Strukturni prelom
Strukturni prelom je v ekonometriji nepričakovana sprememba v parametrih regresijskega modela skozi čas, ki lahko vodi v zelo velike napovedovalne napake in nezanesljivost modela na splošno.[1][2][3] Ta problem je populariziral David Hendry, ki je trdil, da pomanjkanje stabilnosti koeficientov pogosto povzroča napovedovalne napake in je zato potrebno rutinsko testirati prisotnost strukturne stabilnosti. Strukturna stabilnost − časovna nevariabilnost regresijskih koeficientov − je centralni problem v vseh aplikacijah modelov linearne regresije.[4]
Testi strukturnih prelomov
[uredi | uredi kodo]V linearnih regresijskih modelih za testiranje prisotnosti enega preloma aritmetične sredine na znani časovni točki K za K ∈ [1,T] pogosto uporabljamo Chowov test.[5][6] Ta test oceni, če so koeficienti v regresijskem modelu enaki za obdobja [1,2, ...,K] in [K + 1, ...,T].[6]
Izzivi se pojavijo v naslednjih primerih:
1. znano število prelomov aritmetične sredine na neznanih točkah,
2. neznano število prelomov aritmetične sredine na neznanih točkah,
3. prelomi variance.
Chowov test v teh primerih ni uporaben, saj se nanaša samo na modele s prelomi na znanih točkah in kjer napaka variance ostane enaka pred in po prelomu.[7][5][6]
V takih primerih lahko konstantnost koeficientov modela v splošnem testiramo s testoma CUSUM (kumulativna vsota) in CUSUM-sq (CUSUM na kvadrat). Prav tako lahko uporabimo test mej.[6][8] V primerih 1 in 2, ko sta število in kraj strukturnih prelomov neznana, lahko za testiranje nestabilnosti parametrov uporabimo sup-Waldov, sup-LM in sup-LR test, ki jih je razvil Andrews (1993, 2003).[9][10]
Statistični paketi
[uredi | uredi kodo]Iskanje strukturnih prelomov omogoča več statističnih paketov, kot so R,[11] GAUSS in Stata.
Sklici
[uredi | uredi kodo]- ↑ Antoch, Jaromír; Hanousek, Jan; Horváth, Lajos; Hušková, Marie; Wang, Shixuan (25. april 2018). »Structural breaks in panel data: Large number of panels and short length time series« (PDF). Econometric Reviews. 38 (7): 828–855. doi:10.1080/07474938.2018.1454378. S2CID 150379490.
Structural changes and model stability in panel data are of general concern in empirical economics and finance research. Model parameters are assumed to be stable over time if there is no reason to believe otherwise. It is well-known that various economic and political events can cause structural breaks in financial data. ... In both the statistics and econometrics literature we can find very many of papers related to the detection of changes and structural breaks.
- ↑ Kruiniger, Hugo (december 2008). »Not So Fixed Effects: Correlated Structural Breaks in Panel Data« (PDF). IZA Institute of Labor Economics. str. 1–33. Pridobljeno 20. februarja 2019.
{{navedi splet}}
: Vzdrževanje CS1: samodejni prevod datuma (povezava) - ↑ Hansen, Bruce E (november 2001). »The New Econometrics of Structural Change: Dating Breaks in U.S. Labor Productivity«. Journal of Economic Perspectives. 15 (4): 117–128. doi:10.1257/jep.15.4.117.
{{navedi časopis}}
: Vzdrževanje CS1: samodejni prevod datuma (povezava) - ↑ Ahmed, Mumtaz; Haider, Gulfam; Zaman, Asad (Oktober 2016). »Detecting structural change with heteroskedasticity«. Communications in Statistics – Theory and Methods. 46 (21): 10446–10455. doi:10.1080/03610926.2016.1235200. S2CID 126189844.
The hypothesis of structural stability that the regression coefficients do not change over time is central to all applications of linear regression models.
- ↑ 5,0 5,1 Hansen, Bruce E (november 2001). »The New Econometrics of Structural Change: Dating Breaks in U.S. Labor Productivity«. Journal of Economic Perspectives. 15 (4): 117–128. doi:10.1257/jep.15.4.117.
{{navedi časopis}}
: Vzdrževanje CS1: samodejni prevod datuma (povezava) - ↑ 6,0 6,1 6,2 6,3 Greene, William (2012). »Section 6.4: Modeling and testing for a structural break«. Econometric Analysis (7th izd.). Pearson Education. str. 208–211. ISBN 9780273753568.
An important assumption made in using the Chow test is that the disturbance variance is the same in both (or all) regressions. ...
6.4.4 TESTS OF STRUCTURAL BREAK WITH UNEQUAL VARIANCES ...
In a small or moderately sized sample, the Wald test has the unfortunate property that the probability of a type I error is persistently larger than the critical level we use to carry it out. (That is, we shall too frequently reject the null hypothesis that the parameters are the same in the subsamples.) We should be using a larger critical value. Ohtani and Kobayashi (1986) have devised a "bounds" test that gives a partial remedy for the problem.15 - ↑ Gujarati, Damodar (2007). Basic Econometrics. New Delhi: Tata McGraw-Hill. str. 278–284. ISBN 978-0-07-066005-2.
- ↑ Pesaran, M. H.; Shin, Y.; Smith, R. J. (2001). »Bounds testing approaches to the analysis of level relationships«. Journal of Applied Econometrics. 16 (3): 289–326. doi:10.1002/jae.616. hdl:10983/25617.
- ↑ Andrews, Donald (Julij 1993). »Tests for Parameter Instability and Structural Change with Unknown Change Point« (PDF). Econometrica. 61 (4): 821–856. doi:10.2307/2951764. JSTOR 2951764. Arhivirano (PDF) iz spletišča dne 6. novembra 2017.
- ↑ Andrews, Donald (Januar 2003). »Tests for Parameter Instability and Structural Change with Unknown Change Point: A Corrigendum« (PDF). Econometrica. 71 (1): 395–397. doi:10.1111/1468-0262.00405. S2CID 55464774. Arhivirano iz prvotnega spletišča (PDF) dne 6. novembra 2017.
- ↑ Kleiber, Christian; Zeileis, Achim (2008). Applied Econometrics with R. New York: Springer. str. 169–176. ISBN 978-0-387-77316-2.