ECON5007 Assignment (25%) Empirical Appraisal of the CAPM 1. Overview This assignment requires you to empirically test the Capital Asset Pricing Model (CAPM) using

ECON5007 Assignment (25%)

Empirical Appraisal of the CAPM

1. Overview

This assignment requires you to empirically test the Capital Asset Pricing Model (CAPM) using a sample of five stocks from the ASX200. You will implement both time-series and cross- sectional analyses, extend the model to include additional factors, and interpret results in terms

of both model validity and temporary disequilibria (i.e., which assets appear overpriced or

underpriced relative to CAPM predictions).

2. Tasks

(A) Data Collection & Preparation

Collect monthly total returns for 5 ASX200 stocks, the ASX200 index (market proxy),

and the Australian 3-month Bank Accepted Bill (BAB) yield (risk-free rate).

Compute monthly excess returns for each asset and the market.

Define excess return as follows:

!,# = !,# $,# %,# = %,# $,#

where !,# is the raw return of asset

in month , %,# is the raw market index return,

and $,# is the monthly risk-free rate.

(B) CAPM Time-Series Regressions

Estimate CAPM for each stock: !,# = ! + !%,# + !,# .

Report estimated α, β, t-statistics, and R2.

Discuss whether α differs significantly from zero. Interpret α as evidence of

disequilibrium (mispricing).

(C) Joint Test of Alphas

Apply the GRS test to evaluate whether all alphas are zero simultaneously.

Interpret the result in terms of overall support/rejection of CAPM.

2

(D) Cross-Sectional Regression (Security Market Line)

Regress average returns across assets on their CAPM betas: ! = & + ‘! .

Plot the SML, mark each stock, interpret slope and intercept.

Identify any asset points lying significantly above (underpriced, high return vs predicted)

or below (overpriced) the fitted SML.

(E) CAPM Extension: Fama–French 3-Variables

Re-estimate regressions including SMB and HML as additional variables. !# = ! + ‘!%# + (!# + )!# + !#

Compare α’s and R2 with CAPM.

Discuss whether the additional variables are significant and whether they reduce

mispricing.

(F) Fama–MacBeth Cross-Section

Run cross-sectional regressions each month of asset returns across assets on:

o β (CAPM beta),

o β2 (to test linearity of the SML)

o s (idiosyncratic volatility of CAPM residuals).

Report average slopes (λ), their t-stats, and cross-sectional R2.

Interpret:

o Is β priced

o Is there evidence of non-linearity

o Is idiosyncratic risk priced

3. Deliverables and Assessment Weighting

A report of max 1,500 words (excluding tables/figures/captions) with the following structure:

1. Introduction (background, objectives, brief on CAPM). [10%]

2. Data

and Method (description of data and method) [15%]

3. Analysis

a. CAPM Time-Series Analysis (alphas, betas, R2, interpretation ). [15%]

b. Cross-Sectional Analysis (SML plot, pricing errors). [15%]

c. Fama–French 3-Variables Analysis (comparison with CAPM). [15%]

d. Fama–MacBeth Cross-Section Analysis (β, β2, s , implications). [15%]

4. Discussion & Conclusion (Does CAPM hold Which assets are mispriced Insights).

[15%]

Include tables and figures generated from estimation, with short captions.

3

4. Key Questions to Address

Are CAPM alphas significantly different from to zero How about the other parameter

What does this imply about mispricing

Does the joint hypothesis test reject CAPM overall

Is the SML slope positive and significant Which stocks lie above/below the line

Do additional variables improve the model’s explanatory power

Does Fama–MacBeth show that β is priced Is the SML linear Is idiosyncratic risk

priced

Based on your results, which stocks appear overpriced and which are underpriced

5. Important Notes

You may use MATLAB, R, Stata, Python, etc., but MATLAB starter code

(capm_assignment.m) is provided.

When working in MATLAB, use monthly return (or excess return) for all series,

in decimal form (0.01 = 1%). Interpretation in report could be in percent.

Risk-free rate use 3-month BAB yield, converted to monthly return. If unavailable, the

RBA cash rate may be used as an alternative.

Clearly label tables/figures (e.g., Table 1: CAPM regression results).

Word limit is 1,500 words (excl. tables/figures/captions).

6. Due Dates