Kap-Young Jeong and
Robert T. Masson
"A new methodology linking concentration dynamics to current and steady state profits: examining Korean industrial policy during take-off"
JEL codes: L11, L52, O10, O53
: We model the dynamics of industrial structure and market power using Korean manufacturing data during the take-off period (1978-82). StructureÞConductÞPerformance [SCP] methodologies have been criticized for using accounting data and because a few superior firms may have greater shares and greater profits. Both are argued to present possible spurious correlations between concentration and profits. This paper follows earlier work which shows that market structure responds to observed accounting profits as if these were accurate indices of real profits, and not as if leading firms are perceived to have unmatchable advantages.
The methodological contribution of this model is a new latent variable for steady state profits derived from the speed of structural adjustment. Long run profits are identified by the hypothesis that structural adjustment will be more rapid when industries are farther from steady state levels. We analyze the long run profits latent variable, finding strong support for this hypothesis. The speed of adjustment is greater for positive and negative deviations from steady state structure. We show that the profits and structural adjustment relationship is non-monotone. The SCP criticisms above are based on spurious correlations which are monotone in profits and structure. Positing spurious correlations which are at the same time monotone in structure and non-monotone in structural adjustment seems less plausible than accepting the SCP results which are consistent with both.
Our analysis also is new in that it is the first direct econometric analysis of Korean industrial policy during its take-off years. We note that in Korea, unlike elsewhere, Industrial Policy was hypothesized to lead to concentration and market power in the popular press and by professional economists. Ours is the first direct econometric analysis of this hypothesis, and we find it supported.