THE 1987 INTERNATIONAL CONFERENCE OF THE SYSTEM DYNAMICS SOCITY. CHINA821
ACCOUNTING DYNAMICS
=-- Its Concepts and Model ---
Saburo KAMEYAMA CHUO University
Takahiro KOJIMA SENSHU University
Akira UCHINO YOKOHAMA College cf Commerce
Kinnya MACHIDA ASAHI University
ABSTRACT
Accounting Dynamics (AD) is a-methodology for the modelling and simulation
of accounting ‘using System Dynamics. Accounting identifies and measures
economic transactions of an entity, and communicates these’data te both
internal and external decision makers. Thus it may be said that accounting
controls resource allocations in social systems. In this context a
simulation model described in terms of accounting will be very effective in
order to analyze and project the behavior of social systems. In our model
we emphasize the structure-dependent nature of System Dynamics. The AD
model depicts the structure of accounts in an economic entity. The
structure of accounts reflects charge-and-discharge relationships of
accountabilities. Therefore the AD model seems to have a close
relationship to the real world. The definition of accounting by R.
Mattesich, i.e., the method of quantitative description and projection of
income circulation and of wealth aggregates in macro and micro economy
(Mattessich 1964), will be best met by AD. And accounting principles will
be tested for their validation by the AD model simulation. The AD model
will be able to be built at various levels, e.g., industrial, urban,
national, in social systems. In our first approach, we have focused on the
industrial level, and we have made a prototype of AD corporate model in ‘a
business entity. We will suggest critical issues in the application of the
AD model for higher levels of social systems.
CONCEPTUAL FRAMEWORK
Accounting is a most primitive quantification of social process which are
composed of interactions of people. Here, most primitive means that
accounting is a basic way of symbolizing measurements of relationships
among people, especially those important ones related to day-to-day
transactions. One could say that there is no other system that measures
each and every routine transactions as.much as accounting.
With society becoming more and more complex, many social institutions have
been formed in order to regulate economic relationships. Accounting also
has moved from a primitive level to a level of gigantic social institutions
which regulate resource allocations in our society. However, this does not
mean that accounting is no more the best way to quantify our everyday life.
All economic bodies, whether they be individuals, corporations, municipals,
or governments, still process their transaction data through generally
accepted accounting procedures. And through this institutionally accepted
822 THE 1987 INTERNATIONAL CONFERENCE OF ‘THE SYSTEM DYNAMICS SOCITY. CHINA™
accounting information, social relationships relating to resource
allocations among people are decided.
Looking from a different point of view, accounting is a vivid control
system of resource allocations in our society... As such, it can have a
large influence on the dynamic behavior of social systems. However, it cap
be said that until now, accounting has not had an effective method which
analyzes the resource allocation process and interprets the dynamic
behavior of social systems macroscopically. But it will be necessary to
bring the accounting method of quantifying everyday transactions to a
method of _ operational modelling and simulation of social systems.
Accounting Dynamics (AD) is a new methodology of accounting as social
science in this context, that is, AD is a method of analyzing the dynamic
behavior of social systems through accounting model simulation.
In order to ‘simulate the real state of the resource allocation in the
social system, it is necessary to convert the quantification process of
accounting into an operational model. In other word, the accounting
mechanism of processing transaction data must be built into the structure
of the model. The model also has to be able to withstand simulation tests
operationally, Only through this type of model simulation can we expect to
accomplish our primary research objectives.
In this context, we use the SD method for our model operations. The reason
for using SD is mainly the analogy of two basic variables (level and rate)
in SD to the balance sheet accounts and transaction concepts in AD.
Financial Integration of Financial
Position Rate Transaction Flow Position
B/S Transaction P/L B/S
t. 4.-t, t
ts
B/S (x1, xy x4 =B/S (aye Xr xotet {T (ay Xar Xy)
Figure 1 Basic Concept of AD Corporate Model
THE 1987 INTERNATIONAL CONFERENCE OF THE SYSTEM DYNAMICS SOCITY. CHINA 823
Furthermore, due to the structure-dependent nature of the SD model, an AD
model based on the SD method can include the structure. of account and
institutional ‘constraints which accounting data processing is subject to. A
basic notion of AD is as the following. Corporate financial behavior can be
represented as in figure 1. The income statement (P/L) is the integration
of the transaction flow in a given period of time. By adding this to the
first balance sheet the present balance sheet is obtained. The equation
below figure 1 describes this relationship. The AD model formulates this
relationship using the simulation language "DYNANO".
The structure-dependent nature of AD model can be represented as in figure
2. Both the formal postulates of bookkeeping and the institutional
postulates of accounting can be included in the model. The structure of
balance sheet accounts of double entry bookkeeping forms level variables in
causal loop diagram. Transaction flow which connects balance sheet accounts
makes rate variables. Institutional constraints are formulated as constants
and parameters which are the subconcepts of rate variables. Here it is
noticeable that the structure of accounts reflects charge-and-discharge
relationships of accountabilities. Therefore AD model is expected to have a
close relationship to the real economic world.
Feedback
Real
Systea
Bookkeoping
Formal
Postulates
tr] so [ANY 20
[ SPestuteres | “Moder
Figure 2 Postulates of AD Model
The initial conditions of the AD model are decided by the opening balance
sheet. Through the simulation, we can calculate the B/S at any point of
time and the P/L for any period in the future.
Of course the building of the AD model does not confine to the corporate
level. The AD model can be considered in urban, national economics and all
824THE 1987 INTERNATIONAL CONFERENCE OF THE SYSTEM DYNAMICS SOCITY. CHINA
other higher levels in social systems. But, as can be seen in figure 3, it
is important to notice that the AD model is built on the second order
information space. The space, in which mutual economic transactions by
economic bodies such as individuals and corporations are carrying out, can
be called the real space. It is clear that an accounting system exists due
to the necessity of a total optimization in this real space. Therefore, the
accounting system is formed at the first order information space. AD tries
to simulate the real condition of the control of the economic activities,
which performed by the accounting system. Due to this, AD must be a concept
in which both the.real space and the first order information space are its
objects. Also it can be said that all accounting theory which focuses on
accounting in practice is based on the second order information space just
as AD. Furthermore, a meta theory, whose object is an accounting theory
including AD based on the second order space, can be well conceived.
Feedback
Real First Order Second Order
Space Information Space Information Space
Economic Accounting AD
Activity System Concept
AD
Business Business Corporate
Enterprise Accounting Modol
amaneaeiea AD
Ravi Poumenit Social Social
Accounting Model
Figure 3 Relationship of Economic Activity,
Accounting System and AD Model
We think, through the accounting modelling and simulation based on the
second order information space, as explained above, accounting has, for the
first time, a positive accounting methodology such as said by Friedman
(Friedman 1953). As Friedman says, normative research is not possible
without positive research. However, the most accounting theories are
THE 1987 INTERNATIONAL CONFERENCE OF THE SYSTEM DYNAMICS SOCITY. CHINA825
normative oriented without positive research. . Positive research is also
indispensable for the formation of mutual agreements of the related parties
in our society. At the beginning of the corporate accounting principles,
there is the principle of truth. However, until now, has there really been
an effective method to test the truthfulness? In the light of this, we
expect AD to be a breakthrough that allows accounting to become a _ true
social science.
It goes without saying that the results of simulations of the AD model are
fed back to economic activities that occurs in the real space, and to
accounting system that controls the real space. This is shown by the arrows
in figures 2 and 3. At this point, the nature of the AD model which is
structural dependent must be especially emphasized. The reason for this is
that, through the simulation of AD model which have institutional rules in
their structure, we could have an effective way to feedback simulation
results to the improvement of our social institutions. Also, using the AD
model, so called “what if" simulations are very effective .for "policy"
tests, that is, its applications for the DSS (Decision Support Systems)
model in the managemer:t accounting area are very effective.
CHARACTERISTICS OF THE AD “MODEL
Recently, there exist two paradigms in relation to the modelling of social
systems (Meadows 1980). One being the structural dependent: approach,
represented by SD, and the other being the data dependent approach,
represented by the econometrics model. As mentioned earlier, one
outstanding characteristic of the AD model is that it is structural
dependent. For reasons explained later, we consider further that
essentially all accounting models must be structurally dependent.
When saying that accounting is a most primitive quantification of social
processes, it means that there are at least three essential factors. One
-being the monetary valuation as a measurement scale, the second being the
account mechanism of double entry bookkeeping as a measurement formality,
and the third being the communication of accountability as the measurement
contents. We think that two points are the essence of accounting. One is
to measure economic transactions through the double entry account mechanism
in monetary terms, and the other is to communicate social consequences of
accountabilities. Therefore, the three essential factors mentioned above,
must all be indispensable factors to accounting data.
The concept of quantification in accounting is so different with that of
normally understood in general. For example, Professor Takeuchi says that
the quantification is an application of mathematical logic to observation
of objects. He places more importance on a formal relationship between
objects, which represented by a formal relationship between numbers . and he
calls it as-"structure", rather than the object itself. The formal
relationship between numbers is a prerequisite as an axiom in these
contexts, and it is the end of quantification to investigate all the
results deducted logically from that formal relationship (Takeuchi 1971).
826 THE 1987 INTERNATIONAL CONFERENCE OF THE SYSTEM DYNAMICS SOCITY. CHINA
in the fields such as natural science and econometrics, it is normal for
quantification to be understood as Professor Takeuchi said. Underlying
observed data, what actually has observed exists separately to the data,
and the structure of the model is in all respects something that emphasizes
the formal relationship among. observed data. It cannot be guaranteed that
corresponding structure exists actually in the real system.
Can this type of relationship be applied, as it is, to accounting data? In
this, the unique meaning of quantification of accounting becomes clear. The
relationship, where What actually has observed exists separately behind the
accounting data, cannot be plainly said at the least. According to
accounting data, it is impossible to deny an aspect where the real human
relationships in society are restricted to what the accounting data say.
Therefore it can be said that at least one aspect of the social system is
produced through accounting data processing.
Of course, when looking at accounting models in general, especially the AD
model, we are discussing on the second order space, as explained in figure
3. However, there is a need to pay careful attention to the fact that the
data used here is the result of the unique quantification of accounting as
mentioned above. Only searching for the formal relationship between data
and the operational nature of the model is inappropriate. No matter how the
data is processed, we think that if it is accounting data, it must possess
the three essential factors mentioned above. Behind accounting data there
exist, not separately but identically, the human relationships concerned
with social resource allocations. So the structure of the accounting model
reflects the real social system that controls resource allocations.
By building in this way the real social process into the structure of the
model, AD can effectively feedback the results ef model simulation into
real world. We think that at this point, in a sense, the AD model
strengthens the SD method. A normal modelling approach in SD first pays
attention to the time pattern of the behavior shown by real systems (this
is called the reference mode). Then, the basic mechanism of the system from
which the mode comes is made clear, and the cause and effect loop is
formulated into the model (Randers 1980). However, up to now, there was
criticism that the process was often subjective. In contrast to this, the
AD modelling takes, as its reference mode, financial indexes: such as gross
sales and profit which disclosed in financial statements. By putting the
structure of accounts and institutional regulations that produce these
indexes. into the structure of the models-feedback loop, a model with
extremely high reliability can be available.
Social systems are, by nature, open, and are said to be the systems
essentially lacking in data. It is best to say that there is a limit to
analyzing the behavior of complicated social systems and predicting its
future only through the logical deduction from observed data. In order to
understand the behavior of systems which reflect non-linear, often one time
only, sometimes irrational, and extremely complicated human interactions,
it is most effective to resort to a simulation of a model in which a
institutional mechanism that controls human relationships is built. In this
THE 1987 INTERNATIONAL CONFERENCE OF THE SYSTEM DYNAMICS SOCITY. CHINA 827
respect it must be recognized that accounting, the most primitive
quantification of social process, shows a fundamental research area in
building social system models. The method of AD emphasizes this point.
AD CORPORATE: MODEL
Now we investigate the technical problems that occur when actually building
the AD model.
A very close analogy can be seen between the method of SD and the basic
accounting concepts. Forrester himself has said concerning accounting, "The
balance-sheet variables are levels, giving the financial condition of the
business system at one point in time. The balance-sheet levels show the
effect of accumulating the rates of flow over all past time. The profit-
and-loss statement, by contrast, gives the rates of flow that have existed
since the previous balance sheet. The profit-and-loss rates cause the
changes from the previous balance sheet to the present." (Forrester 1968)
Figure 1 is a summary of the corresponding relationships of basic concepts
in accounting and SD.
Furthermore, it has already been noted that the institutional constraints,
such as accounting principles and the commercial law, which control
transactions have been included in the AD model as subconcepts of rate
variables.
In order to develop the AD prototype model which occurs at corporate. level,
we chose company A which is listed on the first section of the Tokyo Stock
Exchange as an object of our investigation. As in table 1, the main
products of company A are concrete piles and pipes. It has a capital of
1,200 million Yen ($8 million), gross sales are about 20,000 million Yen
.($133 million), and the number of employees is 800. Company A produces on
order, and the production process is relatively simple. As can be seen in
table 1, financial performances of company A show gradual decline as in
the trend of gross sales. The problems mainly have been caused by general
economic crises, especially called structural crises in process industries,
and resulted in operational losses.
As prototype of the AD corporate model, we have developed two models (type
} and type 2) of company A. Type 1 model overviews physical structure and
accounting structure of company A and gives the basic behaviors and
critical issues of company A. The model also shows one possible approach to
cope with actual problems. Type 2 model (Figure 4) covers physical
structure and accounting structure of the company in detail. This model
includes factors occurred in external environments at the same time, and
tries to analyze the correlations of simulation results of the model and
critical problems of company A.
Of course, we are aiming for positive feedback of simulation results to
real problems through use of the AD model. But the detailed simulation
results of two types of models are left. Here we only emphasize that the
models in these types show to be prototype of AD corporate model and that
they have the potential power for developing a useful tool to cope against
828THE 1987 INTERNATIONAL CONFERENCE OF THE SYSTEM DYNAMICS SOCITY. CHINA
pas ~=6AD SECTOR — +
B/S
ACCOUNTING SECTOR
Financial
Index
4 AK
@) (4)
i f
e7L \
Accounts
Tae
Accounting
anal
PHYSICAL SECTOR
Production
-_
R&w \-——
terjial
(voy
KZ
| Customer
Supplier
ee
©)
Pigure 4 Framework
of AD. Corporate Model
(Prototype Type 2)
THE 1987 INTERNATIONAL CONFERENCE OF THE SYSTEM DYNAMICS SOCITY. CHINA829
actual problems. Therefore, it should be noticed that the emphasis is not
feedback to reality in actual case, but its possiblity and methodological
issues.
Table 1 The Company A
Ceramics
¥1,200,000,000
($8,000,000)
(3) Sales : Operating Profit:
(1) Industry
(2) Capital
Year Sales Operating Profit
¥ $ ¥ $
1980] 23,800 (159) 774. ( 5.16)
81] 23,050 (154) 510 ( 3.40)
82) 27,270 (181) 1,031 ( 6,87)
83] 21,870 (146) 8 ( 0.05)
84{| 18,040 (120) -1,001 (-6.67)
(¥.$.millions)
(4) Products :
Pile and Pipe
Construction Work
Steel Framework
(5) Raw Materials :
Cement
Sand,Gravel
Steel wire
(6) Expenses :
Depreciation
Fuel
Repair
CONCLUSION
We have presented a new method of Accounting Dynamics, a new methodology
for the modelling and simulation of accounting using SD. The AD model is
effective in analyzing dynamic behavior of social systems and feedback its
simulation results to real world, because the model has both charge-and-
discharge relationship of accountability and the institutional constraints
to accounting data processing in its structure. We think that the AD.
method is a breakthrough for accounting to become genuine social science.
But we have only presented the prototype model of corporate level.
Our next step is to build the AD model of higher level of social systems,
for example, AD national model, where we will have to treat the difficult
830 THE 1987 INTERNATIONAL CONFERENCE OF THE SYSTEM DYNAMICS SOCITY. CHINA
problem of data aggregation and statistics. Furthermore, it is necessary to
maintain the structure-dependent nature of the model. However we expect
that the AD model of higher social level will integrate the two paradigms
of modelling approach, that is, data-dependent and structure-dependent, as
mentioned earlier. This integration is, as it is likely to be, very
troublesome but it is our fruitful research target.
REFERENCES
Forrester, JAW. (1968) Principles of Systems, Wright-Allen Press,
Cambridge, p 4-11
Friedman, M. (1953) The Methodology of Positive Economics, in Essays in
Positive Economics, The University of Chicago Press, Chicago, pp 3-43
Mattessich, R. (1964) Accounting and Analytical Methods, Richard D. Irwin,
Homewood. P 19
Meadows, D. H. (1980) "The Unavoidable A Priori", in Randers J. ed., Element
of the System Dynamics Method, The MIT Press, Cambridge, P 25
Randers, J. (1980) "Guidelines for Model Conceptualization”, in Randers, J.
ed., op. cit., p 119
Takeuchi, K. (1971) Number and Quantity in Social Science (Shakaikagaku ni
okeru Su to Ryo), The Tokyo University Press, Tokyo, P 5
CC BY-NC-SA 4.0