The robustness of results developed in the study depends upon
all three factors in Eq. (2) (Y, A and R). A main limitation on the
data used in the analysis is the Xiamen 2007 IO table. It is
compiled as a ‘competitive import type’ IO table in which imports
are included in intermediate and final uses [75]. As a result,
production-inducing effects can be overestimated because some
intermediate inputs could be from outside the region (imports).
Via validation of the 2007 IO table, it is noted that all the input
from the ‘Extraction of Petroleum and Natural Gas’ sector and
significant fraction of the input from ‘Agriculture, Forestry, Animal
Husbandry and Fishery’ sector are imports from outside of Xiamen.
Since both sectors affect the estimates of regional economic
impacts as well as the environmental effects, actual economic and
environmental impacts associated with these two sectors within
the Xiamen region are likely lower than those reported here. To
account for such problems caused by regional trade patterns,
additional data are needed to estimate regional supply coefficients
[70]. In addition, the IO model developed in the study does not
include many of the useful policy variables (e.g., income and
employment) as in a social account matrix (SAM) based IO model
(e.g., IMPLAN). Finally, the IO approach itself is subject to a set of
well known limitations (e.g., linear assumption, fixed price and
production coefficients). Thus, a future extension of the study will
include the development of computable general equilibrium (CGE)
models for Xiamen.