Example Data Sets

Several data sets are used across multiple examples, which can help to show different approaches to analyzing a given set of data.

GRUNFELD.DAT

This is a commonly used data set for "small N" panel techniques. It is a "PRN" (labeled ASCII) file with data for 10 large U.S. companies over 20 years of annual data for 1935 to 1954. The series are gross investment in 1947 dollars (INVEST), firm market value at end of year in 1947 dollars (FIRMVALUE) and stock of capital and equipment in 1947 dollars (CSTOCK). The firms are (in order) GM, US Steel, General Electric, Chrysler, Atlantic Refining, IBM, Union Oil, Westinghouse, Goodyear, Diamond Match, American Steel.

This is used in PANEL.RPF and SWAMY.RPF and a smaller version (just two series) is used in SUR.RPF. In a different format, it is used in textbook examples in the Baltagi panel data text, Hill-Griffiths-Lim and Koop's Bayesian methods.

G10XRATE.XLS

This is a long (6237 observation) data set in XLS format of daily exchange rates for the US dollar vs ten currencies (in existence at the time). Series are USXJPN, USXFRA (French franc), USXSUI, USXNLD (Dutch guilder), USXUK, USXBEL (Belgian franc), USXGER (German mark), USXSWE (Swedish krona), USXCAN and USXITA (Italian lira). Note that, for consistent treatment, these are all quoted as USD value of a foreign currency unit whether that is the standard quotation or not. This was used in Pritsker(2006) (for univariate GARCH models).

This is used in most examples of multivariate GARCH and related models such as GARCHMV.RPF, GARCHDECO.RPF and COPULA.RPF, as well as some of the univariate GARCH examples.

OECDSAMPLE.RAT

This is a macroeconomic data set in RATS format with a mix of quarterly and monthly data for the G7 countries, running from (roughly) 1960 through 2006, though what is available depends upon a series. It's used in many of the examples of Vector Autoregressions, such as IMPULSES.RPF, CANMODEL.RPF, CVMODEL.RPF and HISTORY.RPF.

NELSONPLOSSER.RAT

This has a set of long annual U.S. macroeconomic series from a variety of sources used in Nelson and Plosser(1982) to examine different ways of representing a trend in economic data. It has also been used in later papers examining more complicated trend models.