Bill Provencher, "An Investigation of the Harvest Decisions of Timber Firms in the South", Journal of Applied Econometrics, Vol. 10, Supplement, 1995, pp. S57-S74. All data sets are in the zip file provench.zip. The files are in DOS format. Data on Tree Volume and Tree Growth: The original source of all data concerning timber growth and timber volume are the state-by-state forest inventories conducted periodically (about every seven years in the South) by the U.S. Forest Service. In particular, data for Georgia were derived from the inventory completed for Georgia in 1989. Data for Florida were derived from the inventory completed for Florida in 1987. Data for each inventory is available from the Forest Service for a price of $200-300. A description of the available data is found in General Technical Report NC-151, published by the North Central Forest Experiment Station, USFS. Here I provide only a brief description of the variables used in my research. When soliciting this data, I requested that for harvested plots it include the year of harvest (as indicated by the number of years between the inventory date and the previous harvest). This variable is not usually provided and is thus not reported in General Technical Report NC-151 (hereafter, "GTS NC-151"). The data files GASITE.TXT and FLOSITE.TXT contain the initial selection of variables from the 1989 inventory for Georgia and the 1987 inventory for Florida for observations meeting the selection criteria described in Provencher. The data in these files is organized around two record types: record type 20 (RECTYPE=20) and record type 30 (RECTYPE=30). Record type 20 concerns the characteristics of the forest stand (the plot). Record type 30 concerns the characteristics of the trees within the stand. So each observation of record type 20 is followed by a series of observations of record type 30 associated with the trees within the stand identified by the record type 20. Here is the sequence of data for each observation in the data set: REC (I5): Identifies the Forest Service plot; RECTYPE (I4): Identifies the type of record (20 or 30) If RECTYPE=20: CTY (I4) = County code (Census Bureau FIPS code; see GTS NC-151, var. 4); OWN (I3) = Ownership type (see GTS NC-151, var. 6, for a complete description; "20" =forest industry, "80" = NIPF leased by the forest industry); TYPOLD (I3) = Forest type at the previous survey (see GTS NC-151, var. 8); TYPCUR (I3) = Current forest type (see GTS NC-151, var. 7); STDAGE (I4) = Age of the stand in years (see GTS NC-151, var. 9); STORCUR (I2) = Current stand is natural (=1) or planted (=2) (see GTS NC-151, var. 11); STOROLD (I2) = At previous inventory stand is natural (=1) or planted (=2) (see GTS NC-151, var. 12); SITECL (I2) = Site productivity class (see GTS NC-151, var. 13); SI (I3) = Site index (see GTS NC-151, var. 14); GRSTKPC (I5) = Growing stock stocking (see GTS NC-151, var. 26); REMPER (F5.1) = Remeasurement period (years.months; see GTS NC-151, var. 28); if RECTYPE=30: STATUS (I3) = Tree status (see GTS NC-151, var. 8 for record type 30); SPGRP ((I3) = Species group (see GTS NC-151, var. 10 for record type 30); DBHCUR (F5.1) = Current diameter of the tree at breast height (in inches; see GTS NC-151, var. 11 for record type 30); DBHOLD (F5.1) = Diameter of the tee at breast height at the last survey (see GTS NC-151, var. 12 for record type 30); VOLFAC (F6.1) = Volume expansion factor to obtain volume per acre (see GTS NC-151, var. 18 for record type 30); MORTFAC (F6.1) = Mortality Expansion factor to obtain mortality per acre (see GTS NC-151, var. 19 for record type 30); REMVFAC (F6.1) = Removal expansion factor to obtain harvested volume per acre (for harvested plots; see GTS NC-151, var. 20 for record type 30); NETCFVL (F8.4) = Net cubic foot volume in the stem (see GTS NC-151, var. 21 for record type 30); NETCFGR (F8.4) = Net cubic foot growth in the stem (see GTS NC-151, var. 24 for record type 30); CUTTIME (I4) = Time elapsed since harvest (for harvested stems only). For harvested stems, the following years correspond to the value of CUTTIME: Georgia: 1982: CUTTIME=6 1983: CUTTIME=5 1984: CUTTIME=4 1985: CUTTIME=3 1986: CUTTIME=2 1987: CUTTIME=1 Florida: 1981: CUTTIME=6 1982: CUTTIME=5 1983: CUTTIME=4 1984: CUTTIME=3 1985: CUTTIME=2 1986: CUTTIME=1 From the data files GASITE.TXT and FLOSITE.TXT, the following data files were developed: GANOCUT.TXT, FLONOCUT.TXT, GACUT.TXT, and FLOCUT.TXT. The *NOCUT.TXT files contain the observations of the unharvested plots. The *CUT.TXT files contain the observations of the harvested plots. These files are ultimately the source files used to calculate volume state equation #1 (see Provencher). They are also used to determine the observations of volume per stem in each of the 24 quarters used in estimation (see Provencher). For GANOCUT.TXT and FLONOCUT.TXT, each observation is organized as follows: REC (I5) = Identifies the Forest Service plot; STDAGE (I5) = Age of the stand in years (see GTS NC-151, var. 9); MDBH (F8.4) = Average DBH for the plot, calculated by weighting DBHCUR for each admissible stem by VOLFAC; MDBHOLD (F8.4) = Average DBH for the plot at the previous inventory, calculated by weighting DBHOLD for each admissible stem by VOLFAC; MCBVL (F8.4) = Average cubic foot volume per stem, calculated by weighting NETCFVL for each admissible stem by VOLFAC; MCBGR (F8.4) = Average cubic foot growth per stem, calculated by weighting NETCFGR for each admissible stem by VOLFAC; CBVL (F10.4) = Cubic foot volume per acre, calculated by multiplying NETCFVL for each admissible stem by VOLFAC; CBGR (F10.4) = Cubic foot growth per acre per year, calculated by multiplying NETCFGR for each admissible stem by VOLFAC; CUTTIME (I4) = See above; SI (I3) = Site index (see above). FOR GACUT.TXT and FLOCUT.TXT, each observation is organized as follows: REC (I5) = Identifies the Forest Service plot; MDBHCUT (F8.4) = Average DBH for the plot at harvested, calculated by weighting DBHCUR for each admissible stem by REMVFAC; MDBHOLD (F8.4) = Average DBH for the plot at the previous inventory, calculated by weighting DBHOLD for each admissible stem by REMVFAC; MCBVL (F8.4) = Average cubic foot volume per harvested stem, calculated by weighting NETCFVL for each admissible stem by REMVFAC; CBVL (F10.4) = Cubic foot volume harvested per acre, calculated by multiplying NETCFVL for each admissible stem by REMVFAC; SI (I3) = Site index. Data on Timber (Stumpage) Prices: Timber prices were obtained from Timber Mart South, a quarterly timber sales report available by subscription which reports regional prices for stumpage and delivered wood in the South. The source of reported prices are private consultants with access to sealed-bid prices. The price series is generally considered by forest economists to be the best available published price series for the period 1977-present. Monthly reports for the average price of pine pulpwood stumpage in southeastern Georgia were averaged to get quarterly values. The same was done for pine pulpwood stumpage prices in northern Florida. These values were then deflated using the producer price index (base year 1982). The data set GAPRICE.TXT gives Georgia prices for the 24-quarter period extending from the first quarter of 1982 to the last quarter of 1987. The first variable of each observation denotes the period of observation (1 to 24, with "1" denoting the first quarter of 1982, "2" denoting the second quarter of 1982, and so on). The second variable denotes the price per cord in 1982 dollars. The data set FLOPRICE.TXT gives Florida prices for the 24-quarter period extending from the first quarter of 1981 to the last quarter of 1986. The first variable of each observation denotes the period of observation (1 to 24, with "1" denoting the first quarter of 1981, "2" denoting the second quarter of 1981, and so on). The second variable denotes the price per cord in 1982 dollars. For both data sets, the format is: I4,F6.2. Contacting the Author: Anyone with questions about this data, or the results generated by the data as reported in Provencher, should feel free to contact me at: Department of Agricultural Economics 330 Taylor Hall University of Wisconsin, Madison 53706 (608)-262-9494 provencher@agecon.wisc.edu