Geoscience Reference
In-Depth Information
The detailed conversion processes occasionally involved case-by-case
reconciliations. Inevitably, some conversions involved mapping one sector into
more than one. The light-gray cells of Fig.
4.2
represent one-to-one allocations. The
dark-gray cells denote bridge allocations with plausible weights specified on a case-
by-case basis.
A major problem in developing an inter-state interindustrial model stems from
the fact that it is difficult to obtain data describing trade flows among states in the
U.S. Since 1993, however, CFS data have been used, in spite of the fact that there
are still problems such as high sampling variability and disclosure rules limiting the
use of individual company data. The existence of many unreported values requires
relying on other data sources to approximate completeness. It is not surprising,
therefore, that since the work by Polenske (
1980
) and Faucett Associates (
1983
),
there has been no comprehensive inventory of MRIO flows in the U.S.
The CFS reports trade flows between U.S. states for 43 SCTG sectors while the
industry and commodity sectors (the data file of the IMPLAN Version 2) include
509 sector estimates, available for all states. The CFS trade flows data include both
foreign and domestic trade. This means that all commodities coming into a
U.S. port are listed as
outbound from that port
and inbound to the next destination;
likewise, all commodities going to a port from anywhere in the U.S. are outbound
from the origin and inbound to the port. For these reasons IMPLAN are added to the
IMPLAN Total Commodity Output tally. In the current application, the 1997 CFS
data were used as a baseline and updated to 2001 year using 2001 IMPLAN data.
The years are being updated to a more recent period.
Differences among alternative industry classification systems from different data
sources make data reconciliation especially difficult in the absence of standardized
and tested conversion bridges. The estimation of trade flows from CFS, therefore,
required intermediate conversion steps between the SCTG code system used in the
CFS and the IMPLAN system of sectors, not always one-to-one matched pairs.
Figure
4.3
shows the data reconciliation steps when creating a SCTG-IMPLAN
conversion bridge enabling the aggregation of 509 IMPLAN sectors to 43 SCTG
sectors.
The following paragraphs and equations summarize the NIEMO model. The
traditional Leontief demand-side model is expressed as
X
¼
LINV F
ð
4
:
1
Þ
where
X
is the
m
1 total output vector for
m
sectors
F
1 vector of final demand from private consumers, government,
investment, and net exports of outputs from
m
sectors, and
LINV
is the
m
)
1
, where
m
matrix of
technical coefficients that captures interindustry relationships in terms of back-
wards linkages between
m
sectors and
I
is the identity matrix.
The inverse matrix in Eq. (4.1) is referred to as the demand-driven I-O model.
The demand driven version of NIEMO can be expressed similarly as
is a Leontief inverse matrix, (
I A
A
is the
m
