Scribd
Upload
16 views2 pages

Regression Analysis Problem Set 3

This document contains a problem set with 5 multiple choice questions and 5 short answer questions related to regression analysis and hypothesis testing. The multiple choice questions assess understanding of parameters in a regression model and interpreting estimated regression results. The short answer questions involve calculating R-squared, hypothesis testing of individual coefficients and all regressors jointly, testing differences between coefficients, and testing the hypothesis of constant returns to scale using results from two estimated regressions.

Uploaded by

dxd032
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
16 views2 pages

Regression Analysis Problem Set 3

This document contains a problem set with 5 multiple choice questions and 5 short answer questions related to regression analysis and hypothesis testing. The multiple choice questions assess understanding of parameters in a regression model and interpreting estimated regression results. The short answer questions involve calculating R-squared, hypothesis testing of individual coefficients and all regressors jointly, testing differences between coefficients, and testing the hypothesis of constant returns to scale using results from two estimated regressions.

Uploaded by

dxd032
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

Problem set 3

This exercise is based on material found in Section 3.

1. In the regression model Yt =  +  X t +  Zt + Wt +  t , which parameter indicates the


influence that variable Z t has on variable Yt ?
(a)  ;
(b)  ;
(c)  ;
(d)  ;
(e) None of the above.

2. For the following estimated regression, in which we assume that all the parameters are
statistically significant, which of the following statements is correct?

Yˆt = 2.554 − 0.839 X t − 0.530 Z t + 0.027Wt

(a) As variable 𝑋 increases by 1 unit, variable 𝑌 decreases by 8.39 units;


(b) As variable 𝑊 increases by 1 unit, variable 𝑌 increases by 0.27 units;
(c) As variable 𝑍 increases by 1 unit, variable 𝑌 decreases by 0.53 units;
(d) Variable 𝑋 has a positive effect on 𝑌;
(e) None of the above.

3. OLS is used to estimate the following regression

𝑌𝑡 = 𝛽0 + 𝛽1 𝑋1𝑡 + 𝛽2 𝑋2𝑡 + 𝛽3 𝑋3𝑡 + 𝛽4 𝑋4𝑡 + 𝜀𝑡


for which the estimated regression is

𝑌̂𝑡 = 2.837 − 1.481𝑋1𝑡 + 1.181𝑋2𝑡 + 0.186𝑋3𝑡 + 0.257𝑋4𝑡


(2.000) (0.987) (0.690) (0.134) (0.102)
where 𝑅𝑆𝑆 = 0.1277, 𝑇𝑆𝑆 = 0.8531, the sample size is 22 and the standard errors are given
in parentheses.

a) Calculate the R2 and the R 2 statistics.


b) Test the null hypothesis 𝐻0 : 𝛽3 = 0 against 𝐻1 : 𝛽3 ≠ 0 at the 5% level of significance.
c) Test the joint significance of all regressors at the 5% significance level.
d) Test the null hypothesis 𝐻0 : 𝛽2 = 1 against 𝐻1 : 𝛽2 > 1 at the 5% level of significance.
e) Another regression 𝑌𝑡 + 𝑋1𝑡 = 𝛽0 + 𝛽2 𝑋2𝑡 + 𝛽4 𝑋4𝑡 + 𝜀𝑡 is estimated with the results

𝑌𝑡 + 𝑋1𝑡 = −0.738 + 0.799𝑋2𝑡 + 0.261𝑋4𝑡 + 𝜀̂𝑡

where the 𝑅𝑆𝑆 = 0.6788. Test the hypothesis that in the original regression 𝐻0 : 𝛽1 =
−1, 𝛽3 = 0 against 𝐻1 : 𝛽1 ≠ −1 and/or 𝛽3 ≠ 0 at the 5% level of significance.
Q4 applies regression to an economic context and tests hypotheses about this context.

4. A firm uses two main factors of production, capital K and labour L, to produce output Q
according to the following Cobb-Douglas production function

𝑄 = 𝐹(𝐾, 𝐿) = 𝐴𝐾𝛽1 𝐿𝛽2

where 𝐴, 𝛽1 and 𝛽2 are constants.

a) In terms of 𝛽1 and 𝛽2, determine the degree of homogeneity of this production function
(this term is related to the concept of returns to scale).
b) Show that this function can be made into a linear function of the parameters 𝛽1 and 𝛽2
using the natural logarithm.

From this firm an econometrician collects 33 years of data on output and the units of capital
and labour used to produce this output. She estimates this log-linear production function using
Ordinary Least Squares (OLS), producing the following results

qt = 0.596 + 0.255kt + 0.753lt , RSS = 0.0025


( 0.241) ( 0.039 ) ( 0.042 )
where qt is the logarithm of output, k t is the logarithm of capital, lt is the logarithm of
labour, RSS is the residual sum of squares and standard errors are given in parentheses.

c) Determine the estimated value for 𝐴.

d) Test the significance of the coefficient on k t at the 5% level, stating clearly the null and
alternative hypotheses, the test statistic used and the rules upon which you base your
decision.

The econometrician then estimates another regression and produces the following results

qt − kt = 0.367 + 0.728 ( lt − kt ) +  t , RSS = 0.0028

e) Test the hypothesis, at the 5% level, that the firm has constant returns to scale using the
results of the two regressions, stating clearly the null and alternative hypotheses, the test
statistic used and the rules upon which you base your decision.

Common questions

Powered by AI

Taking natural logarithms of the Cobb-Douglas function \( Q = A K^{\beta_1} L^{\beta_2} \), we obtain \( \ln(Q) = \ln(A) + \beta_1 \ln(K) + \beta_2 \ln(L) \). This transforms the model into a linear form suitable for OLS regression, allowing for estimation of \( \beta_1 \) and \( \beta_2 \).

The residual sum of squares (RSS) measures the discrepancy between observed and predicted values. A lower RSS indicates a better fit of the model to the data, reflecting higher model accuracy and explaining more variance .

In the regression model, the influence of the variable \( Z_t \) on \( Y_t \) is represented by the parameter \( \gamma \).

To test if a firm has constant returns to scale, verify if the sum of the coefficients in log-linear production regressions equals one. Use a t-test comparing the sum to 1, considering the standard errors of the coefficients. If the test statistic falls within the critical region, reject the null hypothesis of constant returns .

Returns to scale are determined by summing \( \beta_1 \) and \( \beta_2 \). If the sum is 1, there are constant returns to scale; if greater than 1, increasing returns to scale; and if less than 1, decreasing returns to scale are indicated .

To test the null hypothesis \( H_0: \beta_3 = 0 \) against \( H_1: \beta_3 \neq 0 \), one uses the t-test: \( t = \frac{\hat{\beta}_3}{SE(\hat{\beta}_3)} \). Substituting \( \hat{\beta}_3 = 0.186 \) and its standard error, if \( |t| \) exceeds the critical value for a two-tailed test at 5% level, one rejects the null hypothesis .

In the estimated regression model, the parameter \(-0.839\) on \(X\) indicates that as \(X\) increases by 1 unit, \(Y\) decreases by \(0.839\) units. This suggests a negative relationship between \(X\) and \(Y\).

Use the F-test for joint significance to test if all regressors contribute to explaining variation in \( Y \). Compute the F-statistic based on \( R^2 \), number of parameters, and sample size. If the statistic exceeds the critical value from the F-distribution, reject the null hypothesis .

The coefficient of determination \( R^2 \) is calculated using the formula \( R^2 = 1 - \frac{RSS}{TSS} \), where \( RSS \) is the residual sum of squares and \( TSS \) is the total sum of squares. It conveys the proportion of the variance in the dependent variable that is predictable from the independent variables .

Tests like the t-test assess the significance of regression coefficients in log-linear models. By comparing the estimated coefficient to its standard error, the t-statistic shows if an independent variable has a significant effect. This supports understanding variable impacts in econometrics .

You might also like