英语论文网

And because according to him, if someone wants a result that is not bias of the latter equation using OLS, it must be expected that the effect of disturbances on subject of regressors is zero. Because the individual's choice variables are and , and the value chosen for q depends on , and and will be correlated. Because of this, may create OLS estimates are bias. (ibid: 10).

The choice of how dangerous a job will be both mortal and not fatal, has a connection to maximize the benefits, but again depends on the loss of benefits that are expected and the marginal benefit of accepting a more risky job. If the value of and is higher, this will lead to raising marginal gain in earnings for and , and again this will increase the value of the earnings of and (ibid: 10).

The disturbance that increases the earnings were defined as e will generate effect to earnings and simultaneously acquire a less dangerous job. The variable x1 has an influence on the choice of risk through an earnings effect. Meanwhile, earnings that are not in the form of money through an income effect will influence the choice of job risk, and will also affect the level of individual risk aversion (ibid: 10).

Garen (1988) estimated wage risk based on empirical models, and it is presented with correction for the endogeneity of job risk. Based on his study, he found that OLS method underestimates both the risk of fatality and injury risk premium for the average individual. In addition, based on evidence, it was shown that the return to risk is affected by unobserved heterogeneity (ibid: 11).

Garen (1988) points out there are a problem that is using industry data. Because as he points out, the error in measurement to job hazard and is a common problem in such studies, and the problem he describes is that all workers do not meet the level of risk as the average in the industry. This will be most problematic for those who are called white collar workers, because they have a lot of safe work condition than blue collar (ibid: 12).

Using data from PSID (panel data of the dynamics behind income) and BLS (U.S. Bureau of Labour Statistics) he estimated using a semi-logarithmic equation and Wage and included his risk variables as regressors. Deaths variables were related to 100,000 full-time workers and non-lethal risk measurement was for 1000 full-time workers. Regressors that were included were education, employment and its squared term, experience and its squared, and some dummy variables such as race, occupation, location and unionism (ibid: 12).

The magnitude and sign of the coefficient reported in his paper was largely expected. Both coefficients of fatal and non fatal risk were both positive and significant. He本论文由英语论文网提供整理，提供论文代写，英语论文代写，代写论文，代写英语论文，代写留学生论文，代写英文论文，留学生论文代写相关核心关键词搜索。