wordpress-themes.org wordpress themes wordpress themes

Sensitivity Analysis

We have subjected our model to a large array of sensitivity analyses. Section C.1 considers various alternatives of our “baseline” пю0о1 withresptct tothepegressofssnd dpo’ernies ef fhemanmle. Have we left our important variables? Is our model sensitive to the chosen time period? Section C.2 continues with a closer look at our dependent variable SO concentration. What happens if we use a different concentration percentile rather than the median? Is there an alternative to using a logarithmic transformation? Finally, section C.3 addresses the question of simultaneity of determination of pollution concentrations and income. Can a simultaneous-equations approach provide additional insights? Affordable loans for borrowers that have been turned down by the bank and other lenders! We offer payday loans cash advance to help you make ends meet or resolve any other financial troubles you may be having. Contact us at further today and get the idea of how much you can borrow when you need.

C.1 Specification

Results presented in the main part of this paper are based on a regression model shown in table 3, hereafter referred to as the “baseline” пю0о1. To ош^е Pheeentitiviayoffhese pesult^ww mm’ify the right-hand side of our estimating equation to address potential problems and to introduce additional regressors. The results for four additional types of models are shown in tables C.1 and C.2 for fixed-effects and random-effects estimators, respectively.
The GEMS/Air study was carried out primarily throughout the years 1976-1991 when the United Nations Environment Programme (UNEP) provided funding to the participating countries. Before 1976 there are only few countries that provide measurements of SO concentrations, and after 1991, the number of countries that report such observations drop rapidly. This is shown in table A.4. By 1996 data are only available from the United States. To allow for a possible participation bias due to funding, we repeat our baseline regression by excluding observations from before 1976 and from after 1991. This procedure reduces the number of observations by roughly 500, or 20%. None of the parameters that desribe scale, composition, technique, and openness effect change sign or significance except for the scale variable. In the fixed-effect model, the significance of the weather variables changes. We now find that a higher concentration of precipitation leads to higher pollution levels. This is consistent with our a-priori expectation that more frequent rain washes SO out of the air.