The motivation of this paper is estimate the effects of the minimum legal drinking age (MLDA) on alcohol consumption and mortality. This paper will utilize data from the National Health Interview Sample Adult Files 1997-2007. Two methods including the sharp regression discontinuity design as well as an instrumental variable (IV) variable approach will be used to estimate the magnitude of the MLDA. In the regression discontinuity design, there is a reduction in consumption of alcohol by 9.2 percentage points. Additionally, there is an estimated increase in mortality rates by 8.06 deaths per 100,000 individuals. Using the IV approach can support the results within the regression discontinuity design; however, the validly of the estimate is questionable
…show more content…
Several arguments have been put forth such as how it will stimulate growth for businesses and how the current policy results in the United States are more dangerous than if it were lower. The Amethyst Initiative as well as several others believe age 18 is a common age and lowering the age could be a lesson for teenagers to face before they are independent from parental guardians. This paper will help motivate that age 21 is not the optimal age for the MLDA rather provide empirical evidence that suggests setting the MLDA at 21 clearly reduces the proportion of the population that drinks in the U.S. and how much MLDA reduces mortality. In this paper, two approaches will be used to estimate the effects of the MLDA on alcohol consumption and mortality rates. The first design is a regression discontinuity, which focuses on individuals right above and below the MLDA. The bin width and bandwidth chosen reflects the ranges where the discontinuity is most apparent. Based on the first stage estimate, we will find a 9.2 percentage point …show more content…
Notice that equation (1) provides the regression of our estimation of the first stage. We initially superimposed a linear, quadratic, and cubic line to the data points, but noticed the quadratic provides the most precision. We notice that at age 21, there is a discontinuity, which suggest that the MLDA is a barrier for people under the MLDA to consume alcohol. Table 2 provides regression estimates with different polynomial orders. Columns 1, 3, and 5 present regressions for different order polynomials of the percentage point increase in alcohol consumption for people right at the cut off line. Likewise, column 2, 4, and 6 present similar regression to column 1, 3, and 5 except “Birthday” is added to act as a binary variable for whether or not someone had a birthday when they died. In column 3, we observe an increase in alcohol consumption by 9.2 percentage points upon turning age 21. In row 8, we notice that Birthday has no statistically significant effect. This suggest that including “Birthday” in our first stage (1) has no impact on alcohol
Several arguments have been put forth such as how it will stimulate growth for businesses and how the current policy results in the United States are more dangerous than if it were lower. The Amethyst Initiative as well as several others believe age 18 is a common age and lowering the age could be a lesson for teenagers to face before they are independent from parental guardians. This paper will help motivate that age 21 is not the optimal age for the MLDA rather provide empirical evidence that suggests setting the MLDA at 21 clearly reduces the proportion of the population that drinks in the U.S. and how much MLDA reduces mortality. In this paper, two approaches will be used to estimate the effects of the MLDA on alcohol consumption and mortality rates. The first design is a regression discontinuity, which focuses on individuals right above and below the MLDA. The bin width and bandwidth chosen reflects the ranges where the discontinuity is most apparent. Based on the first stage estimate, we will find a 9.2 percentage point …show more content…
Notice that equation (1) provides the regression of our estimation of the first stage. We initially superimposed a linear, quadratic, and cubic line to the data points, but noticed the quadratic provides the most precision. We notice that at age 21, there is a discontinuity, which suggest that the MLDA is a barrier for people under the MLDA to consume alcohol. Table 2 provides regression estimates with different polynomial orders. Columns 1, 3, and 5 present regressions for different order polynomials of the percentage point increase in alcohol consumption for people right at the cut off line. Likewise, column 2, 4, and 6 present similar regression to column 1, 3, and 5 except “Birthday” is added to act as a binary variable for whether or not someone had a birthday when they died. In column 3, we observe an increase in alcohol consumption by 9.2 percentage points upon turning age 21. In row 8, we notice that Birthday has no statistically significant effect. This suggest that including “Birthday” in our first stage (1) has no impact on alcohol