[pic]

Marshall School of Business

University of Southern California

A207-03

Global Investors, Inc.

I have a basic “gut” discomfort with the proposition that investment management as a profit-making function exists only in New York.

-- Alistair Hoskins, chairman/CEO, Global Investors, London

Bob Mascola, CFO of Global Investors, Inc. (GI), took a last look at his notes as he walked into the conference room where he and the other members of the transfer pricing task force would meet with Gary Spencer, GI’s CEO. The transfer pricing task force supervised by Mascola was meeting with Spencer to discuss the latest transfer pricing models that the task force members had identified. Mascola hoped the meeting would result in a final decision about the transfer pricing method that should be used to recognize profits in GI’s subsidiaries.

Mascola knew that the meeting would be difficult. On repeated occasions, two of the members of the transfer pricing task force, Alistair Hoskins and Jack Davis, had engaged in heated debates about which transfer pricing model should be selected. Hoskins, chairman/CEO of GI’s London office, believed that regional offices—or at least the regional office he led—should be treated as largely autonomous profit centers so that the value created by these offices would be reflected in their financial statements. However, Davis, GI’s corporate vice president of operations, argued that virtually all of the investment strategies used to manage the clients’ funds were designed by the research team located in New York. Consequently, Davis believed that the revenues generated by investment activities should be recognized in New York, even if a few investment services were offered by a regional office. The essence of Hoskins’ reply to Davis was that expressed in the epigraph.

The Company

Global Investors, founded in 1965, was a privately-owned investment management company headquartered in New York. A number of directors and executives based in New York, Spencer among them, held majority ownership of GI’s outstanding stock. GI, started as a domestic equity investment manager, had grown to manage US$160 billion for a variety of clients, including corporations, insurance companies, public and private pension funds, endowments, foundations, and high-net-worth individuals.

GI focused on two activities: investment management (which included research, portfolio management, and trading) and client services (which included marketing and investor advisory services provided to institutional investors and independent brokers/dealers). Although the company initially focused on direct sales to institutional investors (such as endowments and pension funds), it was increasingly selling its investment products through independent brokers/dealers who both served wealthy individual investors and invested in GI funds on their behalf (see Exhibit 1). GI generated its investment management revenues by charging a percentage fee for the amount each of its clients invested in GI funds.

GI’s investment philosophy differentiated the firm from most of its competitors. Since its inception, GI based all of its investment strategies on financial market theories emerging from academic research. The company developed a prominent New York-based research team comprised mostly of PhD-qualified investment experts, who were supported by contracted-for advice from some of the world’s most highly regarded academic financial economists.

The company’s investment philosophy revolved around the theory that markets are affected by judgmental biases of the market participants. That is, that under certain circumstances and for certain time-periods, investors, and hence markets, over-react or under-react to information that is publicly available, regarding companies’ expected risks and returns. Instead of focusing on the valuation of individual securities and actively selecting securities based on their estimated value (as most of its competitors did), GI developed its different funds by focusing more directly on the types of securities that academic research had shown to be under-valued by the market.

As part of its strategy, GI was also committed to lowering its trading costs through economies of scale, technological investments aimed at increasing liquidity, and crossing activities (that is, matching clients’ buy and sell requests).

GI focused mostly on equity investments in countries committed to free markets and with reasonably well-functioning capital markets, but it also invested in fixed income and commodity securities. Over the years, GI had expanded its activities throughout Asia, Europe, and the Americas (as shown in Exhibit 2).

GI’s Subsidiaries

Most of GI’s 415 employees were located in New York, but GI also offered its services through four remote subsidiaries (see Exhibit 3). The largest subsidiaries located in Tokyo and London, employed 52 and 40 employees, respectively. The other two subsidiaries, located in Singapore and San Francisco, employed fewer than a dozen individuals each. About 80% of the personnel employed in these offices were dedicated to trading financial assets following guidelines established by headquarters, 15% were dedicated to selling GI’s funds, and the rest were involved in operations. Recently, three highly competent senior portfolio managers in Tokyo and four in London had started both to act as sub-advisors for the Japanese, Pacific Rim, and European portfolios and to manage a series of trusts for their regional clientele. Many of the sales personnel at those subsidiaries were using these local funds to attract new clients. However, all clients attracted by the subsidiaries, regardless of the location of the subsidiary or the fund they invested in, were assigned a contact person in New York in addition to their local representative. They also received timely information resulting from the internal and sponsored research generated in New York.

GI’s subsidiaries were separately incorporated companies. Their ownership composition resembled that of GI’s parent company, but additional shares were issued to the subsidiaries’ chairmen/CEOs. The chairman/CEO of GI-London (Hoskins) owned 23% of the London subsidiary; the chairman/CEO of GI-Tokyo (Paul Hashi) owned 5% of the Tokyo subsidiary; and the chairmen/CEOs of GI-Singapore and GI-San Francisco each owned 3% of their respective subsidiaries.

GI’s subsidiaries had historically been treated as cost-focused profit centers, while the administrative departments providing support were treated as cost centers. Expenses and revenues were recorded according to the following accounting model:

1. Expenses (see Exhibit 4, presenting the format of a consolidated statement):

- Any expenses that could be traced directly to the subsidiaries or the cost centers were recorded in the Direct Controllable Cost category. When expenses could not be directly traced, an allocation method was followed. (The last column of Exhibit 4 describes the allocation bases.) - Royalty expenses (paid to academics for developing trading strategies) were charged to the New York office, the center of the firm’s investment management activities. - Allocations from the cost centers were based on cost center manager estimates of the proportions of the cost centers’ services that were consumed by each of the center’s internal “clients” (other cost centers or subsidiaries). - After the proportions of each cost center’s expenses were established, GI utilized a reciprocal cost allocation method, using a system of simultaneous equations, to identify the total costs incurred by each cost center and the dollar amount that should be allocated from each cost center to its “internal clients,” both cost centers and subsidiaries.

2. Revenues

- GI-New York retained all of the revenues generated worldwide (from fixed fees charged to the clients based on the amount of money they invested in GI) and assigned GI’s subsidiaries a portion of the revenue, based on local costs (direct controllable costs and other costs allocated to the subsidiaries) plus a 10% mark-up over the direct controllable costs.

A report of GI subsidiaries’ profits for 2006 is presented in Exhibit 5. The cost-plus revenue-allocation method resulted in a small profit for all subsidiaries. This profit guaranteed that the subsidiaries would comply with capital requirements imposed by local financial authorities. The subsidiaries also used these profit reports both as a benchmark to calculate their taxes and as disclosures to institutional investor clients interested in learning about the financial health of the subsidiary holding their investments. On a few occasions where GI executives discussed the possibility of selling the firm or a subsidiary, the executives also used subsidiary profits to obtain a rough estimate of the worth of each subsidiary (calculated as a multiple of their EBITDA).

Subsidiary profits were not explicitly tied to the managers’ compensation. A bonus pool based on GI’s total profits was allocated to each executive based on the relative number of bonus points that each had earned. The compensation committee (comprising three members of the board of directors and the vice president of human resources) assigned these bonus points to each executive at the beginning of the year based on its subjective assessments of both the executive’s performance in the prior year and his or her contribution to the company.

Alternative Transfer Pricing Models

Some subsidiary CEOs expressed discomfort with the way the firm was calculating their units’ profits. During the last semester of 2006, Hoskins had been particularly vocal in pointing out to Spencer that treating GI’s subsidiaries as cost-focused profit centers was wrong. He argued that the resulting profits did not portray a fair picture of the subsidiaries’ performance, which could have an adverse effect on the subsidiaries’ sales prices, if they were ever spun off. The inaccurate profits also could be viewed negatively by financial and tax regulators in the countries where the subsidiaries were located.[1]

Gary Spencer was not convinced that the current structure created problems that were worth fixing. However, in December 2006, primarily to appease Hoskins, he asked Mascola to create a committee to evaluate the situation, to address both Hoskins’ and the tax concerns and, if appropriate, to propose an improved transfer pricing system. Mascola was selected to lead the evaluation process because of his financial expertise, his independence (he did not own any GI stock), and his personality (he was widely regarded as being thoughtful and impartial).

Right after his meeting with Spencer, Mascola began recruiting the people that he believed needed to participate in the process if a new transfer pricing model were to be both well designed and successfully implemented throughout the company. Every person invited accepted the invitation to become part of what became known as the Transfer Pricing Model Task Force. Mascola chaired the task force, which also included Jack Davis (operations vice president), Michael Freeman (research director), plus Hashi (GI-Tokyo) and Hoskins (GI-London).

The task force met periodically over a seven month period. During that time, they evaluated a number of different transfer pricing alternatives. In an early meeting, Hoskins took the initiative by proposing that GI revenues should be allocated to the subsidiaries using “assets under management” as the allocation base and that the subsidiaries pay a royalty of (around) 50% to New York as compensation for the R&D and trading strategies developed by headquarters. According to his model, the London office would receive 20% of total revenues, since it managed $32 of the $160 billion of assets under management at GI (see Exhibit 2). Thus, according to Hoskins, the London office would have been allocated the following revenues in 2006:

| |London Subsidiary Revenues ($ 000) |
|Allocated Revenues |20% * 619,949.1 = $ 123,989.8 |
|Minus Royalty Expense |50% * $123,989.8 = $ 61,994.9 |
|Net Revenues | $ 61,994.9 |

However, Davis and Freeman did not agree that London should record all of those revenues. They argued that the fact that London was managing those funds did not mean they were generating a significant proportion of the value associated with them. Davis argued that, in fact, most of the assets managed by the subsidiaries belonged to New York clients. Further, to manage those assets, the subsidiary employees were just following instructions from headquarters, since the investment management research group in New York was the unit in charge of developing GI’s trading strategies. Instead of allocating revenues based on assets under management, Davis believed a more accurate way of allocating revenues would be based on the origin of the clients (the source of the fixed fee revenues generated at each subsidiary), distributed as follows in 2006:

|Subsidiary |Asset Distribution Based on the Origin of the Clients |
|New York |$ 150.6 billion |
|London |$ 2.2 billion |
|Tokyo |$ 4.2 billion |
|Singapore |$ 1.0 billion |
|San Francisco |$ 2.0 billion |
|Total |160.0 billion |

Under this proposal, GI-London’s 2006 revenue would have declined from its actual $26.5 million to a mere $8.5 million and would have resulted in a subsidiary loss of over $16 million.

Totally dissatisfied with Davis’ counter-proposal, which he considered ridiculous, Hoskins decided to conduct some research to learn whether (and how) GI’s competitors were allocating revenues to their subsidiaries. Hoskins learned that the industry standard was to split fee revenues 50:50 between Client Services and Investment Management. Thus, he proposed, GI’s business units should be split into these two categories. Half of the revenue would be allocated to Client Services (including two business units: Institutional Investor sales, and Independent Broker/Dealer sales) and the other half would be allocated to the Investment Management unit. This would allow GI to treat both activities separately.

Hoskins went on to propose that 50% of the fee revenues be assigned to Client Services based on the revenues generated in each subsidiary (or equivalently, based on the origin of clients which were directly proportional to the revenues in each subsidiary), while the 50% assigned to Investment Management be allocated based on assets under management. He proposed still to pay a 50% royalty to New York. Under this proposed scheme, Hoskins calculated GI-London’s 2006 revenues as follows:

| |London Subsidiary Revenues ($ 000) |
|Client Revenues |1.375% * 50% * 619,949.1 = $ 4,262.1 |
|Investment Management Revenues |20% * 50% * 619,949.1 = $ 61,994.9 |
|Minus Royalty Expense |50% * $61,994.9 = $ 30,997.5 |
|Net Revenues | $ 35,259.6 |

Hoskins’ new proposal also met with the disapproval of most of the task force members. Although most agreed with the concept of the subsidiaries’ recording of revenues from Client Services, Davis reiterated that Investment Management should be considered a New York business unit only, since almost all of the investment strategies were developed at headquarters. Consequently, Davis and Freeman proposed that the fee revenues corresponding to Investment Management (50% of total revenues) be fully recognized by the headquarters office. The subsidiaries, on the other hand, would be reimbursed by headquarters for any expenses related to investment management activities in their units plus a 10% mark-up if these expenses qualified as direct controllable costs. Using the 2006 financial results, Davis and Freeman estimated that the operating incomes recorded for the different subsidiaries under their proposed model, would be those shown in Exhibit 6.

This model was unacceptable to Hoskins. He argued that the London and Tokyo subsidiaries were actively participating in investment management, and they should be rewarded for the value these activities created. He explained

Clearly there is activity under the broad banner of Investment Management in London and Tokyo. The issue is whether our offices add value or not. We are building resources in London on the basis that GI-London is at least responsible for the investment management function for locally-sourced clients. We have established an Investment Committee to oversee policies for our fixed income portfolios in the UK and continental Europe as well as for the Irish funds, and we have initiated the development of a local research function. I accept that the local activity is primarily, though not exclusively, one of policy tailoring and implementation rather than original intellectual capital investment, but most companies would regard this as a source of added value.

Hashi supported Hoskins by adding,

Local value-added is not the same for all products or for all clients. It is clear, for example, that GI-Tokyo adds little when it simply implements programs of trades suggested by GI-New York, but it is also clear that it adds a significant share of value when it is managing money for its own clients in products designed specifically for them using local inputs.

Hoskins also expressed a concern about the effect that not recording the investment management revenues at the subsidiary level would have on external parties. Hoskins believed that local tax authorities might disapprove such treatment, as the profitability from investment management operations would be constrained to 10% or less. He claimed that, in practice, it seemed acceptable that support services (such as those provided by the cost centers) would be transferred at cost (or at a slight markup), but functions that formed part of a group’s offering to clients (in this case, client services and investment management) were expected to be transferred in exchange for a proportion of revenues, following an “arms-length standard.”[2] Departures from arm’s length prices could be interpreted by the local authorities as an attempt to shift taxable income out of their countries. Hoskins explained:

Our main competitors in the U.K. allocate revenues to the location actually carrying out the fund management. The alternative of a cost-plus arrangement, such as we have historically maintained, is probably no longer tenable where we now have local clients from whom we are receiving revenues for local investment activities.

Another external party that Hoskins worried about was his own clients. Hoskins believed that key local clients would be hesitant to appoint GI-London to manage their assets if they knew their funds were considered to be managed in New York.

Davis disagreed with Hoskins’ and Hashi’s contentions. He believed that the contributions made to the local investments managed in London and Tokyo were minimal. Davis argued that the majority of operations at the subsidiaries consisted of selling the investment funds managed in the headquarters or executing a few investment operations, following strategies and guidelines developed by the investment management unit in New York.

Although Freeman agreed that the transfer pricing model should adhere to tax regulations, he believed that the model he and Davis proposed was appropriate. It should not trigger regulators’ concerns since it already allowed the subsidiaries to record revenues for the services provided to institutional investor and independent broker/dealer clients (which he considered the main value-added activity performed by the subsidiaries). Additionally, GI’s executives believed that the model used to prepare the subsidiaries’ financial statements was not all that crucial to other financial regulators since GI was required to report consolidated (rather than subsidiary) financial statements.

The Meeting

As Mascola prepared for the meeting with Spencer, he recognized the tensions among the task force members. He had carefully considered the advantages and disadvantages of the models proposed by Hoskins and by Davis and Freeman. He wondered how he could direct the meeting towards a final selection of a transfer pricing model that would both benefit the firm and be accepted by all, or at least most, members of the task force.

Exhibit 1
Global Investors, Inc.
Total Assets Under Management

[pic]

Exhibit 2
Global Investors, Inc.
Types of Funds
(December 2006)

[pic]

Assets Under Management by Subsidiary
(in billions of dollars, as of December 2006)

|Subsidiary Managing the |Equity |Fixed Income |Commodities |Total |
|Assets | | | | |
|New York |70.2 |36.3 |5.9 |112.4 |
|London |14.5 |14.5 |3.0 |32.0 |
|Tokyo |6.2 |4.0 |2.2 |12.4 |
|Singapore |2.1 |0.3 |0.0 |2.4 |
|San Francisco |0.0 |0.2 |0.6 |0.8 |
|Total |93.0 |55.3 |11.7 |160.0 |

Exhibit 3
Global Investors, Inc.
Organizational Structure
[pic]

[pic]
[pic]
[pic]
-----------------------
[1] Many countries’ tax authorities were concerned that multinational corporations use transfer prices to shift income out of their country to countries with lower income taxes. Consequently, laws in the United States as well as other countries constrained transfer pricing policies. For example, Section 482 of the U.S. Internal Revenue Code required that transfer prices between a company and its foreign subsidiaries equal the price (or an estimate of the price) that would be charged by an unrelated third party in a comparable transaction. Regulators recognized that transfer prices can be market-based or cost-plus based, where the plus should represent margins on comparable transactions. Tax rates varied significantly across countries: In GI’s case, Singapore had the lowest effective tax rate (approximately 20%), followed by the United Kingdom and Japan (25-30%). The United States’ subsidiaries paid the highest tax rates (around 40%).

[2] “Arms length” prices are those charged after bargaining between unrelated persons or those charged between related persons that approximate the result of independent bargaining.