Testing for threshold effect in ARFIMA models:Application to US unemployment rate data

Macroeconomic time series often involve a threshold effect in theirARMA representation, and exhibit long memory features. In this paperwe introduce a new class of threshold ARFIMA models to account forthis. The threshold effect is introduced in the autoregressive and/or thefractional integration parameters, and can be tested for using LM tests.Monte Carlo experiments show the desirable finite sample size and powerof the test with an exact maximum likelihood estimator of the long memoryparameter.[...]

MoreLess

Year of publication:	2008-12-04
Authors:	Lahiani, A. ; Scaillet, O.
Institutions:	Finrisk
Subject:	Zeitreihe \| Grenzwert \| LM-Gitter \| LM test

1 Introduction
2 Threshold ARFIMA models and framework
2.1 Long memory threshold effect
2.2 Short memory threshold effect
2.3 Joint threshold effect
3 TheLMtest
4 Monte Carlo experiments
4.1 Results for long memory threshold effect
4.2 Results for short memory threshold effect
4.3 Results for joint threshold effect
4.4 Spurious detection
5 Modelselection
6 Application to US unemployment rate data
6.1 Output of the threshold ARFIMA tests
6.2 Output of model estimation and selection
6.3 Output of forecasting performance assessment
7 Conclusion
References

Persistent link: https://ebvufind01.dmz1.zbw.eu/10005868836