Who Wins in a Dynamic World: Theory of Constraints vs. Activity-Based Costing?

Robin Cooper* Goizueta Business School Emory University Atlanta, GA 30322 David Bray Goizueta Business School Emory University Atlanta, GA 30322 Michael Parzen Goizueta Business School Emory University Atlanta, GA 30322

Abstract
Two system-based views exist regarding managerial value chain analysis: Theory of Constraints (TOC) and Activity-Based Costing (ABC). There has been considerable debate whether TOC or ABC is the more optimal approach for strategic planning. This study seeks to compare TOC and ABC, while keeping constant the level of environmental turbulence each of the approaches encounter. With regard to organizational systems, literature regarding complex adaptive systems supports the idea that bottom-up approaches are more resilient to volatility. Consequently, this study hypothesizes that the bottom-up ABC approach will prove more agile and less limiting than the top-down TOC approach. This study then performs two computational experiments. The first experiment reveals that the ABC approach generated more PROFIT than the TOC approach, while the TOC produced a larger amount of REVENUE, for all instances of the simulation. The second experiment reveals that a hybrid TOC+ABC approach is the most optimal in the midst of environmental turbulence out of four possibilities. This hybrid TOC+ABC selects a first cut of orders that will generate the highest REVENUES per the TOC approach, and then selects a second cut of orders that will have the lowest COSTS and thus the highest PROFIT per the ABC approach. These results challenge the established literature espousing the TOC approach alone.

* = Corresponding author, (robin_cooper@bus.emory.edu)

Keywords: Theory of Constraints (TOC), Activity-Based Costing (ABC), value chain analysis, strategic planning, bottom-up approaches, environmental turbulence

p. 1 of 6
Electronic Electronic copy of copy this paper available is available at: http://ssrn.com/abstract=962270 at: http://ssrn.com/abstract=962270

Who Wins in a Dynamic World: Theory of Constraints vs. Activity-Based Costing?

Robin Cooper, David Bray, and Michael Parzen

Value chain analysis serves as a powerful tool for economic planning by (1) categorizing the value-adding activities of an organizational system and (2) illuminating areas most likely to increase profitable outcomes. Specifically, value chain analysis identifies the value and cost drivers for each activity performed by an organization. Such analysis then strives to maximize aggregate value creation and minimize costs internal to an organization (Smith & Pretorius, 2003). Within the established literature, two system-based views exist regarding managerial value chain analysis: Theory of Constraints (TOC) and Activity-Based Costing (ABC). Proponents of the TOC worldview subscribe that every profit-making organization must have at least one constraint preventing an organizational system from achieving a higher performance relative to its goal. Such constraints include resource-related, market-related, and policy-related constraints. According to the TOC approach, identifying and addressing these constraints facilitates successful management choices regarding optimization of an organizational system (Goldratt, 1990). Contrasting with the TOC worldview, proponents of the ABC approach subscribe that identifying the causal relationships behind the costs of a profit-making organization facilitates cost assessment more optimally than the TOC approach. This alternative approach first identifies activity-based costs and then attributes such costs to product creation based upon the activities performed during production. Consequentially, the ABC approach illuminates areas of high overhead costs for management to consider and adjust (Cooper & Kaplan, 1991). There has been considerable debate whether TOC or ABC is the more optimal approach for economic planning regarding order acceptance (Baxendale and Gupta, 1995; Holmen, 1995). Though TOC represents the established approach, a few empirical cases examining real-world firms support the ABC approach as more effective in increasing profitability and reducing inventory levels (Kirche & Srivastava, 2005). To formalize and buttress these empirical findings, this study seeks to compare TOC and ABC on a level playing field keeping constant the environmental turbulence each of the approaches encounter. As a hypothesis, this study predicts that the ABC approach will demonstrate itself to be more optimal than the TOC approach in a dynamic world, where volume and costs vary per a set of normal distributions representative of empirical cases. Specifically, in a dynamic world, this study predicts the bottom-up, activity-based attribution of costs associated with the ABC approach will prove more agile and less limiting than the top-down constraint-identifying tact associated with the TOC approach. With regard to organizational systems, literature regarding complex adaptive systems supports the idea that bottom-up approaches are more resilient to volatility (Clippinger, 1999; Eisenhardt & Galunic, 2000). This study hopes to provide similar evidence with regard to strategic planning approaches.

p. 2 of 6
Electronic Electronic copy of copy this paper available is available at: http://ssrn.com/abstract=962270 at: http://ssrn.com/abstract=962270

Methodology
This study developed a computational simulation to evaluate both the TOC and ABC approaches. While unorthodox among the established literature surrounding the TOC and ABC debate, a computational simulation was the only methodology that allowed a level playing field to compare these two approaches. While the primary objective of this study is to investigate the performance of the TOC and ABC approaches, a secondary objective includes introduction and espousal of computational simulations as beneficial methodologies for both long-term value chain analysis and the managerial accounting literature. Microsoft Visual Studio .Net served as the platform of choice for the simulation, allowing composition of intuitive code and comments, should others wish to modify the simulation later. Further, the simulation itself included a friendly graphical interface intended to allow students to explore the differences between TOC and ABC interactively. The simulation included eight normally distributed random variables, each representing some aspect of a value chain associated with creation of a product. These random variables included:
Totally variable cost of a unit Labor content of a unit Batch costs with an order Product costs with an order Order size as an integer Run minutes on a bottleneck per unit Setup minutes on a bottleneck per run of an order Selling price per unit TVC LAB BOE POE OS RM SM SP (mean, stddev) (mean, stddev) (mean, stddev) (mean, stddev) (mean, stddev) (mean, stddev) (mean, stddev) (mean, stddev)

Figure 1: Normally distributed random variables included in the simulation These normally distributed, random variables defined the orders received by a firm. For the eight normally distributed random variables, the simulation utilized means and standard deviations based upon empirical cases. The simulation allowed for two modes: (1) independence of all random variables or (2) correlations between TVC, SP, LAB, and RM, in an attempt to increase fidelity of the simulation. That said, tests revealed no difference in the results of the simulation between the two modes. The simulation itself approximated a hypothetical world in which a firm received more orders than it could possibly produce. A firm then had to select which orders would be the most profitable to accept, given its internal limitations. A single instance of the simulation simulated a firm making its long-term decisions based on the TOC approach first, and then based on the ABC approach second, for the same set of possible orders. These two long-term decisions were independent. Given their theoretical differences, the TOC and ABC approaches should suggest accepting different orders, resulting in different REVENUE, COSTS, and PROFIT for each instance of a firm. For either approach, a firm could only select orders up to its production bottleneck. A firm recorded a profit equal to its overall REVENUE minus COSTS for all the orders it accepted, for a specific approach. For the basis of comparison, an approach (either TOC or ABC) demonstrated itself as the more optimal approach if and only if it made a larger PROFIT than the alternative approach.

p. 3 of 6

The simulation allowed the user to specify four additional variables, to include:
Number of orders generated per run Maximum number of runs per order Approximate percentage of orders accepted per instance Number of instances for the simulation N NR K I (integer) (integer) (percentage) (integer)

Figure 2: Four additional variables include with the simulation Some orders included multiple runs to complete, representing instances where a first run could not complete a larger order. A linear relationship, between the size of an order and the maximum number of runs, determined the number of runs required for each order. Additionally, the approximate percentage of orders accepted per instance of the simulation served to calculate the BOTTLENECK CAPACITY (BC) for a firm. A firm had the same bottleneck capacity regardless of whether it employed the TOC or ABC approach, defined by:
BC = K * N * (RMmean * OSmean + NR * SMmean))

A simulated firm selected orders based on the TOC approach via a ranking function, which ranked orders from high to low based on their attractiveness to a firm. For a specific order, the TOC ranking function was:
TOCRANKING = (OS * (SP - TVC) / ((SM * NRorder) + (RM * OS)))

A simulated firm also selected orders based on the ABC approach via a ranking function, similarly ranking attractive orders from high to low. The ABC ranking function was:
ABCRANKING = ((((SP - (TVC + LAB)) * OS) - (BOE * NRorder) - (POE)) / OS)

A firm accepted orders based on their attractiveness, with the most attractive orders accepted first. A firm could not exceed its BOTTLENECK CAPACITY (BC). If a firm could not accept an order because BOTTLENECK CAPACITY (BC) already was full, it moved on to the next attractive order considering all orders. The firm then processed the accepted orders, allowing calculation of REVENUE, COSTS, and PROFIT for a specific approach.

Experiment 1: TOC vs. ABC

Having verified the formulae and code employed, this study ran the simulation for 250 independent instances and outputted results for both the TOC and ABC approaches. The results are clear: for all of the 250 instances of the simulation consisting of 500 orders each time, the ABC approach generated more PROFIT than the TOC for all 250 instances. An additional 10 repeat simulations confirmed this finding: ABC always generated more PROFIT than the TOC approach did in a dynamic world. That said, in every instance, TOC produced a larger amount of REVENUE than the ABC approach did. This finding is consistent with established literature that the TOC approach identifies and addresses system constraints, maximizing REVENUE received. What this study demonstrates is that TOC does indeed maximize REVENUE, but at the expense of also increasing COSTS, such that PROFIT is no longer larger than the orders the ABC approach would have accepted if faced with the same choices. The ABC approach results in less

p. 4 of 6

REVENUE, but also smaller COSTS and consequentially larger PROFIT than the TOC approach, every time. The fundamental assumptions about the way that operating expenses (OE) behave over time represent the primary difference between the TOC and ABC approaches. The TOC approach treats OE as fixed while the ABC approach treats OE as variable. This study reveals a key finding for the two approaches: OE vary by a greater amount than revenue. An ABC firm, with lower OE, will be more profitable compared to a TOC firm earning higher revenue. That said, if the world were static and OE were held constant (at the level required to support a TOC firm), then the TOC approach would be the dominant approach indeed, earning a higher profit than an ABC firm as a result of its higher revenue.

Experiment 2: TOC, ABC, ABC+TOC, vs. TOC+ABC

This study then performed a second experiment, based on the first set of results. If the TOC approach was optimal in selecting orders that maximized REVENUE, and the ABC approach was optimal in selecting orders that resulting in smaller COSTS, would a hybrid of the two approaches be better than just the ABC approach alone? The revised simulation now included the variable D, which represented the coefficient for the first approximate cut-off point of orders to accept. An ABC+TOC firm accepted a cut of orders equal to this coefficient (D) times its bottleneck (BC) via the ABC ranking mechanism first, and then accepted a second cut of orders from this set equal to its bottleneck (BC) alone via the TOC ranking mechanism second. Inversely, a TOC+ABC firm accepted a cut of orders equal to this coefficient (D) times its bottleneck (BC) via the TOC ranking mechanism first, and then accepted a second cut of orders from this set equal to its bottleneck (BC) alone via the ABC ranking mechanism second. Normal, non-hybrid ABC and TOC firms accepted orders equal to the bottleneck (BC) only, per their respective approaches. Again, this study ran the simulation for 250 independent instances consisting of 500 possible orders each time. This time, the results demonstrated that for 115 instances, the hybrid TOC+ABC approach was more optimal than any other approach. For the remaining 135 instances, the ABC and hybrid TOC+ABC approaches tied, meaning they selected the same orders exactly. An additional 10 repeat simulations resulted in a mean of 114.4 instances and a standard deviation of 3.307 where the TOC+ABC approach was more optimal than any other approach.
# TOC+ABC wins TOC+ABC and ABC tie 01 115 135 02 114 136 03 110 140 04 118 132 05 119 131 06 116 134 07 112 138 08 114 136 09 117 133 10 109 141 114.4 135.6 3.306559 3.306559

Figure 3: of different approaches for the same level of environmental turbulence

p. 5 of 6

Implications
For organizations seeking to thrive in increasing environmental turbulence, where volume and costs frequently fluctuate globally, the first experiment demonstrates that ABC is superior to TOC. The second experiment demonstrates that a hybrid TOC+ABC approach is the most optimal strategy to adopt for long-term value chains. For volatile markets, both experiments challenge the established literature espousing the TOC approach alone, yet this study also attempts to reconcile the debate between TOC vs. ABC by recognizing the value of incorporating elements of TOC into a hybrid approach. In a dynamic world, the second experiment demonstrates that the optimal features of the TOC approach aid by selecting a first cut of orders that will generate the highest REVENUES. Following this first cut with a second cut employing the ABC approach, consequentially selects for orders that have the lowest COSTS and thus the highest PROFIT. This hybrid TOC+ABC solution wins or ties as the most optimal approach every time when compared to the ABC, TOC, or ABC + TOC approach. For firms to recognize the value of either the ABC or the hybrid TOC+ABC approaches in the midst of environmental turbulence, they will need to identify their activity-based costs, and then be able to attribute such costs to product creation. Per the ABC approach, such firms also will need to be more agile at either adopting or ceasing activities dependent upon their PROFIT vs. associated COSTS. In a sense, the global economy is already progressing down this path, particularly with firms becoming less vertically structured and instead more market-based. Deeper ramifications of this study, as aforementioned, involve espousal of computational simulations as beneficial methodologies for both long-term value chain analysis and the managerial accounting literature. Given that these fields are predominantly quantitative in nature, it would seem both natural and rational for these fields to look toward computational simulations as a methodology allowing objective evaluation of both existing and new theories.

References
Baxendale, S. and Gupta, M. Aligning TOC and ABC for Silkscreen Printing. Management Accounting, 79, (1995). Clippinger, J. (ed). The Biology of Business: Decoding the Natural Laws of Enterprise. Jossey-Bass, San Francisco, CA (1999). Cooper, R. and Kaplan, R. Profit Priorities from ABC. Harvard Business Review, 69, 3, (1991). Eisenhardt, K. and Galunic, D. Coevolving: At Last, A Way to Make Synergies Work. Harvard Business Review, 78, 1, (2000). Goldratt, E. What Is This Thing Called Theory of Constraints and How Should It Be Implemented? North River Press, Great Barrington, MA (1990). Holmen, J. ABC vs. TOC: Its a Matter of Time. Management Accounting, 76, (1995). Kirche, E. and Srivastava, R. An ABC-Based Cost Model with Inventory and Order Level Costs: A Comparison with TOC. International Journal of Production Research, 43, 8, (2005). Smith, M. and Pretorius, P. Application of the TOC Thinking Processes to Challenging Assumptions of Profit and Cost Centre Performance Measurement. International Journal of Production Research, 41, 4, (2003).

p. 6 of 6