跨国企业利润核算:境外附属企业收入的重复计算与错误归属

• A fundamental debate about international tax policy has ensued over the past decade, leading to recently enacted (or proposed) reforms that represent radical departures from the status quo. • To guide international tax policy, researchers must measure both the distribution of MNE income across jurisdictions, in a way that faithfully represents where the income was generated, and the rate of tax paid by the MNEs on their income in each different jurisdiction. The accounting methods used to report MNE activity affect the measurement of each of these constructs. The measurement issues we document impact to various degrees all MNE data that are disaggregated by jurisdiction. • Overlooking the accounting conventions that guide MNE data can result in two kinds of mismeasurement – duplicative counting of income and misattribution of income. The first error – duplicative or “double” counting – arises when a researcher counts the same dollar of MNE income more than once by including it in two or more different jurisdictions. The second error – misattribution – arises when a researcher fails to recognize a dollar of MNE income in the jurisdiction where it was earned. • The accounting issues we highlight in this paper affect estimates of the aggregate profits of MNEs, measurement of effective tax rates, the amount of revenue losses governments experience due to profit shifting, as well as estimates of the sensitivity or semi-elasticity of income to taxes. • For example, an influential study published just prior to the passage of the 2017 U.S. Tax Cuts and Jobs Act estimated that between $77 and $111 billion dollars in U.S. revenue was being lost to profit shifting each year. When we reexamine that study ( Clausing 2016 ), correcting for double counting and misattributed MNE income, while otherwise holding constant the study’s methodology, this estimate drops to $11 billion a year. Given the perception that multinational enterprises (MNEs) engage in extensive tax planning, ending base erosion and profit shifting activity is a priority on many national agendas. Yet the actual level of such activity is subject to debate. In this paper, we provide guidance on how to accurately measure country-level MNE income using the datasets commonly used in research on profit shifting. This issue is of global concern, as any economic data that reports profits by jurisdiction must use an established accounting method to report the activity of the MNEs’ indirectly owned foreign affiliates. We explain how the accounting method could lead a researcher to double count income or to attribute it to the wrong jurisdiction. Such errors not only affect measures of the MNEs’ country-level profits, but also bias researchers’ estimates of the sensitivity of income to taxes. Although we focus our analysis on data from the U.S. Bureau of Economic Analysis (BEA), we illustrate the consequences of such mismeasurement across a variety of studies relying on a variety of data sources.