CBN Journal of Applied Statistics (JAS)
Keywords
Bootstrapping, Phillips curve, structural change test, SupF
Abstract
In this study, we describe the problem of testing for the stability and persistence of the Phillips curve for Nigeria when there are nonstationarities in the marginal distribution of the regressors. We test for unknown break dates using the ๐๐ข๐๐น,๐ด๐ฃ๐๐น and ๐ธ๐ฅ๐๐น approaches. After reviewing the relevant asymptotic distribution theory we replicate Hansenโs fixed-regressor bootstraping scheme, which shows that Andrewsโ tabulated critical values for the test statistics are oversized, and are not robust to the presence of nonstationarities in the marginal distribution of the regressors. In search of alternative bootstraping schemes, we experiment with the sieve, wild, and Rademacher schemes to ascertain if there are any possible improvements over the fixed-regressor scheme. Finally, we apply the methodology to test the stability and persistence of the Phillips curve in Nigeria using quarterly data on inflation and the output gap from 1960 to 2009. We find that, unlike Andrews asymptotic p-values, inference based on Hansenโs hetero-corrected bootstrap technique supports the hypothesis of a structural break in the inflation dynamics in Nigeria. One key policy implication is that, within a certain range of the output gap, the central bank could use the policy rate to stimulate demand up to a certain limit with no consequential positive impact on inflation.
Publication Title
CBN Journal of Applied Statistics
Issue
1
Volume
8
Recommended Citation
Chuku, Chuku; Atan, Johnson; and Obioesio, Felix
(2017)
"Testing for the Stability and Persistence of the Phillips Curve for Nigeria,"
CBN Journal of Applied Statistics (JAS): Vol. 8:
No.
1, Article 6.
Available at:
https://dc.cbn.gov.ng/jas/vol8/iss1/6