Affiliations: [a] Department of Economics, University of Pennsylvania, Philadeplhia, PA 19104, USA. Tel.: +1 215 898 8485 | [b] Department of Earth Sciences and Geography, Bridgewater State College, Bridgewater, MA 02325, USA. Tel.: +1 508 531 2633/+1 508 984 5251
Correspondence:
[*]
Corresponding author.
Note: [1] The University of Pennsylvania and Bridgewater State College. Support of the National Science Foundation (SES99-11377) for this work is gratefully acknowledged. An earlier version of this paper was presented at the Joint World Bank-OECD Seminar on Purchasing Power Parities, Washington D.C., January 30–February 2, 2001.
Abstract: This paper examines alternative approaches to multilateral comparisons based on spatial chains. The ground covered includes a review of the consistency problem for countries with multiple-benchmarks and a discussion of what might give rise to it. While much of our recent work has been with respect to the Penn World Table (PWT), most of this paper will be concerned with benchmark comparisons. In particular we argue that serious consideration be given to spatial chaining as a method of multilateral comparison, but there still remain major problems to solve. The case we make is as much empirical as it is methodological. Our application is to the 115 country 1996 reduced benchmark data. For this data set, we examine spatial linking based upon the spanning tree approach of Robert Hill, but using both price similarity as well as the Paasche-Laspeyres spread as criteria. We go on to describe how we might extend this benchmark over time and space along the lines of PWT.
