Roxana Chiriac and Valeri Voev, "Modelling and Forecasting Multivariate 
Realized Volatility", Journal of Applied Econometrics, Vol. 26, No. 6, 
2011, pp. 922-947.

The raw data used in this paper are confidential. All the results in the
paper, however, can be replicated using the realized covariance matrices and
open-to-close returns, which we provide.

The high-frequency data used in this paper are from the Trade and Quotation
(TAQ) database disseminated by the New York Stock Exchange (NYSE). The TAQ
data files contain continuously recorded information on the trades and
quotations for the individual stocks listed on the NYSE, American Stock
Exchange (AMEX), and the National Association of Security Dealers Automated
Quotation system (NASDAQ). The sample period examined in this study is from
January 1, 2000 to July 30, 2008 (2156 trading days). The sample consist of
all quotes of the following six stocks: American Express Inc. (AXP),
Citigroup (C), General Electric (GE), Home Depot Inc. (HD), International
Business Machines (IBM) and JPMorgan Chase & Co (JPM). The daily transaction
record extends from 9:30 EST until 16:00 EST. Each quote consists of a bid
and an ask price along with a "time stamp" to the nearest second.

To construct the subsampled realized covariance matrices, we compute the
midquote ((bid price + ask price)/2) and sample at 5 min frequency by
shifting the grid of observations by 10 seconds, thus obtaining 30 distinct
sets of 5-minute-spaced observations.

The GAUSS procedure rc_ss.gss computes the subsampled realized covariance in
a pairwise fashion. The full matrix is obtained by looping over the six
stocks.

The ASCII dataset otoc_returns.txt (comma-delimited) contains the series of
daily open-to-close returns computed as the difference in the log-midquote
at the end and the beginning of each day. Column 1 is the calendar day,
column 2 the calendar month, and Column 3 the calendar year. Columns 4 to 9
contain the daily open-to-close returns for the six stocks in the order AXP,
C, GE, HD, IBM, JPM.

The ASCII dataset RC_matrices.txt (comma-delimited) contains the series of
daily subsampled realized covariance matrices. Each row corresponds to one
day and the matrix is recorded in a vech form, i.e., by vectorizing the
unique upper triangle of the matrix and transposing it into a row vector.

The two data files are ASCII files in DOS format. They and the program file
rc_ss.gss are zipped in the file cv-files.zip. Unix/Linux users should use
"unzip -a".

Please address any questions to:

  Valeri Voev
  CREATES
  School of Economics and Management
  University of Aarhus
  Building 1326
  Batholins All? 10
  DK-8000 Aarhus C
  Denmark
  email: vvoev [AT] creates.au.dk