Johannes Van Biesebroeck, "Complementarities in Automobile Production",
Journal of Applied Econometrics, Vol. 22, No. 7, 2007, pp. 1315-1345.

The data used in this paper are the property of a private company, Harbour
Consulting (http://www.harbourinc.com/), to whom enquiries concerning access
should be addressed. I collected all statistics from a set of spreadsheets
that basically have the same format as the tables in the annual yearbooks
(Harbour Consulting, The Harbour Report: Competitive Assessment of the North
American Automotive Industry. Rochester, MI: Harbour Consulting. Annual.) 

The data used cover the time period 1994-2004 and were published in 11
consecutive annual volumes for 1995 to 2005. Volumes can be ordered from
Harbour Consulting (information is on their website, and back copies are
quite cheap) and are carried in many libraries (for example, the reference
library of the city of Toronto carries a couple of volumes). The data
covers, especially in later years, the universe of assembly plants in North
America. The construction of the variables used in the analysis should be
obvious from the discussion in the text and the column headings. Most
variables are used directly as published. A few others involve summing up
across columns (for the outsourcing index) or a division of two numbers (for
the flexibility index).

One comma-delimited file (location_of_plants.csv) is provided for each
assembly plant used in the study. It contains the latitude and longitude,
the location (city, state), and nationality of the owner. This information
was used to calculate the distance between each plant and the industry
center (constructed weighting by production quantities) and the
country-specific centers. Distances are calculated using the usual formula
for the "great circle distance":

3962.6*ACOS(SIN(latitude/57.3)*SIN(av_latitude/57.3)+COS(latitude/57.3)
*COS(av_latitude/57.3)*COS(longitude/57.3-av_longitude/57.3))

The final column in the CSV file contains the distance from each plant to
the unweighted North-American industry center, calculated using the above
formula.

The CSV file location_of_plants.csv is an ASCII file in DOS format. It is
zipped in the file jvb-data.zip. Unix users should use "unzip -a".


Johannes Van Biesebroeck
johannes.vanbiesebroeck [AT] utoronto.ca