Quiz 9

We are interested in looking at the effect of beer taxes on traffic fatalities. We have data on the traffic fatality rate (per 10,000 state residents) and beer tax (in dollars per case) in 48 US states, for all quarters between 1982 and 1988. We use these data to estimate the following regression:

$Fatality Rate_{st} = \beta o + \beta i BeerTax_{st} + \epsilon_{st}$

1.1. What is the smallest time unit of the panel?

1. Years
2. Year-Quarters
3. Months
4. Days

1.2. What is the group unit of the panel?

a. Beer Tax Rate

b. Fatality Rate

c. Years

d. States

1.3. Given the set up, which of the following fixed effects could we use to reduce the bias in the pooled model above?

1. Individual Fixed effects (because they are the ones drinking beer)
2. County Fixed Effects
3. Month Fixed Effects
4. State Fixed Effects

2. When using the Least Square Dummy Approach (LSDA), we should include a dummy variable for every single state.

1. True, we should include a binary variable
2. False, we should omit one state
3. False, we should include a continuous variable of the state
4. False, because we would omit a year dummy and that would be enough.

3. How can we turn cross-sectional data into panel data?

1. Merge multiple sets of cross-sectional data.
2. Assume that individuals in different cross-sections are the same if they have similar characteristics.
3. Estimate past and future data for each individual observation.
4. Aggregate observations to a higher unit of observation across multiple years.

4. You’re studying the return to education, where your independent variable is years of education and your outcome is income. You have state-year panel data which includes four states: Virginia, New York, California, and Texas. Which regression(s) correctly control for the fixed effects at the state level?

1. $Income = \alpha + \beta Education + \gamma Virginia + \epsilon NewYork + \delta California + \theta Texas$
2. $Income = \alpha + \beta Education + \gamma Virginia + \epsilon NewYork + \delta California$
3. $Income = \alpha + \beta Education + \gamma Virginia + \epsilon NewYork + \theta Texas$
4. $Income = \alpha + \beta Education + \epsilon NewYork + \delta California + \theta Texas$