copy the linklink copied!Annex A. Methodology – Individuals
copy the linklink copied!Countries covered in the analysis
Global and regional surveys are used in the analysis. They are public opinion surveys that explore the perceptions and public attitudes on different issues of society.
World Values Survey 2010-13
The analysis uses the latest World Values Survey wave (2010-13). To explore differences across regions, countries are grouped into OECD members (excluding Chile and Mexico), Latin America, Emerging Asia, Africa, Eastern Europe and Rest of the World (Table A.1). The country classification is limited to the 56 countries that responded to the tax morale question and each regional group varies, being composed of a range of 5 to 12 countries.
Latinobarómetro 2016
The last round of Latinobarómetro (2016) survey was conducted in 18 LAC countries, representative of the region: Argentina, Bolivia, Brazil, Chile, Colombia, Costa Rica, Dominican Republic, Ecuador, El Salvador, Guatemala, Honduras, Mexico, Nicaragua, Panama, Paraguay, Peru, Uruguay and Venezuela.
Afrobarometer 2015
The 2015 round of Afrobarometer covers 36 countries: Algeria, Benin, Botswana, Burkina Faso, Burundi, Cameroon, Cabo Verde, Côte d'Ivoire, Egypt, Gabon, Ghana, Guinea, Kenya, Lesotho, Liberia, Madagascar, Malawi, Mali, Mauritius, Morocco, Mozambique, Namibia, Niger, Nigeria, São Tomé and Príncipe, Senegal, Sierra Leone, South Africa, Sudan, Swaziland, Tanzania, Togo, Tunisia, Uganda, Zambia, and Zimbabwe.
copy the linklink copied!Methodology
To measure tax morale we use the answers to the question ‘’Do you justify cheating on taxes if you have the chance?’’ Tax morale is defined as a dummy that equals unity if the person rates 10 on a scale from 10 (cheating on taxes is never justifiable) and zero from answers raging between 1 (cheating on taxes is always justifiable) and 6 (low tax morale). Therefore, we define individuals with (high) tax morale as those who strictly do not justify cheating on taxes at all.
The drivers of tax morale are analysed based on a simplified micro-econometric approach. Following the standard specification in the empirical literature, a Probit model is used on an individual basis, where the dependent variable is tax morale measured in three ways, when available. First, the traditional definition that measures the attitudes toward taxation by analysing those who would never justify cheating against those who would cheat on taxes. Second, tax morale is measured as the willingness to pay more taxes for development or a public service. Third, and only for the case of Africa the dependent variable, as a proxy of tax morale, captures the willingness to increase spending on certain public services.
The analysis studies the relationship between citizens’ perceptions of tax morale and several socio-economic and institutional (perceptions-related) variables that vary depending on the availability of data across world and regional surveys. For that purpose, the different Probit regressions include in a first stage the socio-economic factors, in order to grasp the information not dependent from policy decisions. In the second stage, the different variables related to institutions, satisfaction with public services, corruption and tax administration as established in the literature and available in the datasets are included successively, one at a time and alone. In the final stage, the Probit includes all of the socio-economic variables alongside all of the possible determinants under analysis. All regressions include country dummies to capture potential differences in perceptions and average behaviour across countries.
Possible caveats
A caveat to take into consideration is that these regional surveys are not comparable to each other. Additionally, the coverage per country is still limited (we focus on regional results and not a country comparison) and the period of implementation of each survey differs. Finally, similar to the World Values Survey, questions intend to measure perceptions and attitudes; the information corresponds to self-report answers that can bias the results.
Metadata, Legal and Rights
https://doi.org/10.1787/f3d8ea10-en
© OECD 2019
The use of this work, whether digital or print, is governed by the Terms and Conditions to be found at http://www.oecd.org/termsandconditions.