Conundrum 10
Understanding costs

‘That’s ludicrous’ Heather exclaimed. As a GP member of the Professional Executive Committee of the Primary Care Trust she had been lobbying for months to move some of the orthopedic services out of the local hospital trust and into primary care. Initial discussions had foundered because moving hospital staff out to run these community sessions would, it was feared, lead to increased waiting lists as they would be seeing fewer patients. Now the hospital managers had agreed to transfer the funds they would save by not providing this service, so that the PCT could take the service over. However the sums they were talking of turned out to be miniscule. Hence Heather’s anger.

Is Heather’s anger justified? Is the hospital necessarily being unreasonable?

Commentary 10

The hospital and the PCT would provide this service in slightly different ways, with different staff and in different locations. So it is not surprising that the costs calculated by the two organisations are different. But this different? The figures are so far apart that Heather believes the hospital is being devious, supplying figures that don’t; reflect reality, and trying to hold onto money that should be coming out to the PCT.

Is she right to feel so aggrieved? Not unless she understands how costs are arrived at, how costs behave with different levels of activity, and that she is comparing like with like.

The costs of providing a service will include staff, equipment they need, materials that are used, premises from which they operate and the management and governance processes associated with the service. The PCT will need to employ additional staff and purchase some equipment. Whereas the hospital can save money spent on the materials, and the money they spend on the portion of their staff members’ time that is associated with this service, but they can’t free up the equipment because it is used for other services also, as are the premises. So we can see immediately that the money saved by the hospital will be less than it costs the PCT to set the service up. But Heather is concerned that the cost per patient contact appears to be so different. That it looks as though every contact provided by the PCT will be more than twice as expensive as if it is provided by the hospital. This cannot be right, she reckons.

Let’s look at how costs are calculated, because this isn’t as straightforward as it may seem.

To find the cost of a ‘widget’ we are manufacturing or a patient contact we are providing we divide the total costs associated with producing it by the number of ‘widgets’ or patient contacts that we produce. That sounds simple, but let us consider exactly what we should include in those ‘total costs’.

Imagine for a moment that we are producing widgets. Clearly, we must include all the costs of the materials used up in the manufacture of the widgets. Also, the costs of the labour hours spent on making them. But if the machines used in their production also produce other goods, then any costs of maintaining these machines cannot be borne entirely by the widgets. And yet, if the other goods were to go out of production, then the entire machine maintenance cost would have to be recouped from the widgets. The widget-making machine will be housed in factory premises which incur rates – should these be included in our ‘total’ costs? And what about the expense of the staff canteen?

Already we can see that there are different kinds of costs.

The cost of any goods or service can be thought of as made up of three parts: direct costs, indirect costs and overheads. Direct costs are those which are used only for, and entirely by, the product we are trying to cost. In our widget example, these would be the materials and labour. Indirect costs cannot be linked solely to one product but with several. Machine maintenance would fall into this category, as would any holiday pay for the operator. Overheads are incurred on an organisation-wide basis; in our example, the building rates and staff canteen are overheads.

If instead of trying to cost a widget we were attempting to calculate the average cost of all the products made by the widget-making machine, then the machine maintenance switches from being an indirect cost to a direct cost. So whether a particular cost is direct or indirect depends on the cost objective; that is, exactly what it is we are trying to cost. It is therefore not possible to say that study leave, for example, is invariably a direct or an indirect cost, since it will depend on whether the individual taking it is engaged in providing more than one service and on what is the cost objective.

Indirect costs must be allocated to the cost objective, but normally the means of doing so are not contentious and require only the deployment of common sense. Machine maintenance costs, for example, could be split according to the ratio of the number of widgets produced to the number of other products produced on the same machine. Alternatively, the machine hours spent on each could be used. Another possibility would be the weight of the materials used in production. A knowledge of the machine and its propensity for malfunction would be needed in order to choose between them, but to someone with this knowledge the appropriate option would be self-evident.

When it comes to overhead costs, however, the allocation process is almost invariably a two-stage process, the alternatives are greater in number and the option selected can have an impact on the decision-making behaviour of departments and budget-holders. In the widget example, the building rates and the costs of running the staff canteen need to be allocated. A sensible rationale for apportioning the rates would be the floor area each department occupies. Using floor area as the basis for allocating canteen costs would be ridiculous; instead, the number of employees would make more sense. The canteen will also have received its portion of the rates bill and will be passing this on when its total costs are allocated. Suppose that when the rates, the salary of the chief executive and the renovation of the building’s exterior are included in the canteen costs, they exceed the prices on offer elsewhere? Suppose departments could feed their staff more cheaply by giving them an allowance to buy meals at a local cafe? If the canteen closed, the rates, chief executive’s salary and renovation expenses would still have to be paid, so should they be included in the ‘total costs’ of the canteen? Is the answer influenced by the fact that the chief executive spends no time at all worrying about the canteen service or management, or that the canteen is in an interior room with no exterior walls? Management accounting requires careful consideration of behaviours.

The different ways in which overhead costs are allocated may all have their basis in logic but lead to very different figures being produced. The same widgets manufactured in the same way can be presented as costing very different amounts. The same is true for patient contacts. So the hospital and PCT may have made different choices about how they allocate their indirect and overhead costs.

But there is another set of differences that are relevant: the differences in costs when the level of activity changes.

We have just seen how costs can be divided into direct costs, indirect costs and overheads. To investigate cost behaviour, we need to use another classification, this time fixed, variable, semi-fixed and semi-variable costs. You may be tempted to try and equate the two classifications (fixed with indirect and overheads, variable with direct). Remember, though, that they have two different purposes, one to calculate costs at a given level of activity, the other to explore their behaviour at different activity levels.
Variable costs change in direct proportion to your activity level. You double or treble your volume of activity and your variable costs double or treble accordingly. They include all the materials used in the production process. Variable costs therefore remain the same per unit as activity levels rise or fall. The total expenditure on variable costs of course rises or falls (varies). Fixed costs remain the same regardless of activity levels. The salary of the chief executive is a fixed cost. Fixed costs, on the other hand, rise per unit if the activity level falls, because the number of units which must bear them has fallen. Similarly, they fall, per unit, if the activity levels rise. The total amount though remains the same.

Staff salaries are not a variable cost, since a rise or fall in activity levels does not precipitate an immediate and consequent change in salaried hours. But if volume increased or decreased to the point where an extra post could be justified, or an existing one jeopardised, then the salary figure would change. Such costs are called semi-fixed costs, or step-fixed costs, since they remain the same over a given range of activity levels and then increase or decrease in steps. Equipment maintenance costs, which are incurred with greater frequency as activity increases, are not fixed costs, but neither are they variable, since they do not increase or decrease with every unit of activity. These are semi-variable or step-variable costs.

Imagine the range of activity levels changing from zero to ten times the current level, and you will see that almost no cost is completely fixed. Ultimately all staff can be made redundant and premises sold; conversely, staff numbers could increase ten-fold and a new site be acquired. So when we consider the breakdown of costs into fixed and variable, we are doing so over a limited range, otherwise known as the relevant range.

Identifying which elements of the costs are fixed and which are variable enables us to differentiate between the average cost and marginal cost. The average cost is just what it says: the fixed plus variable costs divided by the number of units of activity. The marginal cost is that of providing one more unit. Unless you are right at the limit of a step cost, the marginal cost is the variable cost. Obviously, once you have recovered all your fixed costs, you are able to afford to offer your services at a price lower than the average cost, without losing money, as long as you recover at least your marginal costs.

So where costs look very different it may be that one set gives the marginal cost and the other refers to the average cost. This may be what is happening in our scenario. The hospital may be giving the marginal savings to them of not offering the service, as they will still have to pay for all the overheads, and they are not at a step point where the fixed costs can be saved. Whereas the PCT is calculating the average cost, including all the elements involved in offering the service, including a contribution to the cost of the board etc.

This brings us onto another set of costs: relevant and irrelevant costs. In making the decision whether to provide this service itself or to purchase it the PCT will need to ensure it considers only costs that are relevant. What does this mean?

Whereas when you are calculating costs all elements of the cost are relevant, when you are making decisions about different options some of them become irrelevant. This is because they have been incurred anyway: they are irretrievably committed, they are unavoidable. No matter what you do you cannot influence them. These costs are termed sunk costs. You can ignore them, no matter how big they are, because nothing will change them. Since your decision cannot change them, it must not be influenced by them. The fact that a million pounds has been spent on a piece of equipment would be difficult to ignore. If, however, using that equipment is going to cost you more money for a given outcome than a different alternative (once you have included all purchase and related costs in your calculations), then you must be able to say goodbye to the million pounds. The decision to purchase that machine may have been absolutely the right one at the time, or it may have been a misguided one; but whichever is the case, insisting on its use when there are alternatives that will be cheaper for you now, which give clinical results that are just as good, is indisputably wasteful.

When you are financially appraising different options, you need to consider only the costs which differ between them. If both options require the additional employment of a member of staff, or if they can both be accommodated within the existing establishment with neither permitting any staff reduction, then staffing costs are irrelevant to your decision: they do not need to be included in your calculations. You compare the position with and without that option and write a with/without case.

The staff costs mentioned are irrelevant costs, but they are not sunk costs. Thus all sunk costs are irrelevant, but not all irrelevant costs are sunk.

Irrelevant costs, despite being ignored in your calculations, are real in that they have been spent. There is a class of costs that are relevant even though they are only ever hypothetical, and these are called opportunity costs. Whenever you have any money to spend (or any other resources to allocate for that matter), there is always something else you could choose to spend it on. The benefits of this alternative are closed to you if you choose the option under consideration. An opportunity cost measures the opportunity that you must forego when your selection of one option means you cannot pursue another. If you can pursue both, then there is no sacrifice of opportunity, so opportunity costs are only incurred where there is a scarcity of resources.

Although these relevant and irrelevant costs are not what is bothering Heather at the moment, she and her colleagues on the PCT Board will need a clear view of them before they make any decision about whether to provide this service or whether to purchase it from the hospital Trust. All in all Heather needs to be much better informed before she is in a position to remonstrate with her hospital colleagues about the figures they are presenting.

You can read more about costs and monitoring costs against budgets in Really Managing Health Care.

Please send any comments and suggestions to
conundrum@reallylearning.com