Payroll-Weighted Combined Profit Split with Entity-Specific Beta Coefficients

Purpose

This note derives the combined profit split method when each member of a controlled group has its own profit-per-payroll coefficient. The purpose is to show that variable entity-level coefficients do not defeat a simple payroll-based split. They collapse into one combined, payroll-weighted coefficient once operating profits and payroll are combined at the group level.

The economic antecedent is Sidney Weintraub’s wage-cost-markup theory of the price level, in which price is treated as a markup over unit labor cost. That antecedent supports the general idea that payroll or labor cost can be an economically meaningful cost anchor. The transfer-pricing derivation below is not found in Weintraub’s publications. It is an extension from macro price theory to a controlled-group profit allocation problem.

1. Notation and definitions

Let there be n controlled entities, indexed by j = 1, …, n. For each entity:

\begin{aligned}
P(j) &= \text{operating profit of entity } j \\
W(j) &= \text{payroll of entity } j, \text{ with } W(j) > 0 \\
\beta(j) &= \text{entity-specific profit-per-payroll coefficient}
\end{aligned}

The starting entity-level relation is:

(1)\qquad P(j) = \beta(j) W(j), \quad j = 1, \dots, n

Equivalently, for each entity with positive payroll:

(2)\qquad \beta(j) = \frac{P(j)}{W(j)}

Define combined group operating profit and combined group payroll as:

(3)\qquad P = \sum_{j=1}^{n} P(j)

(4)\qquad W = \sum_{j=1}^{n} W(j)

Define entity j’s payroll share as:

(5)\qquad \omega(j) = \frac{W(j)}{W}

Since W is the sum of all payroll amounts, the payroll shares add to one:

(6)\qquad \sum_{j=1}^{n} \omega(j) = \sum_{j=1}^{n} \frac{W(j)}{W} = \frac{1}{W} \sum_{j=1}^{n} W(j) = \frac{W}{W} = 1

2. The summation with varying beta coefficients

Start with the combined operating profit identity:

(7)\qquad P = \sum_{j=1}^{n} P(j)

Substitute equation (1), P(j) = β(j) W(j), into equation (7):

(8)\qquad P = \sum_{j=1}^{n} \beta(j) W(j)

Expanded term by term, equation (8) is:

(9)\qquad P = \beta(1) W(1) + \beta(2) W(2) + \dots + \beta(n) W(n)

The expression is not β(1) + β(2) + … + β(n). The summation operator applies to the product β(j) W(j). The coefficient β(j) is multiplied by that entity’s payroll W(j) before summation. Thus, the sum is a dollar-profit sum, not a dimensionless sum of coefficients.

Using the payroll-share definition W(j) = ω(j) W, substitute into equation (8)

(10)\qquad P = \sum_{j=1}^{n} \beta(j) (\omega(j) W)

Because W is common to all terms, factor W outside the summation:

(11)\qquad P = W \sum_{j=1}^{n} \omega(j) \beta(j)

Define the combined group payroll-weighted beta coefficient as:

(12)\qquad \bar{\beta}(W) = \sum_{j=1}^{n} \omega(j) \beta(j)

Therefore:

(13)\quad P = \bar{\beta}(W) W

Equation (12) is the essential aggregation result. The combined beta is not a simple arithmetic average and not the unweighted sum of the entity betas. It is the payroll-weighted average of the entity-specific profit-per-payroll coefficients.

3. Equivalent formula for the combined beta

From equations (8) and (4):

(14)\quad \bar{\beta}(W) = \frac{\sum_{j=1}^{n} \beta(j) W(j)}{\sum_{j=1}^{n} W(j)}

But since P(j) = β(j) W(j), the numerator is the combined profit P:

(15)\quad \sum_{j=1}^{n} \beta(j) W(j) = \sum_{j=1}^{n} P(j) = P

Therefore, the combined beta also equals:

(16)\quad \bar{\beta}(W) = \frac{P}{W}

Equation (16) is the compact result. The whole group has one observed combined profit-per-payroll coefficient equal to combined operating profit divided by combined payroll.

4. Constant beta as a special case

If every entity has the same coefficient β, then:

(17)\quad \beta(j) = \beta \quad \text{for all } j

Substitute into equation (12):

(18)\quad \bar{\beta}(W) = \sum_{j=1}^{n} \omega(j) \beta = \beta \sum_{j=1}^{n} \omega(j) = \beta

Thus, the constant-beta case is only a special case of the heterogeneous-beta case. The heterogeneous case is more general and still collapses to a simple combined coefficient.

5. The payroll-weighted profit split allocation

Once combined operating profit P and combined payroll W are determined, the proposed arm’s-length proxy allocation to entity j is:

(19)\quad \hat{P}(j) = \left( \frac{W(j)}{W} \right) P = \omega(j) P

Using β̄(W) = P/W, equation (19) can be rewritten as:

(20)\quad \hat{P}(j) = W(j) \left( \frac{P}{W} \right) = \bar{\beta}(W) W(j)

Therefore, every entity receives the same combined group profit-per-payroll coefficient β̄(W) applied to its own payroll. The entity-specific reported coefficients β(j) are not used as allocation coefficients after combination.

The allocation is additive. Summing equation (19) over all entities:

(21)\quad \sum_{j=1}^{n} \hat{P}(j) = \sum_{j=1}^{n} \omega(j) P = P \sum_{j=1}^{n} \omega(j) = P

Thus, the allocated profits exactly exhaust the combined operating profit.

6. Reported profit versus payroll-weighted allocated profit

Reported entity profit is:

(22)\quad P(j) = \beta(j) W(j)

Payroll-weighted allocated profit is:

(23)\quad \hat{P}(j) = \bar{\beta}(W) W(j)

The difference between reported profit and allocated profit is:

(24)\quad D(j) = P(j) - \hat{P}(j) = \beta(j) W(j) - \bar{\beta}(W) W(j) = (\beta(j) - \bar{\beta}(W))W(j)

Summing the differences across all entities:

(25)\quad \sum_{j=1}^{n} D(j) = \sum_{j=1}^{n} P(j) - \sum_{j=1}^{n} \hat{P}(j) = P - P = 0

Therefore, the method reallocates profit among entities without changing combined group operating profit. Entities with β(j) above the group coefficient β̄(W) surrender profit; entities with β(j) below β̄(W) receive profit. The total surrendered amount equals the total received amount.

7. Interpretation for profit shifting

The entity-specific coefficients β(j) may reflect real differences in functions, assets, risks, managerial quality, market position, accounting classification, or tax-motivated profit shifting. In a controlled group, reported entity profits can be distorted by transfer prices, royalties, intercompany services, financing, procurement arrangements, or legal ownership of intangibles. The method does not need to identify which distortion caused each β(j) to differ from β̄(W). It combines operating profit first and then applies one observable payroll-based coefficient to each entity.

This is the core neutralization step: the reported entity-level profit equation P(j) = β(j) W(j) is recognized, but its entity-specific coefficient is not allowed to control the final allocation. The combined group coefficient β̄(W) = P/W replaces all β(j) in the allocation equation.

8. Numerical example

Consider a controlled group with three entities:

The combined beta is:

(26)\quad \bar{\beta}(W) = \frac{P}{W} = \frac{210}{1{,}000} = 0.21

The reported coefficients are 0.10, 0.30, and 0.20. Their simple sum is 0.60, which has no direct allocation meaning. Their simple arithmetic average is 0.20, which ignores the payroll weights. The correct combined coefficient is the payroll-weighted average:

(27)\quad \bar{\beta}(W) = (0.10)(0.10) + (0.20)(0.30) + (0.70)(0.20) = 0.01 + 0.06 + 0.14 = 0.21

The difference between reported and allocated profits is:

The differences sum to zero:

(28)\quad -11 + 18 - 7 = 0

9. General matrix form

Let p, w, and β be n × 1 vectors:

(29)\quad \mathbf{p} = [P(1), \dots, P(n)]^T, \quad \mathbf{w} = [W(1), \dots, W(n)]^T, \quad \boldsymbol{\beta} = [\beta(1), \dots, \beta(n)]^T

Let 1 be an n × 1 vector of ones. Then:

(30)\quad P = \mathbf{1}' \mathbf{p}, \quad W = \mathbf{1}' \mathbf{w}

The entity-level relation can be written as:

(31)\quad \mathbf{p} = \text{diag}(\mathbf{w}) \boldsymbol{\beta}

Combined profit is:

(32)\quad P = \mathbf{1}' \text{diag}(\mathbf{w}) \boldsymbol{\beta} = \mathbf{w}' \boldsymbol{\beta}

The payroll-share vector is:

(33)\quad \boldsymbol{\omega} = \frac{\mathbf{w}}{\mathbf{1}' \mathbf{w}}

The combined beta is:

(34)\quad \bar{\beta}(W) = \boldsymbol{\omega}' \boldsymbol{\beta} = \frac{\mathbf{w}' \boldsymbol{\beta}}{\mathbf{1}' \mathbf{w}} = \frac{P}{W}

The allocated profit vector is:

(35)\quad \mathbf{\hat{p}} = \boldsymbol{\omega} P = \mathbf{w} \left( \frac{P}{W} \right) = \bar{\beta}(W) \mathbf{w}

The reallocation vector is:

(36)\quad \mathbf{d} = \mathbf{p} - \mathbf{\hat{p}}

And the entries of the reallocation vector sum to zero:

(37)\quad \mathbf{1}' \mathbf{d} = \mathbf{1}' \mathbf{p} - \mathbf{1}' \mathbf{\hat{p}} = P - {1}'({\omega}P) = P - ({1}'{\omega})P = P - P = 0

10. Why variable beta coefficients do not defeat the method

The concern is that β(1) + β(2) + … + β(n) does not produce a simple profit split. That concern is correct only if the β(j) coefficients are improperly summed without weights. The algebra never requires Σβ(j) by itself. The relevant object is Σβ(j) W(j), or equivalently Σω(j)β(j). This creates a single group coefficient β̄(W) = P/W. The allocation rule remains simple:

(38)\quad \hat{P}(j) = \left( \frac{W(j)}{\sum_{k=1}^{n} W(k)} \right) \sum_{k=1}^{n} P(k)

This is the payroll-weighted combined profit split. It is a contribution profit split in which payroll is the sole contribution key. The method is especially parsimonious because it requires only combined operating profit and entity payrolls. The heterogeneous β(j) coefficients are absorbed into the pooled coefficient and then displaced from the final allocation.

11. Originality and limits of the claim

The claim should be framed carefully. Weintraub provides an economic antecedent: a wage-cost-markup theory in which price is related to labor cost per unit of output and a markup. Transfer pricing profit split rules, including U.S. Treasury regulation § 1.482-6 and OECD guidance, recognize combined operating profit and a split based on relative contributions. They do not, in the sources reviewed for this draft, derive a payroll-weighted combined profit split from entity equations P(j) = β(j) W(j) with heterogeneous β(j) coefficients, nor do they reduce that system to P̂(j) = (W(j)/W)P through the group coefficient β̄(W) = P/W.

Therefore, based on a preliminary search of Weintraub/WCM materials and profit split guidance, the present derivation appears sui generis: inspired by Weintraub’s unit-labor-cost logic but not contained in Weintraub; consistent with the general profit-split idea of dividing combined operating profit by an economically

defensible contribution key but not found as a standard derivation in the reviewed transfer-pricing sources. The safer formulation is not “no one has ever thought of payroll as a factor,” because payroll, headcount, assets, and costs are sometimes mentioned as possible allocation factors. The stronger and more precise claim is that the algebraic derivation of a combined payroll-weighted beta and the resulting profit split appear to be your original extension.

Essential references with comments

26 C.F.R. § 1.482-6, “Profit Split Method.” Useful legal anchor because it defines the profit split method by reference to combined operating profit or loss and the relative value of each controlled taxpayer’s contribution. It does not supply the payroll-weighted beta derivation.

OECD. 2018. Revised Guidance on the Application of the Transactional Profit Split Method: BEPS Action 10. Paris: OECD. Useful international anchor because it frames the transactional profit split as identifying relevant profits and splitting them on an economically valid basis approximating arm’s-length division. It recognizes the need for profit splitting factors but does not derive the payroll-weighted coefficient system.

European Commission, EU Joint Transfer Pricing Forum. 2019. The Application of the Profit Split Method within the EU. Brussels: European Commission. Useful practical anchor because it discusses application difficulties and subjectivity in profit splitting. It confirms the practical need for administrable split factors.

Weintraub, Sidney. 1958. An Approach to the Theory of Income Distribution. Philadelphia: Chilton Company. Important antecedent for Weintraub’s macro distribution and wage-cost thinking. It is not a transfer-pricing profit split derivation.

Weintraub, Sidney. 1959. A General Theory of the Price Level, Output, Income Distribution, and Economic Growth. Philadelphia: Chilton Company. Core antecedent for Weintraub’s wage-cost-markup theory of the price level. It motivates labor cost as an economic anchor but does not derive a controlled-group payroll profit split.

Weintraub, Sidney. 1960. “The Keynesian Theory of Inflation: The Two Faces of Janus?” International Economic Review 1 (2): 143-155. Relevant to Weintraub’s inflation and price-level theory. It remains macro price theory, not transfer-pricing allocation algebra.

Wallich, Henry C., and Sidney Weintraub. 1971. “A Tax-Based Incomes Policy.” Journal of Economic Issues 5 (2): 1-19. Relevant policy application of wage-cost-markup reasoning. Useful to distinguish payroll-based anti-inflation policy from payroll-based profit allocation.

Seidman, Laurence S. 1978. “Tax-Based Incomes Policies.” Brookings Papers on Economic Activity 1978 (2): 301-348. Useful secondary exposition because it defines markup over standard unit labor cost and links price inflation to wage growth less productivity growth. It supports the Weintraub antecedent but not the transfer-pricing extension.

Compact statement for later use

For a controlled group with entity operating profits P(j) and payroll W(j), suppose P(j) = β(j) W(j). Combining profits gives P = Σβ(j)W(j), not Σβ(j). Defining total payroll W = ΣW(j) and payroll shares ω(j) = W(j)/W, the combined group coefficient is β̄(W) = Σω(j)β(j) = P/W. The payroll-weighted arm’s-length proxy allocation is P̂(j) = ω(j)P = (W(j)/W)P = β̄(W) W(j). Variable entity coefficients are therefore not an obstacle; they are collapsed into the group coefficient and displaced by the common combined coefficient in the final split.