The hypothesis is based on the literature review of related previous researches. Delmas & Pekovic (2013) show that their employees will have higher productivity if the organizations adopt environment standards. Brock & Taylor (2005) also show that better environment equates to better productivity. Since increase in CO2 emissions negatively impacts the environment conditions, it can be hypothesised that productivity should be negatively related to CO2 emissions in the economy.
The estimated model for India is given by the following equation. The coefficients significant at 1%, 5% and 10% are marked by ***, ** and * respectively. The complete results of the model and diagnostic tests are shown in Appendix.
Economic Growth= 0.6986-0.0533log (per capita GDP) -0.2015Emissions level of CO2-0.2097Government spending on education+error
Note that the squared log GDP per capita is omitted here because its inclusion had led to the estimation of a meaningless regression. The corresponding result is shown in appendix.
The above model for India shows that of the given factors, only CO2 emissions is statistically different from zero at confidence level of 90%. The coefficient is negative implying that pollution level impacts the growth of the economy of India negatively. For a unit increase in CO2 emissions, economic growth is expected to decline by 0.2. The other statistically significant parameter is the constant term which shows the growth when all the other variables are zero. The coefficient of log of per capita GDP and government spending on education are both negative and statistically not different from zero. The explanatory power of the above model is 37%.
The estimated model for Canada is given by the following equation. The coefficients significant at 1%, 5% and 10% are marked by ***, ** and * respectively. The complete results of the model and diagnostic tests are shown in Appendix.
Economic Growth= -67.5026+12.4976*log (per capita GDP) *-0.5799(log (per capita GDP) )2+0.5439Emissions level of CO2+0.0323Government spending on education+error
The above model shows that natural log of GDP per capita has a positive estimated coefficient and statistically unequal to zero at 10%. The coefficient of square of log of GDP per capita is negative and statistically different from zero at 10% significance level. Both these estimates follow expectation set based on literatures. However, the coefficient estimate of CO2 emissions is statistically significantly positive which contradicts literature review. The coefficient of government’s spending on education is not significant. The explanatory power of the estimated model for Canada is 24%. The standard errors of the model are robust standard errors to take care of heteroskedasticity.