Premium
CONCENTRATION‐PROFITS MONOPOLY VS. EFFICIENCY DEBATE: SOUTH AFRICAN EVIDENCE
Author(s) -
LEACH DANIEL F.
Publication year - 1997
Publication title -
contemporary economic policy
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.454
H-Index - 49
eISSN - 1465-7287
pISSN - 1074-3529
DOI - 10.1111/j.1465-7287.1997.tb00461.x
Subject(s) - monopoly , economics , profitability index , variables , econometrics , variable (mathematics) , efficiency , regression analysis , dominance (genetics) , microeconomics , statistics , mathematics , chemistry , mathematical analysis , biochemistry , finance , estimator , gene
Using the new official measures of concentration, this study finds a strong and highly significant correlation between concentration and industry profitability for South African manufacturing industries. This correlation is consistent with both the monopoly hypothesis of the traditional structure‐conduct‐performance paradigm and Demsetz's efficiency hypothesis that concentration of industry reflects the dominance of superior low‐cost firms. Using the four‐firm concentration ratio, this study applies Chappell and Cottle's extension of the Demsetz test to distinguish between the monopoly explanation of this correlation and the efficiency explanation. Using a conventional regression specification of the concentration‐profits relationship with no control for efficiency, this study finds a statistically significant coefficient only for the industry and large size class regressions. Because small firms should do at least as well as large firms, under the price umbrella provided by large firms, this result is inconsistent with the monopoly hypothesis but consistent with the efficiency hypothesis. Specifying a variable that measures the efficiency of large plants relative to smaller plants—the cost advantage ratio—yields striking results. Similar to the results obtained by Chappell and Cottle for the United States, this study finds the efficiency variable to be highly significant in the industry and large size class regressions, while concentration becomes insignificant. Moreover, the efficiency variable is insignificant in the smaller size classes Both the statistical insignificance of concentration and the pattern of the efficiency variable coefficients suggest that including concentration in the regression specification without the efficiency variable simply highlights the effect of the efficiency variable. Coming as they do from an environment as different from the United States as the South Africa of the 1980s, these results strongly support the Demsetz efficiency hypothesis.