Académique Documents
Professionnel Documents
Culture Documents
Abstract
This case is about Activity Based Costing (ABC) and throughput accounting (TA) as accounting
tools to “structure” technical (process) insights in an accounting context. The case shows how
working-floor insights and production process data can be used in the computation of income
statements that are relevant managerial decision making. The throughput accounting part of the
case stimulates a discussion on (short run) avoidable costs and hence provides necessary input for
It is essential to understand that ABC is a method that tries to “frame” data in a managerial
decision making context. Models, figures and theories only have value to the extent that they help
us to understand and explain what happens in the company. ABC is an allocation method cum
story: it is an arbitrary way to allocate costs – as arbitrary as any other allocation method – but it
helps in communicating how the (production) process works and, therefore, tries to upgrade the
value of accounting in the decision making process. If, e.g., a good driver is found and if there is
1
statistical evidence that supports the choice of and/or work-floor feeling on such a driver, using
this cost driver in the calculations of product (line) profitability makes the accounting much more
Mr. Van Schenkel, founder and owner of small company producing colour printed T-shirts, has
some ideas on why multi-coloured T-shirts are more expensive in production than T-shirts with
fewer colours printed on them. Process mapping and accounting data are available to set up an
ABC system that will allow to test the validity of Mr. Van Schenkel’s ideas and to come to “better”
accounting calculations of product profitability. The case illustrates how ABC calculations can
support managerial decision making and lead to ABM (activity based management), enabling the
company to better understand how its process works and how this understanding can be put to use
DuoGraphics B.V. is a (non existent) custom screen printing business whose major customers are
local high schools, youth organizations, sporting clubs, and small businesses. The company
produces a variety of T-shirts, silk screened in as many as six different colours. Normally,
customers provide the blank T-shirts. If not, the company buys blank T-shirts and just charges the
customers for them, without a mark-up. Thus, the only “raw material” used by the company is the
ink. Table 1 shows the ink used by each of the products. The raw material (ink) is to be treated as a
direct cost. This means that the cost is incurred only as the product is actually produced (direct
2
There are virtually no entry barriers to the screen printing business: start-up costs are small and the
printing technique is one that is learned easily. Consequently, the business is highly competitive
In its fourth year of operations, DuoGraphics B.V. wants to start with activity-based costing (ABC)
and ABC-management (ABCM) and would like to integrate this cost system with insights from the
theory of constraints (TOC). A first step in this implementation process entailed an analysis of the
activities that are necessary in the custom screen printing business. The company’s founding father,
owner and chief executive, Mr. van Schenkel, has identified the following main activities: (i)
production, (ii) machining, (iii) set-ups, (iv) quality-control, (v) reclaiming, (vi) purchasing and
(vii) sales and marketing. Mr. van Schenkel has also analysed the income statement for the year
ended in order to reclassify the expenses (totalling €129,652) from the traditional account
accomplished through interviews and an analysis of the accounting records together with all types
of supporting documents. The following breakdown was the result of this tracing to activities and is
shown in table 2:
The set-up, quality control, and reclaiming activities had no costs directly traced to them because
analysis revealed that they used resources from the production labour and – to a much lesser extent
As ABCM requires the identification of cost drivers for each of the activities, Mr. van Schenkel has
also gathered all possible and relevant information for “process mapping”. Process mapping means
relating all activities with cost drivers in such a way that it becomes clear why and where resources
3
such as machine and labour hours are consumed in the process of producing, marketing and selling
the company’s products. In other words, a process map shows the relationship of each activity to
other activities and to the products and is the key instrument to determine product costs, cost
drivers and capacity of each activity. Table 3 provides you with the monthly data about the
resources consumed in the production process (weekly data are given in appendix 1). It shows the
result of this mapping process and indicates, where applicable, the capacity of each activity in
Mr. van Schenkel now wants to determine the cost, cost drivers (output measures), and capacity of
each activity, using the information available to him. The company's products are identified
according to the number of colours printed on the shirt. Because the maximum number of colours
printed was six, there are six different products, to be referred to as “clr1” (a single colour T-shirt)
up to “clr6” (a T-shirt with six colours printed on it). The associated EXCEL dataset contains two
worksheets with (i) weekly data for the resources consumed for six out of the seven activities listed
above together with the corresponding weekly outputs of the six different types of shirts and (ii)
monthly “purchasing orders” data. The data concerns the last year of operations. The names of the
The screen printing of six-colour T-shirts is significantly more complex than for one- and two-
colour T-shirts. Thus one would a priori expect e.g. a six-colour T-shirt to require more machine
time, more reclaiming time, more set-up time, more quality control, and more production labour. In
addition, more purchase orders (e.g. for ink) are required for six-colour T-shirts (see also table 1).
Linking up with this, ABC cost calculations require the calculation of “cost driver factors” showing
the extent to which each product makes use of the various activities (i.e. the associated cost drivers)
in the production process. This means that, for all seven cost drivers in our dataset, we can use
multiple regression to estimate the “cost driver factors” from our production data.
4
Concerning these cost drivers, the annual cost of the activity to which the driver is linked can be
calculated. For example, consider the machining activity. The annual cost for machines is €18,200
per year (see table 2). The annual machine capacity can be calculated as follows: 40 hours/week
and 52 weeks per year makes up 2,080 “normal” machine hours per year. Mr. Van Schenkel
suggests that you use the available data to estimate the machine hours required for each (type of) T-
shirt. In fact, mr. Van Schenkel has always been - and still is - in the habit of getting his hands dirty
establishment. He is convinced that the amount of machining does not simply depend on the total
number of shirts produced, but also on the number of colours printed on these shirts. More
precisely, his point is as follows: “Making e.g. a 2-colour shirt means that you have to pull the
shirts through the machine two times, once for each different colour. Logically, this should imply
that to produce a given quantity of such shirts, you need twice as much machine time as for 1-
colour shirts. Similarly for 3-, 4-, 5- and 6-colour shirts, of course!”1. A similar story holds for the
production labour activity. It can be explained and traced to the products in the same way as the
machine activity. The total labour cost amounts to €68,229 per year, the output measure or cost
driver is the number of labour hours, and the annual capacity of fixed labour hours per year is 5,760
labour hours (if calculated on the basis of 3 fte each working 48 weeks per year and 40 hours per
week). Mr. van Schenkel wants you to find out whether (as was the case with machine hours) total
production labour not only depends on the number of shirts produced, but also on the number of
1
Therefore, he wants you to test the following machine cost calculation formula: MACH(x-clr) = a*x, where a is the
parameter to be estimated.
MACH(x-clr) stands for the machine hours needed to produce a single T-shirt with x colours printed on it.
E.g. MACH(1-clr) = the amount of machine hours needed for the production of one single-colour T-shirt.
5
Mr. Van Schenkel also asks you to find out whether it would not be better to “split-up” (total)
production labour into set-up labour, quality control labour, reclaiming labour and “pure”
production labour (i.e. the remainder). Concerning set-up labour, Mr. Van Schenkel reasons as
follows: “To set up the machine requires some basic operations which are always needed, no
matter how many colours we want to print. After that, each additional colour requires the same
operations (e.g. putting in the right ink colour), and should therefore take an equal amount of
labour hours. E.g. if the ‘basic operations’ require a setup lasting for 0.6 minutes (i.e. 0.01 hours),
and each different colour requires an additional 0.3 minute setup, I would expect a ‘cost driver
factor’ pattern like 0.015, 0.020, 0.025, 0.030, 0.035 and 0.040 for 1- up to 6-colour shirts
successively.” This would suggest a formula similar to the one for total labour hours discussed
above. Mr. van Schenkel does, however, point out to you that the formula discussed above might
not hold for the other three types of labour driven activity (i.e. quality control, reclaiming and
Reclaiming, set-up, and quality control activities are performed using production labour. Therefore,
these activities do not have their own annual cost and capacity. They have no cost of their own, and
the capacity is limited by the capacity of the production labour activity. In the case of the machine
and the production labour activity, the capacity represents the number of hours that can reasonably
The specification of the capacities of the administrative activities, i.e. purchasing and sales and
marketing, is more difficult. The line of reasoning is usually as follows. Last year, €28,566 was
spent on the sales and marketing activity and 2,562 sales orders were processed during that year.
Hence, it would be difficult to argue that the capacity was less than these 2,562 sales orders. On the
other hand, it might be possible to argue that the capacity is much greater because there was unused
2
In fact, he suggests to test the following, slightly more complicated formula: LH(x-clr) = a*x + b, where a and b are
the parameters to be estimated (note that the machine cost formula is actually a special case, if we choose b=0).
6
capacity last year. To use last year's experience as the capacity is a conservative approach. Mr. van
Schenkel suggests that you use purchase orders and sales orders as “cost drivers” for the purchasing
Concerning the sales and marketing activity, he argues as follows: “It’s clear that making e.g. a 6-
colour shirt requires more machine time than a 1-colour shirt, because you have to pull the shirts
through the machine six times (once for each different colour). However, it’s not at all clear why a
6-colour shirt would require more sales and marketing time than a 1-colour shirt3.”
Mr. van Schenkel has worked out a basic regression model for the sales & marketing activity. It
looks as follows:
The βx-symbols (x=1,…,6), which have the same meaning as our previous SO(x-clr), are standard
notation in statistics. This model (C), the “complete model”, is an unrestricted model, meaning that
the coefficients can be freely estimated, i.e. they are not thought to obey a particular pattern. The
By means of first impression, Mr. van Schenkel notes that the results look excellent: all the
independent variables (i.e. types of shirts) are significant at the 5% level (the highest p-value is
0.042), and we have a nice R2 of 83.2%. Upon closer inspection, two things strike him, however:
1) The intercept of the model is negative, but very insignificant. this raises the obvious question
3
This would lead to the following formula: SO(x-clr) = b, where b is the parameters to be estimated (note that this is
again a special case of the formula for production labour, if we choose a=0).
4
Using the EXCEL data appendix, this output can be obtained as follows: select the menu Tools > Data Analysis,
choose the “Regression” option, select input Y-range H1:H49, input X-range B1:G49, tick “labels” and click “OK”.
7
2) Mr. van Schenkel also notes that the six “slope” coefficients of the model are not too far apart:
in fact, they all seem to hover around 0.045. This obviously links up with his prior hunch, and it
raises the issue whether those coefficients might actually be the same, as expressed by the
formula SO(x-clr) = b, which can be converted into a null hypothesis in terms of the coefficients
of model (C).
The information in tables 2 and 3 (see also the data in appendix 1) is to be used to calculate the
charging rates for each of the activities for which there are annual costs and annual capacities. The
charging rates are the ratio (annual) cost/capacity for each activity. Together with the information
given in tables 5 and 6, these charging rates can then be used to create an income statement
showing product profitability based on the extent to which each product makes use of the various
Goldratt (1992) introduces three operational measures - throughput, operating expense, and
inventory - to provide feedback to operations management as they seek to make operating decisions
that will enhance profits, cash flow, return on sales and residual income.
Throughput can be defined as the revenue minus the cost of the raw materials based on the number
of units sold (not produced). The Goldratt (1992) definition of operating expense comprises the
expenses that are incurred to convert raw materials to throughput. Thus, he regards all expenses
other than raw materials as operating expenses. Inventory represents the cost of all the raw
materials that have not yet been sold plus the cost of the more traditional assets that are used in the
8
The foregoing is a simplified version of Goldratt's definition: you could argue that more than just
raw materials are to be deducted from revenues to arrive at throughput. So, you could deduct any
resource that is purchased on an as-needed basis and not acquired in lump sum in advance of its
use. Thus, in Goldratt's view, labour is not treated like raw materials. Instead, labour is regarded as
an operating expense because a company typically purchases labour services in a lump sum, and, if
the services are not used, the company has paid for the services anyway.
In Goldratt's view, a process is a series of dependent events (activities), and the goal is the
maximization of the process throughput. The primary challenge to maximizing the throughput is in
identifying the activity or activities that are the constraints to throughput. Once the constraining
activity or activities have been identified, attempts to increase the throughput without increasing
Requirements
(A) Compute a regular Income Statement for the year ended showing the per-product throughput,
return-on-sales and try to define a metric per product that gets as close as possible to a return-
on-investment (ROI) and residual income (RI) metric. (Note: assume a capital charge of 8% p.a.
for the RI metric and allocate costs on the basis of production volume (traditional allocation).
Assess your findings and discuss the pricing policy and product portfolio of DuoGraphics B.V.
(B) Comment on Mr. Van Schenkel’s regression analysis output (model (C)) with respect to the
sales & marketing activity and discuss the two findings that strike him. Is there an alternative
model you would like to test that satisfies Mr. Van Schenkel’s hunch? How would you compare
(C) Mr. Van Schenkel asks you to run regressions for all other activities and asks you to “test his
ideas” as to the what really drives the costs of the activities left to analyse.
9
(D) Use the “process mapping” information to set up an ABC-system that allows you to compute
“charging rates” for each driver and “cost driver factors” per product per activity. Also an
investment” and “residual income” on the basis of your ABC-system. Finally, try to calculate
the “cost of unused capacity” (see Cooper & Kaplan (1992)) and assess your findings. Will
your findings lead to a different pricing policy and product portfolio decision at DuoGraphics
B.V.?
10
TABLES
11
TABLE 1
DuoGraphics B.V.: Ink used in Screen Printing
12
TABLE 2
DuoGraphics B.V.: Tracing Expenses to Activities
Activity
Production labour € 68.229
Machines € 18.200
Purchasing € 14.657
Sales and marketing € 28.566
Total Costs €129.652
13
TABLE 3
Duographics B.V.: Process Mapping Information (Monthly Data)5
Reclaiming
Purchasing
Production
production
Machining
Marketing
Sales &
Quality
Activity:
Set-ups
“Pure”
control
Sales orders
Reclaiming
Purchasing
production
Cost Machine
Quality
Labour
“Pure”
control
labour
labour
labour
labour
Set-up
orders
driver:
hours
hours
hours
hours
hours
hours
Month LH6 MACH SULH QCLH RECLLH PRODLH7 PO SO
Jan 407.50 87.75 67.75 124.75 99.75 115.25 47 146
Feb 385.50 90.25 59.75 121.25 97.25 107.25 47 124
Mar 514.25 107.25 92.00 128.75 137.25 156.25 58 206
Apr 652.75 138.75 103.75 149.00 167.25 232.75 66 235
May 638.00 153.00 115.75 137.50 181.25 203.50 75 238
Jun 647.50 151.00 117.25 131.00 188.50 210.75 80 248
Jul 629.00 148.75 108.25 134.50 179.75 206.50 75 250
Aug 607.75 144.00 103.50 136.75 172.75 194.75 72 227
Sep 584.75 130.50 92.50 148.25 161.00 183.00 67 240
Oct 542.25 122.00 87.00 141.25 148.25 165.75 64 204
Nov 460.50 101.75 74.75 131.50 118.50 135.75 48 171
Dec 652.00 157.40 122.00 127.25 173.50 229.25 89 273
Total 6,721.75 1,532.40 1,144.25 1,611.75 1,825.00 2,140.75 788 2562
5
For weekly data, see appendix 1.
6
Note that 3 full-time equivalents are hired on a fixed basis and used for production. Extra production labour
(excluding the 40 hours per week per fte) is hired on a temporary basis.
7
PRODLH is total production labour minus the set-up, quality control and reclaiming labour hours.
14
TABLE 4
Duographics B.V.: Sales & Marketing – Regression analysis output
Regression Statistics
Multiple R 0.912
R Square 0.832
Adjusted R Square 0.808
Standard Error 5.308
Observations 48
ANOVA
df SS MS F Significance F
Regression 6 5737.991 956.332 33.940 0.000
Residual 41 1155.259 28.177
Total 47 6893.250
15
TABLE 5
DuoGraphics B.V.: Sales
Note: These figures reflect actual sales for the past year.
16
TABLE 6
Duographics B.V.: Sales (Monthly Data)8
Note: These figures reflect actual monthly sales units for the past year, as provided in the
associated EXCEL-datafile, “monthly data” tab.
8
For weekly data, see appendix 2 or the “weekly data” tab of the associated EXCEL-datafile.
17
RELATED LITERATURE
Baxendale, S. J., and M. Gupta. 1997. Integrating TOC and ABCM in a Healthcare Company.
Journal of Cost Management (11).
Baxendale, S. J., and M. Gupta. 1998. Aligning TOC & ABC for Silkscreen Printing. Management
Accounting (April): 39-44.
Campbell, R. 1997. Designing an Information System using ABC and TOC, Journal of Cost
Management (11).
Cooper, R., and R. S. Kaplan. 1992. Activity-Based Systems: Measuring the Costs of Resource
Usage. Accounting Horizons (September): 1-13.
Cooper, R., and R. Slagmulder. 1999. Strategic Cost Management: Integrating ABC and TOC.
Management Accounting (February): 20-21.
Demmy, S., and J. Talbot. 1998. Improve Internal Reporting with Activity-based Costing and
Theory of Constraints. Management Accounting (November): 18-24.
Goldratt, E.M., and J. Cox. 1992. The Goal: A Process of Ongoing Improvement, 2nd (revised)
edition, Croton-on-Hudson, NY: North River Press.
Goldratt, E.M., and R. Fox. 1989. The Importance of a Systems Constraint. The Theory of
Constraints Journal 1 (February): 5-7.
Holmen, J. 1995. ABC vs. TOC: It’s a Matter of Time. Management Accounting (January): 37-40
Noreen, E., D. Smith, and J.T. Mackey. 1995. The Theory of Constraints and Its Implications for
Management Accounting. Great Barrington, MA: The North River Press.
18
APPENDIX 1
DuoGraphics B.V.: Weekly Data Cost Drivers (see EXCEL file, “weekly data” tab)
9
Only monthly data are available.
19
APPENDIX 2
DuoGraphics B.V.: Weekly Sales Data X-colour Shirts (see EXCEL file, “weekly data” tab)
20