World Happiness Report 2023 100 Endnotes 1 Classifications come from UCDP/PRIO Armed Conflict Dataset version 19.1. 2 Besley, T., & Persson, T. (2011) 3 See Weber, M. (1919) 4 See Besley, T., Dann, C., & Persson,T. (2021) for a discussion of data sources. The frequency of civil wars is measured using the UCDP/PRIO Armed Conflict dataset and repression is measured by the presence of political purges in the Banks Cross-National Time-Series (CNTS) Data Archive. 5 The latter according to the UCDP/PRIO data are Israel and the USA. 6 Namely, if V-Dem’s executive constraints variable is greater than 0.8 (corresponding to roughly the top third of the global distribution). 7 See Figure 9 in Besley, T., Dann, C., & Persson,T. (2021) 8 See Figure 9 in Besley, T., Dann, C., & Persson,T. (2021) 9 See, for example, Tilly, C. (1990) 10 See Besley, T., & Persson, T. (2014) 11 https://www.worldbank.org/en/programs/business- enabling-environment/doing-business-legacy 12 See Barro, R. J., & Lee, J. W. (2013) 13 https://databank.worldbank.org/source/world-development- indicators 14 See Besley, T., & Persson, T. (2014) 15 Both these variables take on values between 0 and 1 (with higher values capturing stronger constraints). We create a binary indicator, which we set equal to 1 if the average of V-Dem’s two executive-constraints measures is greater than or equal 0.8, and 0 otherwise. Having an average greater than 0.8 corresponds to being roughly in the top third of the distribution. 16 Besley, T., & Persson, T. (2011) 17 This is also the focus of Bueno de Mesquita et al. (2005) and Persson, T., & Tabellini, G. (2005) 18 Besley, T., & Persson, T. (2011) 19 Besley, T., & Persson, T. (2010) 20 See, for example, Almond, G. A., & Verba, S. (1963); Levy (1989); and Putnam et al. (1994) 21 See Blais, A. (2006) for an overview of the literature on voter turnout and the factors that shape it. 22 Willeck, C., & Mendelberg, T. (2022) for a review and discussion. 23 See Levi, M. (1989) 24 Besley, T. (2020) 25 Besley, T., & Persson, T. (2011) 26 Besley, T., & Persson, T. (2011) 27 This is constructed as one minus the proportion of years a country has been in repression but not civil war since 1975 (multiplied by one half) and the proportion of years that a country is in civil war (but not repression) since 1975. Thus the index gives half as much weight to repression as it gives to civil war. 28 Here we use a min-max normalization so it lies between zero and one. 29 Besley, T., Dann, C., & Persson,T. (2021) for details on the construction of this variable. 30 We use a hierarchical clustering method based on principal components (HCPC) which has two core steps; see Hastie et al. (2009, section 14.3) for further details. First, we use the raw data to create principal components of the variables of interest. This reduces the “dimensionality” of the data so as to find the number of dimensions needed to summarize the underlying variables. Second, we employ an agglomerative hierarchical clustering algorithm (Ward’s criterion) to identify clusters based on the principal components. To confirm the number of principal components, Kaiser’s criterion and the “elbow test” indicate that two components are optimal. 31 A s a reminder, these are income tax as a share of total tax intake, legal quality index, collective capacity index, proportion of years in repression since 1975, proportion of years in civil war since 1975, and GDP per capita. 32 To understand the figure, note that the clustering analysis first takes into account the variation in civil war, repression, income, fiscal capacity, legal capacity, and collective capacity. It then uses a principal-component analysis to construct two core dimensions. One of these distinct clustering dimensions (dimension 2 in the figure) combines civil war and repression into a single component. But it also identifies civil war and repression as distinct factors, giving negative values to repression, positive values to civil war, and values around 0 to peace. 33 Besley, T., Dann, C., & Persson,T. (2021) 34 The weak states are: Algeria, Benin, Bolivia, Côte d’Ivoire, El Salvador, Guatemala, India, Malawi, Morocco, Myanmar, Niger, Pakistan, Paraguay, Peru, Philippines, Senegal, Sri Lanka, Togo, and Turkey. The special-interest states are: Albania, Argentina, Brazil, Bulgaria, Chile, China, Costa Rica, Cyprus, Dominican Republic, Egypt, Greece, Hungary, Iran, Jamaica, Malaysia, Mauritius, Poland, Romania, South Korea, Thailand, and Uruguay. The common-interest states are: Australia, Austria, Belgium, Canada, Finland, France, Iceland, Ireland, Italy, Japan, Luxembourg, Malta, Netherlands, New Zealand, Norway, Portugal, Spain, Sweden, Switzerland, United Kingdom, United States. 35 We have tested the robustness of this core finding by redoing the clustering analysis without including income per capita as a variable. A clear cluster of common interest states still emerges and there is a strong, and statistically significant correlation between being in this group and the average level of life satisfaction. 36 Goff et al. (2018)
RkJQdWJsaXNoZXIy NzQwMjQ=