Cost Classification, Behaviour and Estimation

Introduction and Context

Cost accounting supplies the information needed to plan, control and improve the use of organisational resources. In materials management, this role is especially important because purchasing, inventory, warehousing, production scheduling and supplier decisions create costs that differ in timing, behaviour, controllability and strategic significance. A cost that is useful for inventory valuation may be irrelevant to a sourcing decision; a cost that appears fixed in the short run may become variable over a longer planning horizon.

Cost classification is therefore not merely a matter of labelling accounts. It is a decision-oriented process. The same expenditure can be classified differently for different purposes. For example, the salary of a warehouse supervisor may be a direct cost of the warehouse department, an indirect cost of a particular product, a fixed cost within a monthly capacity range and a relevant cost when evaluating a proposed warehouse expansion.

This chapter develops the principal classifications of cost, explains cost behaviour, and applies high–low, regression and learning-curve methods to cost estimation. It also evaluates the limitations of these techniques and their use in strategic materials decisions.

Learning Objectives

After studying this chapter, the learner should be able to:

1. Distinguish direct and indirect, fixed and variable, product and period costs.
2. Explain relevant, opportunity, sunk and lifecycle costs.
3. Analyse mixed, step-fixed and nonlinear cost behaviour.
4. Estimate costs using the high–low method and least-squares regression.
5. Apply learning-curve analysis to labour and procurement-related activities.
6. Interpret estimation results for planning, sourcing, inventory and capacity decisions.
7. Identify assumptions, limitations, governance risks and appropriate uses of cost information.

1. Cost as a Decision Concept

A cost is the monetary measure of resources sacrificed or committed to achieve a purpose. The purpose may be production, purchasing, storage, distribution, compliance or service delivery. Cost information becomes meaningful only when three questions are specified:

• Cost object: What is being measured—for example, a product, supplier, order, warehouse or customer?
• Time horizon: Is the decision concerned with the next month, the next year or the entire product life?
• Decision context: Is the purpose valuation, control, pricing, outsourcing, capacity planning or performance evaluation?

A cost may be actual, based on resources already consumed; standard, based on a predetermined expected amount; or estimated, based on a forecast of future expenditure. These categories support different managerial purposes and should not be confused.

2. Direct and Indirect Costs

Direct costs

A direct cost can be traced economically and conveniently to a specified cost object. Direct materials incorporated into a product and wages of workers assembling that product are common examples. Direct tracing is not restricted to physical incorporation: a transport charge specifically incurred for one purchase order may be direct to that order.

Indirect costs

An indirect cost cannot be traced to a particular cost object conveniently or economically, or tracing would not justify the administrative effort. Factory security, depreciation of a shared warehouse and the salary of a procurement manager are typical indirect costs. They are assigned using an allocation base such as labour hours, machine hours, purchase orders, storage area or transaction count.

The distinction is relative. A warehouse electricity cost may be indirect to individual products but direct to the warehouse department. The cost of a purchasing department may be direct to a procurement function but indirect to individual suppliers.

Allocation does not create causation. An allocated overhead charge may be required for full-cost reporting, but it should not automatically be interpreted as an avoidable cost. Managers should distinguish:

• Traceability: whether the cost can be linked to a cost object.
• Causality: whether the cost is caused by the activity under analysis.
• Controllability: whether a manager can influence it within the relevant period.

These three properties are related but not identical.

3. Product and Period Costs

Product costs

Product costs are assigned to goods or services and initially become part of inventory when inventory accounting applies. For a manufacturing organisation, they normally include direct materials, direct labour and manufacturing overhead. Product cost is expensed as cost of sales when the related goods are sold.

Period costs

Period costs are charged against income in the period in which they are incurred. Selling expenses, administrative salaries, advertising and head-office finance costs are generally period costs. They are not carried in production inventory merely because they support the business.

The distinction is primarily an accounting timing distinction. It does not mean that product costs are always relevant to a decision or that period costs are always irrelevant. For example, a future advertising expenditure may be relevant to a product launch even though it is a period cost.

4. Cost Behaviour

Cost behaviour describes how total cost changes as an activity driver changes. Typical cost drivers include units produced, purchase orders, deliveries, machine hours, inspection hours and tonnes handled.

Variable costs

A variable cost changes in total in proportion, or approximately in proportion, to the activity level within a relevant range. Variable cost per unit is constant under the simplest model.

In the cost function above, Y is total cost, a is total fixed cost, b is variable cost per unit of activity, and X is activity volume. In a purely variable-cost case, a = 0. The equation assumes a stable relationship, a measurable activity driver and a relevant operating range.

Fixed costs

A fixed cost remains constant in total over a specified range and time period. Rent, a salaried manager’s compensation and straight-line depreciation may be fixed for monthly planning. Fixed cost per unit changes as volume changes:

Thus, fixed cost per unit is not constant. It falls as output increases and rises as output decreases, provided total fixed cost remains unchanged.

Mixed costs

A mixed cost contains fixed and variable components, such as a warehouse utility bill with a standing charge plus a usage charge. The linear mixed-cost model is:

This model is useful for short-term planning but may be inaccurate when prices, capacity, technology or operating conditions change.

Step-fixed costs

A step-fixed cost remains fixed over a narrow activity band but increases when capacity thresholds are crossed. One supervisor may manage up to 10,000 order lines per month; an additional supervisor may be required above that level. Treating such a cost as continuously fixed can cause serious capacity-planning errors.

Nonlinear costs

Costs may be nonlinear because of quantity discounts, overtime premiums, congestion, spoilage, emergency freight or declining productivity. A supplier’s unit price may decrease with order size, while storage and obsolescence costs may increase. Strategic analysis should therefore consider the combined cost rather than focusing on one purchase price.

5. Decision-Relevant Cost Classifications

Relevant and irrelevant costs

A relevant cost is a future cost that differs between alternatives. A cost is irrelevant to a decision when it will be the same under every feasible alternative or has already been incurred and cannot be changed.

Relevance has three conditions:

1. The cost relates to the future.
2. It differs among alternatives.
3. It is measured consistently for the decision.

A fixed cost can be relevant if it is avoidable. A variable cost can be irrelevant if it is identical under all alternatives.

Sunk costs

A sunk cost is a past expenditure that cannot be recovered or changed by current decisions. Historical expenditure on an obsolete inventory-management system is sunk when evaluating whether to replace it. Sunk costs may be important for reporting and learning, but rational incremental decisions should not be based on recovering them.

Opportunity costs

Opportunity cost is the benefit forgone when a resource is used for one alternative rather than its next-best alternative. If warehouse space could generate rental income or store another high-margin product, using it for a slow-moving item has an opportunity cost even if no cash payment appears in the accounting records.

Avoidable and unavoidable costs

An avoidable cost disappears if an activity or alternative is discontinued within the decision horizon. An unavoidable cost continues regardless of the decision. This distinction is often more useful than the conventional fixed-variable distinction in make-or-buy and discontinuation decisions.

Incremental and differential costs

Incremental cost is the additional cost caused by an increase in activity. Differential cost is the difference in total cost between alternatives. In practice, the terms are often used together, although differential cost may include both increases and decreases.

6. Lifecycle Costing

Lifecycle costing evaluates costs from concept and design through acquisition, operation, maintenance, renewal and disposal. For a materials-management decision, the purchase price may be only one component of lifecycle cost. Relevant elements can include:

• supplier qualification and contracting;
• inbound freight, duties and inspection;
• inventory carrying and insurance costs;
• energy consumption and operating labour;
• maintenance, downtime and spare parts;
• quality failures, warranty claims and recalls;
• technology upgrades and training; and
• disposal, recycling or environmental compliance.

A simplified lifecycle cost is:

Here, LCC is lifecycle cost in present-value terms, C₀ is the initial cost at time zero, Cₜ is the cost incurred in period t, Sₜ is any salvage or residual value in period t, and r is the discount rate per period. The model assumes that cash flows can be estimated, the discount rate reflects the decision context and the timing convention is consistent. Non-cash effects such as risk, resilience and environmental harm may require separate analysis or monetary estimates.

Lifecycle costing is especially valuable when a low acquisition price creates high energy, maintenance, failure or disposal costs. It should not be used to produce false precision: uncertain future costs should be represented through scenarios, ranges or sensitivity analysis.

7. Cost Estimation Methods

7.1 High–low method

The high–low method estimates the variable rate from the highest and lowest activity observations, not necessarily the highest and lowest cost observations.

Y_H and Y_L are costs at the high and low activity levels; X_H and X_L are the corresponding activity levels; b is estimated variable cost per activity unit; and a is estimated fixed cost. The method assumes an approximately linear relationship and uses only two observations. It is quick but highly sensitive to unusual observations and therefore should be treated as a preliminary estimate.

Worked example. A distribution centre records the following monthly costs:

The high and low activity observations are 6,000 and 2,000 order lines.

The estimated cost function is:

At 5,000 order lines, estimated cost is ₹260,000. The estimate is valid only if 5,000 lies within a relevant range and the underlying operating conditions remain comparable.

7.2 Least-squares regression

Regression uses all available observations to estimate the line that minimises the sum of squared vertical errors. For simple linear regression:

Yᵢ is observed cost, Xᵢ is observed activity, a is the intercept, b is the slope and εᵢ is the unexplained error for observation i. The least-squares slope and intercept are:

The bars denote sample means. The slope has units of currency per activity unit; the intercept has units of currency. Regression assumes a sufficiently stable relationship, meaningful observations, independent errors for inferential purposes and no serious distortion from omitted drivers or outliers.

Regression is preferable to high–low when several reliable observations exist. However, a high R² does not prove causation. It indicates how much sample variation in the dependent variable is explained by the fitted model, not whether the selected activity driver is economically correct. Managers should inspect residuals, outliers, seasonality, capacity changes and multicollinearity in multiple-driver models.

7.3 Multiple regression and driver selection

A cost may depend on several drivers:

For example, warehouse cost may be related to pallet-days, order lines, inbound receipts and expedited shipments. Multiple regression can improve explanatory power, but it requires more observations and careful interpretation. Highly correlated drivers make individual coefficients unstable. A statistically significant relationship may still be operationally unhelpful if the driver cannot be influenced or forecast reliably.

7.4 Learning-curve estimation

Learning curves model declining average labour time as cumulative output increases. The classical unit-time model is:

Tₓ is the time required for the xth unit, T₁ is the time for the first unit, and b is the learning exponent. For a learning rate LR expressed as a decimal, when output doubles:

If the learning rate is 80%, b ≈ −0.3219. The model assumes that experience, repetition and process improvement reduce time according to a stable pattern. It is most appropriate for labour-intensive, repetitive work and less appropriate when automation, product redesign, workforce turnover or material shortages disrupt learning.

Illustration. If the first unit requires 100 labour hours and the learning rate is 80%, the eighth-unit time is:

This is an estimate of the time for the eighth unit under the unit-time model. It must not be confused with cumulative average time, which is a different learning-curve convention. The convention must always be stated before costs are calculated.

Learning effects matter in new-product procurement, supplier development, warehouse automation implementation and contract pricing. A buyer may negotiate a declining labour component while protecting quality and delivery requirements. Learning should not be assumed to continue indefinitely.

8. Applying Cost Estimates to Materials Decisions

Make-or-buy analysis

Compare the supplier price with the relevant internal costs avoided by outsourcing, plus the opportunity cost of capacity used internally. Allocated fixed overhead that will continue after outsourcing is not an avoidable saving. Qualitative factors—quality, intellectual property, resilience, labour standards and supplier dependence—must accompany the numerical comparison.

Supplier selection and total cost of ownership

A supplier with the lowest quoted price may not have the lowest total cost. Evaluation should include freight, minimum-order quantities, lead-time inventory, inspection, defects, expediting, payment terms and end-of-life obligations. Cost estimates should be prepared by cost driver and tested under demand and exchange-rate scenarios.

Inventory and warehouse planning

Relevant costs include ordering, receiving, storage, insurance, obsolescence, stockout and emergency replenishment costs. Fixed warehouse costs should be separated from volume-sensitive handling costs. Step-fixed capacity costs should be identified before accepting growth assumptions.

Budgeting and flexible control

A flexible budget adjusts expected variable costs to actual activity while retaining the appropriate fixed-cost assumptions. Comparing actual cost with a static budget can produce misleading variance signals when volume differs materially from plan. Standard costs should be periodically updated for supplier prices, process changes and inflation, with changes governed transparently.

9. Illustrative Case: The Valve Sourcing Decision

This is a fictional illustrative case.

Background

A manufacturing firm uses 20,000 specialised valves annually. It currently makes the valves internally. The production department reports an allocated unit cost of ₹470, consisting of direct materials ₹180, direct labour ₹120, variable overhead ₹50 and allocated fixed overhead ₹120. An external supplier offers the valves for ₹420 each. If production stops, ₹1,500,000 of annual fixed overhead will continue; ₹900,000 can be avoided. Internal capacity released by outsourcing can be used to produce a component generating a contribution of ₹300,000 annually.

Problem

Should the firm accept the supplier’s offer solely on cost grounds?

Analysis

The relevant internal cost per valve includes costs avoided by buying externally:

• Direct materials: ₹180 × 20,000 = ₹3,600,000
• Direct labour: ₹120 × 20,000 = ₹2,400,000
• Variable overhead: ₹50 × 20,000 = ₹1,000,000
• Avoidable fixed overhead: ₹900,000

Total avoidable internal cost is ₹7,900,000, or ₹395 per valve. The unavoidable fixed overhead of ₹1,500,000 should not be treated as a saving. The supplier’s annual purchase cost is ₹8,400,000, which is ₹500,000 higher than avoidable internal cost.

However, outsourcing releases capacity worth ₹300,000. The net disadvantage of buying is:

On the stated quantitative assumptions, making internally is preferable by ₹200,000 annually. The reported full cost of ₹470 is not the correct decision comparison because it includes unavoidable allocated overhead. Conversely, the analysis must consider supplier quality, continuity, tooling ownership, lead time, ethical sourcing and the risk that the alternative component’s contribution is uncertain.

Outcome and learning points

The firm should retain internal production unless due diligence shows that the supplier offers superior strategic value or the alternative capacity benefit is understated. The case demonstrates that relevant cost is decision-specific, opportunity cost can change the decision, and allocated overhead should not be mistaken for avoidable cost.

10. Limitations, Governance and Contemporary Practice

Cost models simplify reality. Common limitations include unstable prices, inflation, changing product mix, capacity thresholds, poor activity data, arbitrary overhead allocations and learning assumptions that ignore interruptions. Managers should document the cost object, time horizon, data source, estimation method, relevant range and excluded costs.

Cost information also has ethical implications. Deliberately excluding expected warranty, environmental or labour-compliance costs can make a supplier appear cheaper while transferring harm to workers, communities or future periods. Procurement governance should require transparent total-cost assumptions, conflict-of-interest declarations, audit trails and non-financial risk review.

Enterprise resource planning, warehouse-management systems, process mining and machine learning can improve cost-driver measurement. AI can identify patterns, forecast activity and detect anomalies, but it does not establish causation automatically. Automated estimates require data-quality checks, explainable drivers, access controls, model validation and human approval for consequential sourcing or capacity decisions. Technology should strengthen professional judgement rather than replace it.

Best practice is to combine accounting analysis with operational knowledge. Use high–low for a quick preliminary estimate, regression when reliable historical data exist, learning curves for demonstrably repetitive work, and lifecycle analysis when ownership costs extend well beyond acquisition. Present ranges and sensitivity results where uncertainty is material.

Chapter Summary

Cost classification is purpose-dependent. Direct and indirect costs concern traceability to a cost object; product and period costs concern accounting timing; fixed, variable, mixed, step-fixed and nonlinear classifications describe behaviour within a relevant range. Relevant costs are future and differential, while sunk costs are past and unchangeable. Opportunity costs represent benefits sacrificed, and lifecycle costs extend analysis across acquisition, operation and disposal.

The mixed-cost equation Y = a + bX provides a basic planning model. The high–low method is simple but uses only two observations. Least-squares regression uses all observations but requires careful driver selection and diagnostic review. Learning-curve models estimate declining unit time under stable repetition and must distinguish unit-time from cumulative-average conventions.

In materials management, cost analysis supports sourcing, make-or-buy, inventory, warehousing, budgeting and supplier evaluation. Good decisions compare avoidable and opportunity costs, include lifecycle effects, test uncertainty and integrate financial, operational, ethical and resilience considerations.

Key Terms

Cost object: The product, service, activity, department, order or other object for which cost is measured.

Direct cost: A cost traceable economically to a specified cost object.

Indirect cost: A cost not conveniently or economically traceable to a particular cost object.

Cost behaviour: The manner in which cost changes as an activity driver changes.

Fixed cost: A cost constant in total within a specified relevant range and period.

Variable cost: A cost that changes in total with activity volume.

Mixed cost: A cost containing both fixed and variable components.

Step-fixed cost: A cost fixed over a band but increasing when a capacity threshold is crossed.

Product cost: A cost assigned to inventory and expensed when the related goods are sold.

Period cost: A cost charged to the period in which it is incurred.

Relevant cost: A future cost that differs among alternatives.

Sunk cost: A past cost that cannot be changed by a current decision.

Opportunity cost: The benefit forgone by selecting one alternative over the next best alternative.

Avoidable cost: A cost eliminated by discontinuing an activity within the decision horizon.

Lifecycle cost: The present-value cost of acquiring, operating, maintaining and disposing of an asset or product.

High–low method: A cost-estimation method using the highest and lowest activity observations.

Regression: A statistical method for estimating relationships between cost and one or more drivers.

Learning curve: A model of declining time or cost as cumulative experience increases.

Relevant range: The activity band within which an assumed cost relationship is expected to hold.

Review Questions

5 Short Answer Questions

1. Why is cost classification described as purpose-dependent?
2. Distinguish a direct cost from an indirect cost, using a warehouse example.
3. What are the three conditions for a cost to be relevant?
4. Explain the difference between a sunk cost and an opportunity cost.
5. Why can a step-fixed cost create planning problems?

5 Long Answer Questions

1. Explain fixed, variable, mixed, step-fixed and nonlinear cost behaviour, including their implications for budgeting.
2. Compare product costs and period costs, and explain why the distinction should not be used automatically for decision-making.
3. Describe the high–low and regression methods of cost estimation, including assumptions and limitations.
4. Explain lifecycle costing and its importance in strategic procurement.
5. Discuss how cost information should be governed when used in supplier selection and outsourcing decisions.

3 Analytical / Application Questions

1. A warehouse’s cost is ₹150,000 at 2,500 order lines and ₹270,000 at 6,500 order lines. Use the high–low method to estimate variable cost per order line, fixed cost and cost at 5,000 order lines. State the assumptions underlying your answer.
2. A supplier quotes ₹96 per unit, while internal production incurs materials ₹38, labour ₹24 and variable overhead ₹10 per unit. Annual avoidable fixed cost is ₹400,000 for 20,000 units; unavoidable fixed cost is ₹700,000. Internal capacity could earn an opportunity contribution of ₹250,000 if released. Should the firm buy, based on quantitative factors? Explain.
3. A new warehouse process takes 80 hours for its first implementation and follows a 75% learning rate under the unit-time model. Estimate the time for the fourth implementation and discuss two reasons why the estimate may fail in practice.

Multiple Choice Questions with Answers

Question 1

Which classification describes a cost that can be traced economically to a specified cost object?

A. Direct cost

B. Period cost

C. Sunk cost

D. Opportunity cost

Correct Answer: A. Direct cost

Question 2

A cost remains constant in total within a relevant range but changes per unit as output changes. What type of cost is it?

A. Variable cost

B. Fixed cost

C. Sunk cost

D. Differential cost

Correct Answer: B. Fixed cost

Question 3

Which cost should normally be excluded from a current decision because it has already been incurred and cannot be changed?

A. Opportunity cost

B. Avoidable cost

C. Sunk cost

D. Incremental cost

Correct Answer: C. Sunk cost

Question 4

Using the high–low method, costs are ₹100,000 at 2,000 units and ₹160,000 at 5,000 units. What is the estimated variable cost per unit?

A. ₹15

B. ₹20

C. ₹30

D. ₹50

Correct Answer: B. ₹20

Question 5

In the equation Y = a + bX, what does b represent?

A. Total fixed cost

B. Total activity volume

C. Variable cost per activity unit

D. The regression error

Correct Answer: C. Variable cost per activity unit

Question 6

Which statement about regression is correct?

A. It always proves that the activity driver causes cost.

B. It uses only the highest and lowest observations.

C. It estimates a relationship using multiple observations.

D. It is valid regardless of outliers or changing capacity.

Correct Answer: C. It estimates a relationship using multiple observations.

Question 7

A benefit forgone because warehouse space is used for one product instead of the next-best product is called:

A. Sunk cost

B. Opportunity cost

C. Period cost

D. Allocated cost

Correct Answer: B. Opportunity cost

Question 8

If a learning rate is 80%, what does the rate convention normally indicate?

A. The first unit requires 80% of one hour.

B. Unit time becomes 80% when cumulative output doubles.

C. Total cost rises by 80% when output doubles.

D. Fixed cost falls by 80% at every activity level.

Correct Answer: B. Unit time becomes 80% when cumulative output doubles.

Question 9

Which item is most appropriately included in lifecycle costing?

A. Only the initial purchase price

B. Only direct manufacturing labour

C. Acquisition, operation, maintenance and disposal costs

D. Only costs already recorded in inventory

Correct Answer: C. Acquisition, operation, maintenance and disposal costs

Question 10

A fixed overhead allocation continues after a product is outsourced. For the outsourcing decision, that allocation is generally:

A. An avoidable cost

B. An unavoidable cost

C. A variable cost

D. An opportunity benefit

Correct Answer: B. An unavoidable cost