MSc Accounting and Finance: Applied Econometrics

Assessed Exercise

Instructions

Please complete the Exercise below and submit before 10/12/2017 at 23:59 on The Hub. Submission will only be possible via The Hub. Late submissions will not be considered. Please complete the exercise within your allocated syndicate groups and submit one solution for each group. The maximum number of pages for this exercise is 5, excluding tables and graphs. Papers exceeding the length limit will only be evaluated on the rst 5 pages.

Exercise – Pairs trading

This empirical exercise is based on the work by Gatev, Goetzman and Rouwenhorst (2006).

Portfolios and factors construction

Let's construct tradable portfolios from the given dataset.

1. Upload data ( lename: assignment data17.mat) in Matlab.

2. Data contain daily close prices, standard industrial classi cation (SIC) code and market cap for a large cross-section of stocks in the US. The rst row contains the permno of any given stock, whereas the second row contains the date, the third the SIC code, the fourth the closing price, and the fth market cap.

3. Upload Fama-French, momentum factors and NYSE ME breakpoints monthly data from Kenneth French's website. (careful, ME breakpoints are divided by 1,000,000 on French's database, so you will have to multiply the numbers by 1,000,000 to compare them with the market cap in the

.mat le) http://mba.tuck.dartmouth.edu/pages/faculty/ken.french/data_library.html

2017-2018 Page 1

(Link).

4. Form portfolios of pairs of stocks following the methodology of Gatev, Goetzman and Rouwenhorst (2006), ltering out stocks below the bottom ME quintile. Please only consider portfolios of the top 5 pairs and the top 20 pairs. (10% of mark)

Trading strategies

Calculate the return series of the following trading strategies:

Pairs trading, using only the top 5 and top 20 portfolios, ltering out stocks below the bottom ME quintile.

For any trading strategy, report summary statistics, Sharpe ratios, t-ratios, maximum drawdowns, a plot of the cumulative returns over time.

Comment on the di erent performances and risk-return pro les of the strategies.

(60% of mark)

Risk attribution

Run regressions of each strategy's returns on the three Fama-French factors plus mo-mentum and comment on the sign, size and the signi cance of the estimated alphas and betas.

(30% of mark)

References

Gatev, E., W.N. Goetzmann and K.G. Rouwenhorst (2006). Pairs Trading: Perfor-mance of a Relative-Value Arbitrage Rule. The Review of Financial Studies, Volume 19, Issue 3, Pages 797-827.

2017-2018 Page 2

EDIT:2017.11.27