Contracting Chapter 3

Costing, for all its apparent precision in its use of hard numbers, incorporates many assumptions and conventions. Any report or estimate of costs can be manipulated to advance or to disadvantage particular arguments.

In this Chapter we will look first at the reasons for undertaking cost analysis; the appropriate method will usually depend on the reason we are seeking information. Then we will look at some of the concepts in costing, including the behavior of costs with volume and how costs are shared between different activities.

A key question in management decision-making is often: "How much does X cost?". We may want to know, for example:

This last-mentioned type of decision is central to production decisions in multi-product firms, and tends to dominate costing texts. In the public sector, however, the manager usually has little discretion on whether to expand or discontinue a product or service.

With the exception of this last issue, the public sector generally faces the same costing issues as the private sector. In addition the public sector faces some costing issues of its own. The main ones are:

None of these questions, in management or in policy, has an unambiguous answer. Costing is imprecise, and the answer we get to any costing question will depend on the assumptions we use. Usually the first question we should ask is: "Why do we want to know?", for cost models relevant to one purpose will not necessarily be relevant to another.

Opportunity cost - defined as the value of this resource in its next most productive use. This is a central concept in costing, and, while it has a solid theoretical basis, it can give rise to some tremendous difficulties in measurement. For example, an essential piece of machinery which is not being used to capacity may not have a high opportunity cost. A specialized fire truck may have an extremely high opportunity cost when there are several fires to attend to. When we ask "what is the cost of X if we use it for purpose Y?", we should think "If it were not being used for purpose Y, what would it be doing?" If the fire truck would otherwise be sitting in the depot, then the opportunity cost is small. If it would otherwise be fighting a grass fire, then the opportunity cost of taking it to another fire would have to include the cost of the damage from the grass fire.

Opportunity cost is not always amenable to easy measurement; we are hardly likely to conduct a cost analysis while two fires rage. But it is a useful way of thinking when it comes to allocating scarce resources.

A particular use of opportunity cost relates to the services of volunteers. Many community services use volunteers extensively. Their services do not necessarily enter the financial accountants (though there is no reason they should not be entered at some realistic value as a contribution in kind). Volunteers are scarce resources with an opportunity cost; they are not free. They are not necessarily available for any going task. The volunteer guides in the local art gallery could be used, perhaps, for art education lectures, but probably not for cleaning the gallery.

Perspective of the decision maker - the cost you see depends on where you stand.

Government finance departments often have an obsession with costs which pass though the budget, but little concern for other costs. For example, a local government can save on budgetary costs by resurfacing roads with loose gravel, relying on trucks and cars to do the final rolling. While this saves budgetary costs it shifts some costs on to the community in terms of scratched cars and broken windshields, and the extra costs borne by the community may be greater than the budgetary cost saved.

In the example above, analysis which was confined to the budgetary costs of road surfacing would be referred to as financial evaluation, while analysis which also included the community costs, adding these to the budgetary costs, would be known as economic evaluation. The difference lies in the perspective of the decision-maker.

External costs - costs which are real, but which do not enter the decision maker's accounts.

The most common cases of external costs or externalities relate to pollution. The cost of the open fire in my living room is more than what I pay for building my fireplace and buying wood; there is also the cost, to my neighbors, of soot on their washing and to the community in general of the added pollution. From a parochial perspective it may appear very economical for a city on the Murray river to discharge its effluent into the river, but it is a cost borne by the residents of downstream municipalities.

These are examples of negative externalities. There are also positive externalities, which relate to benefits which do not enter the decision maker's accounts but confer benefits on others. A new university will probably enhance local property values. Our decision to have a guard patrol the city hall after hours will have security benefits for owners and users of nearby buildings. There are external benefits in connection to many utilities; my decision to be connected to sewerage and a telephone line carries benefits for other members of society. (It is because of such externalities that governments often compel people to use certain services, such as trash collection and sewerage, a point to which we return in Chapter 5 where we look at market failure.)

Transfers - a shift of resources from one party to another, but with no change in those resources.

Again the perspective of the decision-maker is important. The decision to sell Qantas is seen by economists as involving no change in resources; all that's happened is that there has been a shift around on the right hand side of the balance sheet. The new owners will have incurred a cost, however, in purchasing the company. It will be a cost to them, but not to the economy.

The notion of transfers as defined by economists, should not be confused with the closely related notion of transfer payments, which are financial transactions within an entity. For example one division in a government department may charge another division for use of its staff. On the department's books this payment will cancel itself out, and will not make a call on any financial resources, but it could still represent a call on real resources.

(In Canberra many years ago, when the Postmaster General's (Telstra) was a government department, and STD calls were new and expensive, Canberra public servants used to make liberal use of their phones for personal calls. The justification was that the cost simply represented a transfer payment between the department and the PMG, but no real cost to the public purse. This argument was valid at the margin, but the service was so heavily used that the PMG had to invest in a new exchange. It was no longer simply a transfer; it had become a real cost.)

Conventional economic wisdom says that costs previously incurred are not relevant to the decision to continue incurring costs. The economists' aphorism is: "let bygones be bygones".

To illustrate the point, imagine we are building a sports stadium, which will yield ongoing net annual benefits of $1.2 million. The quote is $15.0 million, so the stadium's return will be 8.0 percent - just within our government's cut off criterion of 7.5 percent. After the project is started and $4.0 million is spent, it becomes obvious that the sub-surface geology is very different from what was revealed in initial drilling, and that the site will require an extra $3.0 million on drainage and piling. That means the project will cost a total of $18.0 million, and its return will drop to 6.7 percent. Should we proceed or should we abandon the project?

The answer is obvious; if the project is abandoned it will yield no benefits, with a further $14 million it will yield benefits of $1.2 million a year, or 8.6 percent. Had we known at the outset we wouldn't have started, but this far in we are committed. Such reasoning has brought us some expensive projects, however. Understatement of a project's ultimate cost is a common means of manipulation called wedging, used most effectively by New York's Director of Public Works Moses fifty years ago to build some of that city's extravagant projects.⁽⁴⁾ Many prominent public projects, such as Sydney's Opera House, have been developed by wedging. The entrepreneurial bureaucrat will wedge a small appropriation to obtain funds for a project, and will then proceed to ensure that only part of the project is completed - preferably a part which is publicly visible and which is useless without further funds being spent.

Accountants define profit as the difference between revenue and expense. To economists, however, a normal profit (that is, a return on capital) is one of the normal costs of running a business. An economist's definition of cost is the difference between revenues received from the sale of goods and the value of inputs, which includes the opportunity cost of capital, so that profits are economic profits.⁽⁵⁾ In short, the return which assets would normally make, which would be some mixture of interest, capital gain and dividends, is part of the economic cost of the business. It is an example of opportunity cost - what would these assets make elsewhere?

This concept is built into the NSW Pricing Tribunal Act, which refers to the "agency's economic cost of production". Profits above that level, economists call "economic profit". A more colloquial term for economic profit is 'rip off'. For example, if a firm's return on equity was 25 percent in an industry where the norm was 13 percent, 12 percent would be economic profit, with the base 13 percent being normal profit which is part of economic cost.

In general, for any enterprise, costs vary with the level of output. The more we produce the more cost we incur.

Both total and unit costs will be dependent on volume, and, because we live in a world in which there are both economists and accountants, and in which there is a need for both types of information, we consider both types of costs here.

Accountants preparing financial returns, managers preparing budgets, and taxation authorities collecting company tax, are concerned with total costs.

The most basic cause of output dependence is because some costs do not vary with output, while some others do. The costs are known as:

This last concept of avoidable costs requires some explanation. It is an attempt to answer the question "what would we save if we did not produce X", where X is a discrete block of output. What would a university save if it closed a school? What would the State Rail Authority save if it closed the Yass Railway Station? It is of particular relevance in contracting out, where the key question is the cost which our agency would save if we no longer carried out the activity in-house. To this point we will return in Chapter 7 on contracting out.

An alternative way of looking at costs is to look at costs per unit. Economists and others concerned with efficiency, and taxation authorities looking a cost basis for sales tax calculation, tend to be concerned with unit costs.

These four concepts are similar to their counterparts in the total model, except that they take total costs and divide them by output.

The concept of marginal cost needs some explanation. Once we get away from simple linear models it is no longer the same as average variable cost, for, in many enterprises, variable costs actually change as production levels are changed. Marginal costs often follow a U shape as output increases - initially falling, then rising.

To illustrate this point we could think of a small print shop, with four offset machines, each with a capacity of two million pages a year.

To buy inks, photographic plates and papers in small quantities is expensive. There are scale economies in purchasing. Good discounts cannot be obtained until the print room is turning over at least three million pages a year.

But there are also scale diseconomies. The most efficient machine is modern with a low operating cost. The next most efficient is a little older. The other two machines are museum pieces, but are kept for very busy periods. In fact, before swinging these into production the printer prefers to use overtime on the newer machines, but that incurs a labor cost penalty.

We can see how a combination of scale economies and scale diseconomies are at work. The cost curves - MC, ATC, AVC are shown on the graph. (In the interest of keeping the diagram uncluttered AFC is not shown.)

Note the relationship between the AVC and MC curves. They start together, for the MC and the AVC of the first unit of output are identical. As the marginal cost falls, so too does the AVC, but at a slower rate; therefore the AVC curve remains above the MC curve. In other words the marginal is pulling down the average. To use a sporting analogy, if the Sri Lanka cricket team started on 12 runs an over, but slowly dropped to 6 runs an over, then their average would be above 6 but less than 12.

As the marginal cost curve turns up, it catches up with the average, and crosses the AVC curve at its lowest point. To continue the cricket analogy, imagine Sri Lanka now lifts its game. So long as the marginal rate (runs in this over) are below the average so far, the average will keep falling. Once, however, the marginal run rate passes the average so far, the average starts rising.

Consider a community service organization with simply one output, say providing home-delivered frozen meals. It buys the meals from a commercial supplier, and delivers them to clients. It has a refrigerated truck, a refrigerated storeroom, and a small office. It has a staff of three.

Now some costs will be incurred at a certain level regardless of the number of meals supplied. Some costs, on the other hand, will vary with the level of activity. It would be possible, in this organization, to develop a table, dividing these costs into fixed and variable costs.

Fixed costs

Variable costs

Staff salaries

Refrigerated storeroom

Truck standing costs (registration etc)

Office rent

Purchased meals

Truck operating costs (diesel etc)

Some costs will be hard to classify. It may be that after a certain point it is necessary to employ a second driver. It may be hard to classify the depreciation on the truck as fixed or variable - its value will decrease in part with use and in part with the passage of time. It may be that if the agency delivers no meals it can turn off the compressor in the coldroom.

Cost models can be developed to take such refinements into account, but it is usually a good starting point to use a simple fixed/variable classification. Refinements can come later

Suppose that as a result of careful research this organization identifies its fixed costs to be $150 000 a year, and its variable costs $3.00 per meal delivered. Then it can develop the table shown below. (Using a spreadsheet, such as Excel^tm or Lotus^tm, it is very easy to develop a table of fixed and variable costs.)

Number of meals ('000)	Fixed cost $'000	Variable cost $'000	Total cost $'000
0	150	0	150
20	150	60	210
40	150	120	270
60	150	180	330
80	150	240	390
100	150	300	450

It is also possible to present this information graphically, as is shown alongside.

Such a model is useful for budgeting, and, as we will see further on, in break-even analysis. But before we leave this topic, it is useful to look at an alternative way of analyzing this organization's costs, in terms of unit costs and to see that even in such a simple operation there are several different ways to interpret unit cost.

Going back to the table above, we can analyze all costs on a unit basis by dividing the costs by the number of meals. This gives a range of figures depending on the volume and the inclusion or otherwise of fixed costs.

Number of meals ('000)	Average fixed cost $	Average variable cost $	Average total cost $
0		3.00
20	7.50	3.00	10.50
40	3.75	3.00	6.75
60	2.50	3.00	5.50
80	1.88	3.00	4.88
100	1.50	3.00	4.50

Again, it is possible to present this information graphically. In this simple illustration average variable costs and marginal costs are identical because there have no scale diseconomies incorporated into the model.

Although break-even analysis is as much about pricing as it is about costing, it is convenient to introduce it under costing as it is developed from a consideration of fixed and variable costs.

To illustrate, let us come back to our hypothetical frozen meal service, with fixed costs of $150 000 a year and variable costs of $3.00 a meal. If we know the price charged for a meal then we can construct a break-even model of the operation. If the price received per meal is $5.00, then the following table can be developed.

Number of meals ('000)	Fixed cost $'000	Variable cost $'000	Total cost $'000	Revenue $'000	Profit/ loss $'000
0	150	0	150	0	-150
20	150	60	210	100	-110
40	150	120	270	200	-70
60	150	180	330	300	-30
80	150	240	390	400	10
100	150	300	450	500	50

The table and graph are the same as those in the previous section, except that two more columns have been added to the table, relating to revenue and profit/loss, and a revenue line has been added to the graph.

The break-even point comes at a volume of 75 000. This is the volume of sales at which the $2.00 profit per unit is adequate to cover the fixed costs of $150 000. A term for that unit profit is contribution, which is shorthand for "contribution to fixed costs and profit". Note that break-even analysis makes no reference to unit cost. Rather it looks at fixed costs and variable costs in relation to volume.

Such analysis, however done, helps managers and other decision-makers get a feeling of the relationship between cost, revenue and volume.

The costs described above hold only over a given period of time. Accountants and economists point out that over the very short term all costs are fixed and over the long term all costs are variable. For example, over a very short period (an hour or so) an electricity authority cannot vary its costs much, especially when using thermal stations. Over a very long period the authority can decide on different technologies (gas, hydro, nuclear) etc, and will have an optimum technology and scale for each projected level of output. In the medium term, however, the fixed/variable cost breakdown provides us with useful models. Over the short term the meals on wheels service is committed to keep its staff employed. Over the long term it may be able to change its staff number through attrition or recruitment and it may be able to buy a coolroom and a truck which are perfectly matched to the size of the task.

In this regard it is interesting to note that traditionally economists and accountants considered labor, especially factory labor, to be a variable cost. If you didn't produce and sell, you didn't have to pay wages. Labor could be retrenched and re-hired as market conditions demanded. Machinery, on the other hand, was a fixed cost, as were the salaries of managers and others who did not work on the factory floor. A class system, supported by a set of assumptions about the fungibility of labour, tended to become built into accounting systems.

We should use judgement in costing, and not fall back on dated convention. Labor, especially skilled labor, may be a fixed cost, while machinery, if we can lease it, may be a variable cost. And many functions, once considered the preserve of management, can be bought on the market from consulting firms. It is useful to think of fixed and variable cost - an operation with high fixed cost is more vulnerable to a sales downturn than one with costs which vary with volume, for example. Such considerations will help managers take a rational view of investment and risk. But it may be unproductive to draw on dated notions of what costs are "fixed" and what costs are "variable".

Downward sloping unit cost curves and flat variable cost lines reflect economies of scale. They exist in many industries. They are usually easy to model. For a school, for example, the number of support and administrative staff will rise more slowly than the number of students. If we need one cleaner for a school with 300 students, we will probably not need three for a school of 900, and we will still need only one principal.

Diseconomies of scale, however, are less easily measured in any accounting or economic model. Our example of the print room above shows certain easily-traced diseconomies of scale, but there are many others. As an organization gets bigger the tasks of coordination and communication get more complex. In a large school, supervision of trouble some children becomes more difficult. Fewer parents are likely to contribute voluntary resources to a large, impersonal institution (an example of the free rider problem). Staff may find themselves spending more time looking for others, or walking from classroom to classroom. In a small municipal library the librarians may know who the local troublesome users and vandals are; in a larger library such local knowledge is less likely. These diseconomies of scale are real and costly, but are often hard to measure.

When it comes to deciding on the scale of community facilities (swimming pools, libraries) a clear diseconomy of large centralized facilities is the extra travel time compared with dispersed facilities. This is another example of an external cost which would not be covered in a financial evaluation but which should be covered in an economic evaluation.

Single product establishments are the exception rather than the norm. Most establishments produce more than one product. Big manufacturing firms may produce hundreds of different products; government mega departments are responsible for many different programs. Local government has a huge range of activities. Even within one business unit in local government there may be many activities. A water supply authority may have different "products" - water delivered in large but irregular quantities to industrial customers, water produced in regular predictable quantities to domestic customers.

The problem in such establishments is that many costs cannot be traced to particular products. It may be too expensive to trace costs (e.g. tracing electricity use to every machine) or it may be conceptually impossible (e.g. tracing fire protection costs to particular products or tracing the chief executive officer's time to particular products).

Frequently in private industry management wants to know what each product 'costs'. Is each product pulling its weight? Politicians will want to know the costs of various program elements; what does it cost to process each Medicare payment, for example? Often a set of costs developed for one purpose, say inventory valuation for tax purposes, are used, wrongly, for decision-making purposes. Inventory valuation is an important issue. There is an incentive to value inventory low, so that profit is reduced and thus tax is reduced; there may be an incentive to value it high so that a high profit may be reported, leading to a high dividend payout and thus a high share price. Multinational firms have an incentive to report their highest costs in countries with the highest taxation rates. For those reasons there are conventions in costing, particularly inventory valuation, but these 'costs' may not be suitable for decision-making purposes.

Costing conventions come from manufacturing and are applied, often with little change, to other industries. There are many variants, but they are mostly based on building costs up from the factory floor, a "bottom up" approach.

Materials and direct labor (people who actually manipulate the items) are traditionally the main elements of direct cost - that is, costs incurred in the factory which are both variable with respect to volume and traceable to particular products. Bear in mind the tradition of considering labour as a variable cost. Permanent labor is, however, still a variable cost if it can be deployed to different tasks. A worker with multiple skills who can be transferred form one task to another has an opportunity cost.

Some costs, such as electricity and wear on fork lift trucks, may be variable but not practically traceable to particular products. Therefore direct costs, while being variable, are lower than truly variable costs, because there will be some costs which do vary with volume but which cannot be traced.

The figure bwlow shows the combination of these concepts of traceability and variability into a four quadrant diagram. Conventional costing systems use the direct costs (top right hand corner) as a base and count all other costs as overheads. By implication all these other costs ("overheads") are fixed (even though costs in the bottom right corner are really variable). Another term for non-traceable costs is joint costs - that is, costs shared between products.

Overheads are often lumped into two classifications, depending on where they occur, in the factory or in administration. To direct costs are added other factory costs, often called manufacturing or factory overheads, to give a factory cost. This is used to value inventory. Out of these costs must come sufficient contribution to cover managerial overheads, which are assumed all to be all fixed.

Many organizations incorporate such a method in calculating a "full cost" for labour. Once a direct labour cost is calculated, local and corporate overheads are absorbed using standard absorption factors. These factors are calculated from historical accounts; they are set at a level which just covers all costs. Thus, if an organization could classify $10.0 million to direct costs, $4.5 million to local overheads, and $5.5 million to corporate overheads, it would use absorption factors of 45% and 55% for local or factory overheads and corporate overheads respectively.

An example of such a calculation for an employee with an award pay of $35 000 a year is shown in the table below. Annual paid wages are used as a starting point, to which are added all known direct costs. Available hours are used in the denominator, deducting holidays, unproductive time etc. Once an hourly direct cost is obtained the overheads are added in, using standard absorption factors. There are local variants, but the general pattern is similar.

Such an approach to costing labour is used widely. From a financial accounting perspective it works well. All costs can be sheeted home to programs, and the method is time-honored and easily understood.

Calculation of an hourly labour cost

1	Annual wage	35 000

2	Days a year	260
	Less annual and long service leave	-25
	Less public holidays	-8
	Less sick leave (average)	-4
	Less training	-4
3	Adjusted days	219

4	Hours a day	8.0
	Less personal time	-0.5
	Less unavoidable delays	-0.5
5	Adjusted hours	7.0

6	Productive hours a year (3 x 5)	1 533

7	Hourly cost (1 / 6)	22.83
	Plus superannuation 9%	2.05
	Plus workers compensation 6%	1.37
	Plus leave loading 1.2%	0.27
8	Adjusted hourly cost - direct cost	26.53

9	Workplace overheads 45%	11.94
	(supervision, accommodation etc)

10	Corporate overheads 55%	14.59

11	Absorbed hourly cost (8 + 9 + 10)	53.06

While this method may serve the purposes of financial accounting, from a management accounting perspective it is useless.

If the organization were thinking of contracting out the service modelled above, about the only cost it can be reasonably confident about is the $26.53 direct cost. Even that may not have been calculated well. How carefully were the sick leave figures obtained? Can the 0.5 hours of unavoidable delays be reduced?

When it comes to the various overheads there is no logic which dictates that they should respond to the hourly labour cost. Most, by definition, are fixed. Many may be variable in the long run, but even if they are, do they vary with the hourly labour cost? Take, for example, the payroll staff. Their workload may vary with the number of employees, or with the complexity of payment, being highest for those employees with overtime and with fringe benefits. The marginal saving in the payroll area of removing a work team from the establishment may be negligible; there is little saving in producing 90 payslips and group certificates rather than 100. The method of allocating joint or non-traceable costs is usually arbitrary, but different methods can produce very different results.

To illustrate this point further, consider a simplified example, a local authority which handles two functions - water supply and sewerage disposal. Its total annual costs are $2.0 million, of which $655 000 can be traced to water, $515 000 can be traced to sewerage, and the balance is joint or non-traceable.

	Water	Sewerage	Joint (non-traceable)	Total
Basic Operating Figures
Assets $m	4 000 000	2 000 000	1 000 000	7 000 000
Megaliters handled	2 000	160		2 160
Employees	5	8	15	28
Costs $pa
Return on assets (7%)	280 000	140 000	70 000	490 000
Labor costs	170 000	225 000	650 000	1 045 000
Materials	30 000	20 000	5 000	55 000
Depreciation (3%)	120 000	60 000	30 000	210 000
Other	55 000	70 000	75 000	200 000
Total	655 000	515 000	830 000	2 000 000

In this example the $655 000 definitely belongs to sewerage, and the $515 000 definitely belongs to water. That is, $1 170 000 can be allocated to these functions, but $830 000 cannot be traced.

Using labor costs as a basis of allocation, the method would be to allocate proportionately. A fraction of 170/395 of the $830 000 would go to water, and 225/395 would go to sewerage. That is only one way. It could be argued that because these activities are capital intensive joint costs should be allocated on the basis of capital. At one extreme it could be argued that the authority exists primarily to provide water; sewerage is only a secondary function and should not absorb any overheads. The results of such different calculations are shown below, and these do not include some composite allocations (e.g. labour overheads by employment, capital overheads by capital.)

Various Cost Allocations
	Water	Sewerage	Total	Water costs per ML	Sewerage costs per ML
Basic Traceable Costs	655 000	515 000
All joint costs to sewerage	0	830 000
Total	655 000	1 345 000	2 000 000	$328	$8 406
Joint costs by employment	319 231	510 769
Total	974 231	1 025 769	2 000 000	$487	$6 411
Joint costs by payroll	357 215	472 785
Total	1 012 215	987 785	2 000 000	$506	$6 174
By all traceable costs	464 658	365 342
Total	1 119 658	880 342	2 000 000	$560	$5 502
By assets	553 333	276 667
Total	1 208 333	791 667	2 000 000	$604	$4 948
By volume	768 519	61 481
Total	1 423 519	576 481	2 000 000	$712	$3 603
All joint costs to water	830 000	0
Total	1 485 000	515 000	2 000 000	$743	$3 219

Different methods for absorbing non-traceable costs give more than a 2:1 variation in unit costs, and that's before considering whether any of these costs are fixed or variable. It is possible in a production system like water and sewerage, characterized by high investment in long-life assets, that almost all costs are fixed. For example, if the authority's dam is always full, with water going over the spillway, then the marginal cost of water (excluding environmental externalities) may be zero. Conversely, whatever the costing system says, if the authority has a permit to discharge 160 megaliters of treated sewerage, and another permit would be very expensive or unobtainable, then the marginal cost of sewerage may be extremely high.

It will also be evident from the above that choice of a costing system is not a mere technicality; it can be influenced by interest group politics. If the authority is using cost-based pricing, and if its costing method is publicly available, then different interest groups are going to develop arguments very much in their self-interest. Think of the arguments which may be put forward by a hotel owner (many lavatories, little water) contrasted with those which may be put forward by a market gardener with high water use.

Costing often involves bargaining; the bargaining is about the method of costing. Parties will put forward the method which best serves their own interests. For example, a 1989 study of the cost of Telecom's community service obligations found that it was possible to produce annual cost estimates from $148 million up to $800 million.⁽⁶⁾ The low figure ($148 million) was based on avoidable costs; what would Telecom save if it didn't provide these CSOs? This was the favored approach of the Bureau of Transport and Communication Economics. The high figure ($800 million) was generated by Telecom using a system known as stand alone costs; what would it cost Telecom to provide only these CSOs and no other services? In other words, Telecom would exist only to provide these CSOs. One had to imagine a telephone system with nothing except public phone boxes and remote subscribers, with no large metropolitan system or trunk system to provide the basic shared infrastructure. The method may have little connection with reality, but is served as a useful opening gambit in negotiations. Perhaps its main contribution to the discipline of costing is that it assures us that costing is anything but a precise exercise.

These traditional "bottom up" systems have an air of precision and objectivity, and they are still widely used. Indeed, the power of modern computing has made for more elaborate cost systems. But they are getting increasingly out of touch with reality. First, as labor is replaced with machinery, a lower and lower proportion of costs can be traced as 'labor' costs. Second, as people work more in work teams and adopt multiskilling, the whole Taylorist notion of tracing individuals' times to specific functions is becoming less relevant; in fact, more and more people in the workplace are doing work which was once classified as factory overheads. Third, large cost increases have been in the area of managerial overheads and it is hard for middle managers to appreciate that they are, perhaps, adding little value.

Yet many organizations are persisting with systems which were developed in different times, when spans of control were wide, when there were many workers and few bosses, and when most tangible value-added came from the factory floor. Government departments and business enterprises often take on cost systems, developed in manufacturing, in another era, with little modification.

The implication in "bottom up" costing systems is that overheads are non-controllable; the systems focus attention away from managerial costs. This may not appear rational, but it is a reminder that costing systems are designed in those areas of an enterprise classified as "overheads". They don't have a strong incentive to turn the costing spotlight on head office.

The practice now formally known as activity based costing is centered on the basic principle of traceability. It has come to have its own terminology and definitions, and a standard set of processes for implementation. It has also come to incorporate a technique known as process value analysis, which is generally the first stage in implementing an activity based costing system. Good texts on cost accounting should cover activity based costing in some detail - the description below is simply a bare bones description of the underlying principles.⁽⁷⁾

Process value analysis involves looking at each step in the production of a good or service, and asking what value it adds to the final product. This means first checking if there is any value- added at all, by asking the difficult question "would the customer be any worse off if this step was left out of the production process?" If an activity has no value-added it can go. If its value-added is low in relation to its costs, then that activity should be done more efficiently. No area of the enterprise can exempt itself from this exercise, which often turns attention to areas which have escaped scrutiny in the past. Much of the benefit of activity based costing stems from this process value analysis, which can be undertaken as a stand-alone exercise. When people start asking what each activity contributes to ultimate customer welfare many time-honored activities do not stand up to such scrutiny. Complex accountability, paper trails, routine written reports and other bureaucratic processes are often among the victims of process value analysis.

Once all value-adding activities are identified, they can be grouped into activity centers, preferably with measurable inputs and outputs. Some accounting texts suggest four classifications of activities:

Once like work is gathered into "activity centers" - that is, units doing like work for one or more products - costs can be traced to those activity centers. (Unless great care is taken in tracing costs to these activity centers there is a risk of double-counting. For example, rent may be assigned entirely to activity centers, in which case it should not be counted again under facility level activities.)

Then, for each activity center, "cost drivers" are identified. A cost driver is a factor which causes costs to be incurred. In a library, for example, there are several possible drivers, such as number of users, number of borrowings, acquisitions. In general, activity based costing tries to get one driver for each activity center. The driver has to be measurable, and capable of being linked back to the product. It should correlate closely with the costs actually incurred in each activity center. Thus, in a drawing office, the number of drawings may be the appropriate driver, and it is easily linked back to the products (at a product level). A corporate library may use the number of searches as the driver, and may link these back to the type of product.

This last stage of activity based costing can be very expensive, and the allocation of drivers often involves somewhat arbitrary decisions. Often, quantification is impossible, or is forced only after some very simplifying assumptions. Nevertheless it is much more sophisticated than simply using single factor "overhead" absorption. Activity based costing tries to avoid the general "overhead" concept as far as possible.

Activity based costing, like so many accounting systems, grew up in manufacturing. It is probably most suited to batch or mass production manufacturing and service industries. For industries with non-standardised output it is probably too difficult to use a full activity based costing system, but the exercises of process value analysis, identifying activity centers and searching for cost drivers can still be worth the effort in terms of reviewing the cost-effectiveness of all activities.

It is doubtful whether any system can serve all needs. In Chapter 6 we will look at some issues in performance analysis, which go further into the use of costing data. Activity based-costing, while being touted as an innovation, draws on long-established principles. Indeed, organizations centered on projects, such as construction contractors, have tended to use systems with a high level of traceability for many years.

The general principle in costing which underlies activity-based costing and other costing reforms is the principle of traceability. Financial accounting systems, for the sake of simplicity, have done their job when they have absorbed costs fully. The cost analyst has to take apart such costs and trace them back to particular products or services. For example, in looking at a program with a lot of public contact, it may be useful not to accept the accountant's classification of phone calls as "overheads" to be shared by all programs, but to do some analysis of the phone use of the particular program.

Cost awareness does not have to rely on systems. Japanese firms, for example, tend not to use complex cost systems; cost and efficiency awareness is built in more as a corporate culture, and, while they do pay great attention to cost measurement, it is used more for product and process design rather than as a basis for day-to-day control.

The main barrier to costing reform is perhaps that there are many people within organizations who don't really relish the exposure and accountability which come with costing reform.

1. What is the cost of the following resources which could possibly be found in a local government authority:

3. At present a local government sends out rate notices and includes a postage pre-paid envelope. What is the cost saving in requiring residents to pay for their own postage?

5. A government welfare agency has decided to let the private sector compete to provide counselling services. Its own counselling services will be maintained, but they must break even and determine an appropriate pricing structure.

	Counsellor	Psychologist	Total
Interviews	2 000	1 500	3 500
Time per interview	0.33	0.33
Direct cost per hour	$36.00	$63.00
Direct cost per interview	$12.00	$21.00
Annual direct costs	$24 000	$31 500	$55 500
Percentage of annual direct costs	43.24%	56.76%	100.00%
Overheads			$210 000
Allocated proportional to direct costs	$90 811	$119 189
Overheads per interview	$45.41	$79.46
Total cost per interview	$57.41	$100.46

6. The works managers of two adjoining local government authorities have been informally lending each other equipment, keeping a tally in a notebook. Between the two authorities there is often spare equipment.

This practice has come to the attention of the finance areas of the two authorities concerned, and they have directed that such exchanges be on a full cost recovery basis, including overheads. If A lends a backhoe to B, A should charge B $105 per hour, being $60 per hour for direct costs associated with the backhoe, plus $45 to cover overheads. In such a transaction the works department of B would be charged $105 an hour, which would be paid to Authority A. When the $105 is received in Authority A, $45 would be retained by the central administration, and $60 would be credited to the works department to compensate for the direct cost in using the backhoe.

What's wrong with this model of cost recovery? How do such arrangements come about?

4. Robert A Caro The Power Broker - Robert Moses and the Fall of New York (Alfred A Knopf 1974).

6. BTCE Report # 64 Costing Telecom's Community Service Obligations (AGPS 1989).

7. Some of the detail is taken from Ray Garrison and Eric Norween Managerial Accounting - Concepts for Planning, Control and Decision Making (Irwin 1994).