In the classical Transfer Pricing (TP) problem, we usually think of two divisions in a decentralized organization. Division 1 produces an intermediate good and Division 2 transforms it into a final good and sells it in the market. Division 1 ``sells'' a quantity of the intermediate good to Division 2 at a certain price. This price (the transfer price) is used to place a value on the transaction between the two divisions. The total transaction value is considered as income by the selling division and as expense by the buying division. This allows the net profits of the two divisions to be determined. The division managers are evaluated by the profits that their divisions generate, but the organization's objective is to maximize its total profit. The objectives of the organization and those of the division managers are often incompatible. There fore, the problem of the organization is to determine a pricing rule that serves its own goals, taking into account the objectives of the divisions. This is easy in a world of perfect information, where central management can calculate the optimal transfer prices. The problem is more difficult when there is asymmetry of information (i.e., some information is private). Division managers may wish to conceal some information, in order to manipulate the outcome in their own favor.

In practice the problem is still more complicated. In addition to the asymmetry of information, divisions may have multiple products or may face capacity constraints, the product may have to be manufactured by a chain of more than two divisions, some of the intermediate goods may also be sold in the market, etc. Let us briefly discuss the transfer pricing methods in common use today. We list here only some of the more common methods that appear in the accountancy literature. Cost methods: The transfer price is a certain function of the production cost of the selling division. It may or may not include a fixed cost component. There are several variations of this approach such as cost plus fixed fee, cost plus a fixed percentage of the cost, full cost plus markup, variable cost, marginal cost.

Market price methods: If there is a market for the intermediate good, then the market price is used as the transfer price. Often the transfer price is the market price minus the selling expenses.

Dual price methods: The price that the selling division receives is not equal to the price that the buying division pays - and is usually higher. This mechanism generates a deficit, which is set off by central management. Because central management sets the optimal transfer price, and hence sets the transaction volume to be optimal, this mechanism yields higher performance than other methods. However, it is not commonly used because it requires central management to be involved in the complex process of price-setting. The Ronen-McKinney mechanism and the Groves-Loeb mechanism are the motivation behind this method.

Negotiated transfer prices: The transfer price is reached by negotiation between the relevant division managers. The advantage of this method is that it preserves the divisions' autonomy. Its problem is the sensitivity of the outcome to the managers' negotiation skills.

Each method in this list can be applied in various ways. For example, the markup in the cost-plus method can be determined as a percentage of the cost, which equals a certain return on investment of either the division or the organization. The transfer pricing mechanism that an organization applies may have a critical impact on the organization's performance. Although the transfer pricing problem has been studied for many years it is still consider ed an open problem.

Keywords: Transfer Pricing, Mechanism Design, Decentralization, Capacity.

Classification: C69, C79, D81, D82, L22, M11.