闪电代写 -代写CS作业_CS代写_Finance代写_Economic代写_Statistics代写_代码代做_IT代写_加急帮助

Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: daixieit

ECOM20001: Econometrics 1

Assignment 2: Suggested Solutions

1. (1pt) Regression output with HC standard errors:

(1pt) Interpretation:

- A one-year increase in age is associated with a reduction of new consumption by $1.108.

- Age is statistically significant at the 5% level.

- 0.3% of the sample variation in new consumption can be explained by the regressor.

2. (1pt) The change in age is 35-20=15 years. The 95% CI is then 15 × [−1. 108 − (1.96)(0.500), −1. 108 + (1.96)(0.500)] = 15 ×

[−2.088, −0. 128] = [−$31.32, −$1.92] . (Optional: we are ‘95% confident’ that the true change lies between −$31.32 and −$1.92).

(1pt) The null’s hypothesized value, which is −$15, lies inside the 95% CI. Therefore, we fail to reject the null at the 5% sig. level.

3. (1pt: graph, 2pt: interpretation)

The single linear regression is = 0 + 1 + . The omitted variable (income) and u move in the same direction (note: we assumed that higher income is associated with higher consumer confidence, which increases new consumption.) Also, Age is positively correlated with Income. Therefore, Age is positively correlated with u. Therefore, the estimated coefficient on Age should be higher than the true parameter value. (Note: The direction of bias must be precise and clear.)

4. (1pt) The result does not make sense from a causality perspective because it is hard to conceive that an individual’s income (an economic trait) can affect the individual’s age (a demographic trait), other things kept equal. By swapping the variables, we have a model that is consistent with age affecting an individual’s income, which is more reasonable from a causality perspective (even though the regression may still suffer from omitted variable bias) .

5. (1pt: table and interpretation) Regression output with HC standard errors:

Keeping age and income fixed, a 1-year increase in education is associated with a reduction of new consumption by $3.157. Education is statistically insignificant at the 10% level.

(1pt: computation) The offsetting change is

income.

−3. 157 0.409

= 7.719, which translates to a $7,719 increase in

6. (1pt) Regression output with HC standard errors:

(2pts) Keeping age and income fixed, we have on average, relative to

individuals with a postgraduate education level:

$18.698 more new consumption among high school graduates (not

statistically significant at the 10% level)

$8.274 more new consumption among bachelor degree holders (not

statistically significant at the 10% level)

(1pt) This regression is more flexible in that it relaxes the assumption that years of education has a linear effect on newc. Upon examination of the raw data (such as using the “table(mydata$education, mydata$d1)” command to plot frequency tables), we know that the education variable takes three values only (12, 16, 18). Using d1, d2 as regressors, we allow each of the three education categories to have a flexibly different effect on newc.

7. (1pt) 0 : all the coefficients on the regressors are 0;