한국경제에서 필립스 곡선의 안정성 및 비선형성에 관한 연구

Abstract

본 연구는 필립스 곡선의 평탄화 현상에 대하여 검정하고자 하였다. 이를 위해 모형의 안정성과 선형성을 검정하고 검정 결과에 따라 불안정성 및 비선형성을 반영한 모형의 추정이 가능한 TV-STR 모형을 활용하였다. 필립스 곡선의 안정성 및 선형성에 대한 검정은 최근 필립스 곡선 평탄화 논란의 대상이 되고 있는 2000-2018년 기간에 대해 우선 실시하고 동 기간이 필립스 곡선의 구조적 변화를 판단하는 데 충분히 긴 기간이 아닐 수도 있다는 점을 감안하여 1992~2018년으로 확장된 기간에 대해서도 검정을 실시하였다. 먼저 2000~2018년 기간에 대하여는 인플레이션의 수준, 지속성 및 생산갭에 대한 반응 계수에서 안정성이 기각되지 않았다. 또 필립스 곡선의 선형성도 기각되지 않았다. 반면 1992~2018년으로 확장된 기간에 대하여는 인플레이션의 수준, 지속성 및 생산갭에 대한 반응 계수에서 안정성이 모두 기각되었다. 필립스 곡선의 선형성은 일부 기각되는 결과를 나타내었다. 이와 같은 검정 결과에 따라 1992~2018년 기간을 대상으로 TV-STR 모형을 활용하여 모수의 불안정성 및 비선형성이 반영된 필립스 곡선을 추정하였다. 추정결과 첫째, 인플레이션의 생산갭에 대한 반응 계수는 외환위기 이전에는 음(-)의 값을 가져 이론적 예측과 배치되는 결과를 나타내었으나 외환위기 이후 뚜렷한 양의 관계를 보였다. 둘째, 인플레이션의 생산갭에 대한 반응 계수는 비선형성을 가지는 것으로 보이며 비선형성의 형태가 볼록이 아니라 오목의 형태를 가지는 것으로 나타났다.

This study investigates whether the Phillips curve in the Korean economy is flattening. This issue has an important meaning in the implementation of monetary policy. When the Phillips curve is flattening, the amount of output sacrificed to achieve the lower inflation rate becomes large, while the policy of boosting the output will induce the smaller rise of inflation. However, such expectations can be wrong if the cause of flattening is due to the non-linearity of the curve, not due to the instability of the curve. Hence, the test of Phillips curve flattening should be implemented on both aspects of instability and non-linearity of the curve. The time varying smooth transition (TV-STR) model employed in this paper provides the way of testing both of them in one model. After the tests, the model that reflects the instability or the non-linearity is estimated according to the test results. The Phillips curve used for the test and estimation is the triangular Phillips curve which reflects the effects of the past inflation and cost factors as well as the effect of demand factor represented by the output gap. The tests are first implemented for the period of 2000~2018 as the recent contentions about the flattening Phillips curve are focused on this period. And then, the tests are implemented for the period of 1992~2018 as the period of 2000-2018 may not be long enough for deciding the structural change of the curve considering the results of previous literatures. The tests of stability are implemented for the level and persistence of inflation as well as the slope of the curve to investigate overall structural change of the Phillips curve in the Korean economy. The test of linearity is implemented only for the slope of the curve. For the period of 2000-2018, the stabilities are not rejected in all three aspects. The linearity of the curve is not rejected either. For the period of 1992-2018, the stabilities are rejected in all three aspects, The linearity is also partly rejected. Based on these test results, various types of model are estimated for the period of 1992~2018 to investigate the instability and non-linearity of the Phillips curve in Korea. The noteworthy results of the estimation are as following. First, the long term inflation level dropped from 4.8% to 2.2% before and after the currency crisis of 1997. The result does not change regardless of considering the instability of inflation level only or the instabilities of inflation level and slope of the curve together. Second, the slope of the curve was negative before the currency crisis, which contradicts the prediction of the theory, but it evidently turned into positive after the crisis. Third, the slope of the curve kinked at the output gap of -0.16%. The shape of non-linearity was not convex but concave, which contradicts to the usual expectation. The interpretation of non-linearity of the curve needs to be cautious as the statistical significance levels of some coefficients in the model are somewhat low. For the Korean economy, assuming the structural change in the relationship between the output and inflation after the year of 2000 seems to be not persuasive. Intuitively, however, it is hard to deny that dynamics of inflation in Korea has been changed recently. The candidates for the cause of such changes might be sought at the supply side of the economy such and the structural change in the labor market or the intense competition in the global markets.