Professional Documents
Culture Documents
1. The linear relationship between two variables (y and x) can be represented by the equation
. Which of the following statements is true?
(I) Parameter a is termed the intercept
(II) Parameter a is termed the slope
(III) Parameter b is termed the gradient
(IV) Parameter b is termed the constant.
(a) I and IV only
(b)* I and III only
(c) II and III only
(d) II and IV only.
2. Assume that the relationship between a company’s stock price (y) and dividends paid per share
(x) is linear. If the slope of the equation is 0.50 and the intercept is 30, what would be the
expected stock price if the dividend paid was 3?
(a) 33
(b) 30.50
(c)* 31.5
(d) 30.
3. Which of the following values are closes to the roots of the following quadratic equation:
?
(a) 0 and 4
(b) 1 and 4
(c) 0.5 and 3
(d)* 0.3 and 3.7.
a. 21 x × x
5 3
b. 21 x15
c. * 21x 8
d. 21x .
53
a. * –2 and 1
b. –1 and 2
c. –2 and 2
d. –2 (repeated).
(a) –4 and 2
(b) * –3.65 and 1.65
(c) –7.3 and 3.3
(d) Both complex.
c. *x 3
d. 3x × 3x × 3x.
d. x.
12. Writing out all the terms in the expression would lead to:
a. *x +x +x +x
11 12 21 22
b. x ×x ×x ×x
11 12 21 22
c. x +x
1 2
d. x +x .
11 22
d. 5y . 5
14. What is the (first order) derivative of the function ?
(a)
(b)
(c)*
(d) .
a. 4/(4x-2)
b. (4x-2)e 4
c. (4x-2)e 4x-2
d. * 4e .
4x-2
(a)*
(b)
(c)
(d) .
21. For two conformable matrices A and B, expanding the parentheses of (AB) gives:
-1
a. AB
-1 -1
b. *BA -1 -1
c. BA
d. AB.
(a)
(b)*
(c)
(d) .
25. The point where the capital market line is tangential to the efficient frontier is
(a) The point where the portfolio returns are minimised
(b) The point where the portfolio returns are maximised
(c) The point where the portfolio’s Sharpe ratio is minimised
(d)* The point where the portfolio’s Sharpe ratio is maximised.
27. Consider the following data series: 11, 10, 6, 8, 4, 3, 7. What is its semi-interquartile range of
this series?
(a) 6
(b) 5
(c) 4
(d)* 3.
(a) 2 and –4
(b) 0.5 and –2
(c) 2 and 2
(d) * This function has no real roots.
31. (x ) simplifies to
3 2
(a) x
5
(b) * x 6
(c) x
(d) The expression cannot be simplified.
32. Which of the following three equations is/are correct regarding the summation operator?
i=1Kxi+i=1Kzi=i=1Kxi+zi (i)
i=1Kxizi=i=1Kxii=1Kzi (ii)
i=1Kcxi=ci=1Kxi (iii)
a. (iii) only
b. * (i) and (iii) only
c. (i) and (ii) only
d. (i), (ii) and (iii).
35. Using the chain rule or otherwise, what is the (first) derivative of the following function? y =
(2x + 4x – 6)
2 3
(a) 3(2x + 4x – 6)
2 2
(b) (4x + 4) 2
Chapter 2
Correct answers denoted by an asterisk.
5. Data that have been collected on one or more variables at a single point in time is referred to as
(a)* Cross-sectional data
(b) Time-cross-sectional data
(c) Time series data
(d) Panel data.
7. An individual invested £106.40 in the stock market and the value of his investment two years later is £138.22.
What are the simple and continuously compounded returns on his investment?
(a) 26% and 30%, respectively
(b) –29% and -34%, respectively
(c)* 30% and 26%, respectively
(d) 30% and 30%, respectively.
8. An individual has £10000 capital to invest in the stock market. He invests 30% of his capital in stock A, 25% in
stock B and 45% in Stock C. What is the return on his/her portfolio assuming that the simple returns on stocks A, B
and C are 5%, 10% and 12%, respectively?
(a)* 9.0%
(b) 9.7%
(c) 9.3%
(d) 9%.
The average nominal annual rent in the US denominated in dollars and the CPI (2008 levels) are given in the table
below:
Year Average annual rent (US Dollars) CPI (2008 levels)
200 9908 100
8
200 9998 99.7
9
201 10012 101.3
0
201 10180 104.5
1
201 10396 106.7
2
11. The numerical score assigned to the credit rating of a bond is best described as what type of number?
(a) Continuous
(b) Cardinal
(c)* Ordinal
(d) Nominal.
12. Suppose that we wanted to sum the 2007 returns on ten shares to calculate the return on a portfolio over that
year. What method of calculating the individual stock returns would enable us to do this?
(a)* Simple
(b) Continuously compounded
(c) Neither approach would allow us to do this validly
(d) Either approach could be used and they would both give the same portfolio return.
13. If we wish to compare the spread of two series with considerably different mean values, which of the following
measures would be the most appropriate?
(a) The semi-interquartile range
(b) The standard deviation
(c) The range
(d) * The coefficient of variation.
14. For a series with a negative skew in its distribution (a long left tail), which of the following best describes the
relationship between its measures of central tendency?
(a) mean > median > mode
(b) * mode > median > mean
(c) mode > mean > median
(d) median > mode > mean.
15. Which of the following statements is TRUE concerning the correlation between two series?
(a) * It is unit-free
(b) It scales with the product of the units of the two series
(c) It scales with the ratio of the units of the two series
(d) It will take the value –1 if there is no association between the two series.
16. What is the sum of the following infinite set of terms? 5, 2.5, 1.25, 0.625, …
(a) Infinity
(b) 5
(c) 20
(d) * 10.
17. What is the sum of the first 12 terms in the following sequence? 12, 24, 48, …
(a) * 49,140
(b) 24,576
(c) 768
(d) 98,292.
18. If I have £10,000 now and I want it to grow by 50% within eight years, what interest rate, compounded annually,
is required (to one decimal place)?
(a) * 5.2%
(b) 6.2%
(c) 4.6%
(d) 7.4%.
19. If a savings account pays a nominal interest rate of 10% per year, compounded monthly, what is the effective
interest rate to one decimal place?
(a) 11.2%
(b) 9.5%
(c) 10.0%
(d) * 10.5%.
20. If you place £10,000 in a savings account, how long would it take to reach £20,000 assuming an annual interest
rate of 3%, continuously compounded, rounded to the nearest year?
(a) 26
(b) 34
(c) *23
(d) 20.
21. What would be a fair price to pay today, to the nearest dollar, for a zero coupon bond having exactly six years to
maturity and to be redeemed at $1000 if the annual discount rate is 6%?
(a) $1000
(b) * $747
(c) $864
(d) $553.
22. Which of the following statements is FALSE concerning the internal rate of return?
(a) For projects where the cashflow payments change sign, there can be more than one internal rate of return
(b) The internal rate of return is the discount rate that sets the net present value of all of the cashflows to be received
equal to the asset’s purchase price
(c) * In order to calculate an internal rate of return, all of the incoming cashflows must be identical
(d) We cannot calculate a different internal rate of return for each cashflow.
Chapter 3
Correct answers denoted by an asterisk.
2. What does a positive linear relationship between x and y in a simple regression imply?
(a) Increases in the independent variable are usually accompanied by increases in the regressor
(b) The relationship between x and y cannot be explained by a straight line
(c) Decreases in the independent variable is usually accompanied by increases in the regressors
(d)* Increases in the regressor are usually accompanied by increases in the dependent variable.
3. Which of these is NOT a reason for adding a disturbance term to a regression model ?
(a) Some determinants of the effect variable may be omitted from the model
(b) Some determinants of the effect variable may be unobservable
(c)* Some determinants of the independent variable may be omitted from the model
(d) There may be errors in the way that the dependent variable is measured which cannot be modelled.
4. Which of these is not a standard method for estimating econometric models?
(a) Ordinary least squares
(b) The method of moments
(c)* Method of generalised squared moments
(d) Maximum likelihood.
5. The method of estimating econometric models which involves fitting a line to the data by minimising the sum of
squared residuals is the
(a)* Method of ordinary least squares
(b) Method of moments
(c) Method of generalised squared moments
(d) Method of maximum likelihood.
6. Suppose you have 5-year annual data on the excess returns on a fund manager’s portfolio (‘fund ABC’) and the
excess returns on a market index (where is the return on fund ABC, is the risk-free rate and is the
return on the market index):
Year Excess return on fund ABC Excess return on market
t index
1 14.0 16.0
2 32.0 21.7
3 11.6 6.0
4 21.2 16.2
5 17.4 11.0
7. Given the data in Question 6, what is the estimated beta ( ) of Fund ABC?
(a) 3.1
(b) 2.1
(c)* 1.1
(d) None of the above.
8. Suppose that the unbiased estimator of the standard deviation of the disturbance (s) is 5.1. What is the nearest
value to the standard errors of the estimated CAPM alpha ( ) of Fund ABC from Question 6?
(a) 3.5
(b) 4.5
(c) 5.5
(d)* 6.5.
9. The estimated alpha ( ) and beta ( ) of a rival fund, Fund DEF, are 2.3 and 3.1, respectively. If the expected
market risk premium is 12%, what would we expect the excess return of Fund DEF to be?
(a)* 39.5%
(b) 30.7%
(c) 5.4%
(d) 64.8%.
10. What is the most appropriate interpretation of the assumption concerning the regression
disturbance terms?
(a) The errors are nonlinearly independent of one another
(b) The errors are linearly dependent of one another
(c) The covariance of the errors is constant and finite over all its values
(d)* The errors are linearly independent of one another.
11. The estimators and determined by OLS will be the Best Linear Unbiased Estimators (BLUE) if which of
the following assumptions hold?
(I) The errors have zero mean
(II) The variance of the errors is constant and finite over all values of the independent variable(s)
(III) The errors are linearly independent of one another
(IV)There is no relationship between the error and corresponding independent variables
(a) I and II only
(b) I, II and III only
(c) II, III and IV only
(d)* I, II, III, and IV.
13. Using the test of significance approach, what is the test statistic value of a hypothesis to test whether the true
value of statistically different from zero?
(a)* 1.10
(b) 0.91
(c) –0.62
(d) Cannot say without more information.
14. Assuming there are 1000 observations in your sample, what are the test statistic and critical value of a two-sided
hypothesis test of whether the true value of statistically different from zero be given a 5% significance level?
(a)* 1.10 and 1.96, respectively
(b) 0.91 and 1.65, respectively
(c) –0.62 and 1.96, respectively
(d) Cannot say without more information.
15. Consider a bivariate regression model with coefficient standard errors calculated using the usual formulae.
Which of the following statements is/are correct regarding the standard error estimator for the slope coefficient?
i. It varies positively with the square root of the residual variance (s)
ii. It varies positively with the spread of X about its mean value
iii. It varies positively with the spread of X about zero
iv. It varies positively with the sample size T
a. * (i) only
b. (i) and (iv) only
c. (i), (ii) and (iv) only
d. (i), (ii), (iii) and (iv).
16. In a time-series regression of the excess return of a mutual fund on a constant and the excess return on a market
index, which of the following statements should be true for the fund manager to be considered to have ‘beaten the
market’ in a statistical sense?
a. * The estimate for α should be positive and statistically significant
b. The estimate for α should be positive and statistically significantly greater than the risk-free rate of return
c. The estimate for β should be positive and statistically significant
d. The estimate for α should be negative and statistically significant.
18. The type I error associated with testing a hypothesis is equal to
(a) One minus the type II error
(b) The confidence level
(c) * The size of the test
(d) The size of the sample.
19. Which of the following is a correct interpretation of a ‘95% confidence interval’ for a regression parameter?
(a) * We are 95% sure that the interval contains the true value of the parameter
(b) We are 95% sure that our estimate of the coefficient is correct
(c) We are 95% sure that the interval contains our estimate of the coefficient
(d) In repeated samples, we would derive the same estimate for the coefficient 95% of the time.
20. Which of the following statements is correct concerning the conditions required for OLS to be a usable
estimation technique?
a. * The model must be linear in the parameters
b. The model must be linear in the variables
c. The model must be linear in the variables and the parameters
d. The model must be linear in the residuals.
21. Which of the following is NOT a good reason for including a disturbance term in a regression equation?
(a) It captures omitted determinants of the dependent variable
(b) * To allow for the non-zero mean of the dependent variable
(c) To allow for errors in the measurement of the dependent variable
(d) To allow for random influences on the dependent variable.
22. Which of the following is NOT correct with regard to the p-value attached to a test statistic?
(a) * p-values can only be used for two-sided tests
(b) It is the marginal significance level where we would be indifferent between rejecting and not rejecting the null
hypothesis
(c) It is the exact significance level for the test
(d) Given the p-value, we can make inferences without referring to statistical tables.
23. Which one of the following is NOT an assumption of the classical linear regression model?
a. The explanatory variables are uncorrelated with the error terms.
b. The disturbance terms have zero mean
c. * The dependent variable is not correlated with the disturbance terms
d. The disturbance terms are independent of one another.
24. Which of the following is the most accurate definition of the term ‘the OLS estimator’?
(a) It comprises the numerical values obtained from OLS estimation
(b) * It is a formula that, when applied to the data, will yield the parameter estimates
(c) It is equivalent to the term ‘the OLS estimate’
(d) It is a collection of all of the data used to estimate a linear regression model.
25. Two researchers have identical models, data, coefficients and standard error estimates. They test the same
hypothesis using a two-sided alternative, but researcher 1 uses a 5% size of test while researcher 2 uses a 10% test.
Which one of the following statements is correct?
a. Researcher 2 will use a larger critical value from the t-tables
b. * Researcher 2 will have a higher probability of type I error
c. Researcher 1 will be more likely to reject the null hypothesis
d. Both researchers will always reach the same conclusion.
26. Consider an increase in the size of the test used to examine a hypothesis from 5% to 10%. Which one of the
following would be an implication?
a. * The probability of a Type I error is increased
b. The probability of a Type II error is increased
c. The rejection criterion has become more strict
d. The null hypothesis will be rejected less often.
27. What is the relationship, if any, between the normal and t-distributions?
a. A t-distribution with zero degrees of freedom is a normal
b. A t-distribution with one degree of freedom is a normal
c. * A t-distribution with infinite degrees of freedom is a normal
d. There is no relationship between the two distributions.
Chapter 4
Correct answers denoted by an asterisk.
1. Consider a standard normally distributed variable, a t-distributed variable with d degrees of
freedom, and an F-distributed variable with (1, d) degrees of freedom. Which of the
following statements is FALSE?
a. The standard normal is a special case of the t-distribution, the square of which is a special
case of the F-distribution
b. * Since the three distributions are related, the 5% critical values from each will be the same
c. Asymptotically, a given test conducted using any of the three distributions will lead to the
same conclusion
d. The normal and t- distributions are symmetric about zero while the F-distribution takes only
positive values.
(ii) β = 1 and β β = 1
2 3+ 4
(iii) β β = 13 4
(iv) β -β -β = 1.
2 3 4
Suppose that a researcher wishes to test the null hypothesis: β = 1 and β β = 1. The
2 3 + 4
TABULATED value of the F-distribution that we would compare the result of testing this
hypothesis with at the 10% level would be approximately
(a) 19.48
(b) 2.76
(c) * 2.37
(d) 3.11.
6. What is the relationship, if any, between t-distributed and F-distributed random
variables?
a. A t-variate with z degrees of freedom is also an F(1, z)
b. * The square of a t-variate with z degrees of freedom is also an F(1, z)
c. A t-variate with z degrees of freedom is also an F(z, 1)
d. There is no relationship between the two distributions.
7. Which one of the following statements must hold for EVERY CASE concerning the residual
sums of squares for the restricted and unrestricted regressions?
a. URSS > RRSS
b. URSS ≥ RRSS
c. RRSS > URSS
d. * RRSS ≥ URSS.
8. Which one of the following is the most appropriate as a definition of R in the context that the
2
9. Suppose that the value of R for an estimated regression model is exactly one. Which of the
2
iv. The regression F-test will be the same for the two models.
11. Which of the following are often considered disadvantages of the use of adjusted R as a
2
ii. Adjusted R often leads to large models with many marginally significant or marginally
2
insignificant variables
iii. Adjusted R cannot be compared for models with different explanatory variables
2
iv. Adjusted R cannot be compared for models with different explained variables.
2
(a) * I only
(b) I and II only
(c) I and III only
(d) I, II and III.
13. If you are interested in conducting a multiple hypotheses test to determine whether and
are both unity for a regression , what would the restricted
regression be?
(a)
(b)
(c)*
(d) .
14. What would the restricted regression be if you are interested in testing the null hypothesis
and against the alternative hypothesis or for a regression
,?
(a)*
(b)
(c)
(d) .
15. Assuming that the restricted sum of squares of the restricted regression in Question 14 is
436.1 and the unrestricted sum of squares is 397.2, what would the conclusion of the hypothesis
test be? (The significance level is 5%.)
(a)* Reject the null hypothesis
(b) Do not reject the null hypothesis
(c) Reject the alternative hypothesis
(d) Cannot say.
16. Which of these statements is a characteristic of the stepwise regression procedure?
(I) It chooses the jointly most ‘important’ explanatory variable from a set of candidate variables
(II) It can start with no variables in the regression and then it selects first the variable with the
lowest p-value
(III) It can start with no variables in the regression and then it selects first the variable with the
highest p-value
(a) I only
(b) II only
(c) III only
(d)* Both I and II.
17. Trying many variables in a regression without basing the selection of candidate variables on
a financial or economic theory is popularly referred to as
(a) Data fitting
(b) Data clipping
(c)* Data mining
(d) None of the above.
18. Why is R a commonly used and perhaps better measure of how well a regression model fits
2
(c)* The RSS depends on the scale of the dependent variable whereas the R does not
2
(d) The RSS depends on the scale of the independent variable whereas the R does not. 2
19. How can the two models be validly compared to determine the model that better represents
the data y ?
t
20. What is the relevant encompassing model required to compare the two regression models?
(a)*
(b)
(c)
(d) Encompassing models cannot be used to compare these specifications.
Chapter 5
Correct answers denoted by an asterisk.
a. 1.99
b. 2.70
c. * 7.81
d. 8.56.
2. Which of the following would NOT be a potential remedy for the problem of multicollinearity
between regressors?
(a) Removing one of the explanatory variables
(b) * Transforming the data into logarithms
(c) Transforming two of the explanatory variables into ratios
(d) Collecting higher frequency data on all of the variables.
3. Which of the following conditions must be fulfilled for the Durbin–Watson test to be valid?
(i) The regression includes a constant term
(ii) The regressors are non-stochastic
(iii) There are no lags of the dependent variable in the regression
(iv) There are no lags of the independent variables in the regression.
4. If the residuals of a regression on a large sample are found to be heteroscedastic which of the
following might be a likely consequence?
(i) The coefficient estimates are biased
(ii) The standard error estimates for the slope coefficients may be too small
(iii) Statistical inferences may be wrong.
(a) (i) only
(b) * (ii) and (iii) only
(c) (i), (ii), and (iii)
(d) (i) and (ii) only.
5. The value of the Durbin–Watson test statistic in a regression with 4 regressors (including the
constant term) estimated on 100 observations is 3.6. What might we suggest from this?
(a) The residuals are positively autocorrelated
(b) * The residuals are negatively autocorrelated
(c) There is no autocorrelation in the residuals
(d) The test statistic has fallen in the intermediate region.
6. Which of the following is NOT a good reason for including lagged variables in a regression?
(a) Slow response of the dependent variable to changes in the independent variables
(b) Over-reactions of the dependent variables
(c) The dependent variable is a centred moving average of the past 4 values of the series
(d) * The residuals of the model appear to be non-normal.
(a) y = β + β X + β X
1 2 2 3 3
(b) y = β + β X + β X
t 1 2 2t 3 3t
(c) y = - (β / β ) X - (β / β )X
2 1 2 3 1 3
8. Which of the following would you expect to be a problem associated with adding lagged
values of the dependent variable into a regression equation?
(a) * The assumption that the regressors are non-stochastic is violated
(b) A model with many lags may lead to residual non-normality
(c) Adding lags may induce multicollinearity with current values of variables
(d) The standard errors of the coefficients will fall as a result of adding more explanatory
variables.
9. A normal distribution has coefficients of skewness and excess kurtosis which are, respectively,
(a) * 0 and 0
(b) 0 and 3
(c) 3 and 0
(d) Will vary from one normal distribution to another.
10. Which of the following would probably NOT be a potential ‘cure’ for non-normal residuals?
(a) * Transforming two explanatory variables into a ratio
(b) Removing large positive residuals
(c) Using a procedure for estimation and inference which did not assume normality
(d) Removing large negative residuals.
11. What would be the consequences for the OLS estimator if autocorrelation is present in a
regression model but ignored?
(a) It will be biased
b. It will be inconsistent
c. * It will be inefficient
d. All of (a), (b), and (c) will be true.
12. If OLS is used in the presence of heteroscedasticity, which of the following will be likely
consequences?
i. Coefficient estimates may be misleading
ii. Hypothesis tests could reach the wrong conclusions
iii. Forecasts made from the model could be biased
iv. Standard errors may inappropriate.
13. If a residual series is negatively autocorrelated, which one of the following is the most likely
value of the Durbin–Watson statistic?
a. Close to zero
b. Close to two
c. * Close to four
d. Close to one.
14. If the residuals of a model containing lags of the dependent variable are autocorrelated,
which one of the following could this lead to?
a. Biased but consistent coefficient estimates
b. * Biased and inconsistent coefficient estimates
c. Unbiased but inconsistent coefficient estimates
d. Unbiased and consistent but inefficient coefficient estimates.
15. Which one of the following is NOT a symptom of near multicollinearity?
a. The R value is high
2
b. The regression results change substantively when one particular variable is deleted
c. * Confidence intervals on parameter estimates are narrow
d. Individual parameter estimates are insignificant.
16. Which one of the following would be the most appropriate auxiliary regression for a Ramsey
RESET test of functional form?
(a) *
b.
c.
d. .
17. If a regression equation contains an irrelevant variable, the parameter estimates will be
a. * Consistent and unbiased but inefficient
b. Consistent and asymptotically efficient but biased
c. Inconsistent
d. Consistent, unbiased and efficient.
18. Put the following steps of the model-building process in the order in which it would be
statistically most appropriate to do them:
(i) Estimate model
(ii) Conduct hypothesis tests on coefficients
(iii) Remove irrelevant variables
(iv) Conduct diagnostic tests on the model residuals.
19. Test statistics for the LM test and the Wald test are usually constructed to follow a
(a)* χ distribution and F-distribution, respectively
2
(a) I only
(b) I and II
(c)* I, II, and III
(d) I, II, III, and IV.
(A) (B)
25. The graphs above are time series plots of residuals from two separate regressions. Which of
these combinations is true?
(a)* A shows negative autocorrelation and B shows positive autocorrelation
(b) A shows positive autocorrelation and B shows negative autocorrelation
(c) A shows heteroscedasticity and B shows homoscedasticity
(d) A shows homoscedasticity and B shows heteroscedasticity.
26. Assuming a researcher runs the following regression where is residual from
a regression. If the researcher conducts a hypothesis test with null hypothesis of
against an alternative hypothesis of , what type of test is he or she conducting?
(a) Test for heteroscedasticity
(b)* Test for autocorrelation
(c) Test for non-normality
(d) Test for homoscedasticity.
(a) I only
(b) I and II only
(c)* I, II, and III only
(d) I, II, III, and IV.
30. Which of the following statements are true about parameter stability tests?
(I) Parameter stability tests test the assumption that the estimated parameters of a model are
constant for the entire sample
(II) Chow test and predictive failure tests are two types of parameter stability tests
(III) Backward and forward predictive failure tests are two types of parameter stability tests
(IV) Parameter stability tests examine violations of the classical linear regression model
assumptions.
(a) I only
(b) I and II only
(c)* I, II, and III only
(d) I, II, III, and IV.
Chapter 6
Correct answers denoted by an asterisk.
(a) * 0.6
(b) 0.3
(c) 0.0
(d) 0.4.
If we believe that the true DGP can be approximated by the exponential smoothing model, what
would be an appropriate 2-step-ahead forecast for X? (i.e., a forecast of X made at time t)
t+2
(a) 0.2
(b) * 0.23
(c) 0.5
(d) There is insufficient information given in the question to form more than a one-step- ahead
forecast.
a. 0.4
b. 0.0
c. * 0.07
d. –0.1.
4. Which of the following sets of characteristics would usually best describe an autoregressive
process of order 3 (i.e., an AR(3))?
5. A process, x , which has a constant mean and variance, and zero autocovariance for all non-
t
6. Which of the following conditions must hold for the autoregressive part of an ARMA model
to be stationary?
(a) * All roots of the characteristic equation must lie outside the unit circle
(b) All roots of the characteristic equation must lie inside the unit circle
(c) All roots must be smaller than unity
(d) At least one of the roots must be bigger than one in absolute value.
for y?
a. * The current value of y
b. Zero
c. The historical unweighted average of y
d. An exponentially weighted average of previous values of y.
9. Consider a series that follows an MA(1) with zero mean and a moving average coefficient of
0.4. What is the value of the autocorrelation function at lag 1?
a. 0.4
b. 1
c. *0.34
d. It is not possible to determine the value of the autocovariances without knowing the
disturbance variance.
11. Consider the following picture and suggest the model from the following list that best
characterises the process:
a. An AR(1)
b. An AR(2)
c. * An ARMA(1,1)
d. An MA(3).
The acf is clearly declining very slowly in this case, which is consistent with their being an
autoregressive part to the appropriate model. The pacf is clearly significant for lags 1 and 2, but
the question is: does it them become insignificant for lags 2 and 4, indicating an AR(2) process,
or does it remain significant, which would be more consistent with a mixed ARMA process?
Well, given the huge size of the sample that gave rise to this acf and pacf, even a pacf value of
0.001 would still be statistically significant. Thus an ARMA process is the most likely candidate,
although note that it would not be possible to tell from the acf and pacf which model from the
ARMA family was more appropriate. The DGP for the data that generated this plot was y_t = 0.9
y_(t–1) – 0.3 u_(t–1) + u_t.
12. Which of the following models can be estimated using ordinary least squares?
i. An AR(1)
(ii) An ARMA(2,0)
(iii) An MA(1)
(iv) An ARMA(1,1).
a. (i) only
b. * (i) and (ii) only
c. (i), (ii), and (iii) only
d. (i), (ii), (iii), and (iv).
13. If a series, y, is described as ‘mean-reverting’, which model from the following list is likely
to produce the best long-term forecasts for that series y?
a. A random walk
b. * The long term mean of the series
c. A model from the ARMA family
d. A random walk with drift.
14. Consider the following AR(2) model. What is the optimal 2-step-ahead forecast for y if all
information available is up to and including time t, if the values of y at time t, t-1 and t-2 are –
0.3, 0.4 and –0.1, respectively, and the value of u at time t-1 is 0.3?
y = –0.1 + 0.75y – 0.125y + u
t t-1 t-2 t
a. –0.1
b. 0.27
c. * –0.34
d. 0.30.
15. What is the optimal three-step-ahead forecast from the AR(2) model given in Question 14?
a. –0.1
b. 0.27
c. –0.34
d. * –0.31.
16. Suppose you had to guess at the most likely value of a one hundred-step-ahead forecast for
the AR(2) model given in Question 14 – what would your forecast be?
a. -0.1
b. 0.7
c. * –0.27
d. 0.75.
Use the following to answer Questions 19 and 20. Suppose that you have estimated the first five
autocorrelation coefficients using a series of length 81 observations and found them to be
19. Which autocorrelation coefficients are significantly different from zero at the 5% level?
(a) The first and fifth autocorrelation coefficient
(b) The first, second, third, and fifth autocorrelation coefficient
(c)* The first, third, and fifth autocorrelation coefficient
(d) The second and fourth autocorrelation coefficient.
21. Consider the following MA(2) process where the errors follow a
standard normal distribution. What is the variance of ?
(a)
(b)
(c)
(d)* All of the above.
22. A model where the current value of a variable depends upon only the values that the variable
took in previous periods plus an error term is called
(a)* An autoregressive model
(b) An autoregressive moving average model
(c) An autoregressive integrated moving average model
(d) A periodic lag model.
25. Which of these is an appropriate way to determine the order of an ARMA model required to
capture the dynamic features of a given data?
(a) Graphically plotting the time series of the data
(b) Determining the number of parameters that maximises the information criteria
(c)* Determining the number of parameters that minimises the information criteria
(d) None of the above.
28. What are the closest to the mean squared errors for model A and B’s forecasts?
(a) 0.58 and 0.98, respectively
(b) 0.98 and 0.58, respectively
(c)* 0.45 and 1.95, respectively
(d) 1.95 and 0.45, respectively.
29. What are the closest to the mean absolute errors from models A and B?
(a)* 0.58 and 0.98, respectively
(b) 0.98 and 0.58, respectively
(c) 0.45 and 1.95, respectively
(d) 1.95 and 0.45, respectively.
30. Based on the MAE and MSE forecast evaluation metrics, which of these statements are true?
(a)* Model A outperforms Model B at forecasting the house price index
(b) Model A underperforms Model B at forecasting the house price index
(c) Model A and Model B perform equally well at forecasting the house price index
(d) We cannot tell which model does best.
Chapter 7
Correct answers denoted by an asterisk.
Assume that the Y’s are endogenous and the X’s exogenous variables, and that the error terms are
uncorrelated.
(a) Equations that are part of a recursive system can be validly estimated using OLS
(b) Unnecessary use of two-stage least squares (2SLS) – i.e., on a set of right hand side variables
that are in fact exogenous – will result in consistent but inefficient coefficient estimates
(c) 2SLS is just a special case of instrumental variables (IV) estimation
(d) * 2SLS and indirect least squares (ILS) are equivalent for over-identified systems.
5. Which of the following could be viewed as a disadvantage of the vector autoregressive (VAR)
approach to modelling?
(a) We do not need to specify which variables are endogenous and which are exogenous
(b) Standard form VARs can be estimated equation-by-equation using OLS
(c) * VARs often contain a large number of terms
(d) VARs can be expressed using a very compact notation.
Which of the following coefficient significances are required to be able to say that y Granger-
1
7. Which of the following statements is TRUE concerning VAR impulse response functions?
(i) Impulse responses help the researcher to investigate the interactions between the variables in
the VAR
(ii) An impulse response analysis is where we examine the effects of applying unit shocks to all
of the variables at the same time
(iii) Impulse responses involve calculating the proportion of the total forecast error variance of a
given variable that is explained by innovations to each variable
(iv) If the ±2 standard error bars around the impulse responses for a given lag span (i.e., include)
the x-axis, it would be said that the response is statistically significant.
9. Comparing the information criteria approach with the likelihood ratio test approach to
determining the optimal VAR lag length, which one of the following statements is true?
a. The choice of stiffness of penalty term will not affect the model choice
b. The validity of information criteria relies upon normal residuals
c. * Conducting a likelihood ratio test could lead to a sub-optimal model selection
d. An application of the univariate information criteria to each equation will give identical
results to the application of a multivariate version of the criteria to all of the equations
jointly.
10. The second stage in two-stage least squares estimation of a simultaneous system would be to
a. Estimate the reduced form equations
b. * Replace the endogenous variables that are on the RHS of the structural equations with
their reduced form fitted values
c. Replace all endogenous variables in the structural equations with their reduced form
fitted values
d. Use the fitted values of the endogenous variables from the reduced forms as additional
variables in the structural equations.
11. Which of these assumptions is violated when an equation is estimated using OLS when it is
in fact part of a simultaneous structural system?
(a)
(b)*
(c)
(d) None of the above.
12. A variable x is defined as ________ if its value is determined outside of the equation or
system of equations. What is the blank?
(a) Endogenous
(b)* Exogenous
(c) Homogeneous
(d) Heterogeneous.
13. Which of these is not an appropriate method of estimating equations that are from a
simultaneous system?
(a) Indirect least squares
(b) Two-stage least squares
(c)* Aggregate least squares
(d) Instrumental variables.
(a) I only
(b) I and II only
(c)* I, II, and III only
(d) I, II, III, and IV.
15. Which of these is an approach used to determine the appropriate lag lengths of VAR
models?
(a) Graphically plotting the time series of the data
(b) Selecting the number of lags that maximises the information criteria
(c)* Selecting the number of lags that minimises the information criteria
(d) None of the above.
16. Assuming that you have a VAR model with 2 variables (A and B) including many lags, how
can you test whether A cause Granger-causes changes in B?
(a) By observing if the differences in correlation between A and B are statistically significant
(b)* Impose restrictions that all the coefficients of the lags of A are equal to 0 in the equation for
B of the VAR model and test the joint hypothesis within the F-test framework
(c) Impose restrictions that all the coefficients of the lags of B are equal to 0 in the equation for
A of the VAR model and test the joint hypothesis within the F-test framework
(d) None of the above.
Chapter 8
Correct answers denoted by an asterisk.
1. Which of the following are probably valid criticisms of the Dickey–Fuller methodology?
(i) The tests have a unit root under the null hypothesis and this may not be rejected due to
insufficient information in the sample
(ii) The tests are poor at detecting a stationary process with a unit root close to the non-stationary
boundary
(iii) The tests are highly complex to calculate in practice
(iv) The tests have low power in small samples.
2. Which of the following are problems associated with the Engle–Granger approach to
modelling using cointegrated data?
(i) The coefficients in the cointegrating relationship are hard to calculate
(ii) This method requires the researcher to assume that one variable is the dependent variable and
the others are independent variables
(iii) The Engle–Granger technique can only detect one cointegrating relationship
(iv) The Engle-Granger technique does not allow the testing of hypotheses involving the actual
cointegrating relationship.
(a) Johansen’s test for cointegration centres on the rank of the matrix Γ 1
5. You have the following data for Johansen’s λ rank test for cointegration between 4
max
0 40.03 30.26
1 26.81 23.84
2 13.42 17.72
3 8.66 10.71
(a) 0
(b) 1
(c) * 2
(d) 3.
6. Which criticism of Dickey–Fuller (DF)-type tests is addressed by stationarity tests, such as the
KPSS test?
(a) * DF tests have low power to reject the null hypothesis of a unit root, particularly in small
samples.
(b) DF tests are always over-sized.
(c) DF tests do not allow the researcher to test hypotheses about the cointegrating vector
(d) DF tests can only find at most one cointegrating relationship.
Which one of the following most accurately describes the process for y ?
t
9. If there are three variables that are being tested for cointegration, what is the maximum
number of linearly independent cointegrating relationships that there could be?
(a) 0
(b) 1
(c) *2
(d) 3.
10. If the number of non-zero eigenvalues of the pi matrix under a Johansen test is 2, this implies
that
(a) * There are 2 linearly independent cointegrating vectors
(b) There are at most 2 linearly independent cointegrating vectors
(c) There are 3 variables in the system
(d) There are at least 2 linearly independent cointegrating vectors.
11. If a Johansen ‘max’ test for a null hypothesis of 1 cointegrating vectors is applied to a system
containing 4 variables, which eigenvalues would be used in the test?
(a) The largest 1
(b) * The second largest
(c) The second smallest
(d) The smallest.
12. Consider the testing of hypotheses concerning the cointegrating vector(s) under the Johansen
approach. Which of the following statements is correct?
(a) If the restriction is (are) rejected, the number of cointegrating vectors will rise
(b) If the restriction(s) is (are) rejected, the number of eigenvalues will fall
(c) Whether the restriction is supported by the data or not, the eigenvalues are likely to change at
least slightly upon imposing the restriction(s)
(d) * All linear combinations of the cointegrating vectors are themselves cointegrating vectors.
(a) I only
(b) I and II only
(c) I, II, and III only
(d)* I, II, III, and IV.
19. A researcher would like to test for a unit root in a series. She runs the regression
. What should her null hypothesis be assuming that she adopts the Dickey–Fuller
test approach?
(a)*
(b)
(c)
(d) .
20. Assuming the researcher in Question 19 would like to run an augmented Dickey–Fuller test
instead. What is the appropriate regression she would have to run and the null hypothesis of the
test?
22. Assume that you are trying to model the relationship between house prices and rents. If you
find that both series are non-stationary and a linear combination of the two series is stationary,
which of the following is true?
(I) Regressing the levels of house prices on the levels of rents could lead to spurious regressions
(II) House prices and rents are cointegrated
(III) An appropriate linear combination of house prices and rents is I(1)
(IV) House prices and rents are not cointegrated.
(a) I only
(b)* I and II only
(c) I, II, and III only
(d) I, II, III, and IV only.
Chapter 9
Correct answers denoted by an asterisk.
(a) I only
(b)* I and II only
(c) I, II, and III only
(d) I, II, III, and IV.
(a) I only
(b) I and II only
(c) I, II, and III only
(d)* I, II, III, and IV.
7. Which of these is an appropriate technique used in estimating models from the GARCH
family?
(a)* Maximum likelihood
(b) Instrumental variables
(c) Indirect least squares
(d) Ordinary least squares.
8. What are the steps required to estimate an ARCH/GARCH model?
(a) First specify the appropriate equations for the correlation and the variance, then specify LLF
and the computer will generate parameter values that maximise the LLF
(b) First specify the appropriate equations for the median and the variance, then specify LLF and
the computer will generate parameter values that maximise the LLF
(c)* First specify the appropriate equations for the mean and the variance, then specify LLF and
the computer will generate parameter values that maximise the LLF
(d) None of the above.
9. GJR and EGARCH are types of GARCH models that allow for:
(a) An asymmetric response of returns to positive and negative shocks in the dependent variable
(b) An asymmetric response of returns to positive and negative shocks to its lagged values
(c) A symmetric response of volatility to positive and negative shocks
(d)* An asymmetric response of volatility to positive and negative shocks.
10. Assume that you have estimated a GJR model of monthly stock returns and you obtain the
following equations:
Suppose that , what would be the fitted conditional variance for time t if
and then if ?
(a) 1.62 and 1.67, respectively
(b) 1.64 and 1.59, respectively
(c)* 1.59 and 1.64, respectively
(d) 1.67 and 1.62, respectively.
11. Suppose that a researcher estimates a GARCH(1,1) model and obtains a log likelihood
function (LLF) value of 71.22. She is interested in testing whether an ARCH(1) model is a better
model at describing volatility. If she estimates a model which imposes the necessary restrictions
and obtains an LLF value of 68.21, what would be the conclusion of her likelihood ratio test
(assuming a 5% significance level)?
(a) Statistical evidence suggesting that ARCH(1) is better than GARCH(1,1)
(b)* Statistical evidence suggesting that ARCH(1) is not better than GARCH(1,1)
(c) Statistical evidence suggesting that GARCH(1,1) is better than ARCH(1)
(d) We cannot say because we would need to know the number of observations.
12. What would typically be the shape of the news impact curve for a series that exactly followed
a GARCH(1,1) process?
(a) It would be asymmetric, with a steeper curve on the left than the right
(b) It would be asymmetric, with a steeper curve on the right than the left
(c) * It would be symmetric about zero
(d) It would be discontinuous about zero.
14. Which of the following would represent the most appropriate definition for implied
volatility?
(a) * It is the volatility of the underlying asset’s returns implied from the price of a traded option
and an option pricing model
(b) It is the volatility of the underlying asset’s returns implied from a statistical model such as
GARCH
(c) It is the volatility of an option price implied from a statistical model such as GARCH
(d) It is the volatility of an option price implied from the underlying asset volatility.
15. Suppose that a researcher wanted to obtain an estimate of realised (‘actual’) volatility. Which
one of the following is likely to be the most accurate measure of volatility of stock returns for a
particular day?
(a) The price range (high minus low) on that day
(b) The squared return on that day
(c) * The sum of the squares of hourly returns on that day
(d) The squared return on the previous day.
16. Which of the following is the most plausible test regression for determining whether a series
y contains ‘ARCH effects’?
a.
b. *
c.
d. .
17. Consider the following conditional variance equation for a GJR model.
h =α +α
t 0 +βh +γu I
1 t-1 t-1
2
t-1
where I = 1 if u < 0
t-1 t-1
= 0 otherwise
For there to be evidence of a leverage effect, which ONE of the following conditions must hold?
a. α positive and statistically significant
0
18. Consider the three approaches to conducting hypothesis tests under the maximum likelihood
framework. Which of the following statements are true?
i. The Wald test is based on estimation only under the null hypothesis
ii. The likelihood ratio test is based on estimation under both the null and the alternative
hypotheses
iii. The Lagrange multiplier test is based on estimation under the alternative hypothesis only
iv. The usual t- and F-tests are examples of Wald tests.
19. Which one of the following problems in finance could not be usefully addressed by either a
univariate or a multivariate GARCH model?
a. Producing option prices
b. Producing dynamic hedge ratios
c. Producing time-varying beta estimates for a stock
d. * Producing forecasts of returns for use in trading models
e. Producing correlation forecasts for value at risk models.
Chapter 10
Correct answers denoted by an asterisk.
To check for seasonality (day-of-the-week effect) in stock returns of South Korea, Malaysia, the
Philippines, Taiwan, and Thailand, Brooks and Persand (2001) regress daily returns in each of
these countries’ stock market on five dummy variables D1 to D5 representing each day of the
week – i.e., D1 for Mondays, D2 for Tuesdays, D3 for Wednesdays, D4 for Thursdays and D5
for Fridays:
Their results were:
2. Which market(s) did not display any evidence of day-of-the-week effect?
(a) Thailand, Malaysia and Taiwan
(b) Philippines only
(c) South Korea only
(d)* South Korea and Philippines.
4. The unknown parameters of a Markov switching model are usually estimated using:
(a)* Maximum likelihood
(b) Instrumental variables
(c) Indirect least squares
(d) Ordinary least squares.
5. The key difference between threshold autoregressive and Markov switching models is that:
(a) The latter can be estimated using ordinary least squares while the latter is estimated using the
indirect least squares estimation technique
(b) Under the latter, the state variable is assumed to be known and observable, while it is latent
under the former
(c)* Under the former, the state variable is assumed to be known and observable, while it is
latent under the latter
(d) None of the above.
(b) if
if
(c) if
if
(d) if
if .
7. To compare the goodness of fit of Markov switching and threshold autoregressive models with
linear models, one can compare the residual sums of squares of the two types of models using an
F-test. Is the statement true?
(a) Yes
(b)* No
(c) If the autoregressive model is restricted
(d) Cannot say without knowing the number of regimes in the regime switching models.
8. Suppose that a researcher wishes to test for calendar (seasonal) effects using a dummy
variables approach. Which of the following regressions could be used to examine this?
i. A regression containing intercept dummies
ii. A regression containing slope dummies
iii. A regression containing intercept and slope dummies
iv. A regression containing a dummy variable taking the value 1 for one observation and
zero for all others.
a. * (ii) only
b. (i) and (ii) only
c. (i), (ii), and (iii) only
d. (i), (ii), (iii), and (iv).
10. Consider the following two equations in a state space model, where y is the observed series,
t
11. The ratio of the variance of the error term η to the variance of the error term u is used as the
t t
12. Consider the following two equations in a state space model, where y is the observed series,
t
(c) The estimated values of the noise terms (i.e., the values of u and η )
t t
(d) * The variances of the noise terms (i.e., the variances of u and η )’
t t
Chapters 11-14
Correct answers denoted by an asterisk.
(a) I only
(b) I and II only
(c) I, II, and III only
(d)* I, II, III, and IV.
(a) I only
(b)* I and II only
(c) I, II, and III only
(d) I, II, III, and IV.
3. Entity fixed effects models
(a)* Allow the intercept in the regression model to differ cross-sectionally but not over time,
while all of the slope estimates are fixed both cross-sectionally and over time
(b) Allow the slope in the regression model to differ cross-sectionally but not over time, while
the intercept estimates are fixed both cross-sectionally and over time
(c) Allow the intercept in the regression model to differ over time, while all of the slope
estimates are different both cross-sectionally and over time
(d) Any of the above could be true depending on the model specification.
8. To test for unit roots in panel data, Levin, Lin and Chu (2002) develop a test based on the
equation . What is the appropriate null hypothesis
for this test?
(a) *
(b)
(c)
(d) .
9. Logit and probit models are more appropriate than linear probability models because:
(a) Logit and probit can estimate probabilities that are negative
(b) Logit and probit cannot estimate probabilities that are greater than one
(c) Logit and probit cannot estimate probabilities that are negative but not greater than one
(d)* Logit and probit cannot estimate probabilities that are negative or greater than one.
10. Which of the following statements about logit and probit models is true?
(I) They cannot be estimated by ordinary least squares
(II) They can be estimated using maximum likelihood
(III) They can be estimated using non-linear least squares
(IV) They can be estimated using instrumental variables.
(a) I only
(b) I and II only
(c)* I, II, and III only
(d) I, II, III, and IV.
11. If the maximised value of the log-likelihood function for a logit model is 34.55 and for a
restricted model where all of the slope parameters are set to zero is 30.67, what is the pseudo-R ?
2
(a) 0.13
(b)* –0.13
(c) 0.11
(d) –0.11.
12. Appropriate modelling of limited dependent variables that are assigned numerical values
having a natural ordering can be done using:
(I) Probit models
(II) Logit models
(III) Ordered probit models
(IV) Ordered logit models.
(a) I only
(b) I and II only
(c) II and III only
(d)* III and IV only.
14. Which of the following statements is TRUE concerning the calendar time methodology
sometimes used in event studies?
(a) It will weight all the firms in the sample that underwent the event equally
(b) It can involve the calculation of a buy-and-hold abnormal return
(c) If the slope parameter in the test regression is positive and significant, this will provide
evidence of an abnormal return in the event study
(d) *It will give more weight in the sample to firms which underwent the event at a time when
few other firms did so.
16. In the Fama–MacBeth regressions, the parameter estimates in the second stage are
interpreted as:
(a) Factor loadings
(b) * Factor risk premia
(c) Average returns for each stock
(d) The volatilities of returns for each stock.
17. In the Fama–MacBeth regressions, the parameter estimates in the first stage are interpreted
as:
(a) * Factor loadings
(b) Factor risk premia
(c) Average returns for each stock
(d) The volatilities of returns for each stock.
18. In Fama–French (1993)- and Carhart (1994)-type models, for there to be no evidence of
outperformance by a fund manager, we would require:
(a) The intercept is positive but not statistically significant
(b) The intercept is not positive and significant and slope estimates are all insignificant
(c) * The intercept is not positive and significant
(d) The slope estimates are all insignificant.
19. If we use the block maximum approach to estimating the parameters of a member of the
extreme value family of distributions and we select a large number of short blocks, which of the
following is a likely disadvantage?
(a) * A number of data points would be classified as extreme when they are not, leading to bias
in the shape parameter estimate
(b) Too few data points would be classified as extreme, leading to excessive noise in the shape
parameter estimate
(c) Too few data points would be classified as extreme, leading to bias in the shape parameter
estimate
(d) A number of data points would be classified as extreme when they are not, leading to
excessive noise in the shape parameter estimate.
20. Which of the following distributions would be most appropriate for modelling the central
part of the distribution of a set of stock returns?
(a) Gumbel
(b) Fréchet
(c) Weibull
(d) * Normal.
21. If we use the peaks-over-threshold approach to estimating the parameters of an extreme value
distribution, if we use a value of U, the threshold, that is too high (i.e., too far into the tail),
which of the following is a likely disadvantage?
(a) A number of data points would be classified as extreme when they are not, leading to bias in
the shape parameter estimate
(b) * Too few data points would be classified as extreme, leading to excessive noise in the shape
parameter estimate
(c) Too few data points would be classified as extreme, leading to bias in the shape parameter
estimate
(d) A number of data points would be classified as extreme when they are not, leading to
excessive noise in the shape parameter estimate.