The life expectancy data found in the other possibility, the US Census Bureau's International Data Base are generated by a combination of past data and curve fitting models, which yield life expectancy estimates that we judged to be less consistent than the methods used by the Institute for Health Metrics Evaluation (2012). However, the former is heavily weighted toward highly developed countries because of the interest in having high quality data over long periods of time and the life expectancy data and we needed data for a broader range of countries.

There were other choices for the life expectancy and population data such as the Human Mortality Database ( Human Mortality Database, 2009) and the International Data Base assembled by the U.S. Similarly, counting gross values disregard industrial emissions, in the context of climate change. Nor does the predominating measure depict family and community cohesion and individual's inner states, nor trust in politicians and confidence in the economy. It further neglects household duties such as taking care of dependents, considered noneconomic. It neglects net values such as advantages added by knowledge, social skills, group memberships, conservation actions, and participation in civic affairs. “Productivity growth” consists of gross values. There are several oft-noted reasons that make GDP a very misrepresented proxy for progress or life satisfaction. The problem is that over time, productivity growth has become the normative proxy of human progress. There was no original intention to assess people's experiences or progress in their lives. Its narrow intention had only been to count the money value of goods and services that a country produces. With the advent of the MDGs in year 2000, it was increasingly evident that when applied to tackling poverty, GDP was far too abstract and unfocused on peoples’ vulnerabilities. GDP as the measure that has prevailed since the 1970s was shaped around a mathematics of productivity growth (productivity of labor). James Chalmers, in Handbook of Statistics, 2018 4 Unsupported Assumptions About the Existing Metric The troubling thing is that the Stata using hausman is saying to use Random (p>0.05) while R using phtest is telling me to use fixed effects (pchi2 = 0.Udoy Saikia. I ran a Hausman test to compare random and fixed effects regression for a panel dataset in both R and Stata and found that the two are showing different results.

Hausman fixed random stata 14.0