The reciprocal method of cost allocation distributes service department costs by writing one equation per service department, with each equation stating that a department’s total cost equals its own direct costs plus a share of every other service department’s total cost. Solving the equations simultaneously produces fully loaded totals that account for the fact that support departments serve each other, not just production. Those totals are then pushed out to production departments using the same usage percentages.
It is the only allocation method that captures every internal exchange between service departments, including a department’s consumption of its own services. That makes it the most accurate of the three standard methods, and the most work to run.
Data You Need Before You Start
Two categories of information drive the whole exercise: each service department’s direct costs, and the usage relationships among all departments.
Direct Departmental Costs
Pull the total costs directly incurred by each service department during the period: salaries, supplies, equipment depreciation, and similar expenses. These figures come from the general ledger and should tie back to supporting records such as payroll registers and purchase orders. They form the constants in the simultaneous equations. If they are wrong, every downstream number will be wrong too.
Usage Percentages and Cost Drivers
You need to measure how much each department consumes of every other department’s services, expressed as a percentage. The measurement tool, the cost driver, should reflect the actual mechanism of resource consumption. Common choices include machine hours or work orders for Maintenance, employee headcount for Human Resources, and device count, help-desk tickets, or server hours for Information Technology.
A poor cost driver choice can undermine the whole exercise. Allocate IT costs by square footage instead of device count and a large warehouse with two computers absorbs more IT overhead than a small office running fifty workstations. The driver has to bear a logical relationship to how costs are actually generated.
Once drivers are chosen, build a usage matrix showing every department’s consumption of every other department’s output, including what service departments consume from each other. That matrix is the backbone of the equations.
Setting Up the Simultaneous Equations
Each service department gets one equation. The equation says: this department’s total cost equals its own direct costs plus a share of every other service department’s total cost, based on usage percentages. “Total” here means the cost after accounting for all reciprocal exchanges, which is exactly what you are solving for.
With two service departments (call them S1 and S2), the system looks like this:
S1 = Direct costs of S1 + (S2’s percentage consumed by S1 × S2)
S2 = Direct costs of S2 + (S1’s percentage consumed by S2 × S1)
Each equation has two unknowns, so you need both equations to solve the system. Two approaches work.
Substitution
Rearrange one equation to express S1 in terms of S2, then plug that expression into the second equation. You now have a single equation with one unknown, which you solve with basic algebra. Substitute the answer back into the first equation to get the other value.
Elimination
Multiply the equations by constants chosen so that adding or subtracting them cancels one variable. Solve for the remaining variable, then back-substitute. This scales more cleanly to three or more service departments, though at that point most practitioners shift to matrix algebra and use spreadsheet functions like MINVERSE and MMULT to solve the system in one step.
The solved values will always exceed the original direct costs, because they now include the value of services received from other support units. Those inflated totals are the correct numbers to use for the final distribution.
A Two-Department Worked Example
Suppose a company has two service departments, Maintenance and Administration, along with two production departments, Department 1 and Department 2.
- Maintenance has $8,000 in direct costs and allocates based on machine hours. Administration uses 2,000 of the total 6,000 hours (33.3%), Department 1 uses 1,500 hours (25%), and Department 2 uses 2,500 hours (41.7%).
- Administration has $4,000 in direct costs and allocates based on employee count. Maintenance has 100 of the total 500 employees (20%), Department 1 has 250 (50%), and Department 2 has 150 (30%).
Let M = total Maintenance cost and A = total Administration cost:
M = $8,000 + 0.20A
A = $4,000 + 0.333M
Substitute the first equation into the second:
A = $4,000 + 0.333 × ($8,000 + 0.20A)
A = $4,000 + $2,667 + 0.067A
0.933A = $6,667
A = $7,145 (rounded)
Back-substitute to find M:
M = $8,000 + 0.20 × $7,145 = $9,429
The original $8,000 Maintenance budget is now $9,429, and the original $4,000 Administration budget is now $7,145. Those inflated totals reflect the embedded cost of services each department received from the other.
Now distribute the totals to all departments, including between the two service departments themselves:
- Maintenance allocates $9,429: $3,143 to Administration (33.3%), $2,357 to Department 1 (25%), and $3,929 to Department 2 (41.7%).
- Administration allocates $7,145: $1,429 to Maintenance (20%), $3,573 to Department 1 (50%), and $2,144 to Department 2 (30%).
The service departments net to zero after receiving allocations from each other and distributing their totals. Department 1 absorbs $5,930 in overhead ($2,357 + $3,573), and Department 2 absorbs $6,073 ($3,929 + $2,144). Combined with their own direct costs, you now have fully loaded production costs that reflect the true support infrastructure behind each unit.
The Iterative Alternative
Not everyone wants to solve simultaneous equations. The iterative method, sometimes called repeated distribution, reaches the same answer through a series of back-and-forth allocations that converge on the correct totals.
The process works like this. Allocate the first service department’s costs to all other departments using the usage percentages. Then allocate the second service department’s now-updated costs back to all departments, including the first. The first department now has a new balance from the costs it just received, so you allocate that balance out again. Each round produces smaller amounts bouncing between departments. After several iterations the remaining balances become immaterial and you stop.
The iterative and algebraic methods are two ways of solving the same system, so they produce the same result. The iterative approach is more intuitive for people uncomfortable with algebra, but it takes more steps and involves rounding decisions about when to stop. For departments with heavy reciprocal usage, convergence can take many rounds. The algebraic method gets there in one pass.
Pushing the Totals to Production
Once you have the reciprocal totals, apply each production department’s usage percentage to get the final overhead charge. The key distinction from a naive allocation: production departments receive their share of the reciprocal total, not the raw departmental budget. In the example above, Department 1 gets 25% of $9,429 from Maintenance rather than 25% of $8,000. That $357 difference in a single allocation is the embedded cost of Administration’s support of Maintenance that would otherwise vanish.
These allocated figures flow into departmental cost reports, job-order costing systems, or activity-based costing pools depending on the accounting framework. They are the foundation for product pricing, make-or-buy decisions, and profitability analysis by production line. If overhead allocation is wrong, every per-unit cost built on top of it is wrong, and the distortion compounds as it moves through financial statements.
How It Differs From the Direct and Step-Down Methods
Three methods exist for allocating service department costs. The trade-off is always accuracy versus complexity.
The direct method sends each service department’s costs straight to production departments and ignores any services exchanged between support units. If Maintenance spends 15% of its time helping IT, that interaction vanishes from the numbers. Easy to run, easy to explain, and consistently inaccurate about what production departments actually consume.
The step-down method is a partial improvement. Service departments are ranked in a sequence, and each one allocates costs to departments below it on the list. Once a department’s costs have been distributed, it never receives costs back from departments allocated later. This captures some interdepartmental activity but still misses the feedback loop. If IT serves Maintenance and Maintenance serves IT, only one direction gets recorded.
The reciprocal method captures every one of those exchanges simultaneously. It is the only method that also accounts for self-consumption, where a department uses its own services.
When the Reciprocal Method Is Worth the Effort
The reciprocal method is not always the right call. For a small manufacturer with one service department there is nothing to reciprocate, and the direct method handles the allocation perfectly. Even with two service departments, if one barely uses the other’s services, the step-down method may get close enough that the extra accuracy does not justify the setup cost.
The method earns its keep when service departments are deeply intertwined, when product costing precision drives competitive pricing, and in two specific regulatory contexts described below.
Section 263A Inventory Costing
Manufacturers and resellers must capitalize certain indirect costs into inventory under Section 263A of the Internal Revenue Code rather than deducting them immediately.1eCFR. 26 CFR 1.263A-1 – Uniform Capitalization of Costs For allocating mixed service costs (expenses from departments that support both production and non-production activities), the IRS permits a direct reallocation method, a step-allocation method, or any other reasonable allocation method.2Internal Revenue Service. Section 263A Costs for Self-Constructed Assets The reciprocal method qualifies under that “any other reasonable” category and may produce more defensible results during an audit because it captures costs that simpler methods miss.
An election called the Simplified Service Cost Method is available as an alternative for inventory, property held for sale, and self-constructed tangible personal property produced on a routine and repetitive basis, meaning mass-produced items with a recovery period of three years or less.3Federal Register. Guidance Regarding the Simplified Service Cost Method and the Simplified Production Method If your products do not fit those criteria, you are back to choosing among the detailed allocation methods.
Government Contracting
The Federal Acquisition Regulation requires indirect costs to be accumulated in homogeneous cost pools and allocated to cost objectives in reasonable proportion to the benefits received. Contractors cannot double-count by including the same cost as both a direct charge and an indirect allocation, and once an allocation base is accepted, the contractor cannot fragment it by removing individual elements.4eCFR. 48 CFR 31.203 – Indirect Costs The reciprocal method’s treatment of interdepartmental services fits these requirements because it produces cost pools that reflect actual resource flows.
Cost Accounting Standard 418 adds further structure. Contractors must use allocation bases that reflect a beneficial or causal relationship to cost objectives, and acceptable bases include resource consumption measures, output measures, or surrogates representative of resources consumed.5eCFR. 48 CFR 9904.418-40 – Fundamental Requirements Machine hours for Maintenance or headcount for HR both qualify. If business conditions change significantly, the FAR requires the allocation method to be revised, so the system needs periodic review even after initial implementation.4eCFR. 48 CFR 31.203 – Indirect Costs
The reciprocal method is harder to run, harder to explain to non-accountants, and harder to audit. It is also the only method that tells you what your support infrastructure actually costs.