Spoilage, rework and scrap

No production process is perfect. Castings crack as they cool, seals leak under pressure testing, and offcuts of steel bar fall on the floor of the machine shop. The process costing reports in lesson 2 assumed every unit that entered a department came out as a good unit. This lesson shows how cost systems handle units that do not: spoiled units that are discarded or sold for salvage, reworked units that are repaired and sold as good units, and scrap material of minimal value. The accounting matters for two reasons. It determines how much cost good units carry into inventory, and it determines whether the cost of poor quality is visible to managers or buried inside product costs where no one can see it.

Definitions

Spoilage is units of production, whether fully or partly completed, that do not meet the specifications required by customers and are discarded or sold for reduced prices. Rework is units that do not meet specifications but are repaired and sold as acceptable finished goods. Scrap is residual material left over from making a product, with low sales value compared with the product itself, such as metal shavings, offcuts and sprues from the foundry. The three categories call for different treatments, and within spoilage and rework the key distinction is between normal and abnormal amounts.

Normal spoilage is spoilage inherent in a particular production process that arises even when the process is operated efficiently. Some loss is unavoidable given the materials, equipment and tolerances chosen. Because it cannot be avoided in the short run, normal spoilage is treated as part of the cost of making good units. Abnormal spoilage is spoilage that should not arise under efficient operating conditions: a machine set up wrongly, a batch of defective alloy, an untrained operator. Because it is avoidable and controllable, abnormal spoilage is not a cost of good units. It is recognized as a loss in the period in which it occurs, so that managers see it.

Accounting standards reach the same conclusion. Under IAS 2 Inventories, and under ASPE Section 3031, abnormal amounts of wasted materials, labour or other production costs are excluded from the cost of inventory and recognized as expenses when incurred. Normal losses remain part of the cost of the good units produced.

Setting the normal rate and the inspection point

Normal spoilage is usually expressed as a percentage of good units passing inspection, not of total units, because the inspection point is where good and bad units are separated. Halden's assembly department pressure-tests every S-100 pump at the end of assembly. Engineering studies show that, with the current seal design and test rig, about 2 per cent of the good units passing the test will be accompanied by spoiled units that fail it. So if 19,000 good pumps pass in a month, normal spoilage is 2% × 19,000 = 380 units. Any spoilage beyond that is abnormal.

The inspection point matters for two reasons. First, spoiled units are counted in equivalent units only up to the inspection point, because no further work is done on them. Second, the cost of normal spoilage is assigned only to the good units that have passed the inspection point during the period. Units in ending work in process that have not yet reached inspection carry no share of normal spoilage, because they may yet turn out to be spoiled themselves.

Spoilage in a process costing report

November data for the assembly department are shown below. The department uses the weighted-average method; the same logic applies under FIFO, with equivalent units counting only current-period work.

# Assembly department, November: spoilage under weighted average (illustrative)
good, spoiled, end, end_pct = 19_000, 600, 1_400, 0.40
normal_rate = 0.02
costs = {"TI": 86_000 + 1_720_000, "components": 47_000 + 900_000,
         "conversion": 13_800 + 470_000}
eu = {"TI": good + spoiled + end,
      "components": good + spoiled + 0,          # ending WIP before the 50% point
      "conversion": good + spoiled + end * end_pct}
rate = {k: costs[k] / eu[k] for k in costs}
unit = sum(rate.values())
normal = round(normal_rate * good)
abnormal = spoiled - normal
for k in costs:
    print(f"{k:<11} cost {costs[k]:>10,}  EU {eu[k]:>7,.0f}  per EU {rate[k]:8.4f}")
print(f"Cost of a complete unit: {unit:.4f}")
good_cost = good * unit
normal_cost = normal * unit
abnormal_cost = abnormal * unit
ending = end * rate["TI"] + end * end_pct * rate["conversion"]
print(f"Good units {good_cost:,.0f} + normal spoilage {normal_cost:,.0f} "
      f"= transferred out {good_cost + normal_cost:,.0f} "
      f"({(good_cost + normal_cost) / good:.2f} per good pump)")
print(f"Abnormal spoilage loss {abnormal_cost:,.0f}")
print(f"Ending WIP {ending:,.0f}")
print(f"Check: {good_cost + normal_cost + abnormal_cost + ending:,.0f} "
      f"= {sum(costs.values()):,}")

TI          cost  1,806,000  EU  21,000  per EU  86.0000
components  cost    947,000  EU  19,600  per EU  48.3163
conversion  cost    483,800  EU  20,160  per EU  23.9980
Cost of a complete unit: 158.3143
Good units 3,007,973 + normal spoilage 60,159 = transferred out 3,068,132 (161.48 per good pump)
Abnormal spoilage loss 34,829
Ending WIP 133,839
Check: 3,236,800 = 3,236,800

Each complete unit costs about $158.31: $86.00 of casting cost, $48.32 of components and $24.00 of assembly conversion. The 19,000 good pumps cost about $3,007,973 on their own; adding the $60,159 cost of the 380 normally spoiled units gives about $3,068,132 transferred to finished goods, or about $161.48 per good pump. The extra $3.17 per pump is the cost of the normal spoilage that the process inevitably generates. The 220 abnormally spoiled pumps, costing about $34,829, are charged to a Loss from Abnormal Spoilage account and reported to the plant manager. Ending work in process of about $133,839 carries no spoilage cost because it has not yet been inspected.

If the spoiled units have a salvage value, for example because the motors can be recovered, the recoverable amount is recorded as inventory (or as a receivable if the units are sold immediately) and deducted from the cost of spoilage before the remainder is assigned. Suppose each spoiled pump yields a motor worth $40. Normal spoilage cost would then be 380 × ($158.31 − $40) ≈ $44,958, and the abnormal loss would be 220 × ($158.31 − $40) ≈ $26,029.

When inspection happens part-way through

Many plants inspect at an earlier point to avoid adding more work to units that are already bad. Suppose Halden moved the pressure test to the 50 per cent point, just before the components are added. Spoiled units would then count as 50 per cent complete for conversion and would receive no components. More importantly, ending work in process at 70 per cent complete would have passed inspection, so it would share in the cost of normal spoilage along with the units completed. The normal spoilage cost is spread over all good units that passed inspection this period, in proportion to their number.

Moving inspection earlier is often worthwhile even if it adds testing cost, because every spoiled unit caught early saves the components and conversion that would otherwise be wasted on it. In Halden's case each pump caught before components are added saves about $48 of components and about $12 of assembly conversion. With 600 spoiled pumps a month, that is roughly $36,000 a month, which could pay for a second test rig. Cost systems that report spoilage separately make this kind of analysis possible.

Spoilage in job costing

Job costing distinguishes three cases. When normal spoilage arises because of the specific requirements of a particular job, such as unusually tight tolerances requested by a customer, its cost (net of any disposal value) stays with that job. When normal spoilage is common to all jobs, because it is inherent in the process, it is budgeted into the predetermined overhead rate, and actual normal spoilage is charged to manufacturing overhead control rather than to the job on which it happened to occur; every job then bears a share through applied overhead. Abnormal spoilage is charged to a loss account in both cases.

Rework

Rework follows the same three-way logic. Halden's P-300 pumps occasionally leak at a seal joint during the pressure test; a technician disassembles the joint, replaces the seal and retests the pump, at an average rework cost of $45. Normal rework common to all P-300 production is charged to manufacturing overhead and spread through the overhead rate. Rework caused by a particular job's specifications is charged to that job. Abnormal rework, such as a whole batch reworked because the wrong seal compound was used, is charged to a loss account.

Rework raises a decision question as well as an accounting one: is it worth repairing a defective unit, or should it be scrapped? The relevant comparison is the additional cost of rework against the additional value of a good unit compared with a defective one. Costs already incurred on the unit are sunk and irrelevant to this choice, although they matter for evaluating the process.

# Rework or scrap? Relevant analysis for a leaking P-300 pump (illustrative)
rework_cost = 45          # seal, technician time and retest
good_value = 390          # net selling price of a good P-300 to a dealer
salvage_value = 110       # recovered motor and housing if scrapped
sunk_cost = 304           # cost already in the pump (irrelevant to the decision)

gain_from_rework = (good_value - rework_cost) - salvage_value
print(f"Rework: receive {good_value - rework_cost} net; scrap: receive {salvage_value}")
print(f"Advantage of reworking: ${gain_from_rework}")
print(f"(Sunk cost of ${sunk_cost} is the same either way and is ignored.)")

# Capacity matters: if rework uses a bottleneck test rig, include the
# contribution of the good pump that could have been tested instead.
lost_contribution = 120
print(f"Advantage if the test rig is a bottleneck: ${gain_from_rework - lost_contribution}")

Rework: receive 345 net; scrap: receive 110
Advantage of reworking: $235
(Sunk cost of $304 is the same either way and is ignored.)
Advantage if the test rig is a bottleneck: $115

Reworking a leaking P-300 is clearly better than scrapping it when capacity is plentiful: the company gains $235 compared with selling the parts for salvage. If the test rig is a bottleneck, however, each hour spent retesting a reworked pump is an hour not spent testing a new one, and the contribution of the displaced pump becomes an opportunity cost of rework. The analysis then narrows to $115 in favour of rework. ACCT-102 lesson 9 developed this kind of constrained-resource reasoning; it applies equally to quality decisions.

Scrap

Scrap has some value, but much less than the main product. The accounting questions are when to recognize its value and how to credit it. There are two broad approaches. Under the first, scrap is recognized when it is sold: the proceeds are credited to the specific job if scrap can be traced to it, to manufacturing overhead if it is common to all jobs, or simply to other income if amounts are immaterial. Under the second, scrap is recognized when it is produced: it is recorded as inventory at its net realizable value, with the same credit to the job or to overhead, and production costs are reduced immediately. The second approach is preferred when scrap is valuable or held for a long time, because it puts the scrap under inventory control.

Halden's foundry produces runners, risers and sprues each time it pours castings, along with machining chips from the machine shop. The foundry melts most of its own returns, which reduces purchases of alloy, while machining chips are sold to a metals recycler. For the S-100 line, chips are worth about $1.80 per pump at current metal prices. Because the amounts are material over a year (about $36,000 for 20,000 pumps) and the chips are stored in bins until a full truckload is ready, Halden records them as inventory at estimated net realizable value when they are produced and credits manufacturing overhead.

Making the costs of poor quality visible

The most important managerial point about spoilage accounting is visibility. If all spoilage were folded into the cost of good units, a plant could gradually lose control of quality while reported unit costs crept up for no apparent reason. Separating normal from abnormal spoilage, and reporting abnormal spoilage as a distinct loss, puts the problem in front of the managers who can fix it. Setting the normal rate is therefore a management decision with consequences: a generous normal rate hides waste, while a rate set at a world-class level that the plant cannot yet reach labels routine losses as abnormal and may demoralize supervisors.

Many companies now aim for the normal rate itself to fall over time, through better design, mistake-proofing and process control. Lesson 11 places spoilage and rework within a broader cost of quality report, which groups prevention, appraisal, internal failure and external failure costs and shows how spending on prevention can reduce total quality costs. Halden's pressure test is an appraisal cost; the spoiled and reworked pumps are internal failure costs; and pumps that fail in customers' buildings, generating warranty calls for Halden Service, are external failure costs, often the most expensive of all.