Relevant costing: make-or-buy, outsourcing and special orders

Many of the most important decisions a management accountant supports are choices between alternatives: make a component or buy it, accept an order or decline it, keep a product line or drop it. The reported cost figures from the accounting system are often the wrong figures for these decisions, because they include costs that will not change whichever alternative is chosen and omit costs that will. This lesson develops the idea of relevant costs carefully and applies it to two decisions at Northgate: whether to keep making gearbox housings or buy them from a foundry, and whether to accept a large export order at a discounted price. ACCT-222 introduced the basic distinction between relevant and irrelevant costs; here the emphasis is on the harder parts: opportunity costs, avoidable fixed costs, capacity, and the qualitative and strategic factors that numbers alone cannot settle.

What makes a cost relevant

A relevant cost is a future cost that differs between the alternatives being considered. Both conditions matter. Past costs, however large, are sunk: they have been incurred and cannot be changed by any decision now. Future costs that are the same under every alternative cannot help choose between them. What remains are incremental or differential costs and revenues, plus opportunity costs, the benefit given up by using a resource in one way rather than in its best alternative use.

Two simplifications are worth stating. First, relevant costing usually ignores the time value of money because the decisions are short term; when cash flows extend over several years, discounting is needed, as lessons 9 and 10 show. Second, relevance depends on the time horizon. Over a few months, most fixed costs are fixed. Over several years, almost every cost can be changed, and a decision that seems to save money in the short run may commit the organization to costs or capabilities it will regret later.

Make or buy

Industrial Drives makes about 24,000 cast and machined gearbox housings a year for its own drive units. A foundry in another province has offered to supply finished housings for 88 dollars each, delivered. The costing system reports a full cost of 104 dollars per housing, so at first sight buying looks 16 dollars cheaper. Grace's job is to test that first impression.

She starts by asking which of the 104 dollars would actually disappear if Northgate stopped making housings. Direct materials of 38 dollars and variable overhead of 14 dollars would stop. Direct labor of 22 dollars would also stop, because the machinists would move to other work that currently uses overtime and temporary staff. Of the 30 dollars of fixed overhead allocated to each housing, only 8 dollars is avoidable: a supervisor's position and a lease on a machining center. The rest is plant rent, depreciation and shared services that would continue regardless. Buying also brings costs of its own: freight from an out-of-state foundry beyond what is included in the price, and incoming inspection, together about 2.50 dollars a housing.

The second question is what the freed space and machines could be used for. The plant manager has a proposal: if housings were bought in, the machining area could be converted to assemble an additional line of drive units that Industrial Drives currently cannot supply because of capacity limits. The sales team estimates this would add contribution of 420,000 dollars a year.

# Make or buy: gearbox housings for Industrial Drives (invented data, dollars).
units = 24000
make = {                        # cost per housing as reported by the costing system
    'Direct materials':        38.00,
    'Direct wages':            22.00,
    'Variable overhead':       14.00,
    'Fixed overhead allocated': 30.00,
}
avoidable_fixed = 8.00 * units   # supervisor and machine lease end if Northgate stops making
buy_price = 88.00
freight_inspection = 2.50        # extra cost per bought housing
freed_capacity_contribution = 420000  # contribution from extra drive units if space is freed

full_cost = sum(make.values())
relevant_make = (make['Direct materials'] + make['Direct wages'] + make['Variable overhead']) * units \
                + avoidable_fixed
relevant_buy = (buy_price + freight_inspection) * units
print(f"Reported full cost per housing: {full_cost:7.2f}  (buy price {buy_price:.2f})")
print(f"Relevant cost to make:  {relevant_make:12,.0f}")
print(f"Relevant cost to buy:   {relevant_buy:12,.0f}")
print(f"Advantage of making, capacity idle:      {relevant_buy - relevant_make:12,.0f}")
print(f"Advantage of making, capacity has a use: {relevant_buy - relevant_make - freed_capacity_contribution:12,.0f}")
# the buy price at which Northgate is indifferent, with and without the alternative use
for alt in (0, freed_capacity_contribution):
    p = (relevant_make + alt) / units - freight_inspection
    print(f"Indifference buy price (alternative use {alt:,}): {p:6.2f}")

Reported full cost per housing:  104.00  (buy price 88.00)
Relevant cost to make:     1,968,000
Relevant cost to buy:      2,172,000
Advantage of making, capacity idle:           204,000
Advantage of making, capacity has a use:     -216,000
Indifference buy price (alternative use 0):  79.50
Indifference buy price (alternative use 420,000):  97.00

The answer depends entirely on whether the capacity has another use. If the machining area would simply sit idle, making housings is 204,000 dollars a year cheaper than buying, despite the reported full cost: the relevant cost of making is 82 dollars a housing, below the 90.50 dollars of buying. If the space can be used for extra drive units, buying becomes 216,000 dollars cheaper, because the 420,000 dollars of contribution from the new line is an opportunity cost of continuing to make housings. The indifference prices express the same result differently: Northgate should buy only if the foundry's price is below 79.50 dollars when capacity is idle, but it can afford to pay up to 97.00 dollars when the capacity has a profitable alternative use.

Look at what did not matter. The reported full cost of 104 dollars played no part in the right answer. Comparing it with the 88-dollar price would have reached the correct conclusion in the second case only by accident, and the wrong conclusion in the first. Allocated fixed overhead that does not change cannot be saved by buying, and the extra contribution from the alternative use does not appear anywhere in the costing system.

Qualitative and strategic factors

Make-or-buy decisions are rarely settled by the numbers alone. The relevant cost analysis tells management what the decision is worth in the short run; strategic considerations may outweigh it. For the housings, the team lists several.

The team recommends buying, conditional on the drive unit expansion going ahead, with a dual-sourcing clause allowing Northgate to qualify a second foundry within a year. Outsourcing an activity that is not a core competence, in order to redeploy scarce capacity to one that is, fits the logic of the resource-based view from BUS-450. The analysis also shows the trap of outsourcing purely on the basis of full cost: if the capacity freed up had no use, outsourcing would have raised costs by about 200,000 dollars a year while looking like a saving.

Special orders

A special order is a one-time order, usually at a price below the normal selling price, that the organization can accept or reject without affecting its regular sales or prices. The classic analysis compares the incremental revenue from the order with its incremental costs, and asks whether the order would displace regular business.

In autumn 2026, an agricultural distributor in Australia offers to buy 600 Auger A10 units for delivery during Northgate's quiet winter months at 3,150 dollars each, against a normal dealer price of 4,200. The variable cost of an auger is 2,610 dollars. The order would require export packaging at 40 dollars a unit and a one-time engineering change to fit a different motor mount, costing 55,000 dollars. Whether the order makes sense depends on how much spare capacity Northgate has in the delivery window.

# Special order: an export distributor offers to buy 600 auger units at 3,150 each (invented data).
normal_price, variable_cost = 4200, 2610
order_qty, order_price = 600, 3150
extra_per_unit, one_time = 40, 55000      # export packaging; engineering for a different motor mount

def special_order(spare_capacity):
    displaced = max(0, order_qty - spare_capacity)       # regular sales lost if capacity is short
    gain = order_qty * (order_price - variable_cost - extra_per_unit) - one_time
    lost = displaced * (normal_price - variable_cost)
    return gain, lost, gain - lost

for spare in (900, 400, 0):
    gain, lost, net = special_order(spare)
    print(f"Spare capacity {spare:4d}: order contribution {gain:10,.0f}  "
          f"lost regular contribution {lost:10,.0f}  net {net:+11,.0f}")
# minimum acceptable price when capacity is fully used
min_full = variable_cost + extra_per_unit + one_time / order_qty + (normal_price - variable_cost)
min_idle = variable_cost + extra_per_unit + one_time / order_qty
print(f"Minimum price with spare capacity: {min_idle:8.2f}")
print(f"Minimum price with no spare capacity: {min_full:8.2f}")

Spare capacity  900: order contribution    245,000  lost regular contribution          0  net    +245,000
Spare capacity  400: order contribution    245,000  lost regular contribution    318,000  net     -73,000
Spare capacity    0: order contribution    245,000  lost regular contribution    954,000  net    -709,000
Minimum price with spare capacity:  2741.67
Minimum price with no spare capacity:  4331.67

With 900 units of spare capacity in the winter, the order adds 245,000 dollars of contribution and should be accepted on financial grounds. With only 400 units of spare capacity, 200 regular sales would be displaced, each losing 1,590 dollars of contribution, and the order would reduce profit by 73,000 dollars. With no spare capacity, it would cost 709,000 dollars. The minimum acceptable prices summarize the logic: about 2,742 dollars when capacity is idle, which covers incremental costs, and about 4,332 dollars when every unit displaces a regular sale, which also covers the lost contribution.

The capacity question is subtle in a seasonal business. Northgate's plant is busy from February to June and again in late summer, but has slack in November and December. An order delivered in December uses spare capacity; the same order delivered in April would displace dealer sales at the busiest time of year. The analysis must use capacity in the relevant period, not average annual capacity.

Risks of special-order pricing

Even when the numbers favor a special order, several risks need weighing. If regular customers learn that the same product was sold at a 25 percent discount, they may demand similar terms, and the long-run cost of eroded prices can far exceed the order's contribution. Northgate's Australian order carries less of this risk than a discount to a domestic dealer, because the markets are separate and the product is modified. Pricing below full cost repeatedly can also become a habit: if every quiet period is filled with marginal-cost orders, the business may never cover its fixed costs.

Competition law sets limits too. In the United States, the Robinson-Patman Act restricts price discrimination between competing purchasers of commodities of like grade and quality where the effect may be to lessen competition, although sales for export are generally outside its scope. Lesson 8 discusses pricing and competition law further. For exports, anti-dumping rules in the importing country may apply where goods are sold abroad below their normal value at home, a point the team checks before accepting the Australian order.

Behavioral issues in relevant costing

Relevant cost analysis can conflict with how performance is measured. A division manager evaluated on full-cost product margins may reject a special order that is profitable for the company because it reports a loss on a full-cost basis. A plant manager whose budget includes the machining area may resist outsourcing housings even when it frees capacity for a more valuable use, because the plant's reported costs per unit would rise as fixed costs are spread over fewer products. ACCT-321 discussed how goal congruence depends on measures that reflect controllable, decision-relevant costs; here the practical lesson is that the accountant should present relevant cost analysis alongside a reconciliation to the reported figures, so that managers can see why the two differ.

There is also a human tendency to give sunk costs weight they do not deserve. Managers who championed a past investment are often reluctant to abandon it, even when the future costs exceed the future benefits, because doing so feels like admitting a loss. This sunk cost bias is well documented in behavioral research. A useful habit is to begin every decision analysis by listing the costs already incurred and setting them aside explicitly.