On the regionalization of input–output tables with an industry-specific location quotient

Abstract

The purpose of this study is to contribute to the ongoing debate about the performance of several location quotients (LQs), including Flegg’s location quotient (FLQ) and the industry-specific FLQ (SFLQ), which was newly developed by Kowalewski (Reg Stud, 240–250, 2015). Unlike FLQ, which has one unique $$\delta $$ -value for all industries in a region, SFLQ allows for a variation in the exponential parameter value of $$\delta $$ for each industrial sector in order to capture specialization and to reduce the tendency for overestimation or underestimation of each of the industries in a given region. Since the national and regional survey data are available for South Korea, this study evaluates and compares the results for SFLQ against those of other non-survey techniques. The findings prove that SFLQ is superior to other LQs, including FLQ and augmented FLQ. In addition, empirical evidence shows that the optimal values for the exponential parameters vary drastically among industries. SFLQ provides more detailed information on regional industries with individual $$\delta $$ values, which is definitely helpful when analyzing regional specialization and shifts in industrial structure. Furthermore, to determine the optimal exponential $$\delta ({\delta _j })$$ value in FLQ and SFLQ in advance, Flegg’s and Kowalewski’s regression models are tested with the use of Korean data. The statistics indicate that the regression equations are debatable and still need major improvement.