Transfer Pricing

When a multi-division firm sells an intermediate good from one division to another, the internal price at which that transaction occurs—the transfer price—determines how each division perceives its costs and revenues. Set the transfer price wrong, and the downstream division distorts its output decision, destroying firm-wide profit. This page develops the classic theory of transfer pricing, proves that the optimal transfer price equals upstream marginal cost, and explores the conditions under which that elegant rule breaks down.

The Transfer Pricing Problem

Large firms routinely organize into semi-autonomous divisions. A chemical company’s upstream division produces a bulk intermediate, which the downstream division converts into a specialty product sold to final customers. A semiconductor manufacturer fabricates wafers in one division and packages finished chips in another. In each case, headquarters must decide the price at which the upstream division “sells” to the downstream division.

This problem was first formalized by Hirshleifer (1956), who showed that if both divisions operate in competitive external markets, the transfer price should simply equal the market price. The more interesting case—and the one most relevant to pricing theory—arises when no external market exists for the intermediate good. Here the firm must set the transfer price administratively, and the choice has direct consequences for firm-wide profit.

The central tension is between divisional autonomy and firm-wide coordination. If headquarters can directly dictate quantities, the transfer price is irrelevant—it becomes a mere accounting entry. But when divisions are genuinely decentralized, each division manager optimizes against the transfer price as if it were a real cost or revenue. A transfer price above marginal cost causes the downstream division to restrict output, just as an inflated wholesale price causes a retailer to mark up excessively in the supply chain pricing problem.

Hirshleifer’s Result

Hirshleifer (1956) proved a remarkably clean result: when the upstream division has constant marginal cost $c_u$ and no external market exists for the intermediate good, the transfer price that maximizes total firm profit is:

The intuition is straightforward. When $t = c_u$, the downstream division’s profit calculation aligns perfectly with the firm’s overall objective. Every unit the downstream division considers producing appears to cost it $c_d + c_u$, which is exactly what it truly costs the firm. The downstream division therefore makes the same output decision that a centralized planner would.

Note that under this rule, the upstream division earns zero profit: it sells at cost. All firm profit accrues to the downstream division. This creates obvious incentive problems—why would the upstream division manager exert effort if all profit goes downstream?—which we address in the departures section below.

Formal Model

Consider a firm with two divisions. The upstream division produces an intermediate good at constant marginal cost $c_u$ per unit. The downstream division converts this intermediate into a final product at an additional marginal cost of $c_d$ per unit and sells it to consumers. Inverse demand for the final product is linear:

where $a$ is the demand intercept and $b$ is the demand slope. We normalize $b = 1$ without loss of generality, so $P(Q) = a - Q$.

Integrated Benchmark

A centralized firm maximizes total profit:

The first-order condition $\partial \Pi^I / \partial Q = 0$ yields:

with integrated profit:

Decentralized Divisions

Under decentralization, headquarters sets a transfer price $t$. The downstream division treats $t$ as its cost of the intermediate good and maximizes:

The downstream division’s optimal output is:

The upstream division earns $(t - c_u) \cdot Q^D(t)$. Total firm profit under decentralization is:

Substituting $Q^D(t)$ and simplifying:

This is maximized when $Q^D(t) = Q^I$, which requires$t = c_u$. At any other transfer price, the downstream division produces too little (if $t > c_u$) or too much (if $t < c_u$, though this is rarely realistic since it would require the upstream division to sell below cost).

Interactive Transfer Price Optimizer

The chart below plots upstream, downstream, and total profit as functions of the transfer price. The vertical dashed line marks the Hirshleifer optimal at $t^* = c_u$, and the horizontal dashed line marks the integrated optimum profit. The red shaded region shows the deadweight loss from departing from marginal-cost transfer pricing. Adjust the cost and demand parameters to explore how the profit curves shift.

Double Marginalization Redux

The transfer pricing problem is intimately connected to the double marginalization problem in supply chains. In both cases, a sequence of independent markup decisions inflates the final price above the integrated optimum. The key structural difference is one of ownership: in the supply chain setting, the manufacturer and retailer are separate firms with conflicting profit objectives; in the transfer pricing setting, both divisions belong to the same firm, and headquarters has the authority to set the internal price.

This authority is precisely what makes the transfer pricing problem solvable in principle. By setting $t = c_u$, headquarters eliminates the upstream division’s markup entirely, preventing the double marginalization that would otherwise arise. In the supply chain analog, the manufacturer has no incentive to set its wholesale price at marginal cost—doing so would yield zero manufacturer profit. Achieving the same outcome requires more elaborate mechanisms such as revenue-sharing or two-part tariff contracts.

When the transfer price exceeds $c_u$, the downstream division behaves exactly like a retailer facing an inflated wholesale price. It restricts output, raises the final price, and total firm profit falls. The magnitude of the loss depends on how far $t$ deviates from $c_u$: under linear demand, the deadweight loss is quadratic in $(t - c_u)$.

Interactive Division Profits

The stacked bar chart below decomposes total profit into upstream and downstream components at each transfer price level. At the Hirshleifer optimal, the upstream division earns zero and all profit accrues downstream. As the transfer price rises, profit shifts from downstream to upstream—but the total shrinks because the downstream division curtails output.

This visualization highlights a core tension in transfer pricing practice: the transfer price that maximizes total firm profit (Hirshleifer’s $t^* = c_u$) gives the upstream division zero profit, undermining the upstream manager’s incentive to reduce costs or invest in quality. Real firms must balance allocative efficiency against managerial incentives, which is one reason actual transfer prices frequently depart from the textbook optimum.

Departures from the Rule

Hirshleifer’s marginal-cost rule is a first-best benchmark that assumes perfect information, no capacity constraints, and no external markets. In practice, several forces push transfer prices away from upstream marginal cost.

Tax-Motivated Transfer Pricing

Multinational firms face different tax rates across jurisdictions. By setting transfer prices strategically, a firm can shift reported profit from high-tax to low-tax divisions. If the upstream division is in a low-tax country, the firm benefits from setting a high transfer price, even though this distorts the downstream division’s output decision. The optimal tax-adjusted transfer price trades off the allocative efficiency loss against the tax savings. Tax authorities worldwide use “arm’s length” rules to constrain such profit shifting, requiring that inter-company transactions be priced as if between unrelated parties.

Capacity Constraints

When the upstream division faces a binding capacity constraint, the opportunity cost of the intermediate good exceeds its marginal production cost. The appropriate transfer price becomes the marginal cost plus the shadow price of the capacity constraint. Hirshleifer (1956) himself noted this extension: the transfer price should equal the marginal opportunity cost of the intermediate good, which includes any foregone external sales or binding resource constraints.

Asymmetric Information

When headquarters cannot observe divisional costs precisely, setting $t = c_u$ requires knowing $c_u$. Baldenius, Reichelstein, and Sahay (1999) analyze transfer pricing when divisional managers have private information about their costs. They show that negotiated transfer prices—where the two divisions bargain over $t$—can outperform both cost-based and market-based rules when information is asymmetric. The negotiated price typically falls between marginal cost and the market price, balancing information revelation against allocative efficiency.

Managerial Incentives

As the interactive charts above reveal, the Hirshleifer optimal leaves the upstream division with zero profit. In practice, firms often set transfer prices above marginal cost—using cost-plus formulas such as $t = c_u(1 + \mu)$ for a markup rate $\mu$—to give both divisions a share of the total profit. This deliberately accepts some allocative inefficiency in exchange for better divisional incentives. The firm-wide cost of this distortion is the deadweight loss visible in the red-shaded region of the optimizer chart.

References

• Baldenius, T., Reichelstein, S. & Sahay, S. A. (1999). “Negotiated versus Cost-Based Transfer Pricing.” Review of Accounting Studies, 4(2), 67–91.
• Hirshleifer, J. (1956). “On the Economics of Transfer Pricing.” Journal of Business, 29(3), 172–184.