## Economics 2557

Final Exam

美国经济学Final exam代考 There is a controversy among economists about whether the Social Security Program leads to a ecrease in private savings in the US.

Use points as an indicator of time (Total Points: 100). Pace yourself. Show all your work. Don’t panic! Good luck!

### 1. (21 points) 美国经济学Final exam代考

There is a controversy among economists about whether the Social Security Program leads to a ecrease in private savings in the US. Martin Feldstein says that it does but Robert t Barro says that there is no effect of Social Security on private savings.

Barro estimated a model where Real Per Capita Consumption (Con) is the dependent variable.The independent variables were:

YD_{t}, real per capita disposable income in the current year

YD_{t-1 } real per capita disposable income in the year before the current year

RE_{t }real per capita retained earnings of corporations in the current year (corporate

savings 美国经济学Final exam代考

SUR_{t} real per capita government surplus in the current year (lately of course this has been a negative number)

K_{t }real per capita estimated Capital Stork

ssW_{t } real per capita measure of Social Security wealth

a.Write out the structural model if we assume that there is a caausal linear relationship between the independent variables and the dependdent variable. Use ß’s for the cocfficiennts in the model.

b.Write out Feldstein’s hypothesis in the symbolic notation used in part a.

c.Write out Barro’s hypothesis.

d.How can we test between them? Be specific on the kind of test you would do.

e.The results ofBarГro’s estimation are listed below:

Complete the test you outlined in part d. The data used was annual data from 1939 to 1984. State your conclusion. Be specific about your conclusion about the Social Security/Private Savings controversy.

f.Interpret the coefficient on YD_{t}. What does it tell us about the savings rate?

g. Calculate a 90% confidence interval for K. Interpret.

### 2.(12 points) 美国经济学Final exam代考

For each of the following experiments, decide whether the random variable X is a binomial random variable. Just indicicate if the experiments is BINOMIAL or NOT BINOMIAL. No explanation is needed.

a.From past records, it is known that 5% of laptops manufactured by a certain company will need major repairs within 3 months. Your company has just purchased 8 laptops from this firm. Let X be the number of these las”ps which will require major repairs within 3 months.

b.Let X be the number of fair coins t tossed before the first head is observed. 美国经济学Final exam代考

c.Four of the members of the Goveernor?s Council on Youth Fitness are woomen and six are men. Three of the ten members of this committee will be selected randomly to appear before the state legislature to request funds to support physical education in the elementary schools. Let X be the number of females chosen.

d.Calculate the expected value of X from part a.

e.What is the probability from part c that X = 1?

### 3. (25 points) A researcher was interested in the determinants of the level of alcohol consumption among young adults. 美国经济学Final exam代考

After careful consideration he included the following independent variables.

WHITE Dummy which equals 1 i he young adult is White and O for nonwhite

NUMDEPS Number of dependents that the young adult has

FMPOV Family poverty dummy which equals 1 if his/her family income is below the poverty line

HGCOMP Highest grade completed

WRKING Employed in the labor market 1==yes 0==no

TLDOFF Temporarily laid off 1=yes 0-no

UNEMp Currently unemployed l=yes 0= no

KHOUIS Ikeeping house (not in labor force) 1==yes 0= =no

DVAL500 value of debt over $500

The dependent variable was HNGMON, the number of days “hung over” last month.

The results of this regression arere attached on the next two page. 美国经济学Final exam代考

a.Fill in the STATA comnmands that I have blanked out.There are 12 commands or parts of commands missing.

b.What percent of the sample lives below the poverty line?

c.Iwould say the results of this analysis are fairly disappointing if our goal is to be able: to determine the characterisics that lead to heaavy drinking among young adults.How can I test the hypothesis that heavy drinlking ocurs randomly in he population?Go ahead and conduct the test. What do you conclude from this test?

d.Using α= 0.05, test the hypothesis that race has no effect on drinking against the hypothesis that there are racial differences in the level of heavy drinking. Show the test and state the conclusion.

e.Using a = 0.05,test thehypothesis that race has no effect on drinking against the alternative hypothessis that Whites drink more than Nonwhites.Show the test and state the conclusion.

f.Discuss how and why (d) and (e) differ.

### 4. (10 points) A researcher investigates the relation between miles per gallon of vehicles (mpg) and their weight (in pounds). 美国经济学Final exam代考

She obtains the following fitted regression:

МPG = 45 – .005 * weight. Thus s 45 is the intercept and -.005 is the slope.

a.The researcher is warned that she has omitted a key variable: the horsepowerr of the engine of the car. Higher horsepower engines burn more gas and are more likely found in larger cars. If horsepower is added to the regression, what will happen to the coefficientnt on weight? Be explicit: will the coefficient on weight becone more negative or less negative?Explain.

b.If the rescarcher had run a regression with weight measured in 1000s of pounds, how would the coefficient on weight change? Find its new numerical value.

### 5. (20 points) The Stata output below comes from a regression of hourly wage on education (in years), work èxperience (in years), and whether the person is married (ms, a 0/1 variable with a 1 for married) for a sample of 528 people.

a. Interpret the meaning of the coefficieient on education.

b. Interpret the meaning of the coefficiet on marital status.

c.The R^{2 }is not provided above. Calculate it from the information given. 美国经济学Final exam代考

d. Suppose that a new variable “female” is made available that tells whether the person was female (0/1). You assume that education and expeperience do not have different coefficients for men “]d women. However, you believe that women may have a different intercept from men AND that the impact of marriage on wages may differ for men and women.Write down the specification for a regression equation that would allow you to estimate these differences. ( I want only one equation.)

e.If I had used log(wage) as the dependent variable the coefficients would have changed. Suppose that the coefficients on education became .08 and that on ms became.1respectively. Interpret the meaning of these new values.

### 6.(12 points)

a.What do we mean by the statement “_{β1}, is a descriptive statistic” What does _{β1}, describe?

b.What do we mean by the statement “, is a random v variable”? Why is it random? What is the shape of its distribution? 美国经济学Final exam代考

c.we e use OLS regression to give us the best fitting line (given that the OLS assumptions are met).What do we mean by best-fitting line?

d.what does it mean to say that _{β1}, is an unbiased estimate of _{β1}?