One of the most striking features of the modern global economy is the fact that Apple, Google, and other multinational enterprises (MNEs) avoid taxes by shifting profits to subsidiaries in tax havens like Ireland and Bermuda. International profit shifting costs governments in North America, Europe, and other high-tax countries hundreds of billions of dollars in tax revenues per year. Policymakers in these countries have recently begun to coordinate to design international tax reforms to make profit shifting more difficult. These reforms could reduce MNEs' incentives to invest and innovate, however, and this could materially affect the economy as a whole because these firms account for such a large portion of overall economic activity. There is urgent need to determine how profit shifting affects the macroeconomy and whether policies designed to address it would have unintended adverse side effects. The objective of this research project is to develop new theoretical models to analyze how profit shifting affects MNEs' investment decisions, and to develop new quantitative and computational methods to solve these models and measure the macroeconomic consequences of these effects.
The specific objectives of this project are:
- Develop a theory that links multinational firms' profit-shifting decisions to their decisions about investment and employment
- Integrate this theory into an equilibrium model of the overall economy and design a procedure to solve and calibrate this model so that it matches key data on profit shifting
- Use the calibrated model to quantify the positive macroeconomic consequences of profit shifting and policies designed to address this phenomenon
- Use the calibrated model to conduct a normative analysis of the optimal way to tax multinational corporations in the presence of profit shifting
- Use the theory and/or quantitative model and assess the incentives of low-tax countries to cooperate with policy changes designed to reduce profit shifting
- Develop a computational toolbox consisting of numerical methods and algorithms used in the project to tackle challenges associated with solving macroeconomic models with heterogeneity and analyzing optimal policy within this class of models
The key computational object in this project is a structural macroeconomic model designed to achieve project's objectives. The structure of the model can be summarized as follows. There are N regions in the world. They differ in population, aggregate productivity, trade costs, foreign direct investment costs, and corporate income taxes. Each region has a representative household, a government, and a continuum of firms. Firms are heterogenous in productivity, tangible capital (e.g., buildings and structures) and intangible capital (e.g., patents and software). Firms in each region decide the following: where to export and where to open foreign subsidiaries; how much labor to hire in the parent division and each subsidiary; how much to invest in tangible and intangible capital; and how to allocate tangible capital across subsidiaries. Intangible capital is nonrival and is used simultaneously in all of a firm's divisions, whereas tangible capital is rival and thus it has to be allocated across subsidiaries subject to a feasibility constraint. Multinational firms (firms with foreign affiliates) use transfer pricing to allocate the costs of producing intangible capital across their foreign affiliates in proportion to the scale at which these affiliates use this capital. Affiliates license the right to use intangible capital from the division that owns this capital, and MNEs can shift profits by selling their intangible capital to affiliates in lower-tax regions.