Download the country’s main stock index, its GDP (seasonally adjusted), its CPI (seasonally
adjusted), its 10 year government yield (or 5 year, if 10 years is not available), and its 3m (or 1y
if not available) government bond rate. Use quarterly data for all series.
Transform the data to your series of interest: GDP growth rates, inflation, stock market returns,
government bond yields, short-term interest rate.
Provide the description of the data (one paragraph), its descriptive statistics (including pairwise
correlations), and a figure.
Task 2
Determine the appropriate lag order using the AIC and BIC (show intermediate results in a table
with AIC/BIC values), and settle on a specific choice. Motivate your choice briefly.
Task 3
Provide an interpretation of your VAR model. (1 paragraph)
Task 4  
Make a (stand-alone) figure of the impulse responses. [so not a screen dump, but a figure as
you would have it in your own Journal of Finance publication]
Task 5  
State your 2013Q1 10Y gov yield, and your 2014Q1 yield. Assume that the 9Y and 10Y yield are
the same, compute your annual return on a 10Y government bond position (show derivation).
Your pension fund has to manage risks both at a 1-year and a 10-year horizon. For this case, we
concentrate on the 1 year horizon.
Consider the following three portfolios. (A) 10% short term government bonds, 45% equity,
45% long term government bonds; (B) 30% equity, 70% long term bonds; (C) 30% equity, 70%
short term bonds.
Assume your current funds are A1
 bn. Your funds after 1 year, say A0
, depend on the asset mix
you hold, as well as on the realizations of the returns.
Task 6 (4pt)
Using excel, STATA, VBA, Matlab, or whatever package you like, generate 1,000 scenarios 1 year
after your end-of-sample. Describe how you do this, and make graphs of the distributions of
these for each of your 5 variables.
Task 7 (1pt)
Provide the correlation matrix for your 5 1-year simulated variables, and compare this with
your correlation matrix in Task 1.