Stochastic Imputation of Missing Physical Commodity Trade Information Using Monetary Trade Data

International trade had conventionally been expressed in monetary values until the recommendations by the UN Statistics Division. Because international trade in monetary terms alone is not entirely representative of global trade from a transportation and logistics perspective, the physical dimension...

Full description

Saved in:
Bibliographic Details
Published in:Transportation research record 2013, Vol.2354 (1), p.122-132
Main Authors: Ong, G. P., Farhan, J., Chin, A. T. H.
Format: Article
Language:English
Subjects:
Classification
Commodities
International trade
Mathematical models
Statistics
Stochasticity
Transportation
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:International trade had conventionally been expressed in monetary values until the recommendations by the UN Statistics Division. Because international trade in monetary terms alone is not entirely representative of global trade from a transportation and logistics perspective, the physical dimension of international trade is attaining increased importance. However, missing physical quantities in the trade flow databases are observed; these quantities are due largely to (a) noncompliance of reporter countries with the standard units of measurement or classification, (b) confidentiality issues, and (c) erroneous collection and reporting of certain data. This paper, therefore, first presents the existing methods used in the literature to treat the issue of missing commodity weight information in international physical commodity trade databases and then proposes a stochastic multivariate imputation model using auxiliary variables such as monetary trade data and the price index to impute missing physical quantities. The relative performance of those methods in resolving the issue of incomplete physical commodity trade data is then evaluated and compared through a case study. It is concluded that the proposed approach outperforms the existing approaches for commodity flow data imputation.
ISSN:0361-1981
2169-4052
DOI:10.3141/2354-13