Fixed and variable costs

Possibly of even greater importance than the distinction between capital costs and operating costs is the distinction between fixed costs and variable costs. You will see that the main argument for the cost-efficiency of ODL - the expectation that distance learning generates economies of scale - rests on this distinction.

The distinction relates to volume of activity. The principal activities of an ODL institution consists in teaching students. Hence fixed costs are those which do not increase with the volume of activity, i.e. the number of students taught. Development costs of teaching materials are fixed costs in this sense. You need only develop the study guides once for a particular course and they can be used by many learners. Of course, you would have to reprint them as new students enroll but the development costs remain a one-off. The costs for reprinting and posting the materials to the students are examples of variable costs. Ideally, all costs can be classified as either fixed or variable. However, some costs are classified as semi-variable. These are costs, which are fixed up to a certain threshold. Typical semi-variable costs are costs of class tutors.

Activity A8:
Classifying cost-drivers as fixed or variable costs
1. Open the spreadsheet Activity A08.
2. Classify each cost driver in the list as follows:
select an item from the list
copy and paste it onto the stars of one of the categories
the evaluation will tell you whether you have correctly categorized the item.
Click here


Total cost equation

For the time being we will focus on fixed and variable costs only. In this slightly simplified case the total costs are the sum of the fixed and variable costs:

Total Costs
=
Fixed costs
+
Variable costs

TC

=

F

+

V x N

Where:  
TC =
Total Costs
F =
Fixed costs
V =
Variable costs
N =
number of students

Note well that the VxN for Variable costs is a composed term in which V stands for Variable cost per student and N for Number of students. (Example: if it costs US$ 4 per student to replicate some content on a CD-ROM and post it to the respective student, it costs US$ 400 to do the same for all hundred students you may have in the same course. The respective Variable costs come to US$ 400, or US$ 4 per student times 100, the number of students.)

The Total Costs equation is a function of N (total costs depending on the number of students). In order to denote this mathematicians would write: TC(N) = F + VxN.

Activity A9:
Exploring the Total Costs equation
This activity allows you to vary the fixed and variable parts of the cost equation so that you can see how the graph of the costs varies.
1.
Open the Excel spreadsheet A09.
2.
Try changing the variable costs per student. What happens?
3.
Try changing the fixed costs. What happens?
(If you want to understand the maths behind this, you may recall that the equation
  TC(N) = F + V × N
is a linear equation of the form f(x) = kx + c where c is the intersection with the y-axis and k is the gradient (slope) of the graph. In our case the constant is F which identifies the starting plateau of the costs, while V (variable cost per student) is the gradient. The higher the value of V, the steeper the gradient.)
Click here

You may have noted that the factor V affects the slope of the graph of TC. The higher the value of V, the steeper the curve. F affects the plateau, i.e. the starting level of the curve. Generally educational planners try to include as many students as possible but at the same time to keep total costs down as much as possible. The important observation in this context is that eventually V is more decisive than F. You may start at a higher initial cost level, but if V is lower, there will be an intersection point beyond which the function with the lower gradient (i.e. value for V) will have a lower unit cost per student. This can be seen more clearly when we look at the average cost function.


Average cost equation

The total costs equation leads to the other important equation about average costs. Average costs are total costs divided by the number of students N.

Average costs per student = Total Costs / Number of students

AC = TC/N

AC = (F + V x N) / N = (F/N) + (V x N) / N

AC = F/N + V


Where:  
AC =
Average Costs
F =
Fixed costs
V =
Variable costs per student
N =
number of students


It is important to understand this equation clearly, because it provides important guidelines for cost-efficient course planning. The important point is what happens to average costs as N increases.

As N increases, AC decreases, other things being equal.

This is what is meant by economies of scale.

The right-hand side of the equation is made up of two components. The first, F/N decreases as N increases. The second, V, does not change as N increases since it is the variable cost per student.

This result is often described as the fixed costs are 'spread over more students'. Each student is charged a part of the fixed costs. The more students there are, the less each has to pay.

This is what educational planners want: falling unit costs. Mathematicians like to look at extremes and ask what would happen if we were to increase the number of students ad infinitum. The answer is that, in this case, the first term (F/N) approaches zero or, in mathematical language, the average costs 'converge to' V.

Activity A10:
Exploring the Average Costs equation
The average costs graph (Figure 1) shows how cost-effectiveness arises from increasing student numbers.
1.
Open the spreadsheet Activity A10.
2.
Try different values of F and V to see how the graph changes.
Note on the maths
The AC equation is an algebraic transformation of the total cost equation (TC).
You can see from the graph that the line is asymptotic - that is, it tends towards the long-term value V but never quite reaches it.
AC is only defined for N> = 1. AC(1) = F + V. Since often V is almost negligible, AC(1) is approximately F while AC(N) for very big N is approximately V. This differential between F and V identifies the scope for any economies of scale.
Click here


Figure 1: Average costs and economies of scale

In the above example the red lines refer to conventional education (CE) and the blue lines to distance education (DE).
AC DE stands for the Average Costs per Student in Distance Education and V DE for the Variable Costs per Student in Distance Education.
Similarly AC CE stands for the Average Costs per Student in Conventional Education and V CE for the Variable Costs per Student in Conventional Education.
The interpretation of the figure is the following: Since V DE is smaller than V CE and AC CE cannot fall below V CE, the graph of AC DE will, if student numbers are large enough, fall below the graph of AC CE (towards V DE).
At the intersection point you find a downward pointing arrow which marks the break-even point. The break even point marks the number of students, beyond which (in this case) AC DE undercuts AC CE.

It is important to note that AC cannot fall below V, however large the number of students becomes. The variable costs per student therefore represent a bottom line, below which the average costs per student cannot fall. This implies an important strategic guideline:

To lower average costs per student keep variable cost per student low.

The other strategic guideline has already been mentioned:

To bring the average cost per student down, increase the number of students.

This guideline needs some additional comment. The cube in Figure 2 shows that economies of scale vary according to the number of students already enrolled. The higher the number of students, the lower the average costs reduction effect of including another student. Hence you have to judge whether it makes sense to increase student numbers when economies of scale are already largely exhausted.

Of course, the number of students cannot be increased at will. You need to attract students and marketing efforts themselves are a cost factor and may not succeed in increasing numbers at an economic cost. Students may prefer multi-media courses with high level of learner support. If you want to offer this in order to attract further students you will need to increase V or F or both. Hence, it is important to keep in mind that N, F, V are not independent. The behavior of N is influenced by V and F. If you lower one of these parameters, students may walk away from your course and the plan to lower average costs may backfire, because the lower number of students means that fixed costs can be spread only over fewer numbers such that average cost per student will rise, possibly initiating a vicious circle.


Perraton's Costing Cube

Perraton (2000, p.137) has portrayed the relation between volume, media sophistication and the interactivity. His cube (Figure 2) has three dimensions and the arrows show the direction of reducing cost per student.

Figure 2: Perraton's Costing Cube


In fact, Perraton's visualization is very close to the average cost formula. The fixed costs in distance education are generally related to media sophistication, the variable cost per student is strongly influenced by the level of interactivity. (The cube is slightly modified. The original cube speaks only about face-to-face tuition. However, meanwhile we can sustain teacher student interaction at a distance (e.g. videoconferencing, online conferencing). But all these forms of interactivity between students and teacher, irrespective of the technology used, claim the teacher's time and increase variable costs. ) The Internet and videoconferencing influence the cost per student as much as face-to-face tuition does. The number of students, varying from few to many, is explicitly referred to in both models.


Marginal costs

We need to include a definition of marginal costs since the term is part of the language of ODL.

Strictly speaking marginal cost means the cost of including one more student in your system.

We can express this like this in mathematical language:


MC = TC (N + 1 )- TC (N) = [FC + V x (N + 1)] - [FC + V x N] =
FC - FC + V x (N + 1) - V x N = V

Where MC: Marginal Costs


The equation shows that the cost of including one more student in your system is equal to the variable cost per student. The interesting point here is that fixed costs do not impact on marginal costs. Offering something at marginal costs therefore strictly speaking means to offer it at a price that makes no contribution to fixed costs.

Fixed costs in ODL are mainly related to development costs. Saying you offer something at marginal costs often implies that you are willing to write-off the development costs.


Semivariable costs

We have treated fixed costs and variable costs so far as a binary distinction. This means any cost driver can either be treated as fixed cost or as variable cost per student, as F or as V. This is a little unrealistic. In practice, many costs are semivariable Such costs are fixed up to a certain threshold volume.

For example, you can increase the number of students in an online seminar without adding another class as long as the number of students is below the maximum class size. Beyond this size you need to start a new class and employ an additional tutor.

The graph of a semi-variable cost takes the form of a step function: within limits you may increase volume of activity (i.e. number of students) without raising costs. At a certain point costs will jump.

Formally, we define semivariable cost function as follows:

SV = [N/G] x SN

Where:  
SV =
Semivariable Cost
G =
Group size
V =
Variable costs per student
N =
number of students
[N/G] =
Number of groups (the square brackets signify the process of rounding )
SN =
Semivariable Cost per Group

Note that the number of groups (or classes) needed is defined by the number of students in the system and the maximum group size.

Theoretically, it can be argued that all costs are semi-variable Most cost drivers are to some extent affected by an increase in the volume of activity if only the increase is big enough. It may be that the concept of sem-ivariable costs has been ignored in ODL for so long because ODL was largely seen as 'individual studies'. Nowadays it is increasingly possible to teach classes at a distance. In this case the notion of semi-variable cost as distinct from fixed and variable costs per student becomes more and more important.

Total Costs = Fixed costs + Semivariable costs + Variable costs

TC = F + [N/G] x SN + V x N

Where:
TC =
Total Costs
F =
Fixed Costs
SN =
Semivariable Cost per Group
N =
number of students
SN x [N/G] =
Semivariable costs (i.e. Semivariable cost per group x number of groups)
V x N =
Variable costs (i.e. Variable cost per student x number of students)

Activity A11:
Exploring the effects of group size (TC)
This activity explores semivariable costs.
1.
Use spreadsheet Activity A11 for this.
2.
Try changing the group size.
3.
Then try different combinations of fixed cost, variable cost, semivariable cost and group size.
4.
Observe what happens in each case.
In ODL systems with little or no group work, semivariable costs are not very important and can usually be ignored.
When there is a significant amount of group work, semivariable costs become important. You can see why as you change the input values in this spreadsheet.
Click here

This leads to a modification of average costs also:

AC = TC/N

AC = F/N + ([N/G] x SN) / N + (VxN)/N

AC = F/N + SN/G + V

Activity A12:
Exploring the effects of group size (AC)
This activity looks at the effect of group size on the average cost equation.
1. Use the spreadsheet Activity 12 for this.
2. Try different group sizes to see their effect on average cost.
The effects of group size on the graph are generally less visible.
Click here

Unit costs

Another useful concept is unit costs. You have seen that various cost drivers can be seen as variable cost per student, e.g.:

print costs per student

postage costs per student.

In order to keep unit costs low you need to control all the items that contribute to unit cost per student.

The main lessonthat we can draw from our study of semi-variable costs is that larger group sizes lead to greater cost efficiency. The drawback is that this reduces the level of interactivity, a feature, which many see as an important indicator of quality.


The generic costing template and the fixed costs/variable costs distinction

How does the generic costing template relate to the distinction between fixed and variable costs? Table 8 below classifies some cost drivers.

Table 8: Some cost drivers for fixed and variable costs
 
DIRECT COSTS
INDIRECT COSTS
of development
of presentation
overheads
Fixed costs Authoring a text   Director's salary
Variable costs per student   Marking of TMAs Help desk


Summary and caveat:

1. ODL has a different cost-structure than conventional education.
2. Cost-structure in this context refers to the composition of fixed and variable costs in the total or average cost formula.
3. ODL has a generally lower variable cost per student. This is its strategic advantage. Even though often ODL may require a higher up-front investment, these higher costs can be spread across many learners.
4. The high level of fixed costs is often seen as a guarantee for quality. The rationale for expecting ODL to be more cost-effective than conventional teaching is the combination of comparatively low variable costs per student and high fixed costs safeguarding quality (effectiveness). High quality and low costs, according to this line of thinking, can only be achieved in large systems which have a further positive and intended effect: increasing access.
5. One further comment: The efficiency path would lead to lower average cost per students. Given the enormous demand for education (and the 'perverse way' of rising unit costs), the capacity of distance education to bring down average costs per student is closely related to its remit to broaden access to education. Especially, in developing countries coping with large numbers is one of the main reasons to turn to distance education (Perraton, 2000). However, planners should be aware that lowering average costs per student in this model is achieved by expanding the system, which, in turn, raises total costs. (This caveat to any cost-analysis, exclusively singing the praises of distance education for lowering unit costs, is forcefully developed in Butcher & Roberts (2004).):

John Daniel portrays these expectations by his eternal triangles as in Figure 3. According to Daniel (2001) the cost structure of ODL allows costs to be reduced while at the same time increasing access and quality. This reflects our theoretical expectations, but it makes assumptions that may not apply in every context.

Figure 3: The cost-quality-access triangle


oben