Daniela Scidá, "Structural VAR and Financial Networks: A Minimum
Distance Approach to Spatial Modeling", Journal of Applied
Econometrics, Vol. 38, No. 1, 2023, pp. 49-68.

The raw data used in this paper come from the Oxford-Man Institute's
Realized Volatility Library version 0.3 (Heber, G., Lunde, A.,
Shephard, N., & Sheppard, K., 2009). The raw data are publicly
available on Oxford-Man Institute's website and are open source. The
Realized Library is based on underlying high frequency data, which the
authors obtain through Thomson Reuters Tick History. This paper uses a
subset of the available raw data. 
 
The data used in this paper consist of daily realized return
volatility at 5-minute intervals (RV5) of main stock indices of a set
of 16 world leading indices:

#   Index            Country                 Code
1   BVSP             Brazil                  BRA
2   MXX              Mexico                  MEX
3   DJIA             USA                     USA
4   FTSE 100         United Kingdom          GBR
5   EURO STOXX 50    European Union          EURO
6   SSMI             Switzerland             CHE
7   AEX              The Netherlands         NLD
8   DAX              Germany                 DEU  
9   CAC40            France                  FRA
10  IBEX35           Spain                   ESP
11  FTSE MIB         Italy                   ITA
12  Nikkei 225       Japan                   JPN
13  KOSPI            South Korea             KOR
14  HSI              Hong Kong               HKG
15  FTSTI            Singapore               SGP
16  AORD             Australia               AUS

Some of the variables in the raw data had missing values. A version of
the data downloaded in 2015 was first used to reduce the number of
missing values. Then, dates where more than half the variables were
missing were dropped. Finally, the remaining missing values were
inputted with the prior value.

The final data consist of 3,829 observations, spanning the period
01/02/2004 to 12/31/2018. There are 18 variables. The first two
variables correspond to the date and observation count. The next 16
variables correspond to the RV5 for each of the above mentioned 16
world leading indices. The index names in the first row are Reuters
Tickers. The corresponding country or region is provided in the table
above. 

NOTE 1: The EU leading index (EURO STOXX 50) is included to control
for EU countries not readily available in the Oxford-Man data set
(e.g., Greece and Portugal), and it is similar to the DJIA in the US.
The EURO STOXX 50 index is comprised of 50 large companies located in
12 EU countries (Austria, Belgium, Finland, France, Germany, Greece,
Ireland, Italy, Luxembourg, the Netherlands, Portugal, and Spain).
Although many stocks from France, Germany, Spain, and Italy are in the
EURO index, the index is diversified, and no country/component can
dominate. Moreover, the market views this index as a critical gauge of
Europe's overall health. 

NOTE 2: A log transformation should be applied to the RV5 columns to
obtain log-realized volatility.

The file Scida_data.csv is a comma-separated-values file. It is an
ASCII file in DOS format, and it is zipped in the file ds-data.zip.
Unix/Linux users should use "unzip -a".


Daniela Scida
daniela.scida [AT] rich.frb.org