Clements, M. P., and Krolzig, H.-M. (1998). A comparison of the forecast performance of Markovswitching and threshold autoregressive models of US GNP. Econometrics Journal, 1, C47--75. Home/Search Document Details and Download
Summary Related Articles Check |
This paper is cited in the following contexts:
Multiple unemployment equilibria - Do transitory shocks have.. - Akram (1998) (Correct)
....series in a Markov regime switching model with no autoregressive terms, see Hamilton (1989) The degree of persistence in each series is measured once the series are adjusted for the estimated mean rates. The estimates of mean rates may, however, be quite sensitive to model specication, c.f. 3 Clements and Krolzig (1998). This also applies to the estimated number of switches between equilibria. For the Norwegian registered rate of unemployment the model in Bianchi and Zoega (1998) implies two periods of high unemployment equilibrium of 4.72 , in 1982:3 1984:4 and then from 1988:1 to the end of their sample in ....
....structural changes. This problem of spurious switches , low transition probabilities and small deviations between the estimated equilibria may be ascribed to the non rejection of the linearity hypothesis. The problem of spurious switches between regimes has, however, also been faced by Clements and Krolzig (1998) who nd stronger evidence of nonlinearity in a Markov regime switching framework. 4.1.1. Comparison with existing studies Bianchi and Zoega (1996, 1998) are studies of unemployment rates in OECD countries based on annual and seasonally adjusted quartely data, respectively. These are conducted in ....
[Article contains additional citation context not shown here]
Clements, M. and Krolzig, H. M. (1998): A comparison of the forecast performance of Markov-switching and threshold autoregressive models of US GNP. Econometrics Journal, 1, C47-C75.
Forecasting And The UK Business Cycle - Osborn, Sensier, Simpson (1999) (Correct)
....underlying the model. For a full description of the approach taken and the detailed UK results on which we draw, see Simpson et al. 1999) While applications of this type of regime switching model generally relate to the US experience (including Filardo, 1994, Diebold and Rudebusch, 1996, Clements and Krolzig, 1998), other studies in the UK context include Acemoglu and Scott (1994) and Krolzig and Sensier (1999) The Markov switching model is linear within a specific regime, but the properties are different between the two regimes. As already argued, to capture the key feature of the business cycle we ....
Clements, M.P. and Krolzig, H.-M (1998). "A Comparison of the Forecast Performance of Markov-Switching and Threshold Autoregressive models of US GNP", Econometrics Journal, vol. 1, pp.C47-C75.
Long Memory and Structural Change - Diebold, Inoue (1999) (6 citations) (Correct)
....for any fixed sample size, long memory seems to appear, and the implied d estimates begin to grow. The Granger Tersvirta results, however, are based on single realizations (not Monte Carlo analysis) and no theoretical explanation is provided. 10 In a development that supports this conjecture, Clements and Krolzig (1998) show that fixed coefficient autoregressions often outperform Markov switching models for forecasting in finite samples, even when the true data generating process is Markov switching. 23Second, in contemporaneous and independent work, Granger and Hyung (1999) develop a theory closely related to ....
Clements, M.P. and Krolzig, H.-M. (1998), "A Comparison of the Forecast Performance of Markov-Switching and Threshold Autoregressive Models of U.S. GNP," Econometrics Journal, 1, 47-75.
On SETAR Non-linearity and Forecasting - Clements, Franses, Smith (1999) Self-citation (Clements) (Correct)
....empirical forecast accuracy comparisons when the non linear features that were prominent in sample fail to carry over to the out of sample period. Forecast comparisons of US GNP where the out of sample period is taken to be 1992 onwards is a case in point (see, Clements and Smith, 1997) However, Clements and Krolzig (1998) and Clements and Smith (1999) attempt systematic, simulationbased evaluations of the forecast performance of non linear models that seek to ensure that the future bears the same non linear imprint as the past. The general finding turns out to be that linear autoregressive models are reasonably ....
Clements, M. P., and Krolzig, H.-M. (1998). A comparison of the forecast performance of Markovswitching and threshold autoregressive models of US GNP. Econometrics Journal, 1, C47--75.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC