Macro-Economic Factors for Enhancing Construction Output and Policy Formulation in Kenya

Abstract

Management of construction industry at the micro level is currently a problem in Kenya. This happens in three dimensions: construction demand estimation, construction supply targeting and construction output control in the country. This inefficiency in the management of construction industry has resulted in a number of policy failures regarding the industry and its performance. This report presents a research study on the impact of macro-economic factors on the annual construction output in Kenya, in a bid to address this inefficiency in the macro management of construction industry. The study objectives were: analysis of the trends of construction output and the macro - economic factors, establish relationship between macroeconomic factors and construction output and to establish the way past levels of construction output have been influencing present levels over the period of forty-three (43) years. The variables were: annual construction output, inflation rate, unemployment rate, commercial banks weighted interest rate, population growth rate and US dollar to Kenya shilling exchange rate. Time series data analysis methods were used to analyze data which were obtained from Central Bank of Kenya (CBK) and the Kenya National Bureau of Statistics (KNBS) covering forty-three (43) years; from 1977 to 2019. Eviews for windows version 10; a computer software package, was used to analyze the data. Observations from multiple regression output tables of current macro-economic variables showed insignificant influence of macro-economic factors on annual construction output in Kenya in the current year. It was however noticed from regression of lagged macro-economic values that construction output in Kenya responds to the effects of the factors more than a year after they are implemented. Models to this effect have been developed having coefficients of determination (R2) values of 0.13 and 0.48 respectively. Nonlinear regression results of annual construction output and the macro-economic factors were not significant. However dynamic regression modeling of construction output in Kenya which was as a result of co-integration amongst the variables gave reliable and encouraging results. The coefficient of determination (R2) of the dynamic model was 0.92. A number of vigorous tests were carried out to check/show the reliability of the dynamic model. It was however found that, the model was limited in its application due to the condition of integration of order one (I/1). ARIMA modeling of construction output produced encouraging and reliable results as well. ARIMA (0,1,0) was modeled for annual construction output in Kenya which showed that construction output in the current year depended on the output of the past one year. The developed ARIMA model was evaluated on its predictive power on the basis of an out-of-sample forecast. Therefore, the two adopted measures of accuracy; namely MAPE and RMSE produced fairly good results and hence, the model can be used for forecasting construction output in Kenya. Based on the results obtained, conclusions are drawn that macro –economic factors can be used in managing Kenya’s construction industry as effective policy instruments.