Cost Allocation

The process of assigning shared or indirect costs to specific departments, services, or patients using a defined allocation basis.

Source: Horngren, Datar & Rajan 2015

Costs that cannot be traced to a service are first gathered into pools corresponding to the function that incurred them, such as estates, information technology or pathology. Each pool is then assigned an allocation basis intended to reflect how services consume it, such as floor area, headcount, or number of tests requested. The pool is divided across services in proportion to their measured consumption of that basis. The accuracy of the result depends entirely on whether the basis reflects what actually drives the cost.

Source: Horngren, Datar & Rajan 2015

Support functions consume one another's services, so estates supports information technology and information technology supports estates, and a simple one-pass allocation ignores this. The direct method allocates support costs only to front-line services and accepts the resulting error. The step-down method allocates each support department in a chosen sequence, passing costs to those not yet allocated, which captures part of the interaction and depends on the sequence chosen. Reciprocal methods solve the mutual consumption simultaneously and are more accurate and rarely used in practice.

Source: Horngren, Datar & Rajan 2015

Because the allocated share is frequently a large part of a reported unit cost, and it moves with the basis rather than with anything the service did. Allocating support costs by patient numbers overstates the cost of short simple episodes and understates the cost of complex ones. A service can therefore appear to be losing money under one defensible basis and covering its costs under another. The distortion is invisible in the output, since an allocated amount looks identical to a measured one.

Source: Cooper & Kaplan 1988

Activity-based approaches identify the activities that support functions actually perform, establish what drives each, and assign costs through those drivers rather than through a single organisation-wide basis. Time-driven variants simplify this by costing capacity per unit of time and charging activities for the time they consume, which also makes unused capacity visible as a separate figure rather than burying it in the cost of work performed. Both are more accurate and more demanding to build and maintain than a conventional allocation.

Source: Kaplan & Anderson 2004

Three things need to appear alongside any allocated figure. The basis used for each pool, since the reader cannot otherwise judge whether it reflects consumption. The proportion of the total that was allocated rather than traced, which indicates how much of the figure rests on convention. And, where the analysis supports a decision about stopping an activity, the amount that would genuinely cease, since allocated overhead redistributes to whatever remains. Reporting a single absorbed unit cost without these is what produces business cases claiming savings that never materialise.

Source: healtheconomics.wiki