Productivity and growth decomposition: a novel single-index smooth-coefficient stochastic frontier approach
提出一种新的单指标平滑系数随机前沿方法，利用挪威高科技制造业和知识密集型商业服务业的企业数据，将产出增长和全要素生产率增长分解为技术变化、投入驱动和效率变化等成分，发现技术偏向资本或劳动存在行业异质性。
Abstract Our paper investigates productivity, output growth and total factor productivity (TFP) growth using a novel single-index smooth-coefficient stochastic frontier approach and two firm-level datasets respectively from the high technology (high-tech) manufacturing and Knowledge Intensive Business Services (KIBS) sectors in Norway. The approach considers input productivity and technical inefficiency to be flexible functions of production environmental variables indexed with unknown parameters for more precise estimation of marginal effects of these variables on the frontier and inefficiency. Output growth is decomposed into technical change (TC), input-driven component (IDC) and efficiency change (EC), while TFP growth is decomposed into TC, scale component and EC. The primary objective is to (i) maximise output through the frontier and efficiency channels and (ii) enhance productivity growth through such channels as technical progress and efficiency improvement, specifically tailored for the manufacturing and services industries. The empirical results reveal substantial heterogeneity in technology across firms. Overall speaking, geographical industrial concentration, export intensity and urbanisation positively influence output in both sectors. Technical progress contributes to TFP growth in both sectors; however, TC is biased towards capital in the high-tech sector and driven by labour in the KIBS sector. In addition to TC, TFP growth in the high-tech and KIBS sectors also benefits from EC and IDC, respectively.