R04 Introduction To Linear Regression

Download as pdf or txt
Download as pdf or txt
You are on page 1of 12

Question 1

An analyst believes that a stock has a beta equal to that of the market as a whole. He wishes
to test this belief using a 0.01 level of significance. He takes a sample of 24 monthly returns
for the stock and identifies that the sample beta is 0.6 with a standard error of 0.21. Using a
student's t-distribution table, the critical value of t was determined to be 2.819 for a sample
of this size and for this level of significance. The t-statistic for testing the null hypothesis that
the stock's beta is equal to the market beta is closest to:
a) 1.90, and at a 0.01 significance level, we can reject the null hypothesis that the
stock's beta is equal to the market's beta.
b) 1.90, and at a 0.01 significance level, we cannot reject the null hypothesis that the
stock's beta is equal to the market's beta.
c) 2.86, and at a 0.01 significance level, we can reject the null hypothesis that the
stock's beta is equal to the market's beta.

Question 2
A statistical software package outputs the following information concerning the relationship
between a company's share price returns and market returns:
Covariance 0.00437
Standard deviation of stock 12.5%
Correlation 0.23
Based on the data in the table, which of the following is closest to the variance of the
market?
a) 0.0231.
b) 0.0546.
c) 0.1520.

Question 3
The relationship between exchange rates (X) and sales (Y) is being examined by an analyst
who compiled the following data from a sample.
∑(X−̅X)2 536
∑(Y−̅Y)2 338
∑(X−̅X)(Y−̅Y) 23.8
Sample size (n) 13
The correlation coefficient between X and Y is closest to:
a) 0.052
b) 0.056
c) 0.671

Question 4
An analyst has used a sample of 100 observations to create a regression equation with one
independent variable. The intercept coefficient is 2.1 with a standard error of 0.4 and the
slope coefficient for the regression is 3.4 with a standard error of 0.7. The variance of the
prediction error is 9. If the value of the independent variable is 5.0, the prediction interval
for the dependent variable at a 95% confidence level is closest to:
a) 1.1 to 37.1
b) 13.1 to 25.1
c) 14.2 to 24.0

Question 5
The sum of the squared errors for a regression is 40 and this is based on a sample size of 30.
The sample standard deviation for the dependent variable (Y) is 2.42. The coefficient of
determination for the two variables is closest to:
a) 0.43
b) 0.66
c) 0.76

Question 6
The relationship between income (X) and debt (Y) is being evaluated by an analyst who has
compiled the following data from a sample.
∑(X−̅X)2 434
∑(Y−̅Y)2 266
∑(X−̅X)(Y−̅Y) 199
Sample size (n) 12
Based on the gathered sample data, the covariance between X and Y is closest to:
a) 16.6
b) 18.1
c) 30.9

Question 7
An analyst is investigating whether the systematic risk of a security is different to that of the
market. She performs a linear regression of daily excess returns of the security versus the
daily excess returns of the market over a period of 250 trading days and calculates the slope
coefficient to be 1.3538 with a standard error of 0.1345. Given a critical t-statistic associated
with a significance level of 5% of approximately 2, which of the following statements is most
accurate?
a) The analyst should reject the null hypothesis that the security has the same
systematic risk as the market.
b) The analyst should fail to reject the null hypothesis that the security has the same
systematic risk as the market.
c) The analyst should accept the alternative hypothesis that the security has the same
systematic risk as the market.

Question 8
The linear regression model most likely assumes:
a) the variance of each error term is normally distributed.
b) the variance of the error term differs for all observations.
c) the error term is correlated across different observations.

Question 9
Which of the following is least likely an assumption of linear regression?
a) The independent variable is random.
b) The expected value of the error term is 0.
c) The error term is uncorrelated across observations.

Question 10
Use the following information to answer the next three questions:
An analyst wishes to identify whether there is a significant positive relationship between
free cash flow to equity and net income at the 10% level of significance. She has taken a
sample of 28 returns and identified a correlation coefficient of 0.3. An extract from the
Student's t-distribution table is as follows:
Degrees of Freedom One-Tailed Probabilities
0.10 0.05 0.025 0.01 0.005
25 1.31635 1.70814 2.05954 2.48510 2.78744
26 1.31497 1.70562 2.05553 2.47863 2.77872
27 1.31370 1.70329 2.05183 2.47266 2.77068
28 1.31253 1.70113 2.04841 2.46714 2.76326
29 1.31143 1.69913 2.04523 2.46202 2.75639
30 1.31042 1.69726 2.04227 2.45726 2.74998
Next, the analyst wishes to use an F-test to determine whether the slope coefficient for the
regression is significant. The sum of the squares data from an ANOVA table for the
regression based on a sample size of 28 is as follows:
Sum of the Squares
Regression 12
Error 75
i.
At a 10% level of significance, the t-statistic for the correlation coefficient is closest to:
a) 1.60 and the analyst will conclude that there is a significant relationship between net
income and free cash flow to equity.
b) 1.60 and the analyst will conclude that there is not a significant relationship between
net income and free cash flow to equity.
c) 1.55 and the analyst will conclude that there is not a significant relationship between
net income and free cash flow to equity.
ii.
In testing whether the slope coefficient for the regression is significant, the F-statistic to be
used is closest to:
a) 2.08
b) 4.16
c) 4.32
iii.
When conducting an F-test for a regression with one independent variable, it is most likely
that the:
a) F-test will always come to the same conclusion on the significance of the slope
coefficient as the t-test, if the same level of significance is used for both tests.
b) F-statistic is calculated as the regression sum of the squares divided by the sum of
the squared errors.
c) test can give a different result from a significance test on the correlation coefficient.

Question 11
An analyst performs a regression with monthly returns on a large-cap mutual fund as the
dependent variable and monthly returns on the market index as the independent variable.
He uses monthly returns data over the last year (in %).
Regression Statistics
Multiple R 0.7589
R-squared 0.576
Standard error 3.8921
Observations 12
Coefficient Standard Error
Intercept −0.254 1.2984
Slope coefficient 0.782 0.215
Statistic Market Index Return Large-Cap Fund Return
Mean 2.45% 1.68%
Standard deviation 6.82% 7.21%
Count 12 12
Given a return on the market index of 8.25%, the 95% prediction interval for the expected
mutual fund return is closest to:
a) −3.0981% to 15.4931%
b) −32.5850% to 44.98%
c) −25.4295% to 32.64%

Question 12
Correlation coefficients computed from sample data are valid as long as:
a) The means and variances of the two variables are finite and constant.
b) The covariance between the two variables is finite and constant.
c) The means and variances of the two variables and the covariance of the two
variables are finite and constant.
Question 13
The linear regression model least likely assumes:
a) uncorrelated error terms.
b) a random independent variable.
c) a linear relationship between the two variables.

Question 14
Part of an ANOVA table for a regression based on a sample size of 30 is as follows:
Sum of the Squares Mean Sum of the Squares
Regression 24 24.000
Error 62 2.214
The correlation coefficient for the regression is closest to:
a) 0.96
b) 0.62
c) 0.53

Question 15
The sum of the squared errors for a regression based on a sample size of 40 is 232. The
standard error of the estimate (SEE) is closest to:
a) 0.40
b) 2.44
c) 2.47

Question 16
Part of an ANOVA table for a regression based on a sample size of 22 is as follows:
Sum of the Squares Mean Sum of the Squares
Regression 50 50.0
Error 90 4.5
The F-statistic for the regression is closest to:
a) 0.56
b) 11.11
c) 13.22

Question 17
Which of the following is least likely to be a limitation of regression analysis?
a) The residual errors are homoskedastic.
b) Regression parameters may change over time.
c) Regression relationships may cease to be useful in the future.
Question 18
The width of a prediction interval is most likely:
a) positively related to the standard error of estimate, but negatively related to the
standard deviation of the independent variable.
b) negatively related to the standard error of estimate, but positively related to the
standard deviation of the independent variable.
c) positively related to the standard error of estimate and to the number of observations
in the sample.

Question 19
The width of a confidence interval is most likely negatively related to:
a) The number of observations in the sample.
b) The standard error of the estimated parameter.
c) The value of the estimated parameter based on sample data.

Question 20
Which of the following linear regression equations has the steepest slope?
a) Y = 1 + 1.45B
b) Y = 1.5 + 1.50B
c) Y = 1 + 1.55B

Question 21
The relationship between two variables, natural gas consumption (X) and temperature (Y), is
being examined by an analyst who has compiled the following data from a sample.
Covariance between X and Y 7.739
Standard deviation of X 4.672
Standard deviation of Y 3.867
Sample size (n) 24
The coefficient of determination for X and Y is closest to:
a) 0.18
b) 0.43
c) 0.66

Question 22
Use the following information to answer the next five questions.
An analyst regresses the bid-ask spread (dependent variable) for a sample of 1,900 stocks
against the natural log of trading volume (independent variable). The results of the
regression are provided below:
ANOVA SS
Regression 18.395
Residual 47.428
Coefficient Standard Error t-Statistic
Intercept 0.62941 0.026635 23.63094
Slope coefficient -0.05248 0.002941 -17.84427
i.
The coefficient of determination is closest to:
a) 0.2795
b) 0.3879
c) 0.7205
ii.
The correlation coefficient (r) is closest to:
a) 0.6228
b) -0.5286
c) 0.5286
iii.
The standard error is closest to:
a) 0.0984
b) 0.1581
c) 0.0250
iv.
The F-stat is closest to:
a) 0.3879
b) 736.14
c) 0.0014
v.
The lower and upper bounds for a 95% confidence interval for the slope coefficient are
closest to:
Lower Bound Upper Bound
A -0.04672 0.05824
B -0.05824 -0.04672
C 0.04672 0.05824
a) Row A
b) Row B
c) Row C

Question 23
Use the following information to answer the next two questions.
An analyst regresses returns for ABC Stock against the market index using monthly returns
data from January 2008 to December 2012. The results of the regression are shown below:
Coefficients Standard Error t-Statistic
Alpha 0.0058 0.0164 0.3537
Beta 1.1232 0.2985 3.7628
i.
The analyst wants to test whether the stock has the same level of systematic risk as the
overall market and structures the following hypotheses:
H0: βABC = 1 versus Ha: βABC ≠ 1
Given a 5% significance level, which of the following statements is most likely?
a) Since the t-stat is greater than the critical t-value, the analyst can reject the null
hypothesis. He should conclude that the stock does not have the same level of
systematic risk as the overall market.
b) Since the t-stat is lower than the critical t-value, the analyst cannot reject the null
hypothesis. He should conclude that the stock has the same level of systematic risk as
the overall market.
c) Since the t-stat is lower than the critical t-value, the analyst cannot reject the null
hypothesis. He should conclude that the stock does not have the same level of
systematic risk as the overall market.
ii.
The analyst wants to test whether the intercept term is positive and structures the following
hypotheses:
H0: b0 ≤ 0 versus Ha: b0 > 0
Given a 5% significance level, which of the following statements is most accurate?
a) Since the t-stat is greater than the positive critical t-value, the analyst can reject the
null hypothesis. He should conclude that the intercept term is positive.
b) Since the t-stat is greater than the negative critical t-value, the analyst cannot reject
the null hypothesis. He should conclude that the intercept not positive.
c) Since the t-stat is lower than the positive critical t-value, the analyst cannot reject the
null hypothesis. He should conclude that the intercept term is not positive.

Question 24
The following two questions relate to the following information:
An analyst is estimating the relationship between stock market returns and inflation. He
uses a simple linear regression with inflation as the independent variable based on 10 years
of monthly data. The regression yields the following statistics:
Sum of squared residuals of the model 0.003
Sum of squared deviations of stock market returns from their mean 0.016
i.
Which of the following values is closest to the standard error of the estimate for this model?
a) 0.00300.
b) 0.00504.
c) 0.01936.
ii.
Which of the following values is closest to the coefficient of determination for this model?
a) 0.0160.
b) 0.1875.
c) 0.8125.

Question 25
Use the following information to answer the next 3 questions:
An analyst believes that the return patterns of medium-cap growth stocks and medium-cap
value stocks are different. He identifies that, for a sample of size 60, the sample correlation
coefficient for returns on medium-cap value stocks and medium-cap growth stocks is 0.01.
The medium-cap growth stocks and medium-cap value stocks include global securities with
currency exposure. The analyst is interested in the correlations between major currencies
found in the portfolio. Details of a correlation matrix of monthly returns in U.S. dollars to
selected foreign currency returns for different time periods are shown below.
1990 - 1999 British Pound Swiss Franc Yen
British Pound 1.00
Swiss Franc 0.32 1.00
Yen 0.41 0.53 1.00
2000 - 2009 British Pound Swiss Franc Yen
British Pound 1.00
Swiss Franc –0.02 1.00
Yen 0.28 0.34 1.00
The analyst also believes that the returns to medium-cap stocks are positively linked to
growth in real GDP. Using a sample of 27 observations, he estimates that the b1 coefficient
for the regression is 0.4 with a standard error of 0.15.
The analyst decides to test this belief by testing the hypothesis that returns of mid-cap
growth stocks are not linked to growth in real GDP. He opts to complete this test using a 1.0
percent level of significance. He has gathered a portion of the Student's t-distribution to use
in his test:
Degrees of Freedom One-Tailed Probabilities
0.10 0.05 0.025 0.01 0.005
25 1.31635 1.70814 2.05954 2.48510 2.78744
26 1.31497 1.70562 2.05553 2.47863 2.77872
27 1.31370 1.70329 2.05183 2.47266 2.77068
28 1.31253 1.70113 2.04841 2.46714 2.76326
29 1.31143 1.69913 2.04523 2.46202 2.75639
30 1.31042 1.69726 2.04227 2.45726 2.74998
i.
With regard to the analyst's belief that the returns of medium-cap growth stocks and
medium-cap value stocks are different, his belief is most likely:
a) supported by his sample.
b) not supported by his sample.
c) either supported or not supported, but it depends on the time period examined.
ii.
With respect to the correlation matrix of monthly returns in U.S. dollars to holding British
pounds, Swiss francs, or yen for different time periods, it is most likely the case that:
a) correlations between currencies increased over time.
b) diversification benefits from investing in different currencies increased over time.
c) the correlation between the Swiss franc and yen derived returns is lower than for other
currencies.
iii.
With respect to his hypothesis test on the relationship between returns of mid-cap growth
stocks and growth in real GDP, the analyst will most likely conclude that, at a 1.0 percent
level of significance, the returns of mid-cap growth stocks are:
a) linked to growth in real GDP because the calculated t-statistic of 4.00 falls outside the
t-critical range for the null hypothesis that the b1 coefficient could be 0.
b) linked to growth in real GDP because the calculated t-statistic of 2.67 falls outside the
t-critical range for the null hypothesis that the b1 coefficient could be 0.
c) not linked to growth in real GDP because the calculated t-statistic of 2.67 does not fall
outside the t-critical range for the null hypothesis that the b1 coefficient could be 0.

Question 26
When making a prediction of a dependent variable based on a regression with one
independent variable, uncertainty arises most likely due to:
a) the error term only.
b) the parameters only.
c) the error term and the parameters used.

Question 27
Which of the following is most likely regarding the use of confidence intervals to conduct
hypothesis tests?
a) In a confidence interval, we aim to determine whether the hypothesized value of the
population parameter lies within the interval, where the interval is based around the
sample statistic.
b) In a confidence interval, we aim to determine whether the sample statistic lies within
the interval, where the interval is based around the hypothesized value of the
population parameter.
c) In a confidence interval, we aim to determine whether the hypothesized value of the
population parameter lies within the interval, where the interval is based around the
sample test statistic.
Question 28
The next two questions relate to the following information:
An analyst is conducting a linear regression of the total return of a collateralized commodity
futures index against inflation, using 48 monthly returns in percent, deriving the following
results:
R-squared 0.421
Standard error of estimate 2.595
Observations 48

Coefficient Standard error t-statistic


Intercept 0.123 0.231 0.532
Inflation 0.858 0.391 2.194

Statistic Inflation
Mean 0.208
Standard deviation 0.899
i.
What is the predicted total return on the collateralized commodity futures index for an
inflation reading of 0.25%?
a) 0.125%
b) 0.338%
c) 12.515%
ii.
The analyst who derived the regression model makes the following two statements:
Statement 1:
“When establishing a confidence interval for the predicted total return of the collateralized
commodity index given an expected inflation reading, the center of the interval will always
be the predicted total return of the collateralized commodity index using the model.”
Statement 2:
“The width of the confidence interval for the predicted total return of the collateralized
commodity index given an expected inflation reading will increase as the expected inflation
reading moves away from the mean inflation reading.”
How many of the analyst's statements are correct?
a) Neither.
b) One.
c) Both.

Question 29
Which of the following statements is least likely to be an assumption of linear least squares
regression analysis?
a) The independent variable is random.
b) The expected value of the error term in the model is zero.
c) The variance of the error term is constant for all observations.

You might also like