André Kallåk Anundsen, "Econometric Regime Shifts and the US Subprime Bubble," Journal of Applied Econometrics, Vol. 30, No. 1, 2015, pp. 145-169. The data used in the article are collected from various sources (please see detailed descriptions below) and cover the period from 1975q1-2010q4 (144 observations). The attached data sets contain the raw data used in the analysis, and all necessary data transformations used in the analysis of the paper are described below and in Table A.1 in Appendix A of the paper. The data are organized in two .xls files. One contains the data used to estimate the inverted demand equation and the price-to-rent model ("Data to estimate models.xls"), and one contains the data used to test for Granger non-causality between the composite bubble indicator and the various financial instability measures ("Data for GNC tests.xls"). Please find a full description of the data sources below (NSA = not seasonally adjusted, SA = Seasonally adjusted): The two .xls files are zipped in the file anundsen-xls.zip. These are binary files. The same data are also contained in two .csv files, which are zipped in anundsen-cls.zip. Since the .csv files are ASCII files in DOS format, Unix/Linux users should use "unzip -a". "Data-to-estimate-models.xls": - PH = FHFA (former OFHEO) housing price index, all transactions Collected from the website of the Federal housing finance agency: http://www.fhfa.gov/ , NSA - CPI = Consumer price index for all urban consumers, all items Collected from the website of the Bureau of Labor Statistics: http://www.bls.gov/ , NSA - CPI2 = Consumer price Index for all Urban consumers, all items less shelter Collected from the website of the Bureau of Labor Statistics: http://www.bls.gov/ , NSA - Y = Disposable personal income from the Bureau of Economics Analysis Collected from the website of the Federal Reserve Bank of St. Louis: http://research.stlouisfed.org/fred2/ , SA - POP = Total US population from the Census Bureau Collected from the website of the Federal Reserve Bank of St. Louis: http://research.stlouisfed.org/fred2/ , NSA - R = Consumer Price Index for all Urban Consumers, rent of primary residence Collected from the website of the Bureau of Labor Statistics: http://www.bls.gov/ , NSA - tau_p = Marginal tax rate on personal property from the FRB-US model Collected by personal contact with Luke van Cleve at the Federal Reserve Board, NSA - tau_y = Marginal personal income tax rate at twice the median income Collected by personal contact with Luke van Cleve at the Federal Reserve Board, NSA - delta = Depreciation rate on housing from the National Income and Product Accounts Collected by personal contact with Joshua Gallin at the Federal Reserve Board, NSA - i = Weighted average of fixed and flexible mortgage interest rates from the Monthly Interest Rate Survey Collected from the website of the Federal housing finance agency: http://www.fhfa.gov/ , NSA - PJ = Price index for residential structures from the Lincoln Institute of Land policy Collected from the website of Lincoln Institute for Land policy: http://www.lincolninst.edu/subcenters/land-values/price-and-quantity.asp , NSA - H = Aggregate replacement cost of residential structures from the Lincoln Institute of Land policy Collected from the website of Lincoln Institute for Land policy: http://www.lincolninst.edu/subcenters/land-values/price-and-quantity.asp , NSA - sp = The number of new loan originations as a share of total loan originations from the Moody's analytics, NSA. Unfortunately, I cannot legally make these data available for free, but they are available for purchase at a total cost of $50 at www.economy.com. More precisely, the series needed to construct the subprime measure used in the paper (Subprime loans/Total loans) are available at the following webpage: https://www.economy.com/store/shop.aspx?tabs=data&mid=44B5816A-E8E8-49BA-94DB-FDA383F8582F&header=H0002004000520024&src=default.aspx - MT = Dummy variable equal to 1 for the period 1975q1--1982q3, and 0 otherwise. For motivation for this dummy, please consult Section 4 in the paper - CGT = Dummy variable equal to 1 for the period 1997q4--2010q4, and 0 otherwise. For motivation for this dummy, please consult Section 4 in the paper The operative measures used in the econometric analysis are: ph = log(PH/CPI2) = log of real housing prices r = log(R/CPI2) = log of real rents y = log(Y/(POP*CPI2)) = log of real per capita personal disposable income h = log(H/(POP*PJ)) = log of real per capita housing stock UC = (1-tau_y)*(i+tau_p) + delta - inf = Real direct user cost, where inf = (CPI - CPI_4)/CPI_4, with _4 indicating that I have used the L^4 lag operator In the attached data set, I have constructed both inf and UC as well. "Data-for-GNC-tests.xls": - Unemployment = Civilian unemployment rate in percent of the labor force from the Bureau of Labor Statistics, NSA - Industrial production = Industrial production index from the Board of Governors of the Federal Reserve System, SA - Delinquency rates = Delinquency rates on loans secured by real estate (all commercial banks) from the Board of Governors of the Federal Reserve System, NSA - Loan losses = Net loan losses divided by average total loans (all US banks) from the Federal Financial Institutions Examination Council, NSA - Non-performing loans = Non-performing loans in percent of total loans (loans that are 90-days or more past due) from the Federal Financial Institutions Examination Council, NSA - Financial stress index = St. Louis Financial stress index, NSA - Tightened credit standards = Net percentage of banks tightening standards for mortgage loans from the Senior Loan Officer Opinion Survey, NSA - Financial conditions index = Chicago FED national financial conditions index, NSA All the series, with the exception of the series "Tightened credit standards", used in the GNC analysis are available at the website of the St. Louis FED: http://research.stlouisfed.org/fred2/ The series "Tightened credit standards" has been downloaded from the Board of Governors of the Federal Reserve System's website for the Senior Loan Officer Opinion Survey: http://www.federalreserve.gov/datadownload/Choose.aspx?rel=SLOOS These data run through the period 1995q4--2010q4, since the GNC tests are performed on the sample 1995q4--2010q4. The data set "Data-for-GNC-tests.xls" also include the BI measures calculated in the paper: - BI_invdem = Bubble indicator derived from the inverted demand model - BI_ptor = Bubble indicator derived from the price-to-rent model - BI = 0.5*BI_invdem + 0.5*BI_ptor = Average of the two indicators An Ox code for reproducing the results of the paper will be made available at http://www.andre-anundsen.com/. Please address any questions to: André Kallåk Anundsen Department of Economics University of Oslo PO Box 1095 Blindern N0317 OLSO Norway Or by email to: a.k.anundsen [AT] econ.uio.no.