A Rational Interpolation Approach to Least Squares Estimation for Band-TARs

Coakley, Jerry; Fuertes, Ana-Marı́a; Pérez, Mercedes López

doi:10.1007/3-540-45262-1_24

Cited by 3 publications

(1 citation statement)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A normalized AIC is used to compare models with di¡erent delay parameters. For a discussion of TAR estimation issues, see Coakley et al (2001aCoakley et al ( , 2001b. 16 Bearing in mind the potential downward bias of ordinary least squares estimators in AR models we used the median-bias correction in Andrews (1993) for the a parameters (in those cases where p 1) which yielded indistinguishable results.…”

Section: Tar Models For Ermentioning

confidence: 99%

A Non‐Linear Analysis of Excess Foreign Exchange Returns

Coakley

Fuertes

2001

The Manchester School

Self Cite

View full text Add to dashboard Cite

In this paper we explore the dynamics of US dollar excess foreign exchange returns for the G10 currencies and the Swiss franc, 1976^97. The non-linear framework adopted is justi¢ed by the results of linearity tests and a parametric bootstrap likelihood ratio statistic which indicate threshold e¡ects or di¡erential adjustment to small and large excess returns. Impulse response analysis suggests that the e¡ect of small shocks to excess returns inside the no-arbitrage band exhibits most persistence. Large shocks outside the band decay most rapidly and also exhibit overshooting. These phenomena are explained in terms of noise trading strategies and transaction costs. " IntroductionLewis (1995) has highlighted deviations from uncovered interest parity (UIP) or excess foreign exchange returns (hereafter ER) as one of the major puzzles in international ¢nance. ER represent the pro¢t from speculation in foreign currency and are linked by covered interest parity (CIP) to the forward premium. Thus the above puzzle can be restated as the forward premium anomaly in which future exchange rate movements are predicted with the wrong sign (Engel, 1996). Although in theory predicted ER should be zero, the puzzle is that they are signi¢cantly di¡erent from zero, exhibit considerably more variability than spot returns and change sign frequently. Since UIP is a central building block in international ¢nance, it is not surprising that this puzzle has generated a huge literature. One strand investigates the long-run time series properties of both ER and the forward premium. Another focuses on tests for a possibly time-varying risk premium since, assuming rational expectations, ER can be decomposed into a risk premium and a white noise forecast error term.

show abstract

Section: Tar Models For Ermentioning

confidence: 99%