英语论文网

But there is still interest in determining the size of the differentials in terms of input to cost-benefit studies, and by increasing the appropriate coefficient magnitude is less comforting. This could mean that the increase in the coefficient reflects a higher opportunity to the real income. Based on the theory, would safety be a superior good. Compensation differentials ensured by log-linear earnings risk specification to increase proportionally with the level of income but at a higher income than is allowed would this result to a higher coefficient when it is adapted to the real income (ibid: 262).

CONCLUSION

In the last part of the paper examined Arabsheibani and Marin effects by the inclusion of non-lethal risk, and also the decision to allow contents of how dangerous a profession can be endogenous. The results were quite mixed. Unlike the deadly risk, sign, size, and significant of non fatal risk was sensitive to the inclusion of other variables such as dummy variables. It was interesting to see if by removing non fatal risks would this affect the estimated compensation for death risk and whether if there is any significant additional compensation for the risk of nonfatal accidents.

The answer to the first part was negative. The way it looked, either inclusion or exclusion of non-lethal risk would not affect the coefficient of deadly risk. The answer to the second question in this section was less robust. When both non fatal and lethal risk were included in the regression, the coefficient of non fatal changed according to other variables and sample. The coefficient was always negative and a low value of t-statistics, and it was significantly different than zero at the usual levels.

Using OLS estimates, there was no consistent and positive compensation differentials for risk of nonfatal accidents, especially when deaths were included simultaneously. Nevertheless, the coefficient of death risk was not affected either with or without non-fatal risk (ibid: 264).

Another focus was emphasized in this paper, was the possibility that risk is endogenous in the selection of risk. Using the method that was described earlier by Garen (1988), the result increased the twice the coefficient of risk, they also found that the values ??were also very sensitive to precise specifications. But this expansion gave no value on life that was robust (ibid: 265).