Monday, September 30, 2019

Individual Report on Econometric Models

Context This report pertains to the selection of an academic paper (Caporale et al., 2009) which is examined based on its aims and results, citing new insights offered by its analysis. It also involves another paper (Oshio et al., 2011) that cited that of Caporale et al. (2009), explaining the kind of data being used for the analysis. The report also includes a description of the economic model adopted by the authors to analyse the issues, as well as a critical reflection of the issues that are likely to arise with the use of these specific econometric models. a.Summary of the selected paper The selected paper in this report is that of Caporale et al. (2009), which was centred on income and happiness across Western and Eastern Europe. The authors utilised data from the European Social Survey (ESS) for 19 European countries, whereby they examined the association between income and subjective well-being. The aim of their study was to find out whether income can buy happiness. Happiness and life satisfaction served as the dependent variables in the study, which were likewise contained in the ESS data. These variables were tested for whether social comparisons and reference groups wielded a relevant influence on the subjective well-being of individuals in the countries involved, which likewise served as the sample. The findings revealed that for all these countries – despite the positive correlation between income on one hand and happiness and life satisfaction on the other – a negative effect was being exerted by reference income on individual well-being, a result that harmonised with the relative utility hypothesis. A separate analysis was conducted by the authors for some countries in Eastern Europe in which they were able to discover some evidence that supported a so-called ‘tunnel effect’, which was suggestive of a positive impact caused by reference income on subjective well-being. The findings maintained that situations characterising stable income and employment have income serving as the basis for social comparisons. In social comparisons, the authors posited that reference income served as the informant for future status expectations. Further, empirical findings revealed a tendency for males to report lower satisfaction than their female counterparts. The authors also cited that the literature holds robust pattern that mirrors the life-cycle areas of people’s social, economic, and family situations. The findings also suggested that happiness perception was positively related to being married, whilst a negative relationship was suggested for divorce. Moreover, the presence of children had a negative effect on happiness perception whilst good health fostered a significant positive association. There was shaky evidence on the likelihood of university qualifications to cause a negative impact on happiness. In addition, a positive effect was indicated by previous employment occurrence on the perception of happiness, which was more deep-seated for more recent unemployment occurrence. The distress accompanying a recent unemployment occurrence may diminish the aspiration value of the presently employed individua l. As the study’s sample was dominated by countries with liberal social welfarism, the influence of the welfare systems on individual’s happiness may also be indicated by the positive impact of unemployment on life satisfaction during the time of joblessness. The new insights offered by the analysis include an apparent wedge between countries in Western Europe and Eastern Europe, which clearly showed that the pursuit of life satisfaction has become embodied in countries’ political framework, serving as a source of valuable insights in policy development concerning welfare reforms. b. A paper citing the selected paper A paper that cites the selected paper here is that of Oshio et al. (2011) whose aim was to examine the effects of relative income on well-being in China, Japan, and Korea based on nationwide surveys in these countries and comparing them with that of the United States. The various factors for which control was undertaken at the individual level were age, gender, and marital status, to name a few. The results were parallel to past researches the same topic involving Western countries. The study exhibited a significant relationship between a person’s income and that of the reference group on one hand, and perceived life satisfaction on the other. In China, individual income showed stronger relationships between relative income and life satisfaction compared to family income, whilst the opposite was demonstrated for Japan and Korea. The comparisons of income within the reference group were necessary for evaluating life satisfaction, particularly when family income was used. Additi onally, Yitzhaki index was used to determine the relative deficiency within the reference group, which was found to have a negative relationship with life satisfaction. Oshio et al. (2011) used Caporale et al. (2009) as a cited paper in their study in that the former utilised it in its claim that apart from the absolute income levels, the happiness of a person is reliant to comparisons with those of others, particularly those with similar socio-economic distinctiveness. Oshio and colleagues also used Caporale and colleagues’ findings on relative income proposition, whereby both found a positive effect fostered by absolute income on both happiness and life satisfaction. Further, both Caporale et al. (2009) and Oshio et al. (2011) used regression analysis to analyse the findings on income and perceived happiness. The apparent difference that could be cited in their studies was the nature of the countries on which their studies were emphasised – Western and Eastern European countries for Caporale et al. (2009), and Asian countries and the United States for Oshio et al. (2011). The former was also more extensive as it used 19 countries f or a generalisation of income-happiness hypothesis, whilst the latter involved only four. The comparison in Caporale and colleagues’ study was between Western European and Eastern European countries, whilst that of Oshio and colleagues was between three Asian countries and the United States. If the cross-country findings in Caporale and colleagues demonstrated a political agenda (i.e. social welfarism), those of Oshio and colleagues were focused on the level of the individual, such as people’s precautious outlook on individual income than family income in China and the United States, leading to the influence of culture for income perception (e.g. individual-orientedness vs. family-orientedness).Whilst Caporale et al. (2009) asserted a positive relationship between income and happiness, Oshio et al. (2011) declared stronger correlations between individual income (rather than family income) and life satisfaction in China, which was not true in Japan and Korea. c.Data used, structure of the data set, and manner of data collection The ESS data in Caporale and colleagues’ (2009) study contained information about a set of demographic and employment characteristics, which the authors utilised as controls in their regressions. Some of these controls were education, income, and position in the labour force, to name a few. The data set also involved information on previous unemployment encounter, which was used in assessing whether a person’s perception of present economic situation was influenced by previous income shocks, commonly due to unemployment. The structure of the data set included all individuals possessing similar levels of education, age brackets, and current countries of residence. The sample was also limited to employees earning full-time salaries. Alternatively, in Oshio and colleagues’ (2011) study, the data were collected from the countries’ (Japan, China, Korea) General Social Survey (GSS), from which data were downloaded. The GSS of these countries were originally gathered by a research office at the University of Chicago. The basis of their empirical analysis was the GSS of these countries which were recorded in 2006, with widely similar survey questionnaires. The results were compared with those of the United States. All income data were converted into logarithms to enable comparing the results across models and countries. In addition, the authors used sampling weights from GSS and made a computation of standard errors to provide correction for any possibility of heteroscedasticity. It may be inferred that even though both studies aimed at determining happiness and life satisfaction vis-a-vis income, their methodologies and approaches were however different but were appropriate to the nature of their hypoth eses. d. Econometric models used by the authors In this section, the econometric models used by the authors are described to analyse the issue. Caporale et al. (2009) used ordered probit model, which they deemed appropriate for their study, given the ordinal nature of the happiness variables. This model enabled a close correlation between happiness and life satisfaction, in which the authors were able to indicate a high level of happiness or satisfaction amongst the research participants. A clear skewness towards the high end was exhibited by the distribution on happiness. Some variations began emerging with the investigation of the level of happiness across the countries involved, with Denmark achieving the highest score, and Portugal, amongst others, recording the lowest scores. Reference income was used as an explanatory variable in the regressions. The estimated coefficients also showed a general consistency with those of previous studies. On the other hand, Oshio et al. (2011) used the ordered logit model to explain perceived happiness across the three countries involved in their study. The model contained an assumption of care amongst individuals in relation to their individual (or family) income when making a comparison of their income with those of others. Family income was included in both specifications, with an assumption that it represents material living standards. It was expected that the coefficient ? was significantly positive in each specification, given the relative income hypothesis. e. Critical reflection of the issues relating to the econometric models The econometric models used in each of the study of Caporale et al. (2009) and Oshio et al. (2011) were appropriately justified and carried out to suit the aim and expected results of each. The use of the ordered probit model in Caporale and colleagues’ study was suitable as this model was able to deal with the variables in ordered categories, which were present in the dataset. According to Jones (2007) and Gailmard (2014), ordered probit model is utilised in the modelling of a discrete dependent variable with ordered multinomial results. Similarly, Caporale and colleagues also noted fixing some of the threshold values, which Jones also pointed out as a characteristic of the model. The study was able to generate the specific results with the utilisation of this model, which is likewise used in a range of the social and behavioural sciences, as emphasised by Aldrich et al. (2007) and Gill (2008). Using other econometric model for the study of Caporale et al. (2009) might not ge nerate similar results, since the study involved an analysis of more than two outcomes of an ordinal dependent variable (i.e. happiness and life satisfaction). In the same manner, Oshio and colleagues’ (2011) use of the ordered logit model was suitable as the study involved a survey that aimed to find out how well responses to questions can predict a response to one question, to which the model was appropriate. In the study, the model was used for dichotomous dependent variables (happiness and life satisfaction), in which more than two response categories were enabled.Apparently, the study adopted proportional odds assumption, to which the model was applicable, as Sadler (2008) and Allegrezza and Dubrocard (2012) also pointed out. Moreover, the model clearly pursued an estimate for multiple equations in the study, whereby the number of these equations were tantamount to the number of categories minus 1. The study also showed that each equation modelled the odds of being in a category, which is consistent with the ordered logit model (Sadler, 2008; Allegrezza and Dubrocard, 2012). It is important to note that the use of the aforementioned models for the two studies was consistent with their intention to quantify or measure variables in order to come up with objective and systematic results for the relationship between income and happiness. Both studies opted to use data from established data sets (ESS and GSS) to apparently facilitate generalisation of findings, with which they were successful. It may be pointed out that Oshio and colleagues’ (2011) results were congruent with Caporale and colleagues’ (2009), reflecting the validity of the latter’s findings. References Aldrich, J. H.m Alt, J. E., and Lupia, A. (2007) Positive Changes in Political Science: The Legacy of Richard D. McKelvey’s Most Influential Writings. Michigan: University of Michigan. Allegrezza and Dubrocard (2012) Internet Econometrics. Hampshire: Palgrave MacMillan. Caporale, G. M., Georgellis, Y., Tsitsianis, N., & Yin, Y. P. (2009) Income and happiness across Europe: Do reference values matterJournal of Economic Psychology, 30(1), 42-51. Gailmard, S. (2014) Statistical Modeling and Inference for Social Science. NY: Cambridge University Press. Gill (2008) Bayesian Methods: A Social and Behavioral Sciences Approach, Second Edition. NW: Chapman & Hall/CRC. Jones, A. (2007) Applied Econometrics for Health Economists: A Practical Guide. Second Edition. Oxon: Radcliffe Publishing Ltd. Oshio, T., Nozak, K., and Kobayashi, M. (2011) Relative income and happiness in Asia: Evidence from nationwide surveys in China, Japan, and Korea. Social Indicators Research, 104 (3), 351-367. Sadler, A. M. (2008) Determinants of Entrepreneurial Behavior among Immigrant and Non-Immigrant University Scientists in the US: The Impact of Cultural Predispositions and Learning. MI: ProQuest LLC.

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.