한국(韓國) 제조업(製造業)의 기술적(技術的) 효율성(效率性) : 산업별(産業別) 기술적(技術的) 효율성(效率性)의 추정(推定)

Technical Inefficiency in Korea's Manufacturing Industries

본(本) 논문(論文)은 기존의 경험적(經驗的) 산업조직론(産業組織論)에서 소홀히 취급한 기술적(技術的) 효율성(效率性)의 문제를 분석하기 위하여 광공업(鑛工業)센서스의 세세분류제조업별(細細分類制造業別) 사업체(事業體)를 대상으로 한국(韓國) 제조업(製造業)의 기술적(技術的) 효율성(效率性)에 관한 가장 미시적(微視的) 차원(次元)의 경험적(經驗的) 연구결과(硏究結果)를 제공한다. 기술적(技術的) 효율성(效率性)의 산업간(産業間), 국가간(國家間) 격차(隔差)를 성명하기 위해서는 우선 적절한 격차효율성지표(隔差效率性指標)의 선택이 매우 중요할 것이다. 이를 위하여 본(本) 연구(硏究)에서는 확률적(確率的) 생산경계(生産境界)를 수정최소자승법(修正最小自乘法)으로 추정한 후 각 산업별(産業別)로 기술적(技術的) 효율성(效率性)의 네가지 상이한 척도(尺度)에 대한 추정을 시도한다. 추정(推定) 결과 효율성(效率性) 척도간(尺度間)의 상관관계(相關關係)는 부가가치액(附加價値額)보다 생산액(生産額)을 종속변수로 한 경우가 다소 높은 것으로 나타났으나 전자(前者)의 경우가 보다 많은 산업(産業)에 대하여 효율성(效率性)의 추정(推定)을 가능하게 하기 때문에 산업간(産業間) 비교(比較)를 위하여 더 적절하다고 판단된다. 상당수의 산업(産業)에 있어서는 만족할 만한 효율성(效率性) 추정치(推定値)를 구하는 것이 불가능했지만 후속연구(後續硏究)의 관점에서 볼 때 기술적(技術的) 효율성(效率性)의 추정(推定)이 가능한 산업(産業)이 다수를 차지한 사실은 고무적 이라고 판단된다. 또한 종업원규모(從業員規模)가 영세한 사업체(事業體)들이 상대적으로 비효율적(非效率的)임을 알 수 있었다.

Research on technical efficiency, an important dimension of market performance, had received little attention until recently by most industrial organization empiricists, the reason being that traditional microeconomic theory simply assumed away any form of inefficiency in production. Recently, however, an increasing number of research efforts have been conducted to answer questions such as: To what extent do technical ineffciencies exist in the production activities of firms and plants? What are the factors accounting for the level of inefficiency found and those explaining the interindustry difference in technical inefficiency? Are there any significant international differences in the levels of technical efficiency and, if so, how can we reconcile these results with the observed pattern of international trade, etc? As the first in a series of studies on the technical efficiency of Korea's manufacturing industries, this paper attempts to answer some of these questions. Since the estimation of technical efficiency requires the use of plant-level data for each of the five-digit KSIC industries available from the Census of Manufactures, one may consture the findings of this paper as empirical evidence of technical efficiency in Korea's manufacturing industries at the most disaggregated level. We start by clarifying the relationship among the various concepts of efficiency-allocative effciency, factor-price efficiency, technical efficiency, Leibenstein's X-efficiency, and scale efficiency. It then becomes clear that unless certain ceteris paribus assumptions are satisfied, our estimates of technical inefficiency are in fact related to factor price inefficiency as well. The empirical model employed is, what is called, a stochastic frontier production function which divides the stochastic term into two different components-one with a symmetric distribution for pure white noise and the other for technical inefficiency with an asymmetric distribution. A translog production function is assumed for the functional relationship between inputs and output, and was estimated by the corrected ordinary least squares method. The second and third sample moments of the regression residuals are then used to yield estimates of four different types of measures for technical (in) efficiency. The entire range of manufacturing industries can be divided into two groups, depending on whether or not the distribution of estimated regression residuals allows a successful estimation of technical efficiency. The regression equation employing value added as the dependent variable gives a greater number of "successful" industries than the one using gross output. The correlation among estimates of the different measures of efficiency appears to be high, while the estimates of efficiency based on different regression equations seem almost uncorrelated. Thus, in the subsequent analysis of the determinants of interindustry variations in technical efficiency, the choice of the regression equation in the previous stage will affect the outcome significantly.