Happiness and Economics: How the Economy and Institutions Affect Human Well-Being

APPENDIX A

Appendix A This appendix presents in Table A.1 the full microeconometric happiness function for Switzerland, as discussed in this book. Table A.2 shows the results of the DFBETA robustness analysis discussed in chapter 8, section 8.3.3. The empirical analysis in Table A.1 is based on a survey of more than 6,000 residents of Switzerland by Leu, Burri, and Priester (1997). The survey data were collected between 1992 and 1994. The information contained in the data set is based on personal interviews and tax statistics. The single-item measure for people’s subjective well-being is based on answers on a 10-point scale to the following question: “How satisfied are you with your life as a whole these days?” Further information on the data is given in chapter 3, section 3.4.1. Econometric Estimation Method Two statistical approaches are applied to study the correlation between socio-demographic, socio-economic, and institutional factors and people’s reported satisfaction with life. The results for both approaches are presented in Table A.1. First, average satisfaction scores for each differentiated demographic category are calculated. They allow for the assessment of the total effect of a certain demographic characteristic and, compared with means for other categories of the same demographic dimension (e.g., age categories), they offer rough information about simple correlations. Second, partial correlations are presented. They indicate the effect of a certain demographic characteristic independent of other socio-demographic and socio-economic characteristics and independent of the institutional environment . The latter approach uses multiple regression analysis, which is normally applied to estimate microeconometric happiness functions. In the first such equation, a weighted least squares model is estimated. In the second one, a weighted ordered probit model is used in order to exploit the ranking information contained in the originally scaled dependent variable. The weighting variable that is applied allows representative results on the subject level for Switzerland. Throughout the book, we use a robust estimator of variance , because random disturbances are potentially correlated within groups or clusters. Here, dependence refers to residents of the same canton. Ignoring the clustering in the estimation model is likely to produce downward-biased standard errors, due to the effects of aggregate variables on individual data (Moulton 1990). To get unbiased standard errors for the aggregate variable “democratic participation rights” (see chapter 8), the 26 cantons are used as sampling units. The least squares estimation treats happiness as a cardinal variable. This basic estimation technique is applied to facilitate the interpretation of the results. The coefficients are to be read as follows: People belong- 186 APPENDIX A ing to a certain category on average report happiness scores deviating from that of the reference group on the scale of the coefficient. For a continuous variable (such as the measure for democratic participation rights), the coefficient indicates the increase in happiness scores when the independent variable increases by one unit. In the ordered probit estimation, a positive coefficient indicates that the probability of stating happiness greater than or equal to any given level increases. There is no direct quantitative interpretation of the size of the coefficient. Therefore, marginal effects are calculated. The marginal effect indicates the change in the share of persons belonging to a happiness level of 10 when the independent variable increases by one unit. Alternatively , the marginal effect indicates the change in the probability belonging to a happiness level of 10 when the independent variable increases by one unit. In the case of dummy variables, the marginal effect is evaluated in regard to the reference group. APPENDIX A 187 Table A.1 (Part 1) Satisfaction with Life in Switzerland, 1992–94 Descriptive Weighted Statistics Least Squares Weighted Ordered Probit Marginal Coefficient Coefficient Effect for Mean (t-Value) (t-Value) Score of 10 Socio-demographic factors Age 20–29 8
19 Reference group Age 30–39 8
03 −0
142 −0
084 −0
028 −1
06 −0
96 Age 40–49 8
22 −0
001 0
001 −0
3-e3 −0
01 0
01 Age 50–59 7
97 −0
053 −0
012 −0
004 −0
61 0
20 Age 60–69 8
45 0
449 −0
313 0
112 4
53 4
60 Age 70–79 8
41 0
557 0
387 0
141 4
48 4
66 Age 80 and older 8
22 0
435 0
332 0
121 2...