New evidence on the Korean wage curve

An exhaustive empirical and theoretical study conducted by Blanchflower and Oswald (1994) found a new empirical regularity of the negative relationship between wages and regional unemployment, that is, the ‘wage curve.’ The curve is roughly described by – 0.1 of the elasticity of wages with respect to the local unemployment rate with human capital characteristics of individual workers controlled for (ibid, p. 361). This finding is based on individual data on wages and regional unemployment rates within 12 countries, and is preand re-confirmed by their own work and other studies dealing with different countries. However, estimates in Table 9.1 of Blanchflower and Oswald (1994) reveal some variation in the wage curve elasticity with South Korea having the smallest estimate (in an absolute value) of 0.04, suggesting that wages are relatively rigid in Korea than other countries. This observation seems at odds the fact that the union membership rate is lower in Korea than in most developed countries. For example, according to the Labour Force Statistics, organization of economic co-operation and development, the union membership rate of 1997 is lower in Korea (11.1%) than in the United States (14.1%), Japan (22.6%), Australia (30.3%), UK (30.4%), Germany (34.8%) or Sweden (90%). It is generally recognized that, unlike the union sector, the nonunion sector offers competitive wages so that nonunion wages tend to move more flexibly depending on labour market conditions. In addition, wage negotiations are relatively decentralized in South Korea. This article re-examines the Korean wage curve to determine if the Korean labour market obeys the minus-point-one-rule and thereby to re-evaluate the wage flexibility in Korea. In fact, Blanchflower and Oswald (1994, p. 362) state that ‘[S]outh Korea, for which an industry wage curve alone can be estimated, has a low unemployment elasticity of 0.04. Future research, no doubt, will aim to chart divergences from the minus-point-one rule.’ In general, wage flexibility plays an important role in determining a country’s ability to adjust successfully to economic turbulence. Also, wage stickiness is a key feature of many models of the business cycle and is considered to be a major cause of unemployment fluctuations among macroeconomists. We complement Blanchflower and Oswald’s work (1994) on two aspects. First, Blanchflower and Oswald (1994) used the industry unemployment rate instead of the regional unemployment rate. Because the industry unemployment rate is not well defined, measurement error in the explanatory variable would attenuate the true wage curve elasticity. To test if the smaller elasticity is due to the use of the industry unemployment, we use the