For the CFA^{®} Program exam

The Level II CFA^{®} Program exam is often regarded as the hardest among candidates and charterholders. The curriculum is much denser and answering 120 item set questions is not an easy task. In addition, most candidates attempting the Level II exam of the CFA^{®} Program are a lot more serious, and the competition is therefore tougher.

In the 2015 CFA^{®} Program Exam Preparation Survey, it was found that Level II candidates studied on average 315 hours to prepare for the test and most of them use third-party preparation material, which includes questions banks and/or mock exams. The hardest chapters were Derivatives, Alternative investments, and Fixed income. The good thing here is that the chapters are the same as in the level I exam. However, the weights of these chapters in the Level II CFA^{®} Program exam are completely different. Equity valuation is now the most tested chapter, whereas Quantitative methods, Economics, Alternative investments and Portfolio management each represent only 5-10% of the test.

At AnalystPrep, we are cognizant of the fact that passing the Level II CFA^{®} Program exam requires a mix of qualitative and quantitative knowledge as well as problem-solving abilities. AnalystPrep's Level II Question Bank for the CFA^{®} Program mimics the item sets found in the actual exam by covering the most important concepts and calculation steps from in the curriculum. AnalystPrep's item sets contain 6 questions each and are based on past exam questions.

AnalystPrep’s Level II QBank for the CFA^{®} Progam consists of 200 item sets from all of the chapters. Each one of them is made up of a case statement and six multiple choice questions with detailed answers.

There is no Magic Formula

Try out a few test item sets by registering here. You won’t be disappointed! Then, let AnalystPrep’s popular advanced built-in analytics software take over. In your Dashboard, you will be able to see your progress, analyze your results using a customizable timeline, and identify your strongest and weakest chapters. Moreover, you will be able to compare your results with all other candidates that use AnalystPrep’s platform.

The only way to pass the Level II CFA^{®} Program exam is extensive practice. However, you can also greater your chances of passing the exam by studying in a *smart way*. Here is the major difference between Level I and Level II of the CFA^{®} Program:

As shown from the topic area weights given by CFA Institute, Equity Investments, Fixed Income and Financial Reporting and Analysis represent 40-65% of the exam. Therefore, candidates should focus a large proportion of their efforts on these three chapters.

AnalystPrep’s Level II Question Bank for the CFA^{®} Program has been designed by CFA charterholders that understand exactly what the Level II CFA^{®} Program exam is made of. Most item sets include calculation-based questions as well as qualitative questions that focus on the concepts that candidates are assumed to understand after having read in details the curriculum.

**150+**Questions- -
- -
- Performance dashboard
- -
- -

**200**Vignette style cases- Unlimited quizzes
- 2 mock exams
- Performance dashboard
**10**ask-a-tutor questions- -

**200**Vignette style cases- Unlimited quizzes
- 2 mock exams
- Performance dashboard
**Unlimited**ask-a-tutor questions**$30**discount on AnalystPrep's Level III products for the CFA^{®}Program

You represent an organization of 10 or more people? Contact Us to learn more about our group deals.

Example of a Level II CFA

Brittany Ruiz is the chief investment officer (CIO) at Black-Mount Capital Advisory Firm (BMCA), a leading U.S. based money management firm. Last year, BMCA launched the BMCA Equity Fund that invested in stocks of domestic and international markets including those of Russia, Brazil, Canada and U.K. During the fund’s annual performance review, Ruiz discovered that the fund earned a return considerably less than that of its benchmark. Concerned with the outcome, Ruiz called a meeting with the fund’s portfolio management team to discuss key issues related to quantitative analysis of the fund’s investments. Anthony Webb, a leading quantitative expert was invited to chair the meeting. While analyzing the regression models used by the PM team to predict stock returns, Webb requested for the information provided in Exhibit 1. The information relates to a time-series model for annualized monthly returns to a Brazilian stock index.

Lag | Autocorrelation |

1 | 0.1320 |

2 | 0.1856 |

3 | -0.0098 |

4 | 0.0056 |

5 | -0.0062 |

Critical t-value | 1.96 |

Number of observations | 398 |

After his analysis, Webb presented the following conclusions to the portfolio management team:

Conclusion 1: “This time series can be best modeled using a moving average (2) model rather than an autoregressive model.”

Conclusion 2: “A moving average (2) model would be different from a simple moving average because the former would place different weights on the terms in the moving average. And unlike the MA(2), a simple moving average is based on observed values of the time series.”

Since 10% of the BMCA Equity Fund was invested in large-cap Russian stocks, Webb considered it crucial to study the characteristics of the Russian capital markets. While reviewing BMCA’s autoregressive (1) model used to predict the Russian inflation rate, Webb suspected that the model might have a unit root. To accurately test for the presence of a unit root in the time series, Webb decided to regress the first difference of the time series on the first lag of the time series. After obtaining the results of the regression, Webb calculated the t-statistic in the conventional manner for the coefficient. He then used the critical t-values computed by Dickey and Fuller to determine significance. Since he could not reject the null hypothesis at the 5% significance level, Webb concluded that the series did not have a unit root and was stationary.

When talking to Ruiz about ‘random walks’ Webb stated that a time-series that was a random walk could not be analyzed using standard regression analysis. He made the following statements:

Statement 1: “For a random walk, there is no finite mean-reverting level. In addition, the variance of the time-series increases or decreases as we go further into the future without any lower or upper bound. This violates the assumption of a finite variance for the time-series.”

Statement 2: “To model a random walk, we should first-difference it. This results in a mean-reverting level of 0. In addition, the variance of a first-differenced time-series is not only finite, but is also constant. The time-series then becomes covariance stationary.”

Webb continued by stating that if a time-series is a random walk, it is best to model the first-differenced series with an autoregressive model to predict future movements in the time-series. He also stated that the key to choosing the correct model was to analyze each model’s R

Ruiz then asked Webb to help the team decide which model to use to forecast the Canadian interest rate: An AR(1) model or an AR(2) model of Canadian interest rates. Ruiz presented Webb with the following information:

1. The standard errors from the AR(1) model and the AR(2) model were 4.359 and 4.109 respectively.

2. Using a sample period from 1985 to 2010, the AR(1) model’s residuals exhibit serial correlation, but the AR(2) model is correctly specified.

3. For a shorter sample period from 1995 to 2010 both models were correctly specified.

4. The root mean squared errors from the AR(1) and the AR(2) models were respectively 4.213 and 4.250.

5. The Canadian interest rate showed considerable volatility from 1985-2010, with values becoming much larger after 1995.

As an ending note, Webb explained the presence of conditional heteroskedasticity to BMCA’s portfolio management team. He stated that if a time-series model was conditionally heteroskedastic, then the standard errors of the regression coefficients would be incorrect, and the hypothesis tests would be invalid. He added that this was true only for autoregressive models but not for moving-average models or autoregressive moving-average models.

Webb is most accurate with respect to:

A) Conclusion 1 and conclusion 2.

B) Conclusion 2 only.

C) Neither conclusion 1 nor conclusion 2.

-------

The correct answer is A)

Conclusion 1 is correct. The standard error is 1/(398)^{1/2} = 0.05012. Using this standard error and autocorrelations given in Exhibit 1, we can see that only the first two autocorrelations are significantly different from 0 and all autocorrelations beyond that are 0. Hence, an MA(2) model is most appropriate.

Conclusion 2 is correct. A simple moving average gives equal weight to all the periods in the moving average whereas a moving-average time-series model can put different weights. A simple moving average is based on observed values of a time series.*Study Session 3, Reading 11, LOS 11(e): explain how autocorrelations of the residuals can be used to test whether the autoregressive model fits the time seriesStudy Session 3, Reading 11, LOS 11(o): determine an appropriate time-series model to analyze a given investment problem and justify that choice*

Is Webb’s conclusion regarding the presence of a unit root in the inflation rate time-series most likely correct?

A) Yes.

B) No, because the regression variables that he used and his computed t-statistic were incorrect.

C) No, because his application of the regression-based unit root test was incorrect.

-------

The correct answer is C)

If we cannot reject the null hypothesis, then the time series has a unit root and is non-stationary.*Study Session 3, Reading 11, LOS 11(j): describe implications of unit roots for time-series analysis, explain when unit roots are likely to occur and how to test for them, and demonstrate how a time series with a unit root can be transformed so it can be analyzed with an AR modelStudy Session 3, Reading 11, LOS 11(k): describe the steps of the unit root test for nonstationarity and explain the relation of the test to autoregressive time-series models*

Webb is most accurate with respect to:

A) Statement 1 only.

B) Statement 2 only.

C) Neither statement 1 nor statement 2.

-------

The correct answer is B)

Statement 1 is incorrect. The variance of a random walk increases without an upper bound: As ‘t’ increases, variance can approach infinity. There is a lack of an upper bound, not a lower bound. In fact, this explains why we cannot apply standard regression analysis on a random walk time series.

Statement 2 is correct.*Study Session 3, Reading 11, LOS 11(i): describe characteristics of random walk processes and contrast them to covariance stationary processes*

With regards to his comments about random walks, Webb is most likely:

A) Correct.

B) Incorrect about the criterion of choosing the correct model.

C) Incorrect about the use of first-differenced AR models and about the criterion of choosing the correct model.

-------

The correct answer is C)

Modeling the first-differenced series with an AR model does not help predict the future and is not necessarily the best model. Also, we cannot necessarily choose which model is correct solely by comparing the R^{2} of the two models.*Study Session 3, Reading 11, LOS 11(i): describe characteristics of random walk processes and contrast them to covariance stationary processes*

For predicting the Canadian interest rates, which model should most likely be used:

A) The AR(1) model.

B) The AR(2) model.

C) Neither the AR(1) model nor the AR(2) model.

-------

The correct answer is A)

The RMSE for the AR(1) model is smaller than the RMSE for the AR(2) model and hence, the AR(1) model is more accurate out of sample. Although for a longer time period, the AR(1) model is not correctly specified, the vignette clearly states that volatility increased significantly after 1995. Hence, the best approach would be to model each period separately. Since the AR(1) model is correctly specified for the smaller sample period, it can be used for predicting interest rates.*Study Session 3, Reading 11, LOS 11(o): determine an appropriate time-series model to analyze a given investment problem and justify that choice*

With respect to his comment about conditional heteroskedasticity, Webb is most likely:

A) Correct.

B) Incorrect with respect to autoregressive moving-average models.

C) Incorrect with respect to autoregressive moving-average models and moving-average models.

-------

The correct answer is C)

The standard errors of the regression coefficients in AR, MA, or ARMA models will be incorrect, and hypothesis tests would be invalid.*Study Session 3, Reading 11, LOS 11(m): explain autoregressive conditional heteroskedasticity (ARCH) and describe how ARCH models can be applied to predict the variance of a time series*