Anson T. Y. Ho, Kim P. Huynh, and David T. Jacho-Chavez, "Flexible
Estimation of Copulas: An Application to the US Housing Crisis",
Journal of Applied Econometrics, Vol. 31, No. 3, 2016, pp. 603-610.

There are two zip files. The file npcopula-code.zip contains the Stata
and R code files, which are ASCII files in DOS format. The file
npcopula-data.zip contains the data files in CSV format. Unix/Linux
users should use "unzip -a".

##########
## Data ##
##########

o Files: housing_data.csv
  Format: comma separated values
  Info: Original data used in Zimmer (2012). It covers 1975Q2 - 2009Q1,
  available at
  http://thedata.harvard.edu/dvn/dv/restat/faces/study/StudyPage.xhtml?studyId=85444&tab=files

o Files: housing_data_short.csv
  Format: Comma separated values
  Info: Revised data downloaded from Federal Housing Finance Agency web site
  (http://www.fhfa.gov/) on Dec 19, 2013. It covers the same sample period as
  housing_data.csv used in Zimmer(2012).

o Files: housing_data_long.csv
  Format: comma separated values
  Info: Revised and extended data downloaded from Federal Housing Finance
  Agency web site (http://www.fhfa.gov/) on Dec 19, 2013. It extended
  housing_data_short.csv to include data in 2009Q2 - 2013Q1.

o Files: parametric_results.csv
  Format: comma separated values
  Info: Parametric fitted values of the implied conditional probabilities
  based on the original data and model specification used in Zimmer (2012).

o Files: parametric_results_s.csv
  Format: Comma separated values
  Info: Parametric fitted values of the implied conditional probabilities
  based on the revised data downloaded from Federal Housing Finance Agency
  web site (http://www.fhfa.gov/) on Dec 19, 2013. The fit utilizes the
  same model specification used in Zimmer (2012).

o Files: parametric_results_l.csv
  Format: comma separated values
  Info: Parametric fitted values of the implied conditional probabilities
  based on the revised and extended data downloaded from Federal Housing
  Finance Agency web site (http://www.fhfa.gov/) on Dec 19, 2013.
  The fit utilizes the same model specification used in Zimmer (2012).

o Files: all-Transactions_prct-diff.csv
  Format: comma separated values
  Info: Calculated differences between previous series and the revised and
  extended versions.


Note: No other data are used

##############
## Software ##
##############

o Files: do_all.do
  Format: STATA do file (source code with comments)
  Application: Clayton-Gumbel mixture copula estimations.
  Info: This STATA do file filters the series of AR-GARCH effects, estimates
  copula parameters, and calculates conditional probabilities. It calls 6
  other do files (est_m_AZ_NV.do, est_m_AZ_NV.do, est_m_AZ_NV.do,
  est_m_AZ_NV.do, est_m_AZ_NV.do, est_m_AZ_NV.do) to do the estimation for
  each states pair.

o Files: bwx1.R
  Format: R transcript (source code with comments)
  Application: Cross-validated Bandwidth calculations.
  Info: x1 takes values {"0","1","2"}. The code executes all the
  cross-validated bandwidth calculations using "housing_data.csv,"
  "housing_data_short.csv," and "housing_data_long.csv" respectively.
  All bandwidth objects are then saved into "bwx1.Rdata" R objects.

o Files: bootx1.R
  Format: R transcript (source code with comments)
  Application: Bootstrap calculations.
  Info: This R transcript loads the "bwx1.Rdata" files above (where x1 takes
  values {"0","1","2"}) and performs all estimation and bootstrap
  calculations. All these are then saved into "bootx1.Rdata" R objects.

o Files: emp_copulax1.R
  Format: R transcript (source code with comments)
  Application: It plots Figures 2, 3, and 4 in the manuscript.
  Info: This R transcript loads the "bootx1.Rdata" files above (where x1
  takes values {"0","1","2"}) and reads the "parametric_results.csv,"
  "parametric_results_s.csv," and "parametric_results_l.csv" files into
  memory. It then plots the Figures 2, 3, and 4 in the paper, respectively.

o Files: att-Transactions-prct-diff.R
  Format: R transcript (source code with comments)
  Application: It plots Figure 1 in the manuscript.
  Info: This R transcript reads the "all-Transactions-prct-diff.csv" file
  and plots Figure 1 in the manuscript.
 
o The software packages used to obtain the results were:
  
  - STATA, version 13
    http://www.stata.com/
  
  - R system for statistical computing, version 3.0.2
    http://CRAN.R-project.org/
  
  - boot 1.3-9
    http://CRAN.R-project.org/package=boot
  
  - copula 0.999-9
    http://CRAN.R-project.org/package=copula
  
  - fGarch 3010.82
    http://CRAN.R-project.org/package=fGarch
  
  - foreign 0.8-55
    http://CRAN.R-project.org/package=foreign
  
  - np 0.50-1
    http://CRAN.R-project.org/package=np