The impact of tax treaties on revenue collection: A case study of developing and least developed countries

Executive summary

• Foreign direct investment (FDI) by multinational enterprises is given substantial weight by an expanding number of developing countries. Many of them have decided to sign tax treaties (or double taxation agreements) with a number of other – usually richer - countries in the hope that it would attract more investment from these countries. However, tax treaties strongly impact their tax revenues – often at a scale that might not have been intended or foreseen when they were signed. These tax treaties can increase investment and thus the tax base, but they often reduce the applicable tax rates: we focus on the effects of this latter aspect.
• We provide illustrative estimates of the potential costs of tax treaties in developing countries – i.e. if all else remains unchanged, what additional tax revenues the developing countries could have if the standard tax rate applied rather than the tax treaty rate. Due to data restrictions, we are only able to calculate approximate estimates, with reasons pulling the real effects in both directions. On the one hand, we assume that investments are not influenced by the tax treaties and in this important respect we provide upper-bound estimates. On the other hand, we examine only some taxes for some countries and in this sense we provide lower-bound estimates. We estimate the effects of only two types of losses generated by tax treaties, namely, lower withholding taxes on outgoing dividend and interest payments. We estimate the effects for 14 developing countries (for dividends; 11 for interest payments) with information from the ActionAid Tax Treaties Dataset and the International Monetary Fund’s foreign direct investment data.
• Within this group of countries, we estimate the highest potential tax revenue losses for the Philippines (509 million USD) and Pakistan (130 million USD). Relative to their GDP, we estimate that the potential losses are highest for the Philippines and Mongolia (0.17% of GDP for both).
• We find that investor countries Japan, the Netherlands, Switzerland, and Singapore are together responsible for more than half of the estimated losses. The majority of the estimated losses is due to dividends, only around 5% is due to interest payments. We discuss the limitations of these illustrative estimates and how future research could improve their quality as well as coverage.
• We identify four top recommendations for governments of developing countries. First, they shouldcollect, and offer access to, data about the broad range of tax avoidance facilitated by tax treaties.Second, they should consider using our, and other similar, results to identify treaties that wouldbenefit from reviews. Third, governments should consider the UN model treaty tax rates as minimumstandards. Fourth, they should carefully consider whether and under what conditions to sign theOECD’s Multilateral Instrument.