Frequently asked questions

I have a comment/question/suggestion

We welcome input on the datasite and website to info@ictd.ac.

Who made the dataset?

The Tax Treaties Explorer is a project of the International Centre for Tax and Development supported by the World Bank and the G-24, with data visualisations by Agile Collective. An earlier version was published as the ActionAid Tax Treaties Dataset. The research was led by Martin Hearson. Treaties were coded by a team of research assistants: Gulfiya Ashikbayeva, Pablo Bravo-Burnier, Ximena Garcia Torres, Sebastian Gazmuri, Arek Myszkowski, Joy W Ndubai, Daisy Ogembo, Carlos Quevedo, Ana Maria Reyes Villaveces and Johanna Scaro.

What is a tax treaty?

Tax treaties are agreements between states that divide up the right to tax cross-border economic activity. They set limits on when, and in some cases at what rate, signatories can tax the income from that economic activity, primarily by imposing restrictions on countries’ ability to tax the foreign direct investment that they receive. There are over 3000 in force worldwide. Tax treaties are commonly referred to as Double Taxation Agreements, because one of their main functions is to prevent companies and individuals from incurring tax in more than one country on the same income.

Why have we made this dataset?

International efforts to prevent tax treaty abuse and strengthen source taxing rights, together with greater awareness of some of the costs of tax treaties, have led to growing interest in developing countries' tax treaties. Many countries are seeking to assess the costs and benefits of their existing treaty networks, and to develop policies to guide future (re)negotiations. Some argue that tax treaties should be understood and scrutinised as investment incentives, yet efforts to increase transparency over tax incentives have not extended to tax treaties.

The visualisations and dataset on this website provide a means to compare and contrast different treaties in ways that complement analysis of the legal wording. For non-specialist policymakers and others with a stake in tax policy, this is an accessible entry point to understand treaties in comparative context. Part of the reason that tax treaties often receive less scrutiny than other elements of the tax code is that they are obscure technical documents requiring considerable familiarity with domestic and international tax law to interpret them correctly. While analysis using the dataset cannot replace such an analysis, it provides an additional means of investigating how the content of tax treaties varies between countries and over time.

Which treaties are included?

The dataset includes over 2600 bilateral tax treaties, more than 300 amending protocols, 8 multilateral treaties, and certain changes made to these treaties by the Multilateral Convention to Implement Tax Treaty Related Measures to Prevent BEPS (MLI). As far as possible, this includes all treaties signed by 118 countries comprising: those that are or were until recently low and lower-middle income countries, all countries in Africa, and all members of the Intergovernmental Group of 24.

The full list of countries is Afghanistan, Albania, Algeria, Angola, Argentina, Armenia, Azerbaijan, Bangladesh, Benin, Bhutan, Bolivia, Bosnia and Herzegovina, Botswana, Brazil, Burkina Faso, Burundi, Cabo Verde, Cambodia, Cameroon, Central African Republic, Chad, China, Colombia, Comoros, Côte d'Ivoire, Curaçao, Democratic People’s Republic of Korea, Democratic Republic of Congo, Djibouti, Dominican Republic, Ecuador, Egypt, El Salvador, Equatorial Guinea, Eritrea, Eswatini, Ethiopia, Federal States of Micronesia, Gabon, Georgia, Ghana, Guatemala, Guinea, Guinea-Bissau, Guyana, Haiti, Honduras, India, Indonesia, Iran, Iraq, Jordan, Kenya, Kiribati, Kosovo, Kyrgyz Republic, Lao PDR, Lebanon, Lesotho, Liberia, Libya, Madagascar, Malawi, Maldives, Mali, Marshall Islands, Mauritania, Mauritius, Mexico, Moldova, Mongolia, Morocco, Mozambique, Myanmar, Namibia, Nepal, Nicaragua, Niger, Nigeria, North Macedonia, Pakistan, Papua New Guinea, Paraguay, Peru, Philippines, Republic of Congo, Rwanda, Samoa, São Tomé and Principe, Senegal, Seychelles, Sierra Leone, Solomon Islands, Somalia, South Africa, South Sudan, Sri Lanka, Sudan, Syrian Arab Republic, Tajikistan, Tanzania, Thailand, The Gambia, Timor-Leste, Togo, Tonga, Trinidad and Tobago, Tunisia, Turkmenistan, Uganda, Ukraine, Uzbekistan, Vanuatu, Venezuela, Vietnam, West Bank and Gaza, Yemen, Zambia, Zimbabwe.

Some of these countries may not be mentioned in the accompanying dataset because they had not concluded any tax treaties at the time that it was last updated.

Which treaties are not included?

We’ve made every effort to include comprehensive lists of treaties. It has not been possible to code all treaties, because the text was not available, parts vary significantly from the content and structure of the UN and OECD model tax treaties, or they are not published in English, French or Spanish. Treaties that were not coded appear in the results, but without any values attached.

Treaties concluded on or after 15 March 2023 are not included at present, except for MLI modifications. More information is given on the data and documentation page.

How are the treaties coded?

Most tax treaty negotiations use as their starting point the OECD and United Nations model tax conventions, and almost all treaties follow a structure based on these models. Treaties based on the OECD model generally impose greater restrictions on a country’s ability to tax inward investment; the UN model makes amendments to the OECD model that leave more of these rights intact. Each of the fields in this dataset is based on a provision of the model treaties: a difference between the UN and OECD models, a clause that is in both models but does not always appear in negotiated treaties, or a value that the models leave open to bilateral negotiations. This was developed in consultation with an advisory group of tax professionals, some of whom were experienced treaty negotiators. The dataset uses a purposive interpretation, which means that we have tried to take account of the intention of non-standard wording, rather than simply checking for the presence of a specific phrase. Each treaty was coded twice, independently, by two different members of the project team. Any disagreements in the coding were then reconciled by the project lead, consulting with the advisory group.

What are the indices?

As well as using the dataset to look into the detail of negotiated tax treaties, it can also be used to amalgamate the content of a treaty into an expression of the overall bargaining settlement it contains. These indices are useful starting points for comparing treaties, but they are only a very rough approximation, and we recommend a detailed examination of the text before drawing any firm conclusions.

To create the indices, each clause in the treaty was assigned a value between 0 and 1, where 1 represents a greater taxing right over inward investment. Indices are averages of these values over a particular group of clause, as follows.

  • The ‘Source’ index incorporates all fields in the dataset that relate to the balance of taxing rights. It gives a high-level overview of the treaty.
  • The ‘PE’ index includes fields related to Permanent Establishment, which refers to the threshold above which a foreign company’s presence in a country becomes taxable. It is drawn from article 5 of the model treaties.
  • The ‘WHT Rates’ index is an average of the withholding tax rates in each treaty. These are taxes imposed on cross-border investment, which treaties either prevent or limit to a maximum rate. These are articles 10 to 12A of the model treaties.
  • The ‘Other’ index includes the remaining fields, drawn from articles 7, 8, 12, 13, 16 and 21 of the models.
  • The ‘UN’ index employs a strict analysis of only the provisions that vary between the UN and OECD models. It excludes, for example, WHT rates, since these are not specified in the UN model.
What limitations do the data have?

The dataset is designed for broad cross-country and cross-time comparisons. Users should bear in mind that such an exercise has implications for appropriate use of the dataset. This includes:

  • The dataset is not designed to give a precise or comprehensive account of the legal position. The coding structure eliminates nuance and heterogeneity by coding each provision with a single word or number. This approach is appropiate because we have deliberately chosen treaty provisions that appear in a uniform way in the vast majority of treaties, but there are sometimes exceptions. For example, the dataset gives two rates for dividend withholding tax, and the threshold for the recipient's share of ownership that qualifies for the lower rate. Some treaties have three or more rates, or use a different type of criterion to access the lower rate. We have provided a coding guide that explains how the most common variations were treated.
  • Scoring within the dataset follows a purposive interpretation wherever possible, meaning that coders were instructed to look beyond the presence of specific phrases to the underlying meaning. Nonetheless, some coding may not take into account the interaction between provisions included for consideration and other provisions that may have a similar effect if drafted in a non-standard way.
  • Although each treaty has been coded by two people independently, with a third person reconciling disagreements, some coding errors and inconsistencies may remain. This will usually be the case if both coders made the same error. Claims about particular provisions of individual treaties drawn from this dataset should always be checked against the treaty text.
  • The dataset disregards national tax systems. For example, withholding tax rates are compared with each other as absolute values. Yet, a 10% maximum in a treaty may result in a considerable amount of foregone tax revenue for a country whose domestic rate is 20%, whereas it would have no impact on revenues for a country whose domestic rate is 8%. (Of course, the sacrifice of taxing rights is the same in both cases, since neither country has the right to levy a withholding tax on residents of the treaty partner greater than 10%. A country with an 8% rate today may have had a higher rate in the past, just as it may want to increase its rate in the future). We hope that a future update to the dataset will provide for comparisons that take domestic law into account.
  • Certain important clauses are missing entirely from the dataset, due to the wide variation in drafting that make it hard to code reliably. One example is ‘tax sparing’ provisions within the article covering double taxation relief, which significantly influenced many negotiations by developing countries.
  • The indices based on the dataset do not employ any weighting strategy. It is impossible to take into account how important each particular clause is to different countries. Users should bear in mind that defending the right to impose certain taxes will be much more important to some countries than others.