An analyst is interpreting the results of an AR(1) model estimated for a commodity spot price index. The estimated model is: x t = b 0 + b 1 × x t−1 + ε t , where |b 1 | < 1. Which of the following correctly describes the mean-reverting level of this series?

The long-run mean to which the series reverts equals b 0 ÷ (1 − b 1 ), provided the absolute value of b 1 is less than 1

The long-run mean equals the intercept b 0 alone, because the autoregressive coefficient has no influence on the equilibrium level

Mean reversion requires the autoregressive coefficient b 1 to equal exactly 1.0; any deviation from this value eliminates mean reversion

Open App

CFA Level 2 Exam > CFA Level 2 Test > Quantitative Methods > Practice Test: Quantitative Methods - 1 - CFA Level 2 MCQ

Quantitative Methods - 1 - Free MCQ Practice Test with solutions, CFA Level

MCQ Practice Test & Solutions: Practice Test: Quantitative Methods - 1 (30 Questions)

You can prepare effectively for CFA Level 2 Quantitative Methods with this dedicated MCQ Practice Test (available with solutions) on the important topic of "Practice Test: Quantitative Methods - 1". These 30 questions have been designed by the experts with the latest curriculum of CFA Level 2 2026, to help you master the concept.

Test Highlights:

- Format: Multiple Choice Questions (MCQ)
- Duration: 80 minutes
- Number of Questions: 30

Practice Test: Quantitative Methods - 1 - Question 1

An analyst estimates a multiple regression model with monthly portfolio returns as the dependent variable and three macroeconomic factors as independent variables. The regression output is as follows:
Variable Coefficient t-Statistic
Intercept 0.31 1.85
GDP Growth 0.72 2.45
Inflation Rate -0.55 -3.10
Credit Spread 0.18 1.20
The critical t-value at the 5% significance level (two-tailed) is 1.96. Which factor coefficients are statistically significant at the 5% level?

Variable	Coefficient	t-Statistic
Intercept	0.31	1.85
GDP Growth	0.72	2.45
Inflation Rate	-0.55	-3.10
Credit Spread	0.18	1.20

A. GDP Growth, Inflation Rate, and Credit Spread only
B. GDP Growth and Inflation Rate only
C. Intercept, GDP Growth, and Inflation Rate only

Detailed Solution: Question 1

GDP Growth (|2.45|>1.96) and Inflation Rate (|-3.10|>1.96) exceed critical value; others do not.

Practice Test: Quantitative Methods - 1 - Question 2

A portfolio manager estimates a multiple regression with four independent variables. The F-statistic is 18.45, which exceeds the critical value at the 1% significance level. Which of the following is the most accurate interpretation of this result?

A. At least one slope coefficient in the model differs from zero in the population
B. Each individual independent variable is statistically significant at the 1% level
C. The model's R-squared exceeds 0.50, confirming strong explanatory power

Detailed Solution: Question 2

Significant F-statistic means at least one slope coefficient differs from zero; not that each variable is individually significant.

Join for Free

Practice Test: Quantitative Methods - 1 - Question 3

An analyst adds a fifth independent variable to an existing four-variable regression model. The new variable has a t-statistic of 0.45. Which of the following best describes the most likely impact on the model's fit statistics?

A. Both R-squared and adjusted R-squared will decrease because an irrelevant variable introduces noise
B. Adjusted R-squared will always exceed R-squared after the new variable is added
C. R-squared will increase, but adjusted R-squared may decrease because the new variable lacks sufficient explanatory power to offset its degree-of-freedom cost

Detailed Solution: Question 3

R² always increases when a variable is added; adjusted R² may decrease if the variable adds insufficient explanatory power.

Practice Test: Quantitative Methods - 1 - Question 4

An analyst suspects that a regression model of equity returns suffers from conditional heteroskedasticity. Which of the following procedures best describes the Breusch-Pagan test for detecting this condition?

A. Regress the dependent variable on the squared values of each independent variable and test for nonlinearity using an F-statistic
B. Regress the squared residuals from the original regression on the independent variables and use a chi-square test to assess significance
C. Compute the Durbin-Watson statistic and compare it to lower and upper critical bounds to determine whether variance is constant

Detailed Solution: Question 4

Breusch-Pagan regresses squared residuals on independent variables; chi-square test; significant result indicates conditional heteroskedasticity.

Practice Test: Quantitative Methods - 1 - Question 5

A quantitative analyst observes that residuals from an equity return regression grow substantially larger during periods when the CBOE Volatility Index (VIX) is elevated, but are relatively stable during low-volatility regimes. This pattern is best characterized as:

A. Conditional heteroskedasticity, because the variance of the error term is correlated with the level of an independent variable
B. Unconditional heteroskedasticity, because residual magnitude varies across the time series
C. Positive serial correlation, because residual patterns shift systematically across the sample period

Detailed Solution: Question 5

Error variance correlated with an independent variable (VIX) defines conditional heteroskedasticity.

Practice Test: Quantitative Methods - 1 - Question 6

A fixed-income analyst runs a time-series regression and computes a Durbin-Watson (DW) statistic of 0.85. The critical lower bound (d_L) is 1.50 and the upper bound (d_U) is 1.75 at the 5% significance level. Which conclusion is most appropriate?

A. No serial correlation is present because the DW statistic is below the lower critical bound
B. Negative serial correlation is present because the DW statistic is far from 4.0
C. Positive serial correlation is present because the DW statistic falls below the lower critical bound d_L

Detailed Solution: Question 6

DW = 0.85 < dl="1.50;" reject="" null="" of="" no="" serial="" correlation="" in="" favor="" of="" positive="" serial="" correlation.="" dl="1.50;" reject="" null="" of="" no="" serial="" correlation="" in="" favor="" of="" positive="" serial="">

Practice Test: Quantitative Methods - 1 - Question 7

An analyst estimates a multiple regression of fund returns on six macroeconomic variables. The model produces an R-squared of 0.88 and a highly significant F-statistic, yet most individual slope t-statistics are below 2.0. Pairwise correlations among the six independent variables range from 0.88 to 0.95. This pattern most likely indicates:

A. Conditional heteroskedasticity, because residual variance increases with fitted values when many variables are included
B. Multicollinearity, because high inter-correlations among independent variables inflate standard errors and reduce individual t-statistics
C. Positive serial correlation, because the Durbin-Watson statistic would be near zero in models with many correlated predictors

Detailed Solution: Question 7

High R², significant F, low individual t-stats, and high pairwise correlations among regressors are classic symptoms of multicollinearity.

Practice Test: Quantitative Methods - 1 - Question 8

A researcher models seasonal effects in quarterly equity returns using dummy variables to represent four quarters (Q1, Q2, Q3, Q4). To avoid perfect multicollinearity in the regression model, how many dummy variables should be included?

A. Three dummy variables (k - 1), with one quarter serving as the reference category
B. Four dummy variables, one for each quarter, to fully capture all seasonal patterns
C. Two dummy variables, representing only the quarters that differ most from the annual mean

Detailed Solution: Question 8

With k categories, include k-1 dummies to avoid the dummy variable trap (perfect multicollinearity).

Practice Test: Quantitative Methods - 1 - Question 9

An analyst regresses monthly hedge fund returns on the monthly market excess return and a recession dummy variable (D = 1 during NBER-defined recession months, D = 0 otherwise). The estimated model is:
Return_t = 0.80 + 1.10 × Market_t − 1.40 × D_t
Which of the following best describes the model's prediction for recession months, holding the market return constant?

A. The predicted fund return is reduced by 0.80 percentage points, reflecting a shift in the intercept to zero during recessions
B. The recession dummy increases predicted returns because it captures elevated risk premiums during economic downturns
C. The predicted fund return is reduced by 1.40 percentage points relative to non-recession months, all else equal

Detailed Solution: Question 9

Recession dummy coefficient of -1.40 directly reduces predicted return by 1.40 percentage points during recession months.

Practice Test: Quantitative Methods - 1 - Question 10

An analyst estimates an AR(1) model for a monthly economic index:
x_t = 0.50 + 0.75 × x_t−1 + ε_t
What is the mean-reverting level of this time series?

A. 0.50
B. 2.00
C. 0.67

Detailed Solution: Question 10

Mean reversion level = b0/(1-b1) = 0.50/(1-0.75) = 0.50/0.25 = 2.00.

Practice Test: Quantitative Methods - 1 - Question 11

An analyst has estimated the following AR(1) model for quarterly GDP growth rates:
x_t = 0.30 + 0.80 × x_t−1
If the most recent observed value is x₅ = 2.50, what is the one-period-ahead forecast, x₆?

A. 2.30
B. 2.50
C. 2.80

Detailed Solution: Question 11

x6 = 0.30 + 0.80(2.50) = 0.30 + 2.00 = 2.30.

Practice Test: Quantitative Methods - 1 - Question 12

An analyst applies the Dickey-Fuller test to a monthly commodity price series. The test fails to reject the null hypothesis at the 5% significance level. Which of the following conclusions is most appropriate?

A. The time series is stationary and may be used directly in a regression without transformation
B. The AR(1) coefficient is significantly different from 1.0, confirming that the series exhibits trend stationarity
C. The time series contains a unit root and should be first-differenced prior to use in regression analysis

Detailed Solution: Question 12

Failing to reject Dickey-Fuller null (b1=1) means unit root present; series non-stationary; first-differencing recommended.

Practice Test: Quantitative Methods - 1 - Question 13

An analyst estimates an AR(1) model for a stock price index: x_t = 0.02 + 1.00 × x_t−1 + ε_t. A Dickey-Fuller test fails to reject the null hypothesis. Which of the following concerns is most relevant when using this series in a regression?

A. The series is mean-reverting with a long-run equilibrium value of 0.02, so regression results will be reliable
B. The series is non-stationary; using the raw series in a regression with another non-stationary variable may produce spurious results
C. The series is stationary because the positive intercept prevents indefinite drift

Detailed Solution: Question 13

Unit root (b1=1) implies non-stationary random walk; regression with non-stationary series may yield spurious results.

Practice Test: Quantitative Methods - 1 - Question 14

An analyst estimates an AR(1) model on weekly equity index returns and subsequently tests the residuals. The autocorrelation of squared residuals at multiple lags is statistically significant. This finding most likely indicates:

A. ARCH effects are present; the variance of the error term is time-varying and depends on past squared errors
B. Serial correlation is present in the residuals; the Durbin-Watson statistic would be near zero
C. Multicollinearity is present because the independent variable is correlated with lagged values of itself

Detailed Solution: Question 14

Significant autocorrelation in squared residuals signals ARCH effects - time-varying conditional variance in the error term.

Practice Test: Quantitative Methods - 1 - Question 15

A regression analyst detects conditional heteroskedasticity in a model of corporate bond spreads. She decides to apply White-corrected standard errors. Which of the following best describes the effect of this correction on the regression output?

A. The regression coefficients will change to reflect the corrected variance structure, producing new point estimates
B. The model's R-squared will increase because variance is now correctly accounted for in the estimation
C. The regression coefficients remain unchanged, but their standard errors and corresponding t-statistics are adjusted for heteroskedasticity

Detailed Solution: Question 15

White correction adjusts standard errors and t-statistics only; regression coefficients themselves are unchanged.

Practice Test: Quantitative Methods - 1 - Question 16

An analyst is evaluating two model specification errors: (1) omitting a variable that is correlated with included regressors, and (2) including an irrelevant variable with no true relationship to the dependent variable. Which of the following best contrasts the consequences of these two errors?

A. Including an irrelevant variable causes regression coefficients to be biased and inconsistent
B. Omitting a relevant variable correlated with included regressors causes biased and inconsistent coefficient estimates, whereas including an irrelevant variable leaves coefficients unbiased but reduces efficiency
C. Both errors produce equally severe consequences: biased, inconsistent, and inefficient coefficient estimates

Detailed Solution: Question 16

Omitting relevant correlated variable → biased, inconsistent estimates. Including irrelevant variable → unbiased but inefficient (larger SE).

Practice Test: Quantitative Methods - 1 - Question 17

A machine learning analyst is comparing LASSO regression and ridge regression for a predictive model of equity returns. Which of the following most accurately distinguishes between these two regularization techniques?

A. LASSO applies an L1 penalty that can shrink coefficients to exactly zero, performing automatic variable selection; ridge applies an L2 penalty that shrinks coefficients toward but not to zero
B. Ridge regression produces sparse models by eliminating irrelevant variables; LASSO retains all variables with small but nonzero coefficients
C. Both LASSO and ridge regression produce identical coefficient estimates when the regularization parameter λ is held constant across both methods

Detailed Solution: Question 17

LASSO (L1 penalty) can zero out coefficients enabling variable selection; ridge (L2 penalty) shrinks toward but not to zero.

Practice Test: Quantitative Methods - 1 - Question 18

A machine learning analyst builds a model using 500 monthly observations and 40 predictor variables to forecast credit defaults. The in-sample R-squared is 0.95, but when the model is applied to a holdout sample, predictive accuracy deteriorates sharply. The analyst proposes using k-fold cross-validation. Which of the following best describes the purpose and benefit of this approach in this context?

A. The model suffers from underfitting; k-fold cross-validation would identify which additional variables to include to improve in-sample fit
B. K-fold cross-validation primarily reduces in-sample bias by using the entire dataset simultaneously for model training
C. The model is likely overfit; k-fold cross-validation provides a more reliable estimate of out-of-sample performance by systematically rotating training and validation subsets across k folds

Detailed Solution: Question 18

High in-sample but low out-of-sample performance signals overfitting; k-fold CV provides reliable out-of-sample performance estimate.

Practice Test: Quantitative Methods - 1 - Question 19

A credit analyst builds a binary classification model to predict corporate defaults. The confusion matrix results are as follows:
Predicted Default Predicted No Default
Actual Default 40 20
Actual No Default 10 30
Which of the following correctly computes the model's precision, recall, and F1 score?

	Predicted Default	Predicted No Default
Actual Default	40	20
Actual No Default	10	30

A. Precision = 0.67; Recall = 0.80; F1 = 0.73
B. Precision = 0.80; Recall = 0.67; F1 = 0.73
C. Precision = 0.80; Recall = 0.80; F1 = 0.80

Detailed Solution: Question 19

Precision=40/50=0.80; Recall=40/60=0.67; F1=2(0.80×0.67)/(0.80+0.67)≈0.73.

Practice Test: Quantitative Methods - 1 - Question 20

A portfolio manager is reviewing two machine learning approaches: one that uses labeled historical return data to train a model predicting future outperformance, and another that groups securities into clusters based on factor exposures without a predefined target variable. Which of the following best distinguishes these two approaches?

A. The first approach is supervised learning, requiring labeled training data to map inputs to a target output; the second is unsupervised learning, identifying patterns in data without a predefined target variable
B. The first approach is unsupervised learning using clustering; the second is supervised learning using classification algorithms with labeled outcomes
C. Both approaches are forms of supervised learning because they both use historical market data as model inputs

Detailed Solution: Question 20

Predicting labeled outcomes = supervised learning; finding structure in unlabeled data = unsupervised learning.

Practice Test: Quantitative Methods - 1 - Question 21

An analyst estimates an AR(1) model for monthly industrial production growth. After estimation, she examines the autocorrelation function (ACF) of the residuals and finds no significant autocorrelation at lag 1, but significant autocorrelation at lag 2. Which of the following actions is most appropriate?

A. The AR(1) model is adequate; lag-2 autocorrelation in residuals is typical and does not indicate model misspecification
B. The analyst should add a moving average term to form an ARMA(1,1) model, as the autocorrelation pattern reflects a moving average process
C. An AR(2) specification is more appropriate, because significant residual autocorrelation at lag 2 suggests that a second autoregressive term is required

Detailed Solution: Question 21

Significant lag-2 autocorrelation in AR(1) residuals indicates an AR(2) term is needed.

Practice Test: Quantitative Methods - 1 - Question 22

An analyst runs a time-series regression and computes a Durbin-Watson statistic of 0.92. The critical values are d_L = 1.53 and d_U = 1.74. Which of the following most accurately describes the consequences for OLS inference?

A. OLS coefficient estimates are biased and inconsistent; the model must be re-estimated using generalized least squares
B. OLS coefficients remain unbiased but are inefficient; standard errors are understated, causing t-statistics to be overstated and increasing the probability of Type I errors
C. The model is free of specification error; the Durbin-Watson statistic only tests for model fit and does not affect coefficient reliability

Detailed Solution: Question 22

Serial correlation leaves OLS coefficients unbiased but inefficient; standard errors understated → inflated t-stats → excess Type I errors.

Practice Test: Quantitative Methods - 1 - Question 23

An analyst is interpreting the results of an AR(1) model estimated for a commodity spot price index. The estimated model is: x_t = b₀ + b₁ × x_t−1 + ε_t, where |b₁| < 1. Which of the following correctly describes the mean-reverting level of this series?

A. The long-run mean to which the series reverts equals b₀ ÷ (1 − b₁), provided the absolute value of b₁ is less than 1
B. The long-run mean equals the intercept b₀ alone, because the autoregressive coefficient has no influence on the equilibrium level
C. Mean reversion requires the autoregressive coefficient b₁ to equal exactly 1.0; any deviation from this value eliminates mean reversion

Detailed Solution: Question 23

Mean reversion level = b0/(1-b1) when |b1|<1; this="" is="" the="" long-run="" equilibrium.="" this="" is="" the="" long-run="">

Practice Test: Quantitative Methods - 1 - Question 24

A quantitative analyst estimates a multiple regression model with n = 50 observations and k = 4 independent variables, obtaining an R-squared of 0.72. What is the adjusted R-squared?

A. 0.72, unchanged because the large sample size eliminates the degrees-of-freedom penalty
B. 0.71
C. 0.69

Detailed Solution: Question 24

Adj R² = 1-[(1-0.72)(49)/(45)] = 1-[0.28×49/45] = 1-0.3049 ≈ 0.69.

Practice Test: Quantitative Methods - 1 - Question 25

An analyst is building a predictive model for corporate bond defaults using 80 candidate variables, believing that the majority are noise with no true predictive value. Which regularization approach is most appropriate, and why?

A. Ridge regression, because its L2 penalty distributes the influence of irrelevant predictors equally across all coefficients without eliminating any
B. LASSO regression, because its L1 penalty can shrink irrelevant coefficients to exactly zero, producing a parsimonious model that excludes noise variables
C. Ordinary least squares, because adding all 80 variables guarantees the highest possible R-squared and therefore the best predictive model

Detailed Solution: Question 25

LASSO shrinks irrelevant coefficients to exactly zero, producing sparse parsimonious model; ideal when most predictors are irrelevant.

Practice Test: Quantitative Methods - 1 - Question 26

Two independent variables in a regression model have a pairwise correlation of 0.97, causing severely inflated standard errors. An analyst considers several remedies. Which of the following is the most appropriate response to this multicollinearity problem?

A. Remove one of the highly correlated variables from the model, accepting potential omitted variable bias as a necessary trade-off for reducing collinearity
B. Apply White-corrected standard errors to adjust for the inflated variance caused by collinear independent variables
C. First-difference both variables to eliminate the high correlation, which always resolves multicollinearity in cross-sectional data

Detailed Solution: Question 26

Removing one correlated variable is a primary remedy for multicollinearity; White correction addresses heteroskedasticity, not multicollinearity.

Practice Test: Quantitative Methods - 1 - Question 27

A machine learning analyst uses regularized regression to address the problem of overfitting in a high-dimensional financial model. Which of the following most accurately describes the relationship between the regularization parameter λ and model complexity?

A. Regularization increases model complexity by adding penalty terms that allow more coefficients to be freely estimated across the sample
B. Increasing the regularization parameter λ reduces the coefficient penalty, resulting in more flexible and complex models with lower in-sample error
C. Regularization reduces overfitting by penalizing large coefficient values; a higher value of λ produces greater coefficient shrinkage, yielding simpler and more generalizable models

Detailed Solution: Question 27

Higher λ imposes larger penalty on coefficients → greater shrinkage → simpler model → reduces overfitting.

Practice Test: Quantitative Methods - 1 - Question 28

A fraud detection model produces the following confusion matrix results:
Predicted Fraud Predicted No Fraud
Actual Fraud 60 25
Actual No Fraud 15 100
What is the model's recall (sensitivity)?

	Predicted Fraud	Predicted No Fraud
Actual Fraud	60	25
Actual No Fraud	15	100

A. 0.80
B. 0.71
C. 0.60

Detailed Solution: Question 28

Recall = TP/(TP+FN) = 60/(60+25) = 60/85 ≈ 0.71.

Practice Test: Quantitative Methods - 1 - Question 29

An analyst is reviewing two diagnostic tests used to detect problems in time-series regression models: the Breusch-Pagan test and the Durbin-Watson statistic. Which of the following correctly distinguishes the conditions each test is designed to detect?

A. The Breusch-Pagan test detects conditional heteroskedasticity by regressing squared residuals on independent variables; the Durbin-Watson statistic detects first-order serial correlation in regression residuals
B. The Durbin-Watson statistic detects heteroskedasticity by measuring the dispersion of residuals; the Breusch-Pagan test detects autocorrelation through a chi-square test on lagged residuals
C. Both the Breusch-Pagan test and the Durbin-Watson statistic test for the same condition: non-constant variance (heteroskedasticity) in the error term

Detailed Solution: Question 29

Breusch-Pagan tests for heteroskedasticity; Durbin-Watson tests for serial correlation in residuals.

Practice Test: Quantitative Methods - 1 - Question 30

An analyst regresses a non-stationary equity price index (Y) on a non-stationary bond price index (X) over a 10-year period. The regression produces an R-squared of 0.92 and a t-statistic of 12.4 on the slope coefficient. The Dickey-Fuller test fails to reject the null hypothesis for both series. Which of the following concerns is most relevant to the validity of this regression?

A. The results are reliable because the high R-squared of 0.92 confirms a genuine fundamental relationship between the two price series
B. The model is free of specification error because both variables are financial price indices measured in compatible units
C. The regression likely produces spurious results; non-stationary series should be tested for cointegration or first-differenced before regression analysis to avoid false inference

Detailed Solution: Question 30

Regressing two non-stationary series without cointegration test risks spurious regression despite high R² and significant t-stats.

	Quantitative Methods 10 docs\|3 tests

Quantitative Methods

10 docs|3 tests

Join Course

Information about Practice Test: Quantitative Methods - 1 Page

In this test you can find the Exam questions for Practice Test: Quantitative Methods - 1 solved & explained in the simplest way possible. Besides giving Questions and answers for Practice Test: Quantitative Methods - 1, EduRev gives you an ample number of Online tests for practice

	Quantitative Methods 10 docs\|3 tests

Quantitative Methods

10 docs|3 tests

Join Course

About this CFA Level 2 Practice Test

This test contains 30 MCQs on Practice Test: Quantitative Methods - 1 with detailed solutions for each question. It is part of the Quantitative Methods course on EduRev, designed to align with the current CFA Level 2 syllabus. Each solution explains not just the correct answer but also gives you an understanding of the topic — not just to attempt the question but also help you prepare for the exam with practice.

Explore more CFA Level 2 Study Resources

If you found this Practice Test: Quantitative Methods - 1 MCQ Test helpful, you can further enhance your CFA Level 2 preparation with in-depth courses on EduRev, for all the subjects and topics available in Quantitative Methods course. The course offers:

Video Lectures: covering all the topics of Quantitative Methods
Notes for Revision: offering revision notes, important questions, summaries, flashcards, cheatsheets & more for CFA Level 2 preparation
MCQ Test: variety of questions as per the pattern of CFA Level 2 exam

Signup to see your scores go up
within 7 days!

Continue with Google

Takes less than 10 seconds to signup

Quantitative Methods - 1 - Free MCQ Practice Test with solutions, CFA Level

MCQ Practice Test & Solutions: Practice Test: Quantitative Methods - 1 (30 Questions)

A portfolio manager estimates a multiple regression with four independent variables. The F-statistic is 18.45, which exceeds the critical value at the 1% significance level. Which of the following is the most accurate interpretation of this result?

An analyst adds a fifth independent variable to an existing four-variable regression model. The new variable has a t-statistic of 0.45. Which of the following best describes the most likely impact on the model's fit statistics?

An analyst suspects that a regression model of equity returns suffers from conditional heteroskedasticity. Which of the following procedures best describes the Breusch-Pagan test for detecting this condition?

A quantitative analyst observes that residuals from an equity return regression grow substantially larger during periods when the CBOE Volatility Index (VIX) is elevated, but are relatively stable during low-volatility regimes. This pattern is best characterized as:

A fixed-income analyst runs a time-series regression and computes a Durbin-Watson (DW) statistic of 0.85. The critical lower bound (d_L) is 1.50 and the upper bound (d_U) is 1.75 at the 5% significance level. Which conclusion is most appropriate?

A researcher models seasonal effects in quarterly equity returns using dummy variables to represent four quarters (Q1, Q2, Q3, Q4). To avoid perfect multicollinearity in the regression model, how many dummy variables should be included?

An analyst estimates an AR(1) model for a monthly economic index:
x_t = 0.50 + 0.75 × x_t−1 + ε_t
What is the mean-reverting level of this time series?

An analyst has estimated the following AR(1) model for quarterly GDP growth rates:
x_t = 0.30 + 0.80 × x_t−1
If the most recent observed value is x₅ = 2.50, what is the one-period-ahead forecast, x₆?

An analyst applies the Dickey-Fuller test to a monthly commodity price series. The test fails to reject the null hypothesis at the 5% significance level. Which of the following conclusions is most appropriate?

An analyst estimates an AR(1) model for a stock price index: x_t = 0.02 + 1.00 × x_t−1 + ε_t. A Dickey-Fuller test fails to reject the null hypothesis. Which of the following concerns is most relevant when using this series in a regression?

An analyst estimates an AR(1) model on weekly equity index returns and subsequently tests the residuals. The autocorrelation of squared residuals at multiple lags is statistically significant. This finding most likely indicates:

A regression analyst detects conditional heteroskedasticity in a model of corporate bond spreads. She decides to apply White-corrected standard errors. Which of the following best describes the effect of this correction on the regression output?

An analyst is evaluating two model specification errors: (1) omitting a variable that is correlated with included regressors, and (2) including an irrelevant variable with no true relationship to the dependent variable. Which of the following best contrasts the consequences of these two errors?

A machine learning analyst is comparing LASSO regression and ridge regression for a predictive model of equity returns. Which of the following most accurately distinguishes between these two regularization techniques?

A credit analyst builds a binary classification model to predict corporate defaults. The confusion matrix results are as follows:
Predicted Default Predicted No Default
Actual Default 40 20
Actual No Default 10 30
Which of the following correctly computes the model's precision, recall, and F1 score?

An analyst runs a time-series regression and computes a Durbin-Watson statistic of 0.92. The critical values are d_L = 1.53 and d_U = 1.74. Which of the following most accurately describes the consequences for OLS inference?

An analyst is interpreting the results of an AR(1) model estimated for a commodity spot price index. The estimated model is: x_t = b₀ + b₁ × x_t−1 + ε_t, where |b₁| < 1. Which of the following correctly describes the mean-reverting level of this series?

A quantitative analyst estimates a multiple regression model with n = 50 observations and k = 4 independent variables, obtaining an R-squared of 0.72. What is the adjusted R-squared?

An analyst is building a predictive model for corporate bond defaults using 80 candidate variables, believing that the majority are noise with no true predictive value. Which regularization approach is most appropriate, and why?

Two independent variables in a regression model have a pairwise correlation of 0.97, causing severely inflated standard errors. An analyst considers several remedies. Which of the following is the most appropriate response to this multicollinearity problem?

A machine learning analyst uses regularized regression to address the problem of overfitting in a high-dimensional financial model. Which of the following most accurately describes the relationship between the regularization parameter λ and model complexity?

A fraud detection model produces the following confusion matrix results:
Predicted Fraud Predicted No Fraud
Actual Fraud 60 25
Actual No Fraud 15 100
What is the model's recall (sensitivity)?

An analyst is reviewing two diagnostic tests used to detect problems in time-series regression models: the Breusch-Pagan test and the Durbin-Watson statistic. Which of the following correctly distinguishes the conditions each test is designed to detect?

Quantitative Methods

Quantitative Methods

Top Courses for CFA Level 2

Schools

Competitive Exams

International Exams

Quick Links

About this CFA Level 2 Practice Test

Explore more CFA Level 2 Study Resources

Quantitative Methods - 1 - Free MCQ Practice Test with solutions, CFA Level

MCQ Practice Test & Solutions: Practice Test: Quantitative Methods - 1 (30 Questions)

A portfolio manager estimates a multiple regression with four independent variables. The F-statistic is 18.45, which exceeds the critical value at the 1% significance level. Which of the following is the most accurate interpretation of this result?

An analyst adds a fifth independent variable to an existing four-variable regression model. The new variable has a t-statistic of 0.45. Which of the following best describes the most likely impact on the model's fit statistics?

An analyst suspects that a regression model of equity returns suffers from conditional heteroskedasticity. Which of the following procedures best describes the Breusch-Pagan test for detecting this condition?

A quantitative analyst observes that residuals from an equity return regression grow substantially larger during periods when the CBOE Volatility Index (VIX) is elevated, but are relatively stable during low-volatility regimes. This pattern is best characterized as:

A fixed-income analyst runs a time-series regression and computes a Durbin-Watson (DW) statistic of 0.85. The critical lower bound (dL) is 1.50 and the upper bound (dU) is 1.75 at the 5% significance level. Which conclusion is most appropriate?

A researcher models seasonal effects in quarterly equity returns using dummy variables to represent four quarters (Q1, Q2, Q3, Q4). To avoid perfect multicollinearity in the regression model, how many dummy variables should be included?

An analyst estimates an AR(1) model for a monthly economic index:xt = 0.50 + 0.75 × xt−1 + εtWhat is the mean-reverting level of this time series?

An analyst has estimated the following AR(1) model for quarterly GDP growth rates:xt = 0.30 + 0.80 × xt−1If the most recent observed value is x5 = 2.50, what is the one-period-ahead forecast, x6?

An analyst applies the Dickey-Fuller test to a monthly commodity price series. The test fails to reject the null hypothesis at the 5% significance level. Which of the following conclusions is most appropriate?

An analyst estimates an AR(1) model for a stock price index: xt = 0.02 + 1.00 × xt−1 + εt. A Dickey-Fuller test fails to reject the null hypothesis. Which of the following concerns is most relevant when using this series in a regression?

An analyst estimates an AR(1) model on weekly equity index returns and subsequently tests the residuals. The autocorrelation of squared residuals at multiple lags is statistically significant. This finding most likely indicates:

A regression analyst detects conditional heteroskedasticity in a model of corporate bond spreads. She decides to apply White-corrected standard errors. Which of the following best describes the effect of this correction on the regression output?

An analyst is evaluating two model specification errors: (1) omitting a variable that is correlated with included regressors, and (2) including an irrelevant variable with no true relationship to the dependent variable. Which of the following best contrasts the consequences of these two errors?

A machine learning analyst is comparing LASSO regression and ridge regression for a predictive model of equity returns. Which of the following most accurately distinguishes between these two regularization techniques?

A credit analyst builds a binary classification model to predict corporate defaults. The confusion matrix results are as follows:Predicted DefaultPredicted No DefaultActual Default4020Actual No Default1030Which of the following correctly computes the model's precision, recall, and F1 score?

An analyst runs a time-series regression and computes a Durbin-Watson statistic of 0.92. The critical values are dL = 1.53 and dU = 1.74. Which of the following most accurately describes the consequences for OLS inference?

An analyst is interpreting the results of an AR(1) model estimated for a commodity spot price index. The estimated model is: xt = b0 + b1 × xt−1 + εt, where |b1| < 1. Which of the following correctly describes the mean-reverting level of this series?

A quantitative analyst estimates a multiple regression model with n = 50 observations and k = 4 independent variables, obtaining an R-squared of 0.72. What is the adjusted R-squared?

An analyst is building a predictive model for corporate bond defaults using 80 candidate variables, believing that the majority are noise with no true predictive value. Which regularization approach is most appropriate, and why?

Two independent variables in a regression model have a pairwise correlation of 0.97, causing severely inflated standard errors. An analyst considers several remedies. Which of the following is the most appropriate response to this multicollinearity problem?

A machine learning analyst uses regularized regression to address the problem of overfitting in a high-dimensional financial model. Which of the following most accurately describes the relationship between the regularization parameter λ and model complexity?

A fraud detection model produces the following confusion matrix results:Predicted FraudPredicted No FraudActual Fraud6025Actual No Fraud15100What is the model's recall (sensitivity)?

An analyst is reviewing two diagnostic tests used to detect problems in time-series regression models: the Breusch-Pagan test and the Durbin-Watson statistic. Which of the following correctly distinguishes the conditions each test is designed to detect?

Quantitative Methods

Quantitative Methods

Top Courses for CFA Level 2

About this CFA Level 2 Practice Test

Explore more CFA Level 2 Study Resources

A fixed-income analyst runs a time-series regression and computes a Durbin-Watson (DW) statistic of 0.85. The critical lower bound (d_L) is 1.50 and the upper bound (d_U) is 1.75 at the 5% significance level. Which conclusion is most appropriate?

An analyst estimates an AR(1) model for a monthly economic index:
x_t = 0.50 + 0.75 × x_t−1 + ε_t
What is the mean-reverting level of this time series?

An analyst has estimated the following AR(1) model for quarterly GDP growth rates:
x_t = 0.30 + 0.80 × x_t−1
If the most recent observed value is x₅ = 2.50, what is the one-period-ahead forecast, x₆?

An analyst estimates an AR(1) model for a stock price index: x_t = 0.02 + 1.00 × x_t−1 + ε_t. A Dickey-Fuller test fails to reject the null hypothesis. Which of the following concerns is most relevant when using this series in a regression?

An analyst runs a time-series regression and computes a Durbin-Watson statistic of 0.92. The critical values are d_L = 1.53 and d_U = 1.74. Which of the following most accurately describes the consequences for OLS inference?

An analyst is interpreting the results of an AR(1) model estimated for a commodity spot price index. The estimated model is: x_t = b₀ + b₁ × x_t−1 + ε_t, where |b₁| < 1. Which of the following correctly describes the mean-reverting level of this series?

A fraud detection model produces the following confusion matrix results:
Predicted Fraud Predicted No Fraud
Actual Fraud 60 25
Actual No Fraud 15 100
What is the model's recall (sensitivity)?