| |
 |  |
 |  |
|
Business Readings
Questions to Consider
TRANSFER PRICING with ABC Here's the story of how a multinational pharmaceutical company solved its transfer pricing problems by using activity-based costing. BY ROBERT S. KAPLAN, DAN WEISS, AND EYAL DESHEH In the mid-1980s, Teva Pharmaceutical Industries Ltd. decided to enter the generic drug market. Already a successful worldwide manufacturer of proprietary drugs, the Israel-based company wanted to vie globally in this competitive new market, particularly in the United States. The move has proved lucrative so far, as sales have been increasing at an annual rate of nearly 20%. In 1996, Teva's worldwide sales were $954 million and its after-tax net income, $73 million. As part of its new strategy, Teva reorganized its pharmaceutical operations into decentralized cost and profit centers consisting of one operations division and three marketing divisions. The operations division is made up of four manufacturing plants in Israel, which are organized as cost centers because plant managers have no control over product mix or pricing. The plants produce to the orders placed by the marketing divisions, and plant managers are responsible for operational efficiency, quality, cost performance, and capacity management. The marketing divisions are organized into the U.S. market (through Teva's Lemmon subsidiary), the local market (Israel), and the rest of the world. All three have substantially different sales characteristics. The Lemmon USA division handles about 30 products, each sold in large quantities. The Israel division handles 1,200 products in different packages and dosage forms, with many being sold in quite small quantities. The division handling sales to the rest of the world works on the basis of specific orders and tenders [a request from a customer for a price/bid to deliver a specified product or service], some of which are for relatively small quantities. All three divisions order and acquire most of their products from the operations division, although occasionally they turn to local suppliers. The marketing divisions are responsible for decisions about sales, product mix, pricing, and customer relationships. Until the late 1980s, the marketing divisions were treated as revenue centers and were evaluated by sales, not profit, performance. Manufacturing plants in the operations division were measured by how well they met expense budgets and delivered the right orders on time. The company's cost system emphasized variable costs, principally materials expenses...ingredients and packaging...and direct labor. All other manufacturing costs were considered fixed. Teva's managers decided to introduce a transfer pricing system, which they hoped would enhance profit consciousness and improve coordination between operations and marketing. They were concerned with excessive proliferation of the product line, acceptance of many low-volume orders, and associated large consumption of production capacity for changeovers. They proposed a transfer pricing system based on marginal costs, defined to be just materials cost. Direct labor would not be included in the transfer price because the company was not expecting to hire or fire employees based on short-term marketing decisions. High costs were associated with laying off workers in Israel, and, more important, pharmaceutical workers were highly skilled. With Teva's rapid growth, managers were reluctant to lay off workers during short-term volume declines because if new employees had to be hired later, they would need up to two years of training before they acquired the skills of the laid-off workers. But the proposed transfer pricing system generated a storm of controversy. First, some executives observed that the marketing divisions would report extremely high profits because they were being charged for the materials costs only. Second, the operations division would get "credit" only for the expenses of purchased materials. There would be little pressure and motivation to control labor expenses and other so-called fixed expenses or for improving operational efficiency. Third, if Teva's plants were less efficient than outside manufacturers of the pharmaceutical products, the marginal cost transfer price would give the marketing divisions no incentive to shift their source of supply. Finally, the executives concluded that using only a short-run contribution margin approach would not solve the problems caused by treating the marketing divisions as revenue centers. Measuring profits as price less materials cost would continue to allow marketing and sales decisions to be made without regard to their implications for production capacity and long-run costs. An alternative approach had to be found. What Everyone Wanted Teva senior management wanted a new transfer pricing system that would satisfy several important characteristics: 1. The system should encourage the marketing divisions to make decisions consistent with long-run profit maximization. The transfer price should not encourage actions that improved the profit or cost performance of a division at the expense of Teva's overall profitability. 2. The system should be transparent enough so that managers could distinguish costs relevant for short-run decisions...such as incremental, occasional bids for orders...from long-term decisions...such as acquiring a new product line, deleting product lines, and adding to existing product lines. 3. The transfer prices could be used to support decisions in both marketing and operating divisions, including:
Division managers wanted a transfer pricing system with the following characteristics: 1. The transfer prices would report the financial performance of their divisions fairly. 2. Managers could influence the reported performance of their divisions by making business decisions within their scope of authority. That is, the reported performance should reflect changes in product mix, improved efficiency, investments in new equipment, and organizational changes. 3. The decisions made by managers of marketing divisions would reflect both sales revenue and associated expenses incurred in the operations division. 4. The system must anticipate that division managers would examine, in depth, the method for calculating transfer prices and would take actions that maximized the reported performance of their divisions. Finally, the financial staff wanted a transfer pricing system such that: 1. The transfer prices and financial reports derived from them would be credible and could be relied upon for decision making at all levels of the organization without excessive arguments and controversy. 2. The transfer pricing system would be clear, easy to explain, and easy to use. Updating transfer prices should be easy, and the components of the transfer price calculation should promote good understanding of the underlying factors driving costs. 3. The system would be used for internal charging of costs from the operations division to the marketing divisions. Traditional Transfer Price Approaches Wouldn't Work Teva's managers considered but rejected several traditional methods for establishing a new transfer pricing system: market price, full cost, marginal cost, and negotiated price. Market price for the transferred product was not feasible because no market existed for Teva's manufactured and packaged pharmaceutical products that had not been distributed or marketed to customers. A full cost calculation including materials, labor, and manufacturing overhead was rejected because the traditional methods for allocating overhead (labor or machine hours) did not capture the actual cost structure in Teva's plants. Also, the accumulation of all factory costs into average overhead rates could encourage local optimization by each division that would lower Teva's overall profit. For example, manufacturing plants would be encouraged to overproduce in order to absorb more factory overhead into inventory, while marketing divisions might be discouraged from bidding aggressively for high-volume orders and encouraged to accept more low-volume custom orders. Also, this system would not reveal the incremental costs associated with short-run decisions or the relative use of capacity by different products and different order sizes. Using short-run marginal cost, covering only ingredients and packaging materials, was the system proposed initially, which the managers already knew was inadequate for their purposes. And, finally, senior executives believed strongly that negotiated transfer prices would lead to endless arguments among managers in the different divisions, which would consume excessive time on nonproductive discussions. Activity-Based Costing Is the Answer In December 1989, Teva's senior management attended a presentation on the fundamentals of activity-based costing and decided to implement ABC in its largest production plant. They wanted to investigate the use of ABC for calculating transfer prices between that plant and the marketing divisions.  Teva put together a multidisciplinary project team
consisting of managers from the production, finance, and marketing
divisions. The team worked for about six months to develop an
activity dictionary, drive factory costs to activities, identify cost
drivers for each activity, collect data, and calculate ABC-based
product costs. It took the team several more weeks to analyze the
results. Table 1 shows a sample calculation (updated to reflect 1996
data) of the costs to produce 10 tablets of a pain reliever. With
this information, managers believed they now had a defensible,
quantifiable answer to a question about how much it cost to
manufacture a special small batch for a customer.
After seeing how ABC worked at the first plant, in subsequent years the project team rolled out the ABC analysis to the remaining production plants. The ABC models were retrospective, calculating the activity costs, activity cost driver rates, and product costs for the prior year. By the end of 1993, senior managers wanted to use ABC prospectively, to calculate transfer prices for the coming year. In November, Teva built its ABC production cost model for 1994 using data from the first three quarters of 1993. But managers objected to calculating costs for 1994 based on 1993 historical data. The numbers would not incorporate the impact of new products, new machines, and expected changes in production processes. Also, the historical data contained volume and spending variances that occurred in 1993 but that were not expected to be representative of production operations in 1994. The project team took this issue to the company's Financial Control Forum where representatives from the operations and marketing divisions and company headquarters met to discuss costing and financial reporting methodologies. After several meetings, the group decided to use the next year's (1994) forecasted costs...based on budgeted expense data, forecasted volume and mix of sales, and projected process utilization and efficiencies...to calculate the transfer prices. The ABC Transfer Price Model Structure The structure of the early retrospective ABC models and the current prospective model recognizes the ABC hierarchy of unit, batch, product sustaining, and plant-level costs.1 Unit-level costs represent all the direct expenses associated with producing individual product units such as tablets, capsules, and ampoules. These expenses principally include the cost of raw materials, packaging materials, and direct wages paid to production workers. Batch-level costs include the expenses of resources used for each production or packaging batch, mainly the costs of preparation, setup, cleaning, quality control, laboratory testing, and computer and production management. The lot sizes for pharmaceutical production usually are predetermined based on the capacity of containers in the production line,2 but a second batch process, determined by customer orders, occurs for packaging the tablets or syrup. The costs of a production or a packaging batch can vary among different products and, of course, among different plants. For example, a small customer order can trigger the production of a large batch of tablets or syrup of which only a small portion may be packaged for the particular customer order.3 Thus, the batch costs assigned to a particular order include two components: a pro-rata share of the batch cost of the production setup and the full batch cost of the packaging setup. The calculation of batch-level costs for several different types of customer orders is shown in Table 2. Product-specific costs include the expenses incurred in registering the products,4 making changes to a product's production processes, and designing the package. Plant-level costs represent the cost of maintaining the capacity of production lines including depreciation, cost of safety inspections, and insurance, as well as the general expenses of the plant such as security and landscaping. In many ABC applications, machine depreciation would be included in the unit and batch costs associated with producing products and changing from one product to another. Teva decided to treat equipment depreciation as a plant-level cost so the calculated unit and batch costs could be used to estimate more closely the marginal costs associated with producing one more unit or batch of a product. Using ABC Costs for Transfer Pricing Teva bases its transfer price system on a prospective ABC calculation. Prices are set for the coming year based on budgeted data. The company calculates standard activity cost driver rates for each activity. During the year, these costs get charged to products based on the actual quantity of activities demanded during the year.  The use of
standard activity cost driver rates enables product costs to be
calculated in a predictable manner throughout the year. It also
eliminates monthly or quarterly fluctuations in product costs caused
by variations in actual spending, resource usage, and activity
levels.
Transfer prices are calculated in two different procedures. The first one assigns unit and batch-level costs, and the second assigns product-specific and plant-level costs. The marketing divisions are charged for unit-level costs (principally materials and labor) based on the actual quantities of each individual product they acquire. In addition, they are charged batch-level costs based on the actual number of production and packaging batches of each product they order (see examples in Table 2). Now that Teva has the ability to analyze the costs of different presentations, the trend of having a large number of presentations for each product has slowed. For example, the marketing divisions realized that producing special sample packages of six tablets was very expensive and that it was cheaper to give physicians the regular packages of 20 tablets. In general, the procedure has given marketing managers the flexibility to decide when to accept a small order from a customer or how much of a discount to grant for large orders. Table 3 shows a sample calculation of the monthly unit and batch-level charges from a plant to a marketing division. The product-specific and plant-level expenses are charged to marketing divisions annually based on budgeted information (see Table 4). The product-specific costs are easy to assign because each marketing division has specific products for its own markets. No individual product is sold to more than one marketing division. The plant-level (capacity-sustaining) expenses are charged to each marketing division based on the budgeted use of the capacity of the four manufacturing facilities. Activity cost driver rates are calculated based on the practical capacity of each of the four plants. In this way, the rates reflect the underlying efficiency and productivity of the plants without being influenced by fluctuations in forecasted or actual usage. Analysts estimated the practical capacity by noting the maximum production quantities during past peak periods. What about unused capacity? Unused capacity arises from two sources: (1) declines in demand for products manufactured on an existing line, and (2) partial usage when a new production line is added because existing production lines cannot produce the additional quantities requested by one of the marketing divisions. To foster a sense of responsibility among marketing managers for the cost of supplying capacity resources, Teva charges the marketing division that experienced the decline in demand a lump-sum assignment (see Table 4) for the cost of maintaining the unused production capacity in an existing line. When a marketing division initiates an increment in production capacity or manufacturing technology, it bears the costs of all the additional resources supplied unless or until the increment begins to be used by one of the other marketing divisions. At that point, each marketing division would be charged based on its percentage of practical capacity used.  The assignment of the plant-level costs (still referred to as "fixed costs" at Teva because of its long history with the marginal costing approach) receives much attention, particularly from the managers of the marketing divisions. They want to verify that these costs do indeed stay "fixed" and don't creep upward each period. By separating the unit and batch-level costs from the product-sustaining and plant-level costs, the marketing managers can monitor closely the costs incurred in the manufacturing plants. In particular, the marketing managers make sure that increases in plant-level costs occur only when one or more of them requests a change in production capacity. The responsibility for the fixed cost increment is then clearly assignable to the requesting division. The integrated budget process lets marketing managers plan their product mix with knowledge of the cost impact of their decisions. When they propose increases in variety and complexity, they know the added costs they will be charged because of their increased demands on manufacturing facilities. Active discussions occur between marketing and operations personnel about the impact of product mix and batch sizes.  Marketing managers now distinguish between products that cover all manufacturing costs versus those that cover only the unit and batch-level expenses but not their annual product-sustaining and plant-level expenses. Because of the assignment of unused capacity expenses to the responsible marketing division, the marketing managers incorporate information about available capacity when they make decisions about pricing, product mix, and product introduction. One example illustrates the value of assigning product-sustaining and plant-level expenses to individual products in the new transfer pricing system. The initial and subsequent ABC analyses revealed that quite a few of Teva's products were unprofitable; that is, the revenues they earned were below the cost of the unit, batch, and product and plant-sustaining expenses associated with these products. But managers were reluctant to drop these products because many of the expenses assigned to them, including direct labor, would remain for some time even if production of the unprofitable products were to cease. In the early 1990s, however, Teva's growing sales volume led to shortages in capacity. Teva eventually decided to sell 30 low-volume products to another company. These products were not central to Teva's strategy, yet they consumed a great number of resources and managers' attention. By shifting the product mix away from the unprofitable products, Teva was able to use the freed-up capacity of people, machines, and facilities to handle the production of newly introduced products and the expanded sales of existing profitable products. While the debate about selling off the 30 products lasted three years, the ABC system contributed to the final decision by revealing that the cheapest source of new capacity was the capacity released by reducing the production and sales of currently unprofitable products. 
ONGOING BENEFITS FROM ABC TRANSFER PRICING SYSTEM With Teva's continued growth, requests for investments in new production capacity arise continually. ABC's highlighting of unused capacity often reveals where production can be expanded without spending additional money. A second source is the capacity released by ceasing production of unprofitable products...when feasible without disrupting customer relations. Beyond these two sources, investments in a new production line can be assessed by simulating production costs if the line were to be installed. For example, a new line can reduce batch-level costs because of less need for changeovers on both the existing and the proposed production lines. These cost reductions could provide the justification for the investment decision. In addition, the investment decision for a new production line explicitly incorporates the cost and assignment of responsibility for the unused capacity in the early periods while market demand has not yet built to long-term expected levels. Teva executives say that the discipline of recognizing and assigning unused capacity costs of new production lines provides valuable realism to the demand forecasts provided by the marketing divisions. The transfer pricing system also motivates cost reduction and production efficiencies in the manufacturing plants. Managers in the different divisions now work together to identify ways to reduce unit and batch-level expenses. Manufacturing, purchasing, and marketing employees conduct common searches for lower-cost, more reliable, and higher-quality suppliers to reduce variable materials costs. Marketing managers compare Teva's production costs with those of alternative suppliers around the world. They share this information with manufacturing managers who learn where process improvements are required and may concur with a decision to outsource products where the external suppliers' costs are lower than Teva could achieve in the foreseeable future. These actions contribute to increasing Teva's long-term profitability.  The activity-based cost information also helps managers
determine which manufacturing facility is appropriate for different
types of products. For example (see Figure 1), Plant A has a
relatively inflexible (high capital-intensive) cost structure with a
high percentage of plant-level costs and a low percentage of unit
costs. This plant is most appropriate for high-volume production of
standard products. Plant B, with a significantly lower percentage of
plant-level costs and a relatively high percentage of unit costs, is
much more flexible and is appropriate for producing small batch sizes
and test runs of newly introduced products. Thus, ABC information
also is being used to determine operating strategy.
The Best News: Harmony Is Growing An unexpected benefit of the activity-based transfer price system is the ability to measure profit performance under changing organizational structures. Teva, like many other pharmaceutical companies, undergoes periodic organizational changes. By understanding cost behavior at the activity and product level, financial managers can forecast the potential performance of newly created profit centers and reconstruct what the past profit performance history would have been, assuming that the proposed profit center reorganization had existed for the past several years. The ABC system also enables senior executives to measure profit performance across organizational...cost and profit center...boundaries. For example, Table 5 shows the profitability of a significant product family whose individual products are manufactured in different plants and are sold by more than one marketing division. Jacob Winter, Teva's vice president of pharmaceutical operations, commented on the benefits derived from the ABC-transfer price system: In our changing environment, it is important for us to be able to understand and forecast our cost behavior. Some products remain in certain stages of production for a long time. These stages require resources of professional production and quality assurance staff, even when no direct labor is involved. On the other hand, since the supply of these resources is relatively fixed in the short run, we understand that we can use their capabilities for several small batch runs. He also recognized that activity-based costs are not the primary information used for short-term operational decision making: The ABC data provide an indication that must be supported by other information and facts. One cannot rely only on costing information when making operational decisions. Our short-term operational decisions focus on current bottlenecks and lead-time considerations. ABC provides guidance and insights about where we should be looking, but it is not the primary data for operational decisions. Perhaps most important, the introduction of ABC-based transfer prices has led to a dramatic reduction in the conflicts among marketing and manufacturing managers. The managers now have confidence in the production cost economics reported by the transfer price system. Manufacturing managers who "sell" the product and marketing managers who "buy" the product concur with the reasonableness of the calculated transfer prices. Teva's senior executives interpret the sharp reduction in intraorganizational conflicts as one of the most important signs that the use of activity-based transfer prices is succeeding. Robert S. Kaplan is the Marvin Bower Professor of Leadership Development at the Harvard Business School. Dan Weiss is an instructor in the Industrial Engineering and Management Department of the Technion (Israel Institute of Technology) and partner, OIC Technologies Consulting Group. Eyal Desheh, formerly deputy chief financial officer at Teva Pharmaceuticals, is currently vice president and CFO of Scitex Corporation. Comments can be addressed to rkaplan@hbs.edu. 1 R. Cooper, "Cost Classification in Unit-Based and Activity-Based Manufacturing Cost Systems," Journal of Cost Management, Fall 1990, pp. 4-14. 2 Production lot sizes can be expanded, if demand increases to a higher, sustainable level, by making technical changes to the production process and performing a quality control procedure to verify and validate that the product characteristics and quality have not been altered by the larger production batch. 3 At present, the Teva transfer price system does not charge the customer order for the full cost of setting up the production batch nor for the inventory carrying cost of the unused tablets or syrup. This is a refinement that could be added to the system in future years. 4 Registration costs include the costs of gaining and maintaining approval from governmental agencies for the right to manufacture each product.
| ||||||||||||||||||
 | ||||||||||||||||||
