Dynamic interaction between transportation investment and economic development based on stochastic approach

Abstract

Bangladesh has developed a huge road infrastructure at the cost of millions of dollars borrowed from development partners. Lack of policy strategy, inadequate planning and biased regional focus create investment inefficiency in transport sector of Bangladesh. During the last couple of decades transportation attracted more than twenty percent of national investment - most of which is concentrated in road sector. Hence, there exists a scope to examine the role of accessibility, provided by multimodal transportation facilities, in aggregate efficiency of production of different regions. This study establishes a- methodological framework and analysis the impact of transportation infrastructure investment with regional economic efficiency_ of Bangladesh based on c!~oic:.-theoretic Stoc_hastic Data Envelopment Model (SDEM) and compares the result with conventional Data Envelopment Analysis (DEA). (A common measure for relative efficiency is given by the ratio of the weighted sum of outputs and weighted sum of inputv It is assumed that there exists a stochastic error teml in the calculation of efficiency. Each DMU chooses its own production function as .the function yields optimum output. Assuming the error term is lID Gumbel Distributed, the efficiency of the target unit in a set is obtained by solving by conventional logit model, which is non-linear in fonnulation. It is further simplified by using monotonic logarithmic transformation of the efficiency function and assuming Cob-Douglas type multiplicative form of the input and output functions. The parameters of the function are estimated using Maximum Likelihood method having the following Log-Likelihood function. BIOGEME Statistical package is used to estimate the multinomiallogit model. Based on this framework/the time lag between transportation infrastructure investment \, and economic development is evaluated with an objective of assessing priorities in investment and decision making) In the context of Bangladesh, the data for the 20 greater districts or region is analyzed using Choice theoretic SDEM considering each region as an individual production unit. To calculate the relative efficiencies of different regions, effective land area, civilian labour force, industrial asset and accessibility index (as measure of transportation infrastructure and system) are used as~inpu-ts-and the total GDP is considered as-ou-tput. The Accessibility Index has been calculated from the existing transportation network of the country. Results in SDEM suggest that the proposed framework facilitates relaxation of the constraints although it enables incorporation of stochastic error to in assessing production frontiers. All the parameters measured using SDEM are found statistically significant. Analyses show that many regions perform inefficiently and there exists imbalance in input factors of production as demonstrated by the value of exponents. It also shows that there are specific regions where production efficiency is severely constrained by lack of land and industrial capita!. (Analytic framework has been developed to capture time lag between transportation investment and its utilization in economic activities based on stochastic approac~ The analysis reveals that the lag time ranges between 2-3 years for short term, 6-7 years for medium term and 16-17 years for long term in case of Bangladesh. Assessment of factor demand functions demonstrate that price elasticity of accessibility is around -0.90. The analytical procedures for assessing elasticities and demand functions have significant implication on policy formulation and resource allocation at regional level.