In this dissertation, the process of economic growth in Bangladesh under the
administration of the government of Pakistan was studied first. The politicoeconomic forces which caused the rate of economic growth in Bangladesh to be
slower tha.n that of West Pakistan were identified and analyzed. The existence of
technological progress and the form of the returns to scale were studied by
estimating non-linear CES production functions. Strong evidence of only
decreasing returns to scale in Bangladesh was observed.
The rate of economic growth in Bangladesh was found to be not only low but
also unstable owing to the primitive economic structure of production. The purpose
of economic planning in Bangladesh was postulated to make optimum allocation of
resources in order to transform the structure of production for generating high,
stable and selfsustaining economic growth. A macro-economic model representing
the economic structure was then estimated, This model was designed to be useful
for deciding optimum allocations for promoting both short-run and for long-run
policy objectives. The model also allows application of the results for various
growth theories in an aggregate development planning frame work.
The solutions of the model for optimal planning decisions were obtained by
applying the GSEM approach which is flexible enough to yield solutions for single
or multiple objectives. The reduced form matrix of the model was used to group the
variables into target, instrument and data variables and constants. The concept of
the pseudoinverse which enables one to obtain inverses of rectangular or singular
matrices was then used to obtain the final solutions. This method was used to
evaluate the allocation decisions of the Annual Plan announced by the Bangladesh
Planning Commission. The model indicated that the decissions of the Planning
Commission were reasonably correct. The model was then applied to recommend
optimal allocations for the first year of a hypothetical five year plan for Bangladesh.
A few useful guidelines for the next five year plan for were also extracted from
these solutions. The dissertation ends with a discusion of these useful properties as
well as various shortcomings of the model and a hope that the usefulness of the
model is likely to be enhanced with more disaggregation and greater availability of
reasonably accurate data in future.