Global Journal of Human Social Science, E: Economics, Volume 21 Issue 4

argument of the minimum (argmin) of equation 2.4 as follows: ( ) ( ) θ θ T T GMM Q W Q min arg == ∧ .……2.5 The Generalized Method of Moments (GMM) of estimation of DSGE model was employed in analysis of determinants of maritime output which were labor productivity in maritime sector in Kenya as depicted by equation 2.6. t t t Lbr Y ε β β + + = 2 0 .……..….2.6 Where t t t Lbr Y ε , , represents output, labor productivity and error term. Secondary data on port of Mombasa between 2000 and 2019 was sourced from KPA, KFS and various economic surveys among other sources for the analyses. After data collection, analysis was carried out using Stata 13.0 software. Various diagnostic tests were carried out. Inferential statistics were used to understand relationships between different variables. III. R esults This section presents the results and discussions of the examination of the effects of labor productivity on performance of maritime sector using the GMM approach and simple regression model. a) Simple Regression Estimation of Parameters The result for the Simple Regression estimation is shown in table 3.1. Table 3.1: Simple Regression Results In Output ( InOPT ) Coeff. Std. Error Z- Value P > |Z| [95% Conf. Inter. In Foreign Exchange Rate ( InFOREX ) 3.5694 0.6688 5.34 0.000 2.2129 4.9258 Intercept -15.4600 3.0878 -5.01 0.000 -21.7224 -9.1976 Source: Author (2020) The results showed that the intercept coefficient was negative and significantly ( 05.0 000 .0 < = p ) determined the output in the maritime sector in Kenya during the period under study. From the above regression equation it was revealed that holding labor productivity to a constant zero; the intercept coefficient indicated that output produced in maritime sector in Kenya 5,178,365 DWT. With the use of bootstrap which is used to obtain improved estimates and confidence intervals in statistics, the results indicated the range in which the parameter coefficients are normal based on the 95 percentage confidence interval. This implied that the coefficient of InFOREX 3.5694 was within the normal- based 95 percentage confidence interval of 2.2129- 4.9258. In general, transformed models where the dependent variable and independent variables are transformed, a coefficient represent elasticity. The results in this study, the coefficient of foreign exchange rate, 3.5694 was positive and significant at 5% level, 05.0 000 .0 < = p . This implied that every one percent increase in labor productivity, output increased by about 3.5694%. b) GMM Estimation of Parameters The results for the system GMM estimation is shown in table 3.2. Table 3.2: System GMM Estimation GMM Ivregress_GMM Arellano-Bond Betas SE P Betas SE P Betas SE P _cons -6.7142 1.3410 0.000 -13.7809 2.0519 0.000 Indept. Variables In ForeignExhange Rate ( InFOREX) 3.5684 0.6688 0.000 2.9817 0.4146 0.000 Instrumental Variables -14.8527 1.5385 0.000 3.2744 0.3346 0.000 Model Diagnostics R-squared 0.9927 Wald chi2 318.90 458.76 Sig. 0.000 0.000 Source: Author (2020) Volume XXI Issue IV Version I 28 ( E ) Global Journal of Human Social Science - Year 2021 © 2021 Global Journals Effect of Foreign Exchange Rate on Maritime Sector Performance in Enhancing Economic Growth in Kenya