Tests of the Null of Cointegration Using Integrated and Modified OLS Residuals

• 발행기관: 한국계량경제학회
• 저자명: Cheol-Keun Cho
• 간행물 정보: 『JOURNAL OF ECONOMIC THEORY AND ECONOMETRICS』Vol.35 No.4, 55~86쪽, 전체 32쪽
• 주제분류: 경제경영 > 경제학
• 파일형태: PDF
• 발행일자: 2024.12.31

국문 초록

This study develops a KPSS (Kwiatkowski et al., 1992)-type cointegration test utilizing residuals from integrated and modified ordinary least squares (IMOLS) estimation. The test statistic, denoted by KPSS^{Fb} has a pivotal null limit distribution under fixed-b assumption. The proposed test demonstrates reasonable performance in terms of size and power when the Andrews' AR(1) plug-in data-dependent (DD) bandwidth is employed and fixed-b critical values are used. Additionally, two modified IMOLS residuals are proposed to obtain alternative data-dependent bandwidths. In the simulation experiment, these bandwidths deliver improved power properties for the proposed test.

목차

1. INTRODUCTION

2. MODEL SETUP AND ASSUMPTIONS

3. IMOLS-BASED COINTEGRATION TESTS

4. SIMULATION STUDY

5. CONCLUSION

REFERENCES

키워드

Data-dependent bandwidth cointegration fixed-b IMOLS KPSS