Ron Alquist and Lutz Kilian, "What Do We Learn from the Price of Crude
Oil Futures?", Journal of Applied Econometrics, Vol. 25, No. 4, 2010, 
pp. 539-573.

Errata 6/3/13: Bugs fixed in the files that compute the p-values for
the Clark-West test statistic and the files that bootstrap p-values
for the tests of equal predictive accuracy under quadratic and
absolute loss. This changes some entries in Tables 1-5 in the
published version of the paper, but it does not affect the
conclusions.

There are four zip files. They all contain ASCII files in DOS format.
Unix/Linux users should use "unzip -a".

The file ak-data.zip contains files with the data used in the paper. The
first two columns of each file contain the year and the month.

prices.txt	Contains crude oil spot and futures prices. Source:
		PriceData.com.

volume.txt 	Contains the volume of futures contracts traded. Source:
		PriceData.com.

pd.txt		Contains the precautionary demand component of the real
		price of oil. Source: Kilian (2009).

impvol1m.txt	Contains the implied volatility of options on 1-month crude
		oil futures. Source: Commodity Research Bureau.

irate.txt 	Contains the interest rate on US Treasury bills. Source:
		Federal Reserve Board of Governors website.

dslag.txt	Contains the trailing geometric average of past changes in the
		spot price of crude oil. Source: Authors' calculations.

The file ak-programs.zip contains the MATLAB programs used in the production
of the tables and figures:

mspedm.m	Computes the Diebold-Mariano test statistic under quadratic
		loss.

mspecw.m	Computes the Clark-West adjusted p-values for the
		Diebold-Mariano test statistic under quadratic loss.

mapedm.m	Computes the Diebold-Mariano test statistic under absolute
		loss.

nrbpt.m		Computes the Pesaran-Timmeran direction of change test
		statistic.

ols.m		Runs an OLS regression.

olsvarc.m       Estimates a VAR in companion matrix form.

nwvcmat.m	Computes the Newey-West estimator of the variance-covariance
		matrix.

nwesterr.m	Computes the Newey-West standard errors.

recursivefcast.m   Estimates recursive forecasts for the spread regressions.

recursivefcastb1.m Estimates recursive forecasts for the spread regressions
		   assuming beta = 1.

rows.m		Returns the number of rows in a matrix.

nwest.m	        Computes Newey-West robust test statistics for an
		OLS regression.

vec.m		Vectorizes a matrix.

pfind.m		Finds lag-order length.

vline.m		Inserts vertical lines into a figure.

bstrap.m	Returns bootstrapped p-values for test of equal predictive
                accuracy statistic under quadratic and absolute loss.

bsmapeld.m	Returns bootstrapped p-values for test of equal predictive
                accuracy under absolute loss for forecasts based on the
		random walk with local drift.

ar.m		Estimates AR(p) model.

cols.m		Returns the number of columns in a matrix.

The file ak-tables.zip contains the MATLAB files for all the tables. For
Tables 1 through 5, these are:

tablexxsprd.m	Returns the table entries for the forecasts based on spread
		regressions.

tablexxnrbf.m	Returns the table entries for the forecasts that do not use
		regression to generate the forecasts.

tablexxld.m	Returns the table entries for the forecasts based on the
		random walk with local drift model.


For Tables 6, 7, 8a, 8b, 8c, 9, there is just one file per table.

The file ak-figures contain the MATLAB files that return Figure 1, Figure 2,
Figure 4, Figure 5, Figure 7a, and Figure 7b.

  Ron Alquist
  Bank of Canada
  234 Wellington Street
  Ottawa, ON
  K1A 0G9

  ralquist [AT] bankofcanada.ca