Choose two of the research topics below. Answer the following and make sure to justify your assessment. See the sample solution below for an example.

• Imagine that you run an OLS regression of the dependent variable on the regressor.
• Suggest a possible omitted variable.
• Speculate as to whether the OLS estimator of the slope coefficient will be upward or downward biased due to this omitted variable.

Research topics:

1. Does graduating from college cause individuals to live longer?
2. Does a larger number of politicians on a bank's board of governors cause the bank to give larger loans to state-owned firms?
3. Does eating more vegetables during childhood cause individuals to live longer?
4. Does employing more accountants lead firms to pay lower taxes?
5. Does working further away from when you grew up cause individuals to earn higher?

Sample solution using the topic "What is the effect of the number of years of schooling completed by an individual on his or her hourly wages?"

• A possible omitted variable is the individual's level of entrepreneurial spirit. Individuals with a more entrepreneurial personality are more likely to quit school early because of their interest in starting a firm, so that level of entrepreneurial spirit and years of schooling are probably negatively correlated. At the same time, individuals with a more entrepreneurial character are likely to attain better-paid positions, so that level of entrepreneurial spirit probably has a positive effect on hourly wages. This implies that the OLS estimator of the slope coefficient of years of schooling will likely be downward biased due to the omission of this variable.

Important notes:

• Don't suggest a variable that is actually a reason behind a causal relationship between a regressor and a dependent variable. For example, schooling might affect wages because a higher level of knowledge about school subjects leads to higher wages. This doesn't mean that we should worry about 'level of knowledge about school subjects' being an omitted variable, because this is simply how schooling affects wages.

• Make sure that your chosen omitted variable is an ordinal or interval variable (i.e. a variable for which 'higher' or 'lower' is meaningful), because you haven't yet learned how to work with nominal variables (i.e. categorical variables such as gender) in this course.