ARGUE AND REFLECT ON THE FOLLOWING ARTICLE:

ABSORPTION COSTING FOR DECISION-MAKING

Mike Lucas, of the University of Buckingham, describes how absorbing certain fixed overhead costs into unit product costs can complement Kaplan and Cooper's approach to ABC, to improve the quality of planning and decision making

Received wisdom has it that absorption costs should not be used for decisionmaking, i.e. non-volume related costs should not be allocated to the product unit level. If such costs are allocated down to the product unit level, the resultant unit cost will be a function of production volume. Take, for example, the cost of performing a machine set-up. This will be independent of the number of units to be produced. Consequently, if the size of a production run is halved, the per unit cost of a machine set-up is doubled!

Such a cost is therefore only valid for one particular volume of output, whereas most planning or decision-making deliberations involve potential changes in volume.

The supposed raison d'etre of ABC is to improve the quality of planning and decision-making; an ABC unit cost is therefore of little value.

This principle is recognised by Kaplan and Cooper in their advocacy of ABC for planning and decision-making purposes.[1] It is also acknowledged by Drury in his fourth edition,[2] in which he describes Kaplan and Cooper's approach to ABC. The principle is still largely disregarded by a number of textbook writers, who continue to present ABC as primarily a product costing system--a more sophisticated absorption costing system, using a number of absorption bases (cost drivers), rather than just one volume related basis such as labour hours.

Kaplan and Gooper's approach to ABC

Kaplan and Cooper, in their exposition of ABC, explicitly condemn the practice of allocating nonvolume related costs to the product unit level. Consequently, their conception of ABC is as a more sophisticated marginal costing system--employing contribution analysis at several different levels.

In their approach to ABC, the firm is envisaged as a hierarchy of costs with each level representing a different type of cost variability--as shown in Figure 1. The lowest three levels in the hierarchy are concerned with product related costs. Unit level costs are short term variable costs such as direct material, direct labour and so on. These costs vary in direct proportion to the number of units produced.

Many product related costs are not, however, driven by the number of units produced per se, but by the number of batches/production runs undertaken. An obvious example of such 'batch level costs' is machine set-ups.

The third category of product sustaining costs relates to activities which are necessary to maintain the product in the firm's product range. For each product, a firm must maintain a bill of materials, a routing sheet, tooling and so on. Such 'product sustaining' activities will increase as more products (or product variants) are added to the product range. Such costs, however, will be independent of the number of units produced in a particular period.

In addition to the various product related costs, there are some expenses which do not relate to individual products, but to particular customers, e.g. order processing and distribution costs. Finally, there are certain general overheads which do not relate to any particular product or customer, such as the chief executive's salary.

With Kaplan's approach to cost analysis there is no need (or justification) for any allocation downwards from a higher level, of non-volume related costs (in Figure 1) to the product unit level. Rather, contribution should be rolled upward when considering a particular decision. A unit-level operating margin for individual products can be calculated by subtracting unit-level costs from sales revenue. From this unit-level margin can be subtracted batch-related and product-sustaining expenses to arrive at a product-level margin. Such a margin can be calculated for each product in the range and from this can be deducted any customer related expenses to determine the profitability of a particular sales order.

This approach can give managers better insights into the nature of their costs and thus facilitate better informed decisions when bidding for jobs or deciding which orders to accept. The method can also be applied in determining the profitability of a product over a period of time--not by calculating a full unit cost to compare with the selling price, but by identifying the demands the product puts on service and support activities. The estimated cost of these activities is then added to the short-run variable costs of producing the units sold and this total cost is then compared with the total revenue earned by the product. Thus, for example, it may be revealed that low volume products are unprofitable because of the disproportionate demands placed on batch level activities--a machine set-up is necessary whether one unit or one thousand units are being manufactured!

An important implication of Kaplan's analysis is that the non-volume related costs included in any decision analysis are not short-term incremental costs, but proxies for long-run incremental cost-based typically on existing average activity costs. Thus, for example, the total cost of machine setups is divided by the number of set-ups performed to arrive at average cost per set-up. This is then used as an estimate of the long-term incremental cost that will arise by performing machine set-ups. Kaplan and Cooper argue the superiority of this approach on two grounds.

First, they contend that it is probably not feasible to identify all the actual incremental costs of each decision in a complex, multi-product environment. Second, they argue that, during a particular period, numerous decisions are made independently of each other but the collective impact can be to drive up 'fixed' costs over time, even though these longer-term cost increases are not incremental (and hence relevant) to. any one particular decision.

The argument is that, if only actual incremental costs are considered when looking at a particular decision, a positive contribution is likely to lead decision-makers to embark upon actions (e.g. accepting marginal business) which place demands on service and support functions. Over time this consumption of service/support activities will start to drive up costs. The current actual cost of support activities can be viewed as acting as a proxy for these long-run incremental costs. The ABC approach is essentially a long-term one aimed at curbing the excesses which can result from traditional marginal costing--which ignores potential long term increases in cost.

Thus, although Kaplan and Cooper assert that ABC is not a decision-making tool (but an attention directing one), it is difficult to escape the conclusion that it is! It is nearer to the truth to say that ABC is a tool for planning/decision-making, but one which uses proxies for long-term incremental cost rather than strictly decision-relevant incremental costs.

The role of cost allocation

The Zimmerman approach

The Kaplan and Cooper approach is based on the received wisdom that non-volume related costs should not be allocated to the product unit level.

There is, however, a school of thought in favour of such allocations, suggesting that cost allocations have an important role to play in motivating and controlling managers.

Zimmerman,[3] for example, suggested that there are certain service/support costs which are fixed with respect to short-run changes in production volume, but which are related, in the longer term, to the level of primary inputs (e.g. labour). Such costs would typically be classified under the general overhead category in Figure 1. An example would be a works canteen.

Every additional worker employed places additional demands on the canteen. Although actual cost of the canteen may be fixed in the short-term, opportunity costs arise in the form of delays and/or degradation of service to other users. As time goes by, management will tend to adjust the actual level of service provided, in order to reduce delays/degradation of service, and hence costs rise.

Zimmerman argues that these opportunity costs and consequent long-term incremental actual costs should be taken into account by individual managers when deploying labour resources. Such costs are hard to observe/measure, but allocating current average cost can serve to proxy them.

Thus, for example, a production manager deciding on the optimal mix of factor inputs will be induced to take into account the additional costs of using labour--in terms of the demands it places on the works canteen--if labour hours are made to carry the burden of recovering this overhead cost.

The resultant factor input mix, it is argued, will be closer to the optimal one than would be the case if such costs were ignored.

This principle can be illustrated by Figure 2, which shows how the selection of input factors (in this example, capital and labour), to produce a given output, depends on the relative prices of those factors.

The isoquant shows all the possible combinations which could be used to produce a particular output. All points assume technical efficiency (i.e. no waste); the optimal combination (i.e. that producing the lowest unit product cost) will depend on the slope of the isocost line which represents the relative prices of input factors. Isocost line 1 represents relative prices if the demands placed by labour on service departments such as the canteen are ignored and only the actual cost of labour itself is taken into account. The 'true' cost of using labour resource is understated and in consequence more labour hours are used than would otherwise be the case. The manager is, in effect, basing his decision on false price information.

If the true cost of using labour is proxied by means of charging labour hours with service department overhead, the relative price of labour increases, as shown by isocost line 2, and less of it should be used. The resultant input mix (point 2) is closer to the real optimum, because the hidden costs of using labour are recognised. By charging labour hours with fixed overhead burden, overhead is factored into the unit product cost. If a unit requires one labour hour, a corresponding amount of overhead will be included in the unit cost. According to the logic of Zimmerman's argument, this is justified since, in the longer term, actual costs of the service resource will tend to rise in relation to the number of labour hours worked and the allocated fixed cost is a proxy for these additional costs, i.e. one labour hour gives rise to x canteen cost.

The question naturally arises as to which costs should be classified as long-run volume-related costs and thus allocated to the product unit level-after all, virtually all costs change if volume is increased enough! What about, for example, rent? Zimmerman's argument appears to relate to those areas where there is additional consumption of resources as a result of an increase in a primary input such as labour. As a consequence of this consumption of resources, delays/degradation of service to other users occurs. These opportunity costs give rise to additional actual costs in the longer term as management adjusts the level of service provision. This is unlikely to apply to rent, unless a huge increase in labour force was being contemplated-necessitating a new factory!

Therefore an assumption about the likely volume level over the planning horizon is required in deciding whether to allocate or not to allocate. If, over the anticipated volume range, significant delays/ degradation of service--as a result of increasing primary inputs--are considered a realistic possibility, then there is a good case for cost allocation. Where, over the anticipated volume range, it is considered unlikely that delays/degradation of service to other users will result--as a consequence of increasing primary inputs--there is not a good case for cost allocation.

Applying this criterion, it may be appropriate to allocate works canteen, payroll and personnel costs, for example; it will probably not be appropriate to allocate rent and so on.

The Japanese approach

A slightly different rationale for cost allocation is that apparently adopted by a number of Japanese manufacturing firms? These firms load labour hours with overhead burden--not because they believe this will result in an optimal mix of inputs, but to encourage the substitution of capital for labour as part of their advanced manufacturing technology (AMT) strategy. All production overhead--including non-volume related costs (i.e. batch level, product-sustaining level costs etc) are allocated using labour hours. Consequently, the resultant 'unit costs' are of little use for either longterm or short-term decisions--since they are neither the short-term nor long-term incremental costs of

producing a unit! Japanese managers do not believe that the resultant product costs are a reasonable proxy for long run incremental costs; rather, they believe that the behavioral benefits accruing from cost allocation outweigh the disadvantages, namely inaccurate product costs and possible erroneous decisions! Japanese companies tend to use estimated product costs produced outside the management accounting system and based on managers' past experience or intuition.[5] It appears that the primary purpose of management accounting in Japan is to influence behaviour rather than to provide accurate product costs. It is not clear that the West is ready to sacrifice its penchant for accurate product cost information, which has been the cornerstone of western management accounting since its beginning.