Application of simulation technique to
activity-based costing of agricultural systems:
a case study
Tzong-Ru Lee *, Jui-Sheng Kao
Department of Agricultural Marketing, National Chung-Hsing University, Taichung, Taiwan, Republic of China
Received 15 December 1999; received in revised form 20 May 2000; accepted 25 August 2000
Abstract
The aim of this paper is to analyse the operational costs of the Pu-Shin wholesale ®sh market in Taiwan, using both the activity-based costing (ABC) model and the simulation technique. By using simulation results in the calculated model of ABC, allocated resource costs are more accurate and arbitrary allocation is avoided. The objective of this study is to compute the processing cost per kilogram of ®sh. We conclude by providing relevant and accurate information about cost management of the Pu-Shin wholesale ®sh market, compar-ing ABC with traditional costcompar-ing methods, and discusscompar-ing key related issues which may pro-vide opportunities for future research. We believe that the use of the ABC model in conjunction with simulation techniques can also be applied to agricultural systems in other countries.#2001 Elsevier Science Ltd. All rights reserved.
Keywords:Activity-based costing; Simulation technique; Wholesale ®sh market; Cost management
1. Introduction
Activity-based costing (ABC; Hansen and Mowen, 2000) is a system that assigns costs to cost objects by ®rst tracing costs to activities and then tracing costs to cost objects. Cost object is a technical term in cost management and is any item such as products, departments, projects, activities, and so on, for which costs are measured
0308-521X/01/$ - see front matter#2001 Elsevier Science Ltd. All rights reserved. P I I : S 0 3 0 8 - 5 2 1 X ( 0 0 ) 0 0 0 4 2 - 1
www.elsevier.com/locate/agsy
and assigned. Cost objects are used to assign costs and ultimately the objective to compute the total processing cost is achieved.
The use of ABC is still developing. In recent years, many organizations have improved their cost management by utilizing the ABC system, a tool for providing accurate and relevant cost information. To maintain a competitive advantage, organizations must monitor and promptly remove wasted eorts or other non-productive activities. ABC allows organizations to improve their productivity by eliminating non-productive eorts, as well as managing the operational costs through observation and analysis.
In Taiwan, the wholesale ®sh market plays a very important role in ®sh market-ing. To strengthen the business of wholesale ®sh markets, it is necessary to con-tinually modify and improve operational procedures as well as manage operational costs. Hence, this paper discusses activity-based costing in the context of the Pu-Shin wholesale ®sh market in Taiwan.
This paper consists of four sections. First, the general structure of the ABC model is described. Second, the application of system simulation to the ABC system is discussed. Third, a case study is administered. The resource costs are calculated by simulation results. In this way, allocated resource costs are more accurate and arbi-trary allocation is avoided. We conclude by discussing key related issues which may provide opportunities for future research.
2. The general structure of the ABC model
Since the eorts of Robin Cooper in the late 1980s, many industries have suc-cessfully employed ABC to improve operational performance. ABC has continued to provide relevant and accurate information about cost management. In addition, because the ABC system focuses on activities rather than products, it helps prevent distorted product cost information that can arise from the use of traditional costing systems (Gunasekaran and Singh, 1999; Cooper and Kaplan, 1991). The basic assignments of the ABC model are to identify the activities of an organization, cal-culate the cost of each activity, and then cost the product based on activity con-sumption (Gunasekaran and Singh, 1999). Moreover, the ABC approach can be used to allocate various activities to related resources. Costs are appropriately allo-cated to selected cost objects by using the cost driver1of each activity. Therefore,
accuracy of product cost is contingent upon both calculations of activity cost and cost driver volume.
The structure of the ABC model is illustrated in Fig. 1. It contains information relevant to organizational resources, activities, and cost objects. The implication is that the cost object is the cause of activities and that resources exist solely to carry out those activities. After the resource costs have been assigned to their respective
1 Cost drivers are the factors that drive the cost of operational activities. They include such factors as number of parts, number of moves, number of products, number of customer orders, and number of returned products.
activities, they are subsequently allocated to cost objects by means of activity drivers2. By obtaining these measures, activity drivers become a way of assigning the
cost of activities to the actual cost object (Goebel et al., 1998). Hence, in the ABC system, the total cost of a product also includes the cost of all activities required to produce or handle it. In the ABC model, accuracy frequently depends upon the details of the ABC model and the type of activity driver used. There are three types of activity drivers (Cooper, 1990; Spedding and Sun, 1999): (1) Transaction drivers, which count each time an activity takes place; (2) Duration drivers, which represent the time taken for each activity and also takes into account variation; and (3) Intensity drivers, which directly cost the resources used each time an activity takes place. In this study, we employed all three types of activity drivers.
3. The application of simulation technique
The purpose of system simulation techniques is to design a model for a real system, which provides users with the approximated behavior of that real system. Since the simulation technique is capable of establishing dynamic relationships between vari-ables in a dynamic system, it can also presume actual relationships between varivari-ables based on the results, and the results can be used to solve complicated problems. System simulation is also able to locate system bottlenecks for future improvement and create optimum conditions between system inputs and outputs (Lee and Kao,
Fig. 1. The structure of an activity-based costing model.
1999; Chiu, 1997; Chen et al., 1997). Hence, system simulation helps policy-makers make decisions under various circumstances and keep a balance between costs and bene®ts. In other words, simulation results can be used to evaluate the merits and demerits of policies (Karim, 1998).
Through data collection and analysis, we can create an operational simulation model. Continued adjustments are made within the simulation model until results re-semble those of the ``real world''. Afterwards, data from the simulation results (operational time and number of resources used) are applied within an ABC model. The application of the simulation technique in ABC provides users with an enhanced means of assessing the cost-bene®t factors of all activities. Furthermore, by using simulation results in the calculated model of ABC, allocated resource costs are more accurate and arbitrary allocation is avoided.
4. Case study
The Pu-Shin wholesale ®sh market, which has the highest automation level in Taiwan, is the subject. Observations include the product ¯ow from unloading to the completion of the auctioning process. The steps of applying the ABC model are as follows.
4.1. Analysis of activities
There are ®ve sections in the auction area: cultured ®sh section A, cultured ®sh section B, cold-storage polyester box section, cold-storage ®sh-basket section and imported ®sh section. The major activities are: unloading, ordering, billing, grading and weighing, numbering, auctioning (computer auctioning in the cultured ®sh sec-tion A, cultured ®sh secsec-tion B, and cold-storage polyester box secsec-tion; manual auc-tioning in storage ®sh-basket section and imported ®sh section), and administrative operation.
4.2. Allocation of resource costs
After identifying the activities, resource costs are allocated to each respective activity. Allocation can be classi®ed into three categories (Ostrenga, 1990): (1) direct charging, which allocates resource costs directly into the activities; (2) estimation, which allocates resource costs by using resource drivers3; and (3) arbitrary
alloca-tion, which arbitrarily allocates resources into the activities. This study utilizes both direct charging and estimation to allocate the resource costs. The activities, resource costs, and resource drivers are presented in Table 1.
3 Resource drivers are factors that measure the demands placed on resources by activities and are used to assign the cost of resources to activities.
4.3. Computed resource costs
To obtain more accurate allocated resource costs and to avoid arbitrary alloca-tion, this study applies the simulation technique to ABC of the Pu-Shin wholesale ®sh market.
To record the time spent for each activity, a V8 camcorder was installed from 23:00 on 11 August to 06:00 on 12 August 1998. Time study skill was used to meas-ure the duration of each operation. The Stat: Fit statistics software was employed. Table 2 shows the probability distribution function (pdf) and parameters for each operational event. The results of the chi-square/goodness-of-®t test show that all pdf results for operational events are acceptable. Furthermore, the operational time of the simulation model is 7.3 h, which does not dier much from the ``real time'' of 7.5 h. Therefore, the simulation results are compatible to those of the real system. In other words, the simulation model is developed by observing and analysing the actual processing time for the activities and then categorizing their statistical dis-tributions. Random numbers from observed statistical distributions are generated to represent the duration of the activities. The dynamic process of simulation is shown in Fig. 2.
As described, this study also employs a system simulation to estimate the volume of resource drivers. The simulation results are presented in the third column of Table 3. Using these results with estimates of machinery depreciation, we can then compute the resource costs for each activity.
Table 1
Activities, resource costs and resource drivers in the Pu-Shin wholesale ®sh market
Activity Category of Resource Costs Resource Drivers
Unloading Sta wages, cost of electricity resources Operation time
Pallets Volume of use
Ordering Sta wages, cost of electricity resources, Automated guided vehicle (AGV), porter, fork porter, cost of reconditioning machines, cost of machinery fuel
Operation time
Billing Sta wages, cost of electricity resources Operation time Grading and weighing Sta wages, cost of electricity resources Operation time
Fish-baskets Volume of use
Numbering Sta wages, cost of electricity resources Operation time
Numbering machine Direct charging
Auctioning Sta wages, cost of electricity resources, cost of reconditioning machines, cost of machinery fuel
Operation time
Computer auction clock, computer auction machine C, computer auction machine D
Direct charging
Administrative operation Sta wages, cost of electricity resources Operation time Administrative spending, meeting costs,
computers
Table 2
The optimal pdf and its parameter(s) for each operational event in the Pu-Shin wholesale ®sh market
Operational events Distribution Probabilitya(%) Mean Variance Parameters w2testb Test result
a b
Unloading time per box Pearson 5 96 3.15 1.47 5.21 9.05 1.45 Accepted
Unloading time per basket Weibull 100 6.17 3.22 1.84 3.56 0.67 Accepted
Ordering time per pallet Exponential 97 24.20 328.8 ÿc 16.20 2.80 Accepted Grading/weighing time per basket Exponential 100 115.52 20.8 ÿ 8.17 3.56 Accepted
Numbering time per basket Lognormal 100 3.31 2.5 ÿ ÿ 5.10 Accepted
Manual auction time per basket in cold-storage ®sh-basket section
Pearson 5 100 8.53 10.4 2.97 9.63 0.13 Accepted
Manual auction time per box in imported ®sh section
Exponential 100 11.83 25.9 ÿ 5.83 0.67 Accepted
Computer auction time per auction in cultured ®sh section A
Lognormal 100 3.19 0.9 ÿ ÿ 21.6 Accepted
CADT per auction in cultured ®sh section Ad Exponential 100 2.35 1.6
ÿ 1.35 8.40 Accepted Computer auction time per auction in cold-storage
polyester box section
Pearson 6 99 4.19 3.8 a1=4.89,
a2=3.69
1.21 0.79 Accepted
CADT per auction in cold-storage polyester box sectiond
Pearson 6 100 8.32 27.0 a1=1.51, a2=23.78
105.6 100 Accepted
Computer auction time per auction in cultured ®sh section B
Weibull 97 4.26 0.7 2.85 2.36 0.13 Accepted
CADT per auction in culture ®sh section Bd Exponential 100 5.98 8.6
ÿ 2.98 0.13 Accepted
a The probability represents the acceptable level of distribution.
b When the con®dence interval=95%, all operational events pass the chi-square/goodness-of-®t test.
c ``
ÿ'' means that there is no such parameter in that distribution.
d Delay time would appear in each computer auctioning of the Pu-Shin wholesale ®sh market (listed as the computer-auction-delay-time [CADT]).
4.4. Activity cost and cost objective
Activity cost is a measure of the frequency and the intensity of demand placed on each activity by cost objects (Gunasekaran and Singh, 1999). Cost objects are used to obtain accurate product information. The objective in this study is to compute the processing cost per kilogram of ®sh. However, before we calculate the unit cost for each activity, we need: (1) to calculate the volume of ®sh and the units handled in each section (Table 4); and (2) to locate the allocation paths or the links between the activities and the auction area (Fig. 3).
Table 5 shows the activity driver for each activity and the calculated results of the cost objects. The results show that the total processing cost per kilogram of ®sh is NT$2.36 (US$1=NT$31).
As shown in Table 6, the wages of auctioning sta are NT$890.48 per day, which are relatively higher than those of other sta. It implies that the auctioning sta are a key resource. Therefore, the manager in the Pu-Shin wholesale ®sh market must focus on the eciency and eectiveness of the auctioning sta.
Because it would take a long time to improve grading and weighing activities and it is not easy to reduce the cost of administrative operation in a short time, we ana-lyze the cost impact of computer auctioning without manual auctioning. Further-more, the scenario of implementing computer auctioning in all sections is simulated. Fig. 4 shows that manual auctioning is not used when computer auctioning is implemented. Numbering and computer auctioning operations are included in both the cold-storage ®sh-basket section and the imported ®sh section.
Simulation results show that the auctioning operational time (60 min) does not change, but the numbering operational time is 87 min (an increase of 27 min from 60 min). The processing cost for numbering (from $0.09 to $0.07 per kg of ®sh), the processing cost for auctioning (from $0.44 to $0.23 per kg of ®sh) and the total processing cost (from NT$2.36 to NT$2.12 per kg of ®sh) are all reduced/decreased (see Table 7). Even though the cost of manual auctioning is cheaper than that of computer auctioning, however, the total processing cost per kg of ®sh actually goes down from NT$2.36 to NT$2.12 when computer auctioning is implemented in all
Table 3
The resource costs per activity in the Pu-Shin wholesale ®sh market
Activity Category of resource costs (m)a Simulation results Estimation of machinery depreciation (n)
Resource costs for each activityb (NT$/day)c
Unloading Sta(4), cost of electricity resources 105 min ±d 1898.25
Pallets Number used=82 6
Ordering Sta(4), cost of electricity resources, cost of reconditioning machines, cost of machinery fuel
180 min ± 5254.88
AGV(1), fork porter(3) ± 10
Billing Sta(1), cost of electricity resources 66 min ± 850.59
Grading and weighing Sta(5), cost of electricity resources 145 min ± 4372.01
Weighing machines ± 10
Fish-baskets Number used=290 6
Numbering Sta(5), cost of electricity resources 60 min ± 1877.73
Numbering machine(1) ± 10
Auctioning Sta wages(5), cost of electricity resources, cost of reconditioning machine, cost of machinery fuel
102 mine ± Manual auctioning=1657.69
Computer auction clock(1), computer auction machine A(2), computer auction machine B(2)
± 10 Computer auctioning=4826.21
Administrative operations Sta wages(15), cost of electricity resources 150 min ± 15191.37 Administrative spending, meeting costs ± ±
Computers ± 10
a (m):mis the number of sta or machines utilized.
b The resource costs for each activity equal ``the resource costs per day in that activity'' times ``the percentage of resource allocation''.
c US$1=NT$31.
d ``±'' means that there is no such data.
e The manual auction time is 42 min, and the computer auction time is 60 min.
sections. Furthermore, one disadvantage of manual auctioning is that the process may not be fair. By implementing computer auctioning in all sections, these are impartial, public, and equitable contributions in the auctioning process, and the total processing cost per kg of ®sh is reduced.
5. Comparing ABC with traditional costing methods
According to the results of the ABC model, the total processing cost is NT$2.36 per kilogram of ®sh. It is higher compared to NT$2.19, the processing cost calcu-lated by using traditional accounting methods (data supplied by the Pu-Shin wholesale ®sh market). One reason is that Pu-Shin wholesale ®sh market does not calculate machinery depreciation during cost calculation. If we do not add machin-ery depreciation into the ABC model, the total processing cost would be NT$2.28 per kilogram of ®sh, which is more compatible to the calculated results of the tra-ditional accounting methods. However, the ABC model still provides decision makers with relevant information about cost management that the traditional methods does not.
Table 4
The volume of ®sh and the units handled in the Pu-Shin wholesale ®sh marketa
Auction area Volume of ®sh (kg) Units handledb
Cultured ®sh section A 8105 675
Cultured ®sh section B 3571 298
Cold-storage polyester box section 7610 634
Imported ®sh section 4107 342
Cold-storage ®sh-basket section 4618 290
a Source: actual auction data on 11 August 1998.
b The unit handled in the cold-storage ®sh-basket section is the basket (capacity=15 kg), and the unit handled in all other sections is the box (capacity=12 kg).
6. Conclusions
To improve the functions of the wholesale ®sh market and encourage the trends of economic liberalization and internationalization, it is necessary to pay close atten-tion to cost management. For this purpose, this study has applied ABC and simu-lation techniques to analyse the operational costs of the Pu-Shin wholesale ®sh market. The results show that the wages of the auctioning sta are the highest of all activities, an implication that the auctioning sta plays a vital role. Furthermore, the total processing cost per kg of ®sh is reduced by implementing computer auctioning in the auction area. However, the ®ndings are based on a one-day simulation only. For possible future research, we suggest a day-by-day data simulation to obtain
Table 5
The activity drivers and the calculated results of cost objects in the Pu-Shin wholesale ®sh market
Activity Resource costs
Unloading $1898.25 Number of boxes 1949 $0.97 $0.08 3.4
Ordering $5254.88 Number of boxes 1949 $2.69 $0.23 9.7
Billing $850.59 Billing time 1949 $0.44 $0.04 1.7
Grading and weighing
$4372.01 Number of boxes 290 $15.07 $0.94 39.8
Numbering $1877.73 Number of boxes 1607 $1.17 $0.09 3.8
Manual auctioning $1657.69 Number of boxes or baskets
650 $2.55 $0.19 8.1
Computer auctioning $4826.21 Number of boxes 1607 $3.00 $0.25 10.6 Administrative
operation
$15191.37 Minutes 150 $101.27 $0.54 22.9
Total $2.36 100
a The resource costs for each activity is calculated from Table 3.
b The activity driver volume is calculated by referring to the units handled in Table 4 and the allocation paths in Fig. 3. For example, the activity driver volume of unloading is calculated as follows: 1949= 675+298+634+342.
Table 6
The wages of sta in each activity of the Pu-Shin wholesale ®sh market per day
Activity Unloading Ordering Billing Grading and weighing
accurate information about cost management. We believe that the ABC model used with system simulation can certainly be applied to agricultural systems in other countries.
References
Cooper, R., 1990. Cost classi®cation in unit-based and activity-based manufacturing cost system. Journal of Cost Management 4, 4±14.
Fig. 4. The allocation paths in the Pu-Shin wholesale ®sh market after computer auctioning is implemented. Note: The bold lines represent the additive allocation paths after computer auctioning is implemented.
Table 7
The activity drivers and the calculated results of cost items when computer auctioning is implemented in the Pu-Shin wholesale ®sh market
Activity Resource costs for each activity ($ NT/each day)
Activity driver Activity driver volume
Unit cost for each activity
Processing cost per kg of ®sh
%
(X) (Y) (Z=X/Y)
Unloading $1882.79 Number of boxes 1949 $0.96 $0.08 3.7
Ordering $5228.37 Number of boxes 1949 $2.65 $0.22 10.8
Billing $840.87 Billing times 1949 $0.43 $0.04 1.9
Grading and weighing $4350.66 Number of baskets 290 $15.00 $0.94 44.1
Numbering $1887.89 Numbering times 2239 $0.84 $0.07 3.3
Computer auctioning $6468.89 Number of boxes or baskets
2239 $2.88 $0.23 10.8
Administrative operation
$15128.54 Minutes 150 $100.85 $0.54 25.4
Cooper, R., Kaplan, R.S., 1991. The Design of Cost Management System. Prentice-Hall International, London.
Chiu, Y.C., Fon, D.-S., 1997. Simulation of transport operations using a gantry system. ISAMA 971, 465±470.
Chen, L.-H., Shaw, J.-T., Chan, T.-H., 1997. Simulation of rice and handling system. Journal of Agri-culture Machinery 6, 85±96.
Gunasekaran, A., Singh, D., 1999. Design of activity-based costing in a small company: a case study. Computers & Industrial Engineering 37, 413±416.
Goebel, D.J., Marshall, G.W., Locander, W.B., 1998. Activity-based costing accounting for a market orientation. Industrial Marketing Management 27, 497±510.
Hansen, D.R., Mowen, M.M., 2000. Cost Management: Accounting and Contron, 3rd Edition. South-Western College Publishing.
Karim, A.S., Hershauer, J.C., Perkins, W.C., 1998. A simulation of partial information use in decision marking: implications for DSS design. Decision Sciences 29, 53±85.
Lee, T.-R., Kao, J.-S., 1999. The Study of Internal Logistics Process of Wholesale Fish Market by Applying Simulation, 1999 Conference on The Theories And Practices of Commercial Automation. Taipei, Taiwan, pp. 581±595
Ostrenga, M.R., 1990. Activities: The focal point of total cost management. Management Accounting 71, 44.
Spedding, T.A., Sun, G.O., 1999. Application of discrete event simulation to the activity based costing of manufacturing systems. International Journal of Production Economics 58, 289±301.
