, 20, 33, etc.). Principle
repayments are made in the middle of the quarter to even
out the heavy cash flows which occur only once per quarter.
(The other heavy quarterly cash outflow is the payments Of
taxes, which occurs in weeks 13, 26, 39, etc.). When

principle re-payment occurs, the balance sheet variables

99

cash and long term loans are reduced by the amount Of the
re-payment.

Capital Budgeting Routine. The Capital Budgeting
Routine actually consists of two different sub-routines:

1. Monitor and Control Facility Expansion.

2. Equipment Replacement and Addition.

The facility expansion sub-routine will occur only in
the infrequent event that a distribution center outgrows
its physical capacity. It is explained in the Monitor
and Control Sub-System. The equipment replacement and
addition sub-routine is associated with the event YEARLY,
and occurs at the beginning of each year. The description
Of its behavior is as follows:

1. EQREP, the annual investment in replacement

equipment is equal to the straight line
depreciation on materials handling equip-

ment for the previous year, which is

31—5— (DC ) +AMHDP (DC)
15 i j i
2. TINEQ, the total annual investment in
replacement and additional equipment,
is EQREP (1+r) where r is the rate of
expansion Of equipment.

3. TINEQ is subtracted from cash and the
balance sheet variable PLNEQ.

4. The straight line depreciation charges on
TINEQ for each year remaining in solution
and the accelerated depreciation charges
for each year remaining in solution are
calculated and added to appropriate
accumulating variables in the data base,
SDNEQj and ADNEQj

where

100

SDNEQ. = the straight line depreciation
J charges for year j on all equip-
ment added during years (1, 2...
j-l)
ADNEQj = the accelerated depreciation

charges for year j on all
equipment added during years
(1, 2...j-1)

The Dividend Payout Routine. The purpose of the

 

dividend payout routine is to reflect the fact that all
cash generated internally by the firm is not available
for financing future Operations. Dividends are paid

quarterly and are based on the following decision rule:

If quarterly earnings are > C) , dividends
= 0. Otherwise, dividends = .7 quarterly
earnings.

. This routine is called by the event QUARTERLY.

The Operations System: Monitor and Control Sub-System

 

The Monitor and Control sub-system performs two im-
portant general functions:

1. The supervisory function, the organizing
element of the simulator, schedules events
and activities so that each event occurs
at the right time and the necessary infor-
mation flows into and out of the various
routines of the operating system as required.

2. The Controller function provides the capability
of monitoring the value of certain variables
of the model, comparing those actual values
to standard, and adjusting the model when
elements Of the system are out of balance.

The supervisory function is discussed earlier in this

chapter. This section describes the controller function.

101

The Controller Function

 

In the basic LREPS model, LOCATE, a major routine
Of the controller, monitors system performance to determine
when and where new facilities should be brought into solu-
tion. In LREPS-F, the number Of facilities in solution
are held constant within experiments. Therefore, LOCATE
is deactivitated and new facilities are added exogenously
for experimental purposes.
The other activities of this function which require
explanation are inventory management and facility expension.
Inventory Manpgement. The purpose Of the inventory
management routine is to calculate the parameters for the
inventory Control routine in the Operating sub-system.
The critical variable is inventory on hand for each tracked
product, IOH (ITP). The parameters which are necessary
for each tracked product are M (ITP), which is the level
to which inventories for the tracked product are to be
replenished; ROP (ITP) which is the level Of IOH (ITP) at
which inventories should be re-ordered; and M (ITP) —
IOH (ITP), which is the amount which must be re-ordered
to replenish IOH (ITP) to M (ITP).
The equations which are used in monitor and control

to calculate these parameters are:

M = (EOQ + ROLTM) * DEM
and
ROP (ITP) = DEM * (ROLTM + BUF)

where

102

DEM = the daily sales forecast for
this tracked product for DCi'

EOQ = the number of days sales in
the re-order quantity for
this tracked product.

ROLTM = the re-order lead time for

this tracked product at DCi'

BUF = the number of days sales in

safety stock for this tracked
product at DCi.

EOQ and BUF are exogenously inputted based on the
inventory management policies of the sponsor company.
DEM is calculated at the end of each quarter and is an
exponentially smoothed sales forecast based on modeled
sales from previous quarters. ROLTM is the actual ex-
perience in the model with re-orders for each tracked
product at DCi.

Faculty Expansion. The purpose of the facility ex-

 

pansion routine (EXPAND) is to increase the capacity of
a distribution center when the quarterly volume moving
through the facility exceeds 70 percent of the capacity of

the facility. EXPAND is called at the end of each quarter,

DC Sales

and 1f the DC Capacity

 

ratio exceeds .70, an investment
is scheduled for the end of the sixth succeeding month

sufficient to expand the DC to the next facility size.

The Report Generator System

 

The purpose of the Report Generator System is to
arrange and process the output of the Operating System to

generate several management and financial reports and to

103
prepare the data for analysis. Output exists in two
general forms: the Operating results for each time period
and the level of status variables at the end of each
period. Reports which are generated for this research are:

l. The Individual Distribution Center Management
Report.

2. The Total Physical Distribution System
Management Report.

3. The Balance Sheet.

4. The Income Statement.

5. The Financial Ratio Report.

6. The Statistical Analysis Report.

7. The Comparative Incremental Report.

Individual Distribution Center Management Report

 

All items in this report are operating results. The
purpose of the report is to facilitate the measurement of
changes in cost and service between experiments. The
report is printed quarterly for both DC's for all exper-
iments with two facilities and the variables reported are
as follows: total number of stock-outs; average days
delay per stock-out; standard deviation days delay per
stock-out; percent of total dollar sales delivered within
normal order cycle times of four, five, six, seven, and
eight or more days, respectively; outbound transportation
costs; inbound transportation costs; throughput costs; and

communications costs.

104

Total Physical Distribution System Management Report

All items in this report are operating results. The
purpose of the report is to record the quarterly measures
of service and cost for the total system. The variables
reported are the same variables for the total system that
are reported in the Individual Distribution Center
Management Report, plus the economic cost of carrying
inventories; the depreciation expense on private facilities;
the economic cost of financing the private facilities; and

the total cost of the physical distribution system.

The Balance Sheet

 

The balance sheet is a special arrangement of the
levels of various financial status variables and can be
printed as often as daily. It is stored on output tape
weekly and printed quarterly. The items in the balance
sheet and the routines which effect the level of each

item are listed in Table 3.1.

The Income Statement

 

All items in the income statement are operating
results. It is stored on output tape weekly and can be
printed for any integer multiple of one week. For this
research, it is printed quarterly. The items in the income
statement and the routines associated with each item are

listed in Table 3.1.

105

The Statistical Analysis Report

 

The purpose of this report is to summarize the output
of the experiments to facilitate analysis of the results.
The statistics generated by this report include:

1. The mean average and standard deviation of

the twenty quarterly values for each vari-
able in the Total Physical Distribution
System Management Report, the Balance Sheet,
the Income Statement, and the Financial
Ratio Report.

2. A simple linear regression equation describing
the relationship between quarterly sales and
the quantity (Net Income + .05 Interest +
Depreciation).

3. A simple linear regression equation describing
the relationship between quarterly sales and
the quantity (Net Income + Depreciation -
Principle Re-payment).

4. The total, explained, and residual sum of
squares; and the co-efficients of determina-
tion and correlation for the equations in
(2) and (3) .

The mean and standard deviation of the quarterly
values are used to compare the experimental results in
summary form. The equation X = 5 formed by transforming
the regression co-efficients is the break-even level of
quarterly sales, which offers a composite measure of the
extent to which a particular warehouse decision has affected
downside risk of the firm. Changes in total variance of
adjusted net income between experiments measure the shift
in cash flow volatility resulting from each decision
alternative. Changes in explained variance measure the

sensitivity of net cash flows to the shifts in operating

and financial leverage which results from the warehousing

106

decision under study. Changes in residual variance measure

the shifts in unexplained volatility resulting from each

alternative. Unexplained variations in net cash flows

are especially critical because they cannot be anticipated

through sales forecasting.

The Financial Ratio Report

The ratios in this report were selected based on the

following criteria:

1.

The ratios identified by Carlson as effective
predictors of "financial efficienc " in
univariate analysis were included.

Those ratios identified in three or more
of the studies in the literature review
as effective predictors of corporate
failure were selected.

Ratios which measure certain aspects of
financial position not measured by the
ratios identified in (l) and (2) were added
by the researcher.

The selected ratios and the formulas for calculating them

are as follows (see pages 57-64 for definitions of symbols):

 

 

Ratio Formula

Current Assets Cash + ACCRC + INVTY + MKTSC
Current Liabilities ACCPY + TXPAY + STDBT

Acid Ratio Cash + ACCRC + MKTSC

 

ACCPY + TXPAY + STDBT

Cash Turnover Quarterly Sales

 

Average Weekly Cash Balance

 

 

Working Capital Cash + ACCRC + INVTY + MKTSC
Total Assets TOTAS
Net Working Capital Current Assets-Current Liabilities

 

 

Total Assets TOTAS

107

 

 

 

 

 

 

 

Ratio Formula
Fixed Asset Quarterly Sales
Turnover Average weekly level of Fixed Assets
Debt TOTAS - SHEQT
Equity SHEQT
Debt TOTAS - SHEQT
Total Assets TOTAS
Return on Assets EAFIT
TOTAS
Return on Equity EAFIT
SHEQT
Debt Service EAFIT + CSDBT + SLDPX
Coverage CSDBT + Principle Re-payments

Two of the ratios, return on assets and return on equity,
are measures of performance. Each of the other ten ratios

measures some aspect of financial position.

The Comparative Incremental Report

 

The purpose of this report is to compare the output
between experiments. The report is printed quarterly and
consists of the following elements:

1. The difference between each experiment
and the base experiment of its experi-
mental set of the quarterly value of
the 61 variables in the Total Physical
Distribution System Management Report,
the Balance Sheet, the Income Statement,
and the Financial Ratio Report.

2. The quarterly percentage change for
each of the values in (l).

3. The twenty quarter average of the per-
centage changes reported in (l).

4. The standard deviation of that average.
5. A ranking of the 61 variables on the

basis of average quarterly percentage
change.

108
6. A ranking of the 61 variables on the basis

of standard deviation quarterly percentage
change.

Concludinngemarks
The output from the Report Generator System is the
input to the analysis of experimental results. The
findings from the analysis are reported in Chapter V and
the output from the Report Generator is reported there
in summary form as deemed necessary for discussion of the
findings. Before the findings are reported, however, an

explanation of the experimental design is presented.

CHAPTER I I I--FOOTNOTE REFERENCES

1

Merton H. Miller and Daniel Orr, "A Model of the
Demand for Money by Firms,"

Quarterly Journal of Economics,
(August, 1966), p. 413-415.

2

C. Robert Carlson, "A Financial Efficiency Model,"

(unpublished dissertation, Michigan State University,
1971).

109

CHAPTER IV
EXPERIMENTAL DES I GN

The first step in this research is the development of
a dynamic computer simulation model on which necessary ex-
perimentation can be performed. The attainment of this
intermediate stage is fulfilled through LREPS-F which is
performing according to design specification and which is
described in the foregoing chapter. The remaining steps
center on experimentation on the model. The purpose of
this chapter is to explain the design of those experiments.

The chapter is divided into two major sections.
First, a general review of techniques and problems involved
in design of computer simulation experiments is presented.
By necessity the review is limited to general topics
specifically pertinent to the design of experiments for
this research. For a broader treatment, the reader is
referred to Naylor.l In the second section, problems
involved in the design of experiments on LREPS-F are dis-
cussed in the context of the general treatment, and the

experimental design resultantly adopted is described.

110

111

Techniques and Problems in the
Design of Simulation Experiments

 

Motives for Experimentation

Experimentation on simulation models is conducted for
one of two general purposes, according to Naylor.2 The
first and by far most p0pular motivating factor is the desire
to converge toward the optimal state (solution) for some real
world process. A second, less common motive is to make a
rather general investigation of the relationship between
variables in the model in search of the "underlying
mechanisms governing the process under study."3
Experimental Variables: Factors and Responses

Model variables for purposes of experimentation are
divided into factors and response variables.4 Factors,
which are sometimes called independent variables, will
usually be model parameters or exogenous inputs. Response
variables, which are sometimes called dependent variables,
are endogenous or output variables.5 The purpose of ex-
perimental design is to test the effect of variations in
certain factors (called test factors) on a response variable.

Not all independent variables which may affect a re-
sponse variable in the real world are included as test
factors in simulation experimentation. Some factors are in-
cluded and observed in order to improve the precision (degree
of explained variation) of the experimentation. Other in-
dependent variables in the real world which might affect the

response variable are excluded in the interest of model

efficiency. The role which these variables might play in

112

the real world is represented within the model by the
random character of the included exogenous variables.6
However, the inclusion of random, non-test factors
need not distort the measurement of the effects of the
controlled test factors on the response variables. Con-
ditions of ceteris paribus, while unattainable in real
world experiments, can often be achieved in computer
simulation by using a common sequence of pseudorandom
numbers on repeated computer runs. Thus, as Conway ex-
plains, the simulation experimenter can attain "perfect
homogeneity of experimental medium" between experiments,
even though random variations in exogenous variables are

included within experiments.7

Statistical Inference in Experimental Design

 

The purpose of statistical inference from experimental
results is two-fold:8
1. To determine if observed differences in the
response variable between experimental runs
could have occurred by chance rather than as
a result of shifts in the test factors.
2. To generalize the results beyond the range of

exogenous inputs considered.

Determining statistical significance of observed

 

differences. The techniques of statistical inference most

 

often used in the design of experiments on computer

113

simulation are the analysis of variance techniques. However,
the analysis of variance techniques (indeed all classical
statistical inference techniques) are inapprOpriate when a
common sequence of pseudorandom numbers has been used
between experiments.9 The null hypothesis of statistical
inference is that there is no inherent difference between
two or more sets of experimental results, that the observed
differences occurred simply because of chance factors. A
basic assumption is that the various experiments are in-
dependent of each other. In experimentation using a common
sequence of pseudorandom numbers, as many exogenous vari-
ables as possible except for the test factors are held
constant between experiments. By limiting these sources of
residual variation a high degree of dependence between ex-
periments is purposefully established.10 Thus, some actual
difference between the moael outputs is expected and the
failure of a technique of statistical inference to reject
its null hypothesis would demonstrate a power deficiency in
the statistical test, not the absence of output differences
between the experimental runs.11
Even when the statistical test rejects the null hypoth-
esis (recognizes the dependence between experimental out-
puts), the experimentefs objectives have likely not been
served. He is most likely interested in how large a dif-
ference is exhibited between different experiments and
between which experiments the largest differences are ex-

hibited.12

114

Generalizing results beyond the range of test factors

 

considered. Generalization of the results on the response

 

variable of a small sample of values for the test factors
to the larger population technically assumes that the
levels of the test factors studied have been randomly
selected.13 While generalization therefore requires the
use of statistical inference techniques, it is much more
important when research is undertaken to experimentally
search for the optimal solution to some problem. When
the motive of the research is to study the underlying
mechanisms of a process, as is true in this case, the
necessity of generalization through statistical inference
is far less critical. However, the researcher must be
careful in generalizing his conclusions from experimenta-

tion of this nature.

The Many Response Problem

 

The literature on experimental design has established
the capability of analyzing the simultaneous effects of
four or more test factors on a single response variable,
although the math involved becomes greatly complicated as
the number of factors increases. However, experimental
design techniques are virtually non-existent for multiple

. 14
response experiments.

The multiple response problem,
which occurs frequently in computer simulation experimenta-
tion, can exist in either or both of two aspects:15

1. The experimental output may consist of several

response variables.

115

2. The researcher may wish to observe the output of a
particular response variable as a time series (i.e.
observe a different value for each year, month, or
other convenient time period).

It is often possible to avoid the many response problem
by treating an experiment with many responses as many ex-
periments each with a single reSponse. Or several reSponses
could be combined and treated as a single response. However,
it is not always possible to bypass the many response prob-
lem; often, multiple responses are inherent to the situation
under study. Experimental design techniques in such situa-

tions are usually unique to the specific situation.16

Design of Experiments on LREPS-F

 

Motive for Experimentation

 

The underlying motive for experimentation on LREPS-F
is to study the nature of the interaction between ware-
housing decisions and variables in the finance sub-systems
of modeled firms in two industries under specified con-
ditions. It is of the second, less common class of motives
which Naylor describes as "the search...for the underlying
mechanisms governing the (real world) process under study."

Test Factors, Non-Test Factors, and the Homogeniety of

 

Experimentation

 

The test factors in this experimentation are the ware-
housing system alternatives; the sales growth rate bias; and

the modeled industry (firm). The levels of each test factor

116
to be studied are subjectively selected rather than randomly
drawn.

LREPS—F includes several non-test factors which are
included to improve the "precision" of the simulation. The
level of daily sales and the geographic and product charac-
teristies of daily sales are examples of non-test factors
which are included. Random variations in these and other
model non-test factors within experiments also represent the
effect on the response variables in the real world from in-
dependent variables (factors) which are omitted from the
model in the interest of efficiency. However, the effect
of these random non-test factors is held constant across
experiments through the use of a common sequence of
psuedorandom numbers. Although many of the output varia-
ablesand observed exogenous variables vary widely within
experiments, the results are exactly replicated if the
simulation is repeated with no exogenous changes in test
factors. Further, the values of the random non-test factors
are exactly duplicated between experiments except in
situations in which the non-test factors are affected by
changes in test factors. Thus, conditions of perfect
homogeneity of experimental medium are obtained and the
observed differences in response variables between experi-
ments are entirely attributable, either directly or in-

directly, to shifts in the test factors.

117
Response Variables and the Multiple Response Problem

The response variables of primary interest in this
experimentation are the finance sub-system variables dis-
cussed on pages 57-64. Responses of the distribution sub-
system variables to shifts in the test factors are used by
the researcher to identify the effects of factor shifts on
the finance variables. As such, the distribution variables
probably are also response variables, although they might
be classified as non-test factors.

Including distribution variables, there are sixty-one
response variables. In each experiment, the value for each
response variable during each quarter for five years is
reported. Thus, for purposes of experimental design this
research suffers from both aspects of the multiple response
problem. By working with twenty quarter average data, a
general technique recommended by Naylor, the second aspect
of the multiple response problem is rectified. In
attacking the problem of a large number of response vari-
ables, Naylor reports that sometimes each response variable
within an experiment can be treated as a separate experiment.
Across thirty experimental runs, this solution would require
1830 statistical tests of experimental results. The sheer
numbers involved would render the analysis almost meaning-
less. Furthermore, several of the financial variables have
little meaning standing in isolation from each other. For
these two reasons treating the response variables as

separate experiments is rejected as a feasible technique of

118
experimental design. When multiple response variables are
inherent to the process studied, as is true in this case,
there are no generally accepted techniques of experimental
design.

Statistical Inference

 

The use of statistical inference on the experimental
results from LREPS-F is technically inappropriate for two
'reasons:

1. The common sequence of pseudorandom numbers is
used, thus assuring the high degree of dependence
among experiments.

2. The experiments have multiple responses, which is
a condition not conducive to experimental design
techniques.

For measuring the effects of a shift in a single test
factor (alternative warehousing systems, but for a given
growth rate and firm), the failure to meet the requirements
for statistical inference is not critical. When only one
test factor is changed, the differences in response variables
between experiments are entirely the result of the shift in
this factor. The concept of significance in this context
loses its popular inferential meaning and means not whether
the observed differences do in fact exist, but whether they
are sufficiently meaningful to merit reporting and analysis.
In the interest of efficiency, they cannot all be analyzed
and reported and the decision of how large a change should

be to merit discussion is by its nature a subjective one.

119

Testing the significance of differences between ex-
periments is not the only general purpose of statistical
inference in experimental design. The other purpose, of
statistically generalizing the results to levels of the
test factors other than those specifically considered, is
not of concern here because the levels of test factors were
not randomly selected and the results of this research are
therefore not qualified for generalization on technical
grounds.

When both the warehousing system and either the growth
rate assumption or industry have been changed, the observed
differences in response variables might be the result of
either or both of the factor shifts. Thus, statistical
inference might be applicable in testing the consistency of
the observed differences across growth rate assumptions and
industry lines. The analysis and time required to test
each individual response variable by this approach are
prohibitive. Rather, the technique selected for testing
the consistency of results under different growth rate
assumptions and for different industries is that of com-
paring the ranks of the response variables on the basis
of largest percentage changes. Although the assumptions of
the technique are not strictly net, the Spearman Rank
Correlation Co-efficient will be used because it serves the
purpose of reducing the required comparison to a single
measure. This technique has often been used in financial

research for measuring the consistency of real world

120
economic data over time when the assumption of indepen-

. . l
dence between succ0351ve observations could not be met. 7

The Research Design

 

The preceding sections of this chapter present a
general discussion of experimental design techniques and
the specific problems involved in designing experiments on
LREPS-F. This section reports the research design adopted
after due consideration of the problems discussed and
several minor design problems.

Thirty experiments are performed for this research,
with an experiment performed for each possible combination
of five warehousing alternatives, three possible growth
rates, and two modeled firms under study. The following
is a listing of the code numbers used for purposes of

identifying these experiments:

 

 

 

Health Cares Appliances

Warehousing Alternative .05 .15 -.05 .05 .15 -.05
Private at Chicago 1.01 1.06 1.11 2.01 2.06 2.11
Private at Chicago and

Private at Columbus 1.02 1.07 1.12 2.02 2.07 2.12
Private at Chicago and

Public at Columbus 1.03 1.08 1.13 2.03 2.08 2.13
Public at Columbus 1.04 1.09 1.14 2.04 1.09 2.14
Public at Chicago and

Public at Columbus 1.05 1.10 1.15 2.05 2.10 2.15

The five experiments for a given firm and growth rate
constitute an experimental set. For example, experiments
1.01 through 1.05 are an experimental set. 1.01, 1.06, 1.11,

2.01, 2.06, and 2.11 are the base experiments. It is

121

assumed that this warehouse system is the one currently
used by the firm and from which a change is proposed. For
each of the other alternatives within an experimental set,
only one factor (the warehouse system design) is changed
from the base experiment and the changes in response
variables are all directly or indirectly caused by the
shift in the warehouse system. An absolute change of 10%
in either the average or standard deviation value of a
response variable is subjectively established as a
"significant" response. The significantly affected response
variables in each alternative experiment are ranked and re-
ported. At the end of each experiment the researcher
offers an explanation of the observed behavior of the
response variables. These explanations are based on visual
inspection (graphical and tabular) and are essential to the
search for the underlying mechanisms of the interaction
between warehousing decisions and finance sub-system
variables.

The consistency of results between experimental sets
is evaluated using the Spearman Correlation Co-efficient
of the ranks of response variables of each warehousing
system design with the ranks of the variables for the com-
parable alternatives under different growth rate assump-
tions and in the other modeled industry. For example, the
ranks of the significantly affected response variables
between Experiments 1.01 and 1.02 are correlated to the

ranks for 1.06 vs. 1.07, to 1.11 vs. 1.12 and to 2.01 vs.

122

2.02. Although the original design was to explain only
those differences which were statistically significant at
the .01 level, the differences between experimental sets
are visually inspected after each test. Particular
interest is given in those situations in which the
difference is statistically significant at the .01 level.
The purpose of testing the consistency of results is to
gain a better understanding of the observed interactions.
Since it is documented that statistical significance loses
much of its meaning in this type of research, it is con-

sidered but not assigned prime importance.

CHAPTER IV--FOOTNOTE REFERENCES

1Thomas H. Naylor, Joseph H. Balintfy, Donald S.

Burdick, and Kong Chu, Computer Simulation Techniques (New
York: John Wiley, 1966).

2Ibid., p. 338

31bid.

41bid., p. 332

51bid.

 

61bid., p. 323

7R. W. Conway, "Some Tactical Problems in Digital
Simulation," Management Science, Volume X (October, 1963),
p. 52.

8Ibid., p. 53

91bid.

loIbid.

llIbid.

12G. E. P. Box and William G. Hunter, "The Experimental
Study of Physical Mechanisms," Technometrics, Volume 7
(February, 1965), p. 24.

13Naylor, pp. cit., p. 325

 

14Ibid., p. 339

lSIbid., p. 340
16 .
Conway, 9p. c1t., p. 61

17James E. Murphy, "Earnings Growth and Price Changes,"
Financial Analysts Journal, Volume 25(September, 1968),

p. 97.

123

CHAPTER V

EXPERIMENTAL RESULTS AND ANALYSIS

The purpose of this chapter is:

1. To report the response variables which were
significantly affected in each of the
experiments.

2. To offer an explanation for these results.

3. To report the degree of consistency among
the response variables for different growth
rate assumptions and for the different
industries studied.

4. To comment on the significance of these
findings to the general understanding of
the cross-functional interaction between
warehousing decisions and finance sub-system
variables and to report several interesting
by-products of the research.

The changes in reSponse variables from a shift to a
two public warehousing system and to a one private, one
public warehousing system are found to be simply com-
binations of the changes exhibited in shifting to a
single public warehouse and in changing from a one private
to a two private warehouse system. The significance of this
finding is that the crucial questions in warehouse system
design for financial variables appear to be public vs pri-
vate warehousing and the number of warehouses included in

the system. The other two alternatives are variations

on these two questions. Therefore, this discussion centers

124

125
on the significantly affected financial variables by
shifting from one private warehouse to either one public
or two private warehouses. For these purposes "signifi—
cantly affected variables" are those for which the
decision alternative in question causes change in the
average or standard deviation quarterly value of more
than 10 percent or less than -10 percent from the
equivalent value in the base experiment. Twenty quarter
average data are used as a solution to the time series
aspect of the multiple response problem. The standard
deviation is used to detect intermediate variations in
the variables which would be overlooked if only one para—
meter were examined. The changes in distribution
variables are included to help explain the changes in

financial variables.

Experimentation on the Modeled

Health Cares Company

 

The results of experimentation on the modeled health
cares company and observations drawn from those results
are reported in this section and are compared to the
results from experimentation on the appliance company in
the following section.

Addition of a Second Warehouse

.05 Annual Growth Rate in Sales. Those financial

variables which are significantly affected by the

addition of a private warehouse at Columbus when sales

126

are designed to grow at 5 percent per year are ranked on

the basis of largest percentage change in Table 5.1.

The distribution variables which are significantly affected

by this alternative are ranked on the basis of largest

percentage change in Table 5.2.

Analysis of the reasons for these results identifies

three direct effects to which the other changes can be

traced:

l.

2.

The average level of inventories increases
by $299.5 thousand.

The average level of accounts payable
declines by $130.2 thousand.

Average quarterly net earnings drop by $9.8
thousand partly because of the increased
interest expense on the short term loans
required to offset the change in inventories
and accounts payable; and partly because the
cost savings from reduced outbound trans-
portation expenses is not sufficient to
offset the increases in the inbound trans-
portation, throughput, communications, and
facility expenses.

These three basic changes are reflected through

significant changes in the financial ratios and

statistics in the following ways:

1.

Debt/net worth and cash turnover increase
and the acid ratio deteriorates, all re-
flecting the strain on cash and short term
financing brought on by the adverse changes
in inventories and accounts payable.

Debt service coverage deteriorates and return
on both assets and equity decline, all because
earnings drop and short term debt rises.

The levels of sales required to break-even
both for profits and for servicing debt
increase, indicating that the company's
ability to withstand downturns has been
impaired.

127

TABLE 5L1” --Finance Variables Singificantly Affected by
the Addition of a Second Private Warehouse in
the Health Cares Company.

 

 

Quarterlnyercentage Change

 

 

Rank Variable Average Standard Deviation
1 Short Term Loans 94 93
2 Interest Expense 43 46
3 Net Liquid Assets -41 252
4 Inventories 38 70
5 Taxes Payable -34 30
6 Debt Service Coverage -32 29
7 Cash Turnover 32 47
8 Return on Assets -26 73
9 Debt/Equity -23 34

10 Net Earnings —22 28
11 Inventory Turnover 21 33
12 Residual Variance 20 -
13 B Regression Co-efficient -20 -
14 Total Liabilities 17 31
15 Return on Equity -16 76
16 Break-even: Debt

Servicing 16 -
17 Acid Ratio —16 24
18 Accounts Payable -13 20
19 Break-even: Net Income 13 -
20 Cash - 9 30

 

128

TABLE 5.2.—-Distribution Variables Significantly Affected
by the Addition of a Second Private Warehouse
in the Health Cares Company.

 

 

Qparterly Percentage Change

 

 

Rank Variable Average Standard Deviation
l 4 Day ORCT 195 13
2 Stockouts 104 182
3 Inventory Expense 49 62
4 S.D. Stockout

Days Delays -34 87
5 Communications

Expense 31 3
6 Throughput Expense 26 8
7 5 Day ORCT 20 3
8 Avg. Stockout

Days Delays 15 53
9 Inbound Transport Exp. 15 29

10 Outbound Transport Exp. - 8 l

 

129

4. The residual variance of adjusted net

income increases, indicating that an

element of instability has been added

to the profit function. This element

of additional instability is the in-

creased variability of inventories and

accounts payable, which result in wider

swings in cash, short term loans, and

interest expense.
The composite analysis is overwhelming that the financial
position of this firm would be impaired by the decision
to add a second private warehouse if sales grow at
a normal rate. The most interesting observation in
this experiment is the effect which the second ware-
house has on accounts payable and inventories. This
interaction is explored in more detail in relation
to other experiments and in the last section of the
chapter.
Sensitivity Experiments: .15 and -.05 Growth Rates. The
ranks of financial variables which are significantly
changed by the addition of a private facility at
Columbus under the assumption of an annual growth rate
of .15 (Experiment 1.07) and under the assumption of an
annual growth rate of -.05 (Experiment 1.12) are compared
to the results under the .05 growth rate assumption in
Table 5.3. The ranks of the significantly affected
distribution variables are compared in Table 5.4.

The Spearman Rank Correlations between the financial

results of 1.02 and of 1.07 and 1.12 are .89 and .54,

respectively. Thus, the results under the .15 growth

rate assumption do not change substantially from those

130

TABLE 5.3. --Comparison of Financial Variables Signifi-
cantly Affected by the Addition of a Second
Private Warehouse Under Varying Growth Rate
Assumptions for the Health Cares Company

 

 

Avg. Quarterly

 

 

 

Rank % Change
Variable -.05 .05 .15 -.05 .05 .15
Short Term Loans 1 l 2 116 94 112
Interest Expense 9 2 3 37 43 35
Net Liquid Assets 1 3 l - 85 —41 -l33
Inventories 3 4 4 90 38 33
Taxes Payable 5 5 6 - 63 -34 - 23
Debt Service Coverage 18 6 5 - 10 -32 - 24
Cash Turnover 2 7 18 94 32 12
Return on Assets 7 8 9 - 41 -26 - 20
Debt/Equity 12 9 7 20 -28 - 21
Net Earnings 15 10 10 - 12 -22 18
Inventory Turnover 8 11 ll - 41 21 - 18
Residual Variance 20 12 22 6 20 —
B Regression Co-efficient l7 13 21 - 11 20 2
Total Liabilities l3 l4 l6 16 17 14
Return on Equity 16 15 12 - 12 -16 - 18
Break-even: Debt Servicing 19 16 13 8 16 17
Acid Ratio 10 l7 l4 - 30 -16 — 15
Accounts Payable 14 18 8 - 15 -13 - 21
Break-even: Net Income 21 19 15 6 13 15
Cash 6 20 20 - 46 - 9 - 6
Total Variance ll 21 19 23 - - ll

(1 Regression Co-efficient 22 22 17 - 5 - 9 l3

131

TABLE 5.4.--Comparison of Distribution Variables Signifi-
cantly Affected by the Addition of a Second
Private Warehouse Under Varying Growth Rate
Assumptions for the Health Cares Company

 

 

Average Quarterly

 

 

 

Rank % Change
Variable -.05 .05 .15 -.05 .05 .15
4 Day ORCT l l 1 204 195 191
Stockouts 2 2 2 94 104 93
Inventory Expense 3 3 3 89 49 31
S. D. Stockout Days Delays 5 4 6 42 - 34 - 20
Communications Expense 8 5 5 20 30 23
Throughput Expense 6 6 4 21 26 30
5 Day ORCT 7 7 7 21 20 20
Avg. Stockout Days Delays 4 8 9 67 15 9
Inbound Transport. Exp. 9 9 8 19 15 16

Outbound Transport. Exp. 10 10 10 - 9 - 8 - 8

 

132
of the .05 assumption, but the results of the -.05
assumption do reflect an interesting difference. Four
financial ratios have at least a three position change
in rank between 1.02 and 1.12. These are: cash
turnover, inventory turnover, the acid ratio, and debt
service coverage.

These differences in responses between growth rate
assumptions occur basically for two reasons. First, the
exponential smoothing sales forecasting technique causes
forecasted sales and thus inventory levels to be below
actual sales during upturns and above actual sales during
downturns. Splitting the single demand forecast into
two forecasts aggravates this forecast error. The result
is that the difference in inventories between a one and
two warehouse system is larger under a period of sales
decay than during a period of sales growth.

Second, when sales decline within an experiment, a
partial liquidation of inventories and accounts receivable
occurs. This results in a build-up of cash and a partial
re-payment of short term loans. Therefore, the decline
in accounts payable and the increase in inventories
which results from adding the second warehouse can be
financed largely from cash when sales are decaying
(Experiment 1.12). When sales are increasing, the levels
of cash are minimal because inventories and accounts

receivable are building up faster than accounts payable,

133
causing a strain on cash. Thus in 1.02 and 1.07 drops
in accounts payable and increases in inventories
resulting from the second warehouse are financed through
short term debt.

The interesting result is that the addition of the
second warehouse causes a larger increase in inventories
but impairs the debt service coverage ratio less during
a -.05 decay in sales than it does in a .05 or .15 growth
in sales. This is because the firm is in a better
position to finance a second warehouse when current non-
liquid assets are being worked off and serves to illus—
trate that during periods when firms need to undergo
structural changes is often when they are least prepared
to finance them. The implication is that such changes and
the provision for financing them should be planned well
in advance of their actualization.

Observations of Interest in the Distribution Area.

 

The measures of distribution service efficiency offer no
help in explaining the changes in the financial variables.
This is because LREPS-F does not include a feedback

loop between system performance and customer demand.
However, some general observations are interesting con-

cerning the service variables.

134

Under all three growth rate assumptions, the two
distribution variables most significantly affected by the
addition of a second warehouse are percent of sales
delivered within four day order cycle time and the number
of stockouts. The former improves by approximately 200
percent while the latter deteriorates (increases) by
approximately 100 percent. The impressive improvement in
delivery times quite possibly could have two favorable
effects on the financial system which are not demonstrated
here because of the absence of the necessary feedback
mechanisms in the model. First, the improved delivery
times could stimulate sales, and thus bolster the profit
and cash flow picture. Second, quicker delivery could
mean quicker billing and improved turnover in accounts
receivable, thus improving cash flows and the cash balance.
These relationships are highly conjectural, however; and
just as it is difficult to model them, it is dangerous to
speculate on the nature of these improvements.

The adverse effect which the increase in stockouts
would have on customer demand might offset the favorable
effects of improved delivery time. However, the deterior-
ation in stockout performance in this research has little
meaning on an absolute basis. Consider the .05 growth
rate (Experiment 1.02). The increase in per quarter

stockouts from approximately 5.1 to 7.5 represents a shift

Stockouts
Total Orders

percent of orders filled on demand is still .9985 after

 

in the ratio from .0009 to .0015. The

135
the change. That any adverse effects would occur from such
a change is doubtful.

Public Versus Private Warehousipg

 

.05 Annual Growth Rate in Sales. The financial

 

variables in the health cares company significantly affected
by the decision to use a public instead of a private
warehouse at Chicago assuming a 5 percent annual growth
rate in sales are ranked in Table 5.5. The significantly
affected distribution variables are also ranked in Table
5.5. The two direct effects which trigger the changes
in financial ratios and statistics are as follows:
1. The investment in the facility is eliminated
along with the financing, depreciation
charges, deferred taxes, and interest
expense associated with it.
2. Average earnings decline because the
drop in depreciation and interest expense
is not sufficient to offset the additional
throughput expense.
These basic changes are reflected in significant
changes in the following financial ratios and statistics:
1. Working Capital and Net Working Capital
Total Assets Total Assets
increase significantly, and the
coefficient of the adjusted net income

equation declines, primarily reflecting
the drop in fixed assets and fixed expenses.

 

 

2. Debt service coverage improves because of
the elimination of interest and principle
re—payment on long term debt.

3. Return on assets and return on equity
improve, indicating that the drop in
investment was more than proportional to
the drop in net earnings. The additional
earnings of the private facility would offer
a return of only 3.6 percent after taxes, on
the average, over the five year period.

136

TABLE 5.5.--Financia1 and Distribution Variables
Significantly Affected by Using One
Public Instead of One Private Warehouse
in the Health Cares Company

 

 

Quarterly Percentage Change

 

 

Rank Finance Variables Average Standard Deviation
1 Debt Service Coverage 144 97
2 Plant and Equipment -100 -
3 Deferred Taxes -100 -
4 Long Term Loans -100 -
5 Depreciation -100 -
6 Fixed Asset Turnover -100 -
7 Interest Expense - 61 13
8 Working Capital/Total

Assets 45 5
9 Net Working Capital/

Total Assets 44 7
10 Net Worth -33 5
11 Total Assets -32 3
12 (1 Regression

Co-efficient -32 -
13 Total Liabilities -31 6
14 B Regression Co-efficient -22 -
15 Total Variance -20 -
16 Return on Net Worth 20 50
17 Break-even: Debt

Servicing -18 -
18 Return on Assets 16 35
19 Net Liquid Assets -15 86
20 Taxes Payable -15 5
21 Net Earnings -10 5
22 Cash -10 16

Rank Distribution Variables Average Standard Deviation

 

 

 

 

1 Throughput Expenses 105 10
2 Depreciation -100 -
3 Facility Capital Expense -100 -

 

137
4. The total variance of the adjusted net income
equation declines, indicating a strong improve-
ment in the stability of cash flows.
5. The break-even level of sales drops for
adjusted net income and even more so for the
level required to service the firms debt,
indicating a much stronger defensive posture
in case of an economic or market reversal.
That such a change renders the firm less susceptible to
economic downturn or market reversals is seen in Figure
5.1 in which quarterly sales are graphed against the
difference between quarterly earnings for the public
and the private systems. The public system is superior
for lower sales volumes because of the heavier fixed ex-
penses in private systems, while the private system
becomes increasingly superior as the sales volume increases.
This is a classic illustration of the effect of leverage.
A special note on the subject of leverage is in
order. The level of fixed expenses is reduced signifi-
cantly in going from a one private to a one public ware-
housing system. This is seen in the shift in the re-
gression a co-efficient from -$56,400 to —$38,100, a shift
of 32 percent. However, the level of sales required for
adjusted net income to break-even shifts only 7 percent
(from $1.367 million to $1.268 million per quarter), much
less than would have been expected considering that net
earnings decline by only 10 percent. The depreciation
flows generated by the significant increase in fixed

investment are included in adjusted net income and these

depreciation flows keep the level of sales required to

138

 

139

break-even on adjusted net income for 1.01 from being much
higher.

However, depreciation flows are simply a re-capture
of funds expended at the time of the investment in a
private facility. That net liquid assets and break-even
sales for the one private facility system compare so
favorably is largely because of an assumption of this
research. For all experiments, it is assumed that the
original investment in plant and equipment (warehouses)
is made entirely with externally secured funds and that
the funds are obtained 50 percent through debt and 50
percent through equity.

Another assumption which could be made is that the
money needed to finance the private warehouse exists
within the firm as the result of retained earnings; or
if the private facility is not built, that this excess
capital is put into fairly safe marketable securities
yielding 6 percent per year, before taxes. The same
effect can be obtained by assuming the company goes to
the outside for the entire financing through equity sources
and then decides not to build the public facility. Table
5.6 compares the actual effects on selected financial
variables in Experiment 1.04 to an estimate of those
which would have been achieved if the alternate assumption
concerning financing had been used.

Under the alternative assumption, private compares

much more favorably to public than before in return on

140

muma~0© mo mcoflHHHE CH

 

 

 

 

m
mHMHHOU NO mmuqmmDOnu GHH

ema.- mmo\- mom. mom.a ooo.H kmm.a mmaoocH umz "cm>m-xmmum

Hoo.- omo. meo. oma. 45o. ooa. suasvm so cusumm

moo. moo. mmo. omo. vmo. «mo. mummmm so cusumm

m~.~ A. m.4 mo.~ m.m H.m caumm namnuso

~mo.- mmo. mam. mo.H omm. «km. suasam\unmo

m.maau A.H4m- 4.0mkm 0.4Hma o.oomm e.mmma Hanna; umz

4.4mmm n.4Hom o.o~ov H.m~mm Hmummma annoy

m.mmma m.mou m.4m~a H.ova- 0.0G «.mm- Hmummma oflswaq umz

k.m- m.m- m.om m.mm 0.4m m.me H9:85 umz

mumcumua< Hmspoa mumcumuad anabom mumcumuam Hmnuofi manmwum>
mum>flum I Deanna OHHQDm mum>fium

 

 

.mcfloamcflm mo mousom mcflcumocou mcoflumESmm¢ m>HumcuwuH¢ Hews:
mflmaamcd mum>wum m> veabsm map on moanmflnm> mom samuumu mo conflummsounn.o.m mqm<a

141
investment. However, the superiority of public warehousing
as a hedge against economic downturns or market reversals
is magnified. Under the alternate calculations, sales can
drOp 14 percent further before adjusted net income turns
negative if a public instead of a private warehouse is
used. Further, there are $1.228 million dollars more of
net liquid assets to cushion against unprofitable periods
if public is used instead of private.

The implication of this comparison is that private
warehousing is typically a more risky alternative than
public warehousing. If the firm goes to the outside for
financing of the private facility, partly through equity,
the return of capital in the form of depreciation flows
can provide additional protection against reversals.
However, if the firm decides on the public alternative
it likely can go to the outside for the needed liquidity
protection and use only a fraction as much of its outside
financing capacity as would be used under the private
facility alternative.

Sensitivity Experiments: .15 and —.05 Growth Rates.
The ranks of financial variables which are significantly
changed by the use of a public facility instead of a
private facility at Chicago under the assumption of an
annual growth rate of .15 (Experiment 1.07) and under
the assumption of an annual growth rate of -.05
(Experiment 1.12) are compared to the results under the

.05 growth rate assumption in Table 5.7. The ranks of

142

 

 

 

 

TABLE 5.7. —-Comparison of Financial Variables Signifi-
cantly Affected by Using One Public Instead
of One Private Warehouse Under Varying Growth
Rate Assumptions for the Health Cares Company

Avg. Quarterly
Rank % Chapge

Variable -.05 .05 .15 -.05 .05 .15

Debt Service Coverage 1 l 6 155 107 86

Plant and Equipment 2 2 1 -100 -100 -100

Deferred Taxes 3 3 2 -100 -100 -100

Long Term Loans 4 4 3 ~100 -100 -100

Depreciation 5 5 4 -100 -100 -100

Fixed Asset Turnover 6 6 5 -100 -100 -100

Net Liquid Assets 15 7 17 - 15 - 75 - 18

Interest Expense 7 8 7 - 69 - 61 — 48

Working Capital/Total Assets 9 9 8 51 45 39

Net Worth 8 10 10 52 44 33

Total Assets 12 11 11 - 35 - 33 - 28

a Regression Co—efficient 16 12 12 - 12 — 32 - 24

Total Liabilities 11 13 14 - 40 - 31 - 23

B Regression Co-efficient 17 l4 l3 - 12 - 22 - 28

Total Variance 14 15 9 - l7 - 20 — 35

Return on Net Worth l3 16 16 25 20 22

Break-even: Debt Servicing 21 17 21 - 4 - l8 — 6

Return on Assets 10 18 22 50 16 5

Taxes Payable 18 19 19 - 9 - 15 - 17

Net Earnings 19 20 18 - 9 — 10 - 19

Cash 20 21 20 - 6 - 10 - 7

143
distribution variables significantly affected under .15
and -.05 are exactly the same as under the .05 growth
rate assumption.

The relative importance of financial variables
remain strikingly consistent as the growth rate assumption
is changed. The Spearman Rank Correlation of 1.04 and
of 1.09 and 1.14 are .93 and .95, respectively. Among
those variables which do exhibit a change in rank, net
liquid assets rank much higher in 1.04 only because the
absolute value of net liquid assets in the base experi-
ment for the .05 growth rate is much smaller than in the
base experiments for the other two assumed growth rates.
The actual change in net liquid assets is about the same.
The only other inconsistencies were expected. The
superiority of public over private warehouses in both
return on assets and in debt service coverage improves
significantly as the sales growth rate declines from .15
to .05 to -.05 as the theory of leverage suggests it
should.

The most interesting finding from these experiments
is associated with accounts payable, which does not exhibit
a 10 percent change and therefore is not listed as a
significantly affected variable under any of the growth
rate assumptions. However, it does experience a per-
sistent change each time a shift from private to public
warehousing is tested. The reason is that public ware-

housing charges generate accounts payable while many

144
private warehousing expenses do not and therefore the
shift to public warehousing can be a small but permanent
source of financing. This finding is developed in detail
in the last section of this chapter.

Research By-Products to Financial Analysis

 

Comparison of experimental results with different
growth rate assumptions but with no change in the ware-
housing system provides interesting insights to financial
analysis and control. The following is a comparison of
the values of certain key financial ratios in Experiments
1.12, 1.02, and 1.07, which are the experiments with two

private warehouses under the three growth rate assumptions.

 

 

 

Experiments
Ratio 1.12 1.02 1.07
Current Ratio 2.05 1.87 1.70
Acid Ratio 1.12 1.06 .94
Cash Turnover 6.91 10.02 13.32
Debt/Equity 1.07 1.18 1.34
Debt Service Coverage 1.74 2.20 2.42

 

This comparison is an experimental verification of
the classic relationship between financial position and
sales growth. As the rate of sales growth increases,
earnings and return on assets and equity also increase.

Yet the current and acid ratios, cash turnover, and the
debt/equity ratio all experience significant deterioration.

This lends strong support to the financial axiom that

145
profitable, high growth oriented companies often get
into financial difficulty because as sales grow, invento-
ries and accounts receivable grow faster than accounts
payable. The net result is a drain on cash. This
problem is especially crucial for wholesale companies
because their operations revolve around the purchase,
control, and selling of inventories.

The relationships between growth and the debt/equity
(D/E), debt service coverage, and acid test ratiasare
also interesting. As the growth rate in sales increases,
the D/E ratio increases because current assets increase,
causing short term debt to rise. However, the debt
service coverage ratio increases substantially, indicating
that the increase in the D/E ratio has not hurt the
company. Indeed, this type of increase in D/E is one
that redeems itself as sales growth drops back off. A

major implication is that only funded debt should be

Debt
Equity

 

included in the ratio.
Likewise, in calculating the debt service coverage
ratio, including the interest expense on working capital
loans introduces a downward bias into the calculation.
The purpose of this ratio is to determine how much pro-
tection against sales, earnings, or cash flow downturns
is afforded by the current relationship between cash
flows and debt servicing requirements. Since working

capital loans should be redeemed from liquidating current

assets as sales turn downward, liquid protection for

146

this type of debt service is really not necessary. The
implication is that only funded debt should be included
in debt service coverage calculations.

As illustrated in the preceding table, the acid
test ratio improves significantly as sales turn down
and deteriorates when sales are increasing. Thus, the
appearance of liquidity deterioration which is given by
the acid test ratio as sales grow would tend to disappear
if sales were to turn back down. The implication is that
the acid test ratio is a very poor measure of the suscep-
tibility of the firm to economic downturns or market
reversals. Rather, its value is in signaling the danger
of financing too much of the growth of the firm with
short term debt, if the growth pattern is expected to
continue. This analysis is tempered with the reCOgnition
that if the firm were operating improfitably, a deteriora-
ting acid test ratio as sales increase might be a warning

of impending financial difficulty.

Model Changes Required for

 

the Appliances Company

 

The appliance wholesale industry is dissimilar to
the health cares wholesale industry in several major
respect. The demand pattern is seasonal for appliances
and not for most health care products. The appliances
product lines are much less diversified. Transportation

and handling characteristics are very dissimilar between

147
the two industries. Experimentation on the modeled
appliances company therefore offers a good test of the
consistency of research results across industry lines
and this section describes the basic model changes
which are required to model the appliances company.

Product and Demand Information

 

The data in the order file are changed to reflect
the characteristics of fifty products identified in
questionnaire survey reSponses by two companies in the
appliances industry. The fifty products included product
items of various styles and colors in the following
product lines: washers, dryers, refrigerators, and
ranges. Since inventory control is performed on each
of the fifty products, tracked product extrapolation
factors are unnecessary and all routines including
extrapolation factors are modified accordingly.

Demand Forecastingylnformation

 

The mean and standard deviation of the distribution
from which daily sales are randomly generated are left
unchanged from the health cares industry. However, the
seasonal weighting factors are changes from wi = 1.0 for

all i to the following:

148

 

 

4 Week Period wi 4 Week Period wi
1 i7 8 17
2 .7 9 1.2
3 .8 10 1.1
4 1.0 11 .9
5 1.2 12 .7
6 1.3 13 .6
7 1.4

These weighting factors are based on the seasonal demand
patterns identified in the data survey of two appliance
firms.
Distribution System Changes

The two alternative distribution centers (DC's)
remain Chicago and Columbus. However, reflecting the
general location patterns of appliance industry manufac-
turers, the three supplier locations (MCC's) are changed
to:

1. Benton Harbor, Michigan

2. Marien, Ohio

3. Ft. Smith, Arkansas

Changes in CostingiEquations

 

The outbound transportation regression equations
for the appliance industry are calculated based on rates
provided by a transportation consulting firm and on trans-
portation expense data provided by the respondent companies.
The inbound transportation freight rates are from the same

sources. Throughput, communications, and facility

149
investment expense data are based on information provided
by the respondent companies and by an independent pro—
fessional society.

Balance Sheet Initialization and Model Pre-testing

 

The initial values of balance sheet items are
established based on industry average data from the
Robert Morris Associates reports for the appliance whole-
sale industry. Various financial and cost parameters
were adjusted through pre-testing so that the model
output for the base experiment is a reasonable approxima-
tion of the financial performance of the hypothetical

"typical firm" in the industry.

Comparison of Results Between Industries

 

The results from experimentation on the modeled
appliances company are reported and compared to the
results from the health cares company in this section.

Addition of a Second Warehouse

 

The financial variables which are significantly
affected by the addition of a private warehouse at
Columbus for the appliances company under the assumption
of growth rates of .05, .15, or —.05 are listed in
Table 5.8. The ranks and the average quarterly per-
centage changes for those variables under each of the
growth rate assumptions are also listed. The ranks and
the average quarterly percentage changes in distribution

variables under each of the growth rate assumptions are

150

TABLE 5.8. --Comparison of Financial Variables
Significantly Affected by the Addition of
a Second Private Warehouse Under Varying
Growth Rate Assumptions for the Appliances

Company.

 

 

Avg. Quarterly

 

 

 

Rank % Change
Variable -.05 .05 .15 -.05 .05 .15
Net Liquid Assets 1 l l - 54 -l90 -115
Short Term Loans 10 2 2 20 49 57
Interest Expense 7 3 3 23 29 32
Cash Turnover 2 4 13 47 28 8
Plant and Equipment 4 5 4 '26 26 26
Long Term Loans 5 6 5 25 25 25
Depreciation 6 7 7 24 24 24
Deferred Taxes 8 8 6 22 23 25
Cash 3 9 16 - 28 - 21 - 8
Fixed Asset Turnover 9 10 8 - 21 — 21 - 21
Debt Service Coverage 15 ll 9 - 9 - 15 - 19
Accounts Payable 13 12 10 - 15 - 15 - 17
Net Worth 16 13 14 9 11 8
Net Income 14 14 15 - 10 - 4 - 4
Return on Assets 12 15 12 - 16 - 7 - 9
Return on Equity 11 16 ll - 17 - 10 - ll

 

151
listed in Table 5.9.
On balance, the change to a two warehousing system
has the same relative impact on financial variables
in this industry as in the health cares industry. This
is demonstrated by the following Spearman Rank Correlations
between their ranks in the two industries, which would

have been significant at the .01 level at .46.

Correlation of Ranks of Finance

 

 

Growth Rate Variables Between Industries
.05 .50
.15 .57
-.05 .55

However, there are several major differences between
the two companies. The total distribution system cost
changes are substantially smaller in the appliances
company from addition of the second warehouse. Because
of differences in freight rate structures between the two
industries, the addition of a second warehouse generates
much larger transportation savings in the appliances
company. However, this advantage is partially offset by
the heavier additional fixed facility investment for
appliances, which is necessary because of the excess
capacity created by the seasonal pattern of demand in
the appliances industry.

The net result of the cost changes is that while
the addition of a second warehouse is an improfitable
move for the appliances company, it is not as improfitable

as in the health cares company. This simply means

152

TABLE 5.9. --Comparison of Distribution Variables
Significantly Affected by the Addition
of a Second Private Warehouse Under
Varying Growth Rate Assumptions for the
Appliances Company.

 

 

Avg. Quarterly

 

 

 

Rank % Change
Variable -.05 .05 .15 -.05 .05 .15
4 Day ORCT l l l 214 210 199
Throughput Expenses 3 2 3 56 53 50
Communication Expenses 4 3 4 50 50 50
Stockouts 2 4 2 63 40 147
Facilities Capital Exp. 5 5 5 23 25 25
5 Day ORCT 6 6 6 24 24 24
Depreciation 7 7 7 24 24 24
S. D. Stockout Days Delay 9 8 8 l7 — 23 - 18
Outbound Transp. Exp. 10 9 10 - 10 - 10 - 10

AVg. Stockout Days Delay 8 10 9 l8 - 4 l7

 

153

that the profit desirability of adding a second warehouse
will vary depending upon the specific situation, which is
an obvious fact not in question for this research.

Of special interest here is that the addition of
a second warehouse for appliances causes inventories to
increase by only one to four percent while in health
cares inventories increase by 33 to 90 percent. It is
generally accepted in the distribution field that safety
inventories will normally increase with the addition of
a warehouse and that a good approximation of that increase

can be found from the formula:1

35 551‘“)
n __—.__—
“n
where: SSn = safety stock for n locations
= the number of locations
SSl = safety stock for one location

Since safety inventories are designed to be 25
percent of total inventories in the modeled companies,
and since the general formula estimates the increase in
safety stocks from adding a second warehouse at
approximately 40%, the expected increase in total inven-
tories is approximately 10%. The primary reason the
results from this research depart from the general
pattern is that the above formula is based on the
assumption that demand at the two locations are independent
of one another. Since sales in LREPS-F are randonly

generated for the total company, and then randomly assigned

154

to the individual DC's, there is a high degree of
dependence between demand at the two locations.

The inconsistency in research results between the
two industries exists because demand is relatively
stable in health cares and highly seasonal in appliances.
A prime advantage for using a exponential smoothing
inventory forecasting technique such as the one in
LREPS-F instead of using simple trend projection is that
exponential smoothing will detect and forecast inter-
mediate upswings and downturns in the demand pattern.
The temporal swings detected in appliances are usually
actual seasonal swings and the forecast technique is
very accurate. The swings detected and forecasted in
health cares are usually only random noise around the
stable trend line and therefore result in a substantial
amount of forecast error. This forecast error introduces
instability to the level of inventories which is magnified
when daily sales are randomly assigned between two
locations.

Because appliance inventories are not significantly

Debt

affected, the changes in short term loans, the EEG???

ratio and the acid test ratio are much smaller in the
appliances company than in the health cares company.
Furthermore, the risk or defensive posture does not
deteriorate as badly in the appliances company, which is
indicated by the much smaller changes in total variance

of adjusted net income and the level of sales required

155

to break-even on earnings and adjusted net income.

Conspicuous in that they did not change substantially
in importance between industries are accounts payable
(A/P). The downward shifts in A/P remained approximately
the same in the appliances as in the health cares company
deSpite the substantial differences in inventories. The
relationship found between the change in the number of
warehouses, inventories, and accounts payable is the most
striking finding of this research and is examined in
detail later in this chapter.

Public Versus Private Warehousing

 

Those financial variables which are significantly
affected by using a public warehouse instead of a private
at Chicago for the appliances company under the assumption
of growth rates of .05, .15, or -.05 are listed in Table
5.10. The significantly affected distribution,variables
under each of the growth rate assumptions are also listed
in Table 5.10.

The ranks of the distribution variables are exactly
the same as those for the health cares industry. The
relative importance of changes in financial variables
from going to public warehousing are approximately the
same as in the health cares industry. The Spearman Rank
Correlations for comparison of the results of the two

industries, which would have been significant at .41, are:

156

TABLE 5.10 --Comparison of Variables Significantly
Affected by Using One Public Instead of One
Private Warehouse Under Varying Growth Rate
Assumptions for the Appliances Company

 

 

 

 

 

 

Avg. Quarterly
Rank % Change
Finance Variables -.05 .05 .15 -.05 .05 .15
Debt Service Coverage 1 1 1 300 356 204
Plant and Equipment 2 2 2 -100 -100 -100
Deferred Taxes 3 3 3 -100 -100 -100
Long Term Loans 4 4 4 -100 -100 -100
Depreciation 5 5 5 ~100 —100 -100
Fixed Asset Turnover 6 6 6 -100 -100 -100
Interest Expense 8 7 7 - 75 - 71 - 60
Return on Equity 7 8 8 80 59 53
Net Working Capital/Total
Assets 10 9 9 64 55 47
Return on Assets 9 10 11 75 55 42
Working Capital/Total Assets 11 ll 10 59 53 46
Net Worth 12 12 12 - 39 - 37 - 32
Total Assets 13 13 13 - 37 - 35 - 31
Total Liabilities 14 14 14 - 34 - 32 - 30
Taxes Payable 18 15 20 - 6 - 13 4
Total Variance l6 l6 l7 - 10 - l3 - 16
Net Liquid Assets 15 l7 18 17 10 12
Debt/Equity 20 18 16 9 16
<1 Regression Co-efficient 17 l9 l9 - - 9 - 12
B Regression Co-efficient 19 20 15 - - 7 - 19
Distribution Variables
Throughput Expenses 118 147 173
Depreciation Expense -100 -100 -100

Facilities Capital Expense

-100 -100 -100

157

Correlation of Ranks of

 

 

Sales Growth Rate Variables Between Industries
.05 .70
.15 .83
-.05 .97

The few differences between the two industries
which do exist are because the appliance industries
sales patterns are seasonal and the health cares industries
are not, with the result that more excess capacity exists
for appliances and thus requires heavier investment in
private facilities. Resultantly, public warehousing is
relatively more attractive both profit-wise and in terms
of return on assets and equity in the appliances industry.
Because of the seasonal fluctuations in sales, the
variance of net income also is relatively lower for the
appliances industry if public instead of private ware-
housing is used.

The researcher expected that the improvement in
required level of break-even sales for going to public
warehousing would be much stronger in the appliances
industry because of the heavier fixed investment. This
did not occur because the heavier flow of depreciation
generated funds from the private facility in appliances
keeps the break-even level for the private facility
alternative much lower than it otherwise would have been.
The increased flow of depreciation funds is only a re-

capture of additional required investment and lends a

158
misleading appearance of liquidity, as is explained on

pages 129-134.

Warehousing Decisions, Inventories, and

 

Accounts Payable

 

The most significant interactions between ware-
housing decisions and financial variables identified
in this research are:
l. The direct effects of an additional ware—
house on inventories and accounts payable
and the interaction between inventories and
accounts payable.
2. The effects of the increased operation
expenses from warehouse system changes
on the level of accounts payable.
To fully develop the interaction between additional
warehouses, accounts payable and inventories it is
necessary to examine the behavior of inventories and
accounts payable as sales change over time. Therefore
this section is presented in three parts. The first
part is the general treatment of the effects of sales
changes on inventories and accounts payable. The second
part is an analysis of the effect of additional ware-
houses on inventories and accounts payable, of the inter—
action between inventories and accounts payable, and the
implications of these findings to management sciences.
The last part reports the relationship between ware—
housing decisions, changes in operating expenses, and

changes in accounts payable and describes general

principles developed from this relationship.

159

Sales Induced Changes in Inventories and Accounts Payable

 

In the following exhibit, sales, inventories, and
accounts payable are portrayed by their average values
over twenty quarters under growth rate assumptions of
-.05, .05, and .15, respectively, with one private ware-

house in the health cares industry.

Average Quarter Value ($MM)

 

A/P Inventory Net Liquid

 

 

 

 

 

 

Experiment Sales Inven A/P Inv Turnover Assets
1.11 (-.05) 2.001 .718 .688 .95 3.05 .315
1.01 ( .05) 2.540 1.026 .797 .79 2.71 .083
1.06 ( .15) 3.202 1.251 .932 .74 2.63 -.486

As sales rise, inventories increase and accounts payable
also increase, with inventories rising faster than accounts
payable, as demonstrated by the drop in the percent of in—
ventories financed by accounts payable (%£%)

A sales induced increase in safety (buffer) inven-
tories is financed by cash, not by accounts payable.
Such an increase does not repeat itself (it is a some-
what permanent shift in the level of inventories).
Although an additional purchase of safety inventories
immediately causes accounts payable to increase, at the
end of the allowed days of trade credit the account is
paid from cash and accounts payable returns to the
previous level.

But sales induced increases in re-order inventories

are at least partially financed through accounts payable.

160
This increase in inventories repeats itself each inventory
cycle and at the time of each inventory replenishment
will also "replenish" the amount owed on trade account.
The specific percentage of sales induced increases in
re—order quantity which will be financed through accounts
payable depends upon the re-order cycle time, the terms
of trade credit, and the number of days before the
increased sales which stimulate the increased inventories
are converted from receivables to cash.

A simplified illustration of these relationships is
presented in Table 5.11. The illustration is of
hypothetical weekly patterns in inventories, accounts
receivable, accounts payable, and cash when sales are
first $10 per week and then $15 per week. The firm is
one with only one product and to avoid complications to
the presentation is assumed to have no Operating costs
and to make no profit on the re-sale of the inventories.
The inventory re-order cycle is five weeks, and terms on
accounts receivable and accounts payable are two and
four weeks, respectively. The company policy is that
safety stocks are two weeks of expected sales and re—
order quantity is five weeks of expected sales. It is
assumed that the company is on-going, that inventories
have been replenished at the end of the first week and
that the net balance in cash at the end of the first week
is thirty dollars.

In the first ten weeks of operations actual and

eXpected sales are $10 per week. In ordering inventories

161

.meHHouce>cH
Eoum coHueHec ucemeumeu meHem cheeB Ode .cee3 ecu mo cce ecu um eecwHec ecu meumeHccH

 

 

 

 

 

 

uec cam “xee3 ecu mcHHSp mcoHueHec meueoHch I “xee3 ecu mcHHsc coHqucm meueoHch+m
mI mH me O on mH mH mv mH mm
mm mH mm on mH mH om mH ON
Oe mH mm om mH mH mm mH mm
mm mH me om mH mH om mH mm
OH mH mm mm on mH mH mOH mm mH Hm
mI mH me O mm om mH mH mv mH ON
mm mH mm om mH mH OO mH OH
ov mH me om mH mH me mH OH
mm mH me on mH mH om mH OH
OH mH mm mm om mH mH mOH mm mH OH
O mH mm H mm om mH mH mv mH mH
Oh mH mm om mH mH OO mH vH
mm mH mm om mH mH mm mH MH
OH OH mm mm OH mH OO mH NH
om OH mm mm ON OH OH mOH mm OH HH
Om OH om O om ON OH OH om OH OH
OO OH om ON OH OH OH OH O
om OH Om ON OH OH Om OH O
Ov OH om ON OH OH om OH O
Om OH om om ON OH OH on om OH O
ON OH om O om ON OH OH Om OH m
OO OH om ON OH OH ov OH H
om OH om ON OH OH om OH m
OH OH om ON OH OH OO OH N
om OH om om ON OH OH on om OH H
uez cH uso uez I + uez I + uez + meHem xeez
cmeu e.>mm mucdoeec e.eem mucsoood mmeHHouce>cH

 

 

AmumHHoo cue Hmuuamo mcuxuoz can
meHem cH emcmco ceeBuem chmcoHueHem ecu mo coHumHumSHHH ceHmHHmEHm «IIHH.m MHmca

162
during week 11, the firm expects sales to increase to
$15 per week and therefore orders an additional $10 of
inventories for safety stock (2 weeks times the $5
increase) and raises the re-order quantity by $25 (5 weeks
times $5). Thus, inventories received in week 11 total
$85 instead of $50. Actual sales, as the firm expected,
go to $15 per week every week after week 11. Therefore,
the inventories re-ordered each re-order cycle thereafter
total $75, the new re—order quantity.

For one inventory cycle after the increase in inven-
tories, beginning accounts payable go from $50 to $85
and then stabilize at $75. This illustrates the earlier
contention that increases in safety stocks, unlike
re-order quantities, are not repetitive increases and
therefore do not generate a permanent shift in the levels
of accounts payable.

Once the system has stabilized following the change
in the rate of sales, the relationship between sales
induced changes in re-order inventories, working capital,
and accounts payable can be observed. The period of
weeks six through ten is a stable period with sales at
$10 per week and the period of weeks twenty-one through
twenty-five is a stable period with sales at $15 per
week. These periods are A and B, respectively. Consider
the following comparison between the values of the elements
of working capital at the beginning week of each period

(week 6 vs 21), at the ending week of each period (week

163

10 vs 25), and the average value for each element over

each period.

 

 

 

 

 

 

 

A B B - A

Element Beg. End Avg. Beg. End Avg. Beg; End Avg.
Inven-

tories 70 30 50 105 45 75 35 15 25
Accounts

Rec. 20 20 20 30 30 30 10 10 10
Cash 30 20 40 10 -5 25 -20 -25 —15
Accounts

Pay. 50 0 40 75 0 60 25 0 20

The difference in inventories between B and A at the
beginning of the inventory cycle is $35, which consists
of the $10 permanent upward shift in safety stocks and
the $25 of additional re-order inventories. Accounts
receivable (A/R) also shows a permanent increase of $10
and the total increase in non-liquid assets (inventories
plus A/R) is $45. These increases are financed by a
$20 drop in cash (because the $10 respective permanent
increases in safety stocks and A/R both reduce cash
by $10) and by a $25 increase in A/P which occurs because
of the $25 increase in re—order inventories.

At the end of the last week, the additional financing
through A/P has been redeemed with cash, but all except
$5 of the additional re-order inventories have been
worked off. Thus only five additional dollars of cash
are required. Since the increases in safety stocks and

A/R are permanent, their values and their impact on cash

164
remain the same as at the beginning of the cycle.

Based on the average values at the end of each week
in the five week cycle, non-liquid assets increase by
$35, which is broken down as follows: +$10 each for
safety stocks and A/R and +$15 for average re-order
inventories. The increases in safety stocks and A/R
reduce cash by a total of $20. However, the additional
working inventories decline over the entire inventory
cycle and the additional accounts payable are constant
until totally depleted. Thus, the average increase in
accounts payable ($+20) is larger than the average
increase in working inventories ($+15) which it is
financing and therefore takes $5 off the strain placed
on cash by safety stock and accounts receivable.

The analysis of average values over the inventory
cycle is most important for multi-product companies
because the many product cycles tend to balance each
others peaks and valleys, with the total result being
the sum of their averages. As the number of products
approach one, the beginning and ending points of the
cycle become more important. Whichever is the case,
however, this illustration and the experimental results
demonstrate that accounts payable can be a somewhat per-
manent source of financing for increases in re-order

inventories.

165

Additional Warehouses, Accounts Payable, and Inventories

 

Although a definite positive relationship is estab-
lished in the previous section between sales induced
changes in inventories and accounts payable, the experi-
mentation reveals no such relationship between inventory
adjustments resulting from the addition of a warehouse
and changes in accounts payable.

Indeed every experiment involving the change to a
two warehouse system exhibits an increase in inventories
and a decline in accounts payable. That the decline in
accounts payable is approximately the same in theliealth
cares industry and in the appliances industry, in which
the average expansion of inventories are quite different,
is further evidence that the resultant adjustments in
accounts payable and in inventories from switching to a
two warehouse system occur quite independently of one
another.

The effect of an additional warehouse on inventories
is explained on page 153. The decline in accounts payable
occurs because with the addition of a second warehouse,
the number of days required to accumulate to the minimum
allowable replenishment shipment of inventories from
each supplier to the individual distribution centers
increases. Since orders occur less frequently, accounts
payable build up less slowly. At an early stage additions
to and deletions from accounts payable (A/P) begin to

balance out and accounts payable stabilize. Since A/P

166

build up slower in a two warehouse system, they stabilize
at a lower level. The quarterly patterns of accounts
payable in Experiment 1.01 and Experiment 1.02 are compared
graphically in Figure 5.2. As can be seen, when the
two systems stabilize beyond the second quarter of the
first year, accounts payable in the one warehouse system
consistently is at a higher level than accounts payable
in the two warehouse system.
That accounts payable do not increase, but rather
decline when inventories are adjusted upward by the
move to a two warehouse system leads to several interesting
observations of the relationship between accounts payable
and inventories. As was true for a sales induced increase
in safety stocks, the level of accounts payable immediately
increases, but returns to its initial level at the expira-
tion of the allowed credit period, if the move to a two
warehouse system results in an increase in safety stocks.
The effect which increases in re-order inventories
resulting from the addition of a warehouse, has on
accounts payable is more complex. To avoid unnecessary
complications, the following analysis assumes that no
trade discounts are available and that the firm has an
established policy of paying all accounts at the expira—
tion of the allowed credit terms. While this analysis
relates specifically to the effects of the addition of
a warehouse, it pertains to any management action induced

change in re-order inventories. It can be shown that

167

 

168

an increase in re-order quantity inventories, with no
change in actual or expected sales, has two counter-
balancing effects on accounts payable: (1) an additional
amount equal to the increase in re-order inventories is
added to accounts payable at the beginning of each
inventory cycle, and (2) the inventory re—order cycle
is lengthened by the same percentage by which the re-
order quantity is increased. The net result is that a
change in re-order quantity resulting from management
action (that is, not because of a change in the daily
sales rate) leaves the level of accounts payable un-
changed. The proof of this relationship is now given.

Assume that the re-order quantity for a given product
is R with constant daily sales S and inventory (re-order)
cycle time g = i1. The accounts payable cycle time, p,
is the allowable days under credit agreement. Although
the assumption is not critical to the proof, assume that
f1<ilf Then, over the total inventory cycle the amount
in accounts payable is R from day 1 to day p and 0 from
day p to day i1. Thus, the weighted average balance of

accounts payable is

*0=§_*R
1

 

i-p
APl-‘(i-WHi
1 1

Now, assume re-order quantity is increased by management
action (sales are constant) to (l+w)R. Since S is
constant, the new inventory cycle time i2, is (l+w)il,

and the new average balance of accounts payable is

169

AP =Ei’—*(1+w)R=

(1+w)R
2 2

___B___ *
(1+w)il

The change in accounts payable is
_ = * - E. *
APz Apl (175F11— (“wm 11 R

and since the terms (l+w) cancel each other

AP -AP =P—*R-E—*R=o
2 1 11 11

It can be shown that this relationship also holds for
p>iamip=ju

Thus, a tautology is developed that exogenous changes
in re-order inventories, all other factors being constant,
leaves accounts payable unchanged. A corrollary to this
conclusion is that the percentage of the investment in
inventories which is financed through accounts payable,
g , moves inversely to changes in re-order quantity,
all other factors being constant. Since A/P is a cost
free sourse of financing (assuming the company has
already made its decision concerning discounts and credit
reputation), it implicitly follows that the average cost
of the capital invested in inventories is an increasing

function of EOQ. Functionally, the average cost of

capital for inventories (Ic) is

 

in am
Ic R+B 0+” n+3) k
where B = the investment in safety or buffer

stock, which is constant, and as
shown earlier not financed by in-
ventories after the first period.

k = the average capital costs of all
sources of inventory financing
other than accounts payable.

170

Eka
l

—§;§—- = the percentage of inventories

financed through accounts payable.

Since g decreases by the same percentage as R in-
creases, changes in R cause no change in the numerator but
cause the denominator to increase in the term

2 * R
1____
R+B

Thus, the value of the term declines and the value of

2*12
1.

———§—) increases as R increases, demonstrating that

(l - R+

IC increases with increases of R. The relationship
between re-order quantity and Ic is graphed in Figure
5.3 under the assumption that the sales rate is constant
and the average cost of capital not provided by accounts
payable is 10 percent per year after taxes. The weighted
average cost of capital for inventories is shown to be
an asymptotically increasing function for the re-order
quantity.

The traditional formula for calculating economic
order quantities assumes a constant value for the

2 It has been demonstrated that

average cost of capital.
this assumption is not always an accurate one.

Warehousing Decisions, Operating Expenses,
and Accounts Payable

 

 

The comparison between any two experiments on LREPS-F
with the same sales growth rate and the same number of
warehouses demonstrates no change in inventories. For

these comparisons, since the number of facilities does

171

 

172

not change, it was expected that accounts payable would
not change. They do, however, and closer examination
reveals that the reason is the relationship between
Operating expenses and accounts payable. The following
is a listing of changes in expenses which give rise to
accounts payable and the associated changes in accounts
payable between all possible combinations Of experiments
with the same growth rate in sales and the same number

of facilities in the health cares industry:

 

 

 

 

AAverage Quarterly AAverage I

Experiments Expenses A/P

1.01 vs 1.04 $40.9M $39.4M
1.02 vs 1.05 56.2M 38.2M
1.06 vs 1.09 52.4M 32.6M
1.07 vs 1.10 74.3M 45.0M
1.11 vs 1.14 26.3M 23.0M
1.12 vs 1.15 40.4M 26.4M

In each case, the differences reflect the increases
in throughput expenses from changing from a private to
a public warehouse. The increases in accounts payable
occur because of two related changes. First, in the
model throughput expenses occur weekly, but deplete
cash only after a five week credit grace period. There-
fore, for the first five weeks of the increased through-
put expense there is not an associated depletion in cash.
In the sixth week, and in each week thereafter, one
week of additional expenses is depleted from cash. Also,
one weeks additional expenses is both added and subtracted
from accounts payable, with A/P resultantly stabilizing

at the amount five times one weeks additional expenses

173

above the level of A/P prior to the increased expense.

Secondly, when public warehousing is used, 100
percent of all throughput expenses are added to accounts
payable when theyare first incurred. Some of the
throughput expenses incurred when private warehousing
is used are more immediately paid from cash. In this
model, it is assumed that 50 percent of private facility
throughput costs are paid from cash at the end of the
week in which they are incurred and that the other 50
percent generate an account payable. Resultantly, the
same amount of throughput expenses which were incurred
using private warehousing will generate 50 percent more in
accounts payable using public warehousing.

Mathematically, the total change in the level of
accounts payable, after they stabilize, because of the
shift from private to public warehousing is:

AP = (X2 * a)p - (Xl * b)p

Where X X = the weekly amount of throughput

I
1 2 expenses for the private and
public warehousing alternatives,
respectively.

b,a = the percent of the weekly throughput
expenses which are added to accounts
payable for the private and public
warehousing alternatives, respectively.

p = the number of weeks allowable under
trade credit.

174

A note of caution is.hiorder. After accounts
payable stabilize following the change, weekly cash
outflows from the additional expenses are approximately
the same as the weekly additional expenses because
expenses incurred in time period t are paid in t+5.
Cash as a flow entity must be distinguished from cash
as a status variable. The status variable cash at t is
a composite of the cash inflows and outflows in t-l,
t-2, t—3, ...t—n. Therefore, at any week m beyond
the stabilization point (week five), the change in
cash (cash flow) which will have occurred is:

Cashm = m(X2 - X1) - [(X2 * a)p - (Xl * b)p]

where m(X2 - X1) = the total additional expenses

which have been incurred up
through the mth week since the
change was made.

An important part of the art of capital budgeting
analysis is the necessity of estimating the cash (working
capital) requirements needed to support the project under
study. This research has indicated that this estimate
should consider the potential changes in the level of

accounts payable financing when:

l. The project results in a change in an expense
which generates an accounts payable.

2. The project results in trading an expense
which is paid from cash for an expense which
generates an account payable; or vice versa;
or results in trading an expense for one which
also generates an account payable but carries
different trade terms.

 

175
Indeed, this research indicates that any management
decision which involves a shift in expenses should
include explicit consideration of the potential impact

on accounts payable.

CIIAPTE R V--FOOTNOTE REFERENCES

lBowersox, Smykay, and LaLonde, 9p. cit., p. 221.

2

Ibid., p. 204.

176

 

CHAPTER VI

SUMMARY AND CONCLUDING REMARKS

Although financial position is typically managed by
correcting adverse changes after they occur, a more en-
lightened approach to financial management is to anticipate
these adverse changes and prevent their occurrence. This
research is based on two premises:

1. Much of the change in financial position

of the firm over time is the result of

overt management decisions in other

functional areas of the firm.

2. Most operational decisions within the

firm trigger changes, either directly

or indirectly, in many of the variables

of financial position.
A major step toward preventive financial management is
to develop a better understanding of the effects of
decisions in the functional areas on financial variables.
Such an understanding will also facilitate more correct
decisions in the functional areas.

The objective of this research is to study the nature
of the effects of warehousing decisions on financial
variables. With a fuller understanding of this inter-
action financial managers are better able to anticipate

changes in financial position resulting from a particular

warehousing decision and distribution managers are better

177

 

178
able to make correct warehousing decisions.
To accomplish this goal, four steps were taken:

1. A dynamic computer simulation model was built
upon which experimentation and sensitivity
analysis could be performed to study the total
impact of five alternative warehousing decisions
on financial variables in two test industries.

2. The experimentation and sensitivity analysis
on the model were performed, with only one
variable exogenously changed per experiment
and the resultant changes in financial
variables were Observed.

3. Those financial variables which are "significant-
ly" affected under varying economic conditions
in each of the two industries by the alternative
warehousing decisions were identified from the
results of the experimentation.

4. The results were studied for general relation-
ships which would help explain the interaction
between warehousing decisions and changes in
financial variables.

There are several results of the research which

should help financial management to better anticipate
the effects of warehousing decisions. The significant
findings are associated with the addition of a second
warehouse and the use of public instead of private ware-

housing. Further, the research yields several by-products

of interest to the general field of finance.

Addition of a Second Warehouse

 

Whether a particular company will want to add a
second warehouse or not will depend upon the specific
situation and it is not the purpose of this research to
recommend an alternative. The purpose here is to study

in general the significant effects which the addition

179

Of a second warehouse has on financial variables.

Warehouse Additions and Accounts Payable

Every experiment on a two warehouse system experi—
enced a drop in the average level of accounts payable
from 13 percent to 21 percent of the level of accounts
payable in the comparable one warehousing system. These
declines in accounts payable caused adverse changes in
the following variables in varying proportions:

1. Cash dropped, causing a deterioration in

liquidity position as reflected by the
current and acid ratios.

2. Short term loans increased in order to
take up the financing slack and this
increase was reflected in the Debt ratio.

Equity

3. The debt service coverage ratio and
earnings dropped, reflecting the fact
that an interest bearing form of debt
(short term loans) replaced a cost free
source of financing (accounts payable).

The major implication of this finding is that
managers contemplating an additional warehouse should
consider the adverse effects which such a decision will
have on the financial structure and liquidity position
of the firm and include these considerations in the
decision making framework. If the additional warehouse
is still deemed attractive, financial managers can

anticipate the changes instead of taking corrective

action after they occur.

180
Warehouse Additions and the Inventory:Accounts Payable

Interaction

 

The experiments on the two warehouse system in the
health cares company typically experienced an increase in
the average level of inventories ranging from 33 to 38
percent above the level of inventories in the comparable
experiment with only one warehouse. One experiment had a
change in inventories of 90% and the reason for this ab-
normally large change is explained on page 118. For the
appliances company, this change ranged from 1 percent to
4 percent.

It is well documented in the distribution field that
safety stocks normally increase at a decreasing rate as
the number of warehouses is increased.1 The general

relationship is popularly expressed mathematically as:2

ssl(n)

SSn = Tr§:——
where
SSn = safety stock for n locations
= number of locations
SSl = safety stock for one location

The general formula predicts that the increase in
safety stocks from adding a second warehouse will be
approximately 40 percent. Since safety inventories are
designed to be 25 percent of the total in the modeled
companies, the general relationship predicts a 10 percent

increase in inventories for this research. However, the

*0

 

II
'1

181

general formula assumes that demand patterns at the two
locations are perfectly independent of each other. Sales
in LREPS-F are randomly generated daily for the total
company and then randomly assigned to the individual
distribution centers. Thus, there is a high degree of
dependence between sales patterns at the two locations,
which explains the departure of these research findings
from the assumed general pattern. The degree of depen-
dence modeled may be closer to reality than that assumed

in the general formula.

 

‘T

The inconsistency between industries in the effect
of the second warehouse on inventories exists because of
the exponential smoothing inventory forecasting technique
used in this research. One of the prime advantages of
using exponential smoothing rather than trend projection
forecasting techniques is that exponential smoothing can
detect intermediate upturns and downturns in the demand
pattern. Since the appliances industry has a definite
seasonality factor in its demand pattern, the swings
detected and forecasted by the forecast mechanism usually
were the actual pattern. There is no seasonal pattern
to demand in the health cares industry but a substantial
amount of random variation. The swings detected by the
forecast mechanism in this industry were only "noise"
around the trend line and resulted in a substantial amount
of forecast error. This forecast error introduces in-

stability into the inventory levels which is magnified

182

by dividing inventories between two inventory locations.
Although the effect Of the second warehouse on
inventories varied widely between the two industries, the
change in accounts payable was remarkably consistent. The
levels of inventories and accounts payable are clearly
related in most Of the experimentation, but the changes in
inventories resulting from the addition of the second ware-

house have no effect on the level of accounts payable. i

 

Rather, the substantial increases in inventories from the

additional warehouse in the health cares company caused

cash to drOp substantially, short term loans to rise, and
causes deterioration in the liquidity, QEE£—— , and debt
Equity

service coverage ratios. The changes in these variables
are in addition to the adverse effects on them from the
decline in the level of accounts payable. The analysis
of these results leads to several general Observations
concerning inventories and the extent to which they are
financed through accounts payable.

First, increases in safety stock (buffer inventories)
are not repetitive in nature and after the expiration of
the allowed credit period, must be paid from cash. Thus,
accounts payable are not a permanent source of financing
for "safety" inventories. Second, increases in re-order
(working) inventories which are induced by changes in the
sales rate are repetitive in nature and therefore can be
financed at least partially through accounts payable. The

specific prOportion depends upon the relationship between

183
the re-order cycle, i, and the length of the credit period,
p.

Third, increases in re—order inventories which occur
because of management action such as adding a warehouse,
result in an increase in accounts payable at the beginning
of each re-order cycle, causing the ratio § to decline.

The net result of the two effects is that the average level
of accounts payable is left unchanged, meaning that the .
increase in re-order inventories resulting from management 4

action must be financed from some source other than

 

Hf

accounts payable.

Finally, when g declines, the proportion of inven-
tories which is financed through accounts payable declines.
Since accounts payable are (or can be) a cost free source
of financing, in many cases this decline in E results in
an increase in the average financing cost of inventories.
Thus, since management induced changes in re-order inven-
tories cause g to shift, such changes also cause the
average carrying cost of inventories to shift. This
demonstrates that in many cases the average cost of
carrying inventories, C, is an increasing function of the
re-order quantity. The traditional formula for calculating

economic order quantity assumes that C is a constant.3

Public vs Private Warehouse Decisions

 

The results of experimentation indicate that the

relative attractiveness of public vs private warehousing

184

depends upon the individual company and situation. For
example, the use of private warehousing required more
excess capacity in the appliances company because of the
seasonal demand pattern than in the health cares company.
Resultantly, public warehousing is much more attractive
in terms of profitability and liquidity protection in
the appliances company than in the health cares company.

The experimental results offer empirical evidence of
the classic relationship between fixed expenses, contri-
bution margins, and the total leverage function. Each
experiment involving a change from private to public
warehousing resulted in a drop in the level of sales
required to break even on net income and in the variability

of earnings and cash flows. They also resulted in in-

Net Working Capital
Total Assets

coverage ratios. These changes individually and in total

 

creases in the and debt service

reflect a much improved defensive posture against market

and economic reversals if the firm uses public warehousing.

Net WorkingyCapital
Total Assets

ratio also indicate that the firm will not benefit as much

 

The change in earnings variance and the

from boom conditions using public warehousing.

Each experiment involving replacing a private with
a public warehouse exhibited an increase in accounts
payable of from 60 percent to 99 percent of the change in
Operating expenses which generate accounts payable.
Analysis of this finding led to the conclusion that the

level of accounts payable will shift when there is a

 

185

change in the normal level of an expense which is paid on
trade account or when a particular type of expense is
"traded" for a type of expense which carries different
trade credit characteristics.

The expenses incurred through the use of public

warehousing (the warehouseman's charges) all give rise

"‘7

to payables. Many of the expenses incurred through the
use of private warehouses, especially owned private

warehouses, do not give rise to payables. Those which

”vs—ma "f

-_- .. \—

do (wages to labor, for example) often are for much

'M (ii
' K
V

‘I

shorter terms than the payables due public warehousemen.
Thus, accounts payables shift up and are a source of
financing, if private warehousing is replaced with
public warehousing. This increased amount of financing
through payables if public warehousing is used frees
cash for other purposes, which is another reason that
the use of public warehousing puts less strain on the
liquidity position of the firm.

The identification of this relationship should
enable managers to more accurately anticipate the
total financial effects of warehousing decisions and
help distribution managers make more correct warehousing

decisions.

186

Research By-products to Financial Analysis

 

The comparison of experiments with the same number
and types of warehouses but with different sales growth
rates yields several interesting research by-products

concerning the construction and use of financial ratios.

Debt

. . ratio
Equity

These results demonstrate that the

will Often give misleading signals if short term debt is

Debt

included in its calculation. The purpose of the ———H——
Equity

ratio is to measure the extent to which the financial

.' ‘JTH-uu'

structure magnifies the danger of market or economic

‘zvvo

reversals. The experimental results show that short
term loans have a strong tendency to work themselves
off when sales turn downward and to build up when sales
rise. Thus, their importance in case of market reversals
. . . . . . Debt .
is not cruCial and their incluSion in the -——+——-ratio
Equity

can give false danger signals. For the same reasons,
the debt service coverage ratio will have a negative bias
if interest or re-payments of short term loans are in-
cluded in its construction.

Furthermore, a weak acid test ratio is really a
signal that the firm is in a poor liquid position if sales
were to increase, not a danger signal of the susceptibility

of the firms liquid position to downturns as is popularly

believed.

187
Limitations and Implications for Future Research
The limitations of this research and implications for
future research are discussed according to the following
general problem areas:

1. The size and system characteristics of the
model firms.

2. The danger of generalizing results based on pt
distribution wholesellers to manufacturing :
firms.

3. The need for additional and more advanced
system performance — feedback adjustment
mechanisms.

 

4. The limited possible future states of L“
nature tested in this research and the
potential for developing flexibility
as a third parameter in financial
decisions.

5. The potential for using LREPS-F to attack
specific problem areas in the finance field.

6. The possibility of studying the distribution -
finance interaction on a broader scale.

Size and Characteristics of the Model Firms

Generalizing the results of this eXperimentation,
which is on model firms with ten million dollars annual
sales, to larger firms is dangerous. Additional research
should be performed to test the consistency of the
effects of the alternative warehousing decisions as the
size of the firm changes.

A limitation closely akin to the foregoing is that
the results observed in this study might be much different
for a firm with a larger number of existing warehouses.
For example, the impact of a second warehouse on accounts

payable in this study occurs because of the change in the

188

frequency with which inventories are re—ordered in the
total system. The prOportional change in re-order
frequency from adding another warehouse to a six warehouse
system, for example, would probably be much smaller than
those observed in this research. Testing the consistency
of findings as the number of warehouses increases is
another possible extension for this research.
Generalizigg the Results to Manufacturing Companies

Some of the results and conclusions from the study
of wholesellers would not hold for manufacturing concerns.
For example, the inventories purchased by manufacturing
companies serve predominantly as inputs to the production
process. Therefore, the simple relationship observed
between the turnover of inventories and the accounts
payable cycle in the wholesale industries does not exist
in manufacturing firms. The relationship is complicated
by the introduction of the time which the purchased
materials spend as raw material and in-process inventories.

TO gain a full appreciation of the interaction between
two sub-systems (in this case distribution and finance)
in a larger system, the total system must be modeled. The
addition of manufacturing, marketing, and physical supply
sub-systems to LREPS-F are seen as logical next steps in
the general direction of accurate forecasting of the

total effects of management decisions.

189

The Extent of Feedback Mechanisms

The results of experimentation might be more
significant if a more complete set of feedback mechanisms
were included in the monitoring function. Projects must
be kept within a manageable sc0pe, however, and model
building often proceeds in a step-wise manner. Feedback
mechanisms which would be added in a more advanced
version of LREPS-F include modeling sales as a partial
function of the level of service; modeling accounts
receivable as a function of the relative proximity of
customers to the distribution centers; and modeling debt
management as a function of certain key financial ratios
and statistics.

Possible Future States of Nature and the Concept of

 

Flexibility

 

The range of possible future economic and market
conditions were limited by design. Although the different
growth rate assumptions covered a wide spectrum, many
other changes in the state of nature could affect the
performance and financial position of the firm. Such
exogenous factors as changes in population location
patterns, general shifts in the cost structure of one or
more of the cost components of the distribution system and
governmental action on depreciation and tax structures
would probably influence the findings of this research.

That there is a wide range of possible future

states of nature under which the firm may Operate is an

190

implication for future applications of LREPS-F. Sen-
sitivity analysis on the model offers the Opportunity to
add a new dimension to management decision analysis.

The research colleagues on the LREPSproject term this
new dimension "flexibility."

In selecting an alternative from a group of
possible decisions, overt consideration should be given
to how well each decision alternative can adapt to what-
ever the future states of nature will be. The more
flexible a decision is, the more attractive it would be
for possible selection. The use of multiple experimental
runs on LREPS-F is an avenue by which the concept of
flexibility could be added as a third dimension in
financial analysis. Fuller development of this concept
is a major implication for future research.

Other Future Financial Research

 

Cash management has been receiving much attention
recently. Managers and academicians are increasingly
concerned with getting Optimum productivity from liquid
assets without sacrificing liquidity. This model can be
used to compare various quantitative cash management models
for a wide range of situations. It can also be used to
test the often heard criticism that the financial management
of most corporations is unreasonably conservative.4 Instead
of initializing the balance sheet items in the model based
on industry average financial ratios, they could be ini-

tialized based on the median of the poorest quartile for

r. 1.11 fl ll.

191

each ratio. The Objective of experimentation would be to
determine how bad conditions could become before the
companies would be in serious trouble; and to compare these
to the similar results when industry averages are used.
Other Decision Variables in the Distribution Area

An interesting extension of this research would be
to use a model such as LREPS-F to study the total impact
on the finance variables of decisions made in the trans-
portation, inventory, communications, and unitization
(throughput) functions. For example, the existence of
trade-offs between the costs of these various components
is well documented in the distribution literature;5 LREPS-F
could be used to make changes which would result in an
increase in the cost for one of the functions but which
would result in an exactly offsetting cost reduction in
another of the functions and the effects of such a

change on financial variables could be studied.

CHAPTER VI- -FOOTNOTE REFERENCES

lBowersox, Smykay, and LaLonde, gp. cit., p. 221.

21bid, p. 204.

31bid.
4 .
Donaldson, gp. c1t.

5Bowersox, Smykay, and LaLonde, 9p. cit., p. 299.

192

BIBLIOGRAPHY

BIBLIOGRAPHY

Altman, Edward I. "Financial Ratios, Discriminant
Analysis, and the Prediction of Corporation Bank-
ruptcy." Journal of Finance. Vol. 23, No. 4.
Chicago: American Finance Association, September,

 

 

1968.

Bierman, Harold. "Measuring Financial Liquidity."
Accounting Review. Vol. 35, No. 2. Menasha,
WiSconsin: American Accounting Association, April,

Bowersox, D. J., et a1. Dypamic Computer Simulation of
Physical Distribution Systems. Monograph. East
Lansing, Michigan: Division of Research, Michigan
State University, Forthcoming.

 

 

 

Bowersox, D. J.; Smykay, E. W.; and LaLonde, B. J.
Physical Distribution Management. New York: The
MacMillan Company, 1968.

 

Box, G.E.P. and Hunter, William G. "The Experimental
Study of Physical Mechanisms." Technometrics.
Vol. 7, No. 1. Richmond, Virginia: American Society
for Quality Control, February, 1965.

 

Carlson, C. R. "A Financial Efficiency Model."
Unpublished Dissertation, Michigan State University,

1971.
Conway, R. W. "Some Tactical Problems in Digital
Simulation." Management Science. Vol. 10, NO. l.

 

Providence: The Institute of Management Sciences,
October, 1963.

Dewing, Arthur. Corporation Finance. New York: The
Ronald Press Company, 1931.

 

Donaldson, Gordon. Copporate Debt Capacity. Cambridge,
Massachusetts: The Harvard Press, 1964.

 

Foulke, Roy. Practical Financial Statement Analysis.
New York: McGraw-Hill, 1926.

Gillman, Stephen. Analyzing Financial Statements. New
York: The Ronald Press, 1925.

 

194

195

Gilmour, Peter. Develgpment of a Dynamic Simulation Model
for PlanningyPhysical Distribution Systems:
Validation. Unpublished Dissertation, Michigan
State University, 1971.

 

 

 

Graham, Benjamin and Dodd, David. Security Analysis:

Principles and Techniques. New York: McGraw-Hill,
1951.

 

 

Grunewald, A. E. and Nemmers, Erwin. Basic Managerial

Finance. New York: Holt, Rinehart, and Winston,
1970.

 

Guthmann, Harry and Dougall, Herbert. Corporate
Financial Policy. Englewood Cliffs, New Jersey:
Prentice-Hall, Inc., 1955.

 

 

Helferich, O. K. Develgpment of a Dynamic Simulation
Model for Planning Physical Distribution Systems:
Formulation of the Mathematical Model.
Unpublished Dissertation, Michigan State
University, 1970.

 

 

Horrigan, J. O. "A Short History of Financial Ratio
Analysis." The Accounting Review. Vol. 43, NO. 2.
Menasha, Wisconsin: American Accounting
Association, April, 1968.

 

Jaedicke, R. J.; Bonini, C. P.; and Wagner, H. M.
Management Controls: New Directions in Basic
Research. New York: McGraw-Hill, 1964.

 

Johnson, R. A.; Katz, F. E.; and Rosenzweig, J. E.
The Theo£y of Management of Systems. New York:
McGraw—Hill, 1963.

 

Kent, Raymond. Corporate Financial Management.
Homewood, Illinois: Richard D. Irwin, Inc., 1969.

 

Kollaritch, Felix. "Can the Balance Sheet Reveal
Financial Position?" The Accountinngeview.
Vol. 35, No. 3. Menasha, Wisconsin: American
Accounting Association, July, 1960.

 

Lerner, Eugene and Carlton, W. T. "The Integration of
Capital Budgeting and Stock Valuation." American
Economic Review. Vol. 54, No. 5. Menasha,
Wisconsin: American Economic Association,
September, 1964.

 

Marien, E. J. Develgpment of a Dynamic Simulation Model
for Planning Physical Distribution Systems:
FormulatiOn of the Computer Modei. Unpublished
Dissertation, Michigan State University, 1970.

 

 

 

196

Merwin, C. L. Financing Small Corporations. New York:
National Bureau Of Economic Research, 1938.

 

Miller, M. H. and Orr, Daniel. "A Model of the Demand for
Money by Firms." Quarterly Journal of Economics.
Vol. 80, No. 3. Cambiidge: Harvard University,
August, 1966.

 

Murphy, J. E. "Earnings Growth and Price Changes."
Financial Analysts Journal. Vol. 24, NO. 1. New
York: The FinanEial Analysts Federation, January,
1968.

Naylor, T. H., et a1. Computer Simulation Techniques.
New York: John Wiley, 1966.

 

 

Neff, Thomas. "Financial Reports Aimed at Management
Action," Financial Executive. Vol. 35, NO. 1.
New York: Financial Executives Institute,
January, 1967.

 

Optner, Stanford. Systems Analysis for Business and
Industrial Problem Solving. Englewood Cliffs,
New Jersey: Prentice Hall, 1965.

 

Rogers, R. J. Development of a Dynamic Simulation Model
for Planning Physical Distribution Systems: Design
of Experiments. Unpublished Dissertation, Michigan
State University, Forthcoming.

 

Schwartz, Eli. "Theory of the Capital Structure of the
Firm." Journal of Finance. Vol. 14, NO. 1.
Chicago: American Finance Association, March, 1959.

 

Sprouse, Robert. "The Measurement of Financial Position
and Income: Purpose and Procedure." Research in
Accounting Measurement. Jaedicke, Ed. Menasha,
Wisconsin: Amefican Accounting Association,

1966.

 

Solomon, Ezra. The Theory of Financial Management. New
York: Columbia University Press, 1963.

 

Timms, H. L. Introduction to Operations Management.
Homewood, Illinois: Richard D. Irwin, Inc., 1967.

 

Tucker, James. Managerial Control Through Ratio Analysis.
New York: McGraw-Hill, 1963.

 

Van Horne, James. Financial Management and Policy.

Englewood Cliffs, New Jersey: Prentice-Hall, Inc.,
1968.

 

197

Wall, Alexander. How to Evaluate Financial Statements.
New York: Harper and Brothers, 1936.

Walter, James. "Determination of Technical Solvency."
Journal of Business. Vol. 30, No. 1. Chicago:
The University of Chicago, January, 1957.

Weston, J. F. and Brigham, E. F. Managerial Finance.
New York: Holt, Rinehart, and Winston, 1969.