top of page

Level II CFA® Program exam: Quantitative Methods cheat sheet

In this series of revision posts, we ask your AB Maximus CFA® Program exam trainers to give you quick tips and essential advice for different chapters in the curriculum. Handy for revision or simply for a last minute review to make sure you’re thoroughly prepared – don't miss the chance to brush up on your knowledge and do a little extra prep!

3 must-know concepts are: How values in the ANOVA table relate to each other, the three violations of regression assumptions, and the different Autoregressive Model tests.

1. How values in the ANOVA table relate to each other

Candidates should be able to understand all the different elements of this table and how they relate to each other. If a candidate can read an ANOVA table for multiple regression, then he or she should be able to understand it for linear regression.

Candidates should be able to interpret all the different values in the table. To practice, hide one column or one row and derive the values based on the rest of the table. Once a candidate understands this concept, he or she will have covered at least 25% of the Quantitative Methods programme.

2. The three violations of regression assumptions

Be familiar with the three violations of regression assumptions:

  • Heteroscedasticity: heteroscedasticity is any set of data that isn’t homoscedastic. It refers to data with unequal variability (scatter) across a set of second, predictor variables.

  • Serial correlation: also known as auto correlation, serial correlation is a term used to describe the relationship between observations of the same variable over specific periods of time. If it is measured to be zero, it means there is no correlation, and each of the observations are independent of one another.

  • Multicollinearity: this refers to a situation where a number of independent variables in a multiple regression model are closely correlated to one another.

3. The different Autoregressive Model tests

Autoregressive Model (AR): This is a model where a value from a time series is regressed on previous values from that same time series.

The different tests to be performed to ensure that an AR can be used to derive reliable estimates for out-of-sample data: first ensure that the model is covariance stationary, then test for serial correlation, seasonality, and heteroscedasticity. Only when a model passes all these tests successfully can the model be used on out-of-sample data.

Students often forget: how covariance stationarity affects the validity of results from a time series.

Students often fail to understand the concept of covariance stationarity. A time series must display covariance stationarity in order to derive results that are valid. A time series without covariance stationarity will have economically invalid results, because the regression leads to spurious results, the estimate of b1 is biased, and hypothesis tests will be invalid.

The key words to remember for covariance stationarity are constant and finite. The expect value in all periods, the variance in all periods, and the covariance with all lagged versions of the time series, is constant and finite.

A practical tip is: know how to perform simple calculations well and quickly.

When using the ANOVA table, you should be able to manipulate the different values in the table quickly. To do this, focus on a set of practice questions. Think about how the inputs in the questions can be combined or transformed to derive other values.

Instead of trying to solve as many questions as possible, try to focus on understanding selected questions and mastering key concepts fully. The calculations in themselves are straightforward and do not involve long equations.

About the Author

Marie-Laure Caille, CFA, MSc is the AB Maximus trainer for Economics. She has worked in the Boston Consulting Group, United Nations and OECD. Marie currently lectures at the ESSEC Business School and is the Director of The Human Factor (APAC).​

Related Posts
bottom of page