Abstract
Multiple structural change tests by Bai and Perron (Econometrica 66:47–78, 1998) are applied to the regression by Demetrescu et al. (Econ Theory 24:176–215, 2008) in order to detect breaks in the order of fractional integration. With this instrument we tackle time-varying inflation persistence as an important issue for monetary policy. We determine not only the location and significance of breaks in persistence, but also the number of breaks. Only one significant break in U.S. inflation persistence (measured by the long-memory parameter) is found to have taken place in 1973, while a second break in 1980 is not significant.
Similar content being viewed by others
Notes
Such a feature is sometimes called “mean-reversion” although Phillips and Xiao (1999) argue that this is a misnomer given the nonstationarity.
The model in (6) introduces a nonlinearity in \(\Delta ^d y_t\) which is not present under the null in (1). Baillie and Kapetanios (2007) and Baillie and Kapetanios (2008) found evidence in favour of nonlinearity in addition to long memory in many economic and financial time series. Contrary to (6), however, they instead assume a smooth transition autoregression or a similar nonlinear \(I(0)\) model for \(\Delta ^d y_t\). An investigation of their tests under breaks in memory is beyond the scope of the present paper.
The critical values are available from an unpublished appendix to Bai and Perron (2003b) posted on the homepage of Pierre Perron.
Tables containing corresponding information as reported in Table 1 are available for all variations to the simulation set-up reported in this subsection.
Seasonality is accounted for by twelve monthly dummies (\(dum_{seas}\)), the break in mean is accounted for by a mean dummy (\(dum_\mu \)) taking on the value one before and the value 0 after \([\tau _0 \,T]\). The variable \(y_t\) is the residual of the regression of \(p_t\) on \(dum_\mu \) and \(dum_{seas}\).
As an alternative to the sequential procedure we also allow for two mean shifts simultaneously and obtain similar break points and \(p\) values.
Bai and Perron (1998) also investigate a double maximum test, not considered in this paper. The number of break points is found by taking the maximum over all \(sup F(m)\) test statistics, where \(m={1,2,...,5}\). This maximum value is then compared to critical values in order to determine the significance. This suggests that in our analysis there is only one break point.
As becomes evident in Fig. 5, the power of the test increases with the difference in the order of integration before and after the break. Another factor is the total number of observations in the whole sample, see Fig. 6. If the difference in the order of integration is at least 0.3 and the sample size is 500, the test has a rejection rate of more than 80 % if there are at least 150 observations left before and after the break.
This interpretation is inferred directly from Proposition 3 in Demetrescu et al. (2008) and the derivations of this paper.
The order of integration was estimated to be 0.22 for \(B=T^{0.70}\) with a 90 % confidence interval \([0.01, 0.43]\). However, the estimation depends heavily on single observations.
The events are not described in order to indicate causality but rather in order to integrate the break date into its historical background.
References
Andrews DWK (1993) Tests for parameter instability and structural change with unknown change point. Econometrica 61:821–56
Bai J, Perron P (1998) Estimating and testing linear models with multiple structural changes. Econometrica 66:47–78
Bai J, Perron P (2003a) Computation and analysis of multiple structural change models. J Appl Econ 18:1–22
Bai J, Perron P (2003b) Critical values for multiple structural change tests. Econ J 6:72–78
Baillie RT, Chung C-F, Tieslau MA (1996) Analysing inflation by the fractionally integrated ARFIMA–GARCH model. J Appl Econ 11:23–40
Baillie RT, Kapetanios G (2007) Testing for neglected nonlinearity in long-memory. J Bus Econ Stat 25:447–461
Baillie RT, Kapetanios G (2008) Nonlinear models for strongly dependent processes with financial applications. J Econ 147:60–71
Beran J, Terrin N (1996) Testing for a change of the long-memory parameter. Biometrika 83:233–233
Beran J, Terrin N (1999) Correction: testing for a change of the long-memory parameter. Biometrika 86:627–638
Breitung J, Hassler U (2002) Inference on the cointegration rank in fractionally integrated processes. J Econ 110:167–185
Busetti F, Taylor AMR (2004) Tests of stationarity against a change in persistence. J Econ 123:33–66
Cogley T, Sargent TJ (2001) Evolving post-world war II U.S. inflation dynamics. NBER Macroecon Ann 16:331–373
Cogley T, Sargent TJ (2005) Drift and volatilities: monetary policies and outcomes in the post WWII U.S. Rev Econ Dyn 8:262–302
Demetrescu M, Hassler U, Kuzin V (2011) Pitfalls of post-model-selection testing: experimental quantification. Empir Econ 40:359–372
Demetrescu M, Kuzin V, Hassler U (2008) Long memory testing in the time domain. Econ Theory 24: 176–215
Gadea MD, Mayoral L (2006) The persistence of inflation in OECD countries: a fractionally integrated approach. Int J Central Banking 2:51–104
Geweke J, Porter-Hudak S (1983) The estimation and application of long memory time series models. J Time Ser Anal 4:221–238
Gil-Alana LA (2008) Fractional integration and structural breaks at unknown periods of time. J Time Ser Anal 29:163–185
Hassler U, Breitung J (2006) A residual-based lm-type test against fractional cointegration. Econometric Theory 22:1091–1111
Hassler U, Kokoszka P (2010) Impulse responses of fractionally integrated processes with long memory. Econ Theory 26:1855–1861
Hassler U, Rodrigues PMM, Rubia A (2009) Testing for general fractional integration in the time domain. Econ Theory 25:1793–1828
Hassler U, Scheithauer J (2011) Detecting changes from short to long memory. Stat Pap 52:847–870
Hassler U, Wolters J (1995) Long memory in inflation rates: international evidence. J Bus Econ Stat 13: 37–45
Hidalgo J, Robinson PM (1996) Testing for structural change in a long-memory environment. J Econ 70:159–174
Hsu C-C (2005) Long memory or structural changes: an empirical examination on inflation rates. Econ Lett 88:289–294
Kew H, Harris D (2009) Heteroskedasticity-robust testing for a fractional unit root. Econ Theory 25: 1734–1753
Kim J-Y (2000) Detection of change in persistence of a linear time series. J Econ 95:97–116
Kumar MS, Okimoto T (2007) Dynamics of persistence in international inflation rates. J Money, Credit Banking 39:1457–1479
Lavielle M, Ludeña C (2000) The multiple change-points problem for the spectral distribution. Bernoulli 6:845–869
Leeb H, Pötscher BM (2005) Model selection and inference: facts and fiction. Econ Theory 21:21–59
Lobato IN, Savin NE (1998) Real and spurious long-memory properties of stock-market data. J Bus Econ Stat 16:261–268
Martins LF, Rodrigues PMM (2012) Testing for persistence change in fractionally integrated models: an application to world inflation rates. Comput Stat Data Anal doi:10.1016/j.csda.2012.07.021
Meller B, Nautz D (2012) Inflation persistence in the euro area before and after the European Monetary Union. Econ Model 29:1170–1176
Meltzer AH (2006) From inflation to more inflation, disinflation, and low inflation. Am Econ Rev 96:185–188
Phillips PCB, Xiao Z (1999) A primer on unit root testing. J Econ Surv 12:423–470
Pivetta F, Reis R (2007) The persistence of inflation in the United States. J Econ Dyn Control 31:1326–1358
Ray BK, Tsay RS (2002) Bayesian methods for change-point detection in long-range dependent processes. J Time Ser Anal 23:687–705
Robinson PM (1991) Testing for strong serial correlation and dynamic conditional heteroskedasticity in multiple regression. J Econ 47:67–84
Robinson PM (1994) Efficient tests of nonstationary hypotheses. J Am Stat Assoc 89:1420–1437
Robinson PM (1995) Gaussian semiparametric estimation of long range dependence. Ann Stat 23:1630–1661
Robinson PM (2005) The distance between rival nonstationary fractional processes. J Econ 128:283–399
Schwert GW (1989) Tests for unit roots: a Monte Carlo investigation. J Bus Econ Stat 7:147–159
Shimotsu K, Phillips PCB (2005) Exact local whittle estimation of fractional integration. Ann Stat 33:1890–1933
Sibbertsen P (2004) Long memory versus structural breaks: an overview. Stat Pap 45:465–515
Sibbertsen P, Kruse R (2009) Testing for a break in persistence under long-range dependencies. J Time Ser Anal 30:263–285
Stock JH (2001) Evolving post-world war II U.S. inflation dynamics: comment. NBER Macroecon Ann 16:379–387
Stock JH, Watson MW (2007) Why has U.S. inflation become harder to forecast? J Money, Credit Banking 39:3–33
Tanaka K (1999) The nonstationary fractional unit root. Econ Theory 15:549–583
Tillmann P (2012) Has inflation persistence changed under EMU? Ger Econ Rev 13:86–102
Yamaguchi K (2011) Estimating a change point in the long memory parameter. J Time Ser Anal 32:304–314
Acknowledgments
In particular we thank Dieter Nautz for his valuable comments. We are grateful to two anonymous referees who helped to improve the paper. Financial support by the German Research Foundation (Deutsche Forschungsgemeinschaft, DFG) through HA-3306/2-1 and by the Frankfurt Graduate Program in Finance and Monetary Economics is gratefully acknowledged. The paper represents the authors’ personal opinions and does not necessarily reflect the views of the Deutsche Bundesbank or its staff.
Author information
Authors and Affiliations
Corresponding author
Additional information
Earlier versions of this paper were presented at the 15th International Conference on Computing in Economics and Finance (July 2009, Sydney) and the European Society Econometric Meeting (August 2009, Barcelona).
Appendix
Appendix
1.1 Proof of Proposition 1
Under Gaussianity and Assumption 2 the pseudo-log-likelihood function becomes
with \(\varepsilon _t = \Delta ^{d+\theta _{j-1}} y_t\) for \(t=T_{j-1}+1, \ldots , T_j \) using (3) and (4), or \(\varepsilon _t = \Delta ^{d+\theta _{j}} y_t\) for \(t=T_{j}+1,\ldots ,T_{j+1}\), such that (\(j=1,\ldots ,m\))
With \(\log (1-L)=-\sum _{j=1}^{\infty }j^{-1}L^{j}\) we obtain for the score vector evaluated under the null (where \(\Delta ^d y_t = \varepsilon _t\))
where \(\{\varepsilon _{t-1}^{**}\}\) is a stationary process with variance
To construct the LM statistic we compute the Fisher information as the outer product of gradients,
Hence, we obtain
Since the LM statistic is evaluated under \(H_0\), we replace \(\varepsilon _t\) with \(\Delta ^d y_t\). Given the starting value assumption in Assumption 2, this coincides with \(x_t\) defined in (7). Consequently, \(\varepsilon _{t-1}^{**}\) equals \(x_{t-1}^*\) from (9 ), and the LM statistic becomes \(LM\) from (8) as required. \(\square \)
1.2 Proof of Proposition 2
Write the regression equation (13) in obvious matrix notation, \( y = X \, \widehat{\beta }+ \widehat{\varepsilon }\), with
and \(X\) containing \((x_1^*,\ldots ,x_{T-1}^*)^\prime \) as the first column, while the other columns contain zeros and segments of \((x_{T_j}^*,\ldots ,x_{T_{j+1}-1}^*)^\prime \). Under Assumption 2, we have \(x_{t}=\varepsilon _{t} \sim iid (0,\sigma ^{2})\). The required limiting distributions can be obtained as set out by Robinson (1991) or Tanaka (1999), see also Hassler and Breitung (2006, Lemma A):
where \(\overset{p}{\rightarrow }\) stands for convergence in probability, and
and
Consequently, \(\sqrt{T} \, \widehat{\beta }\) follows a limiting normal distribution with \(\Sigma = \frac{6}{\pi ^{2}} \left( \Lambda ^0\right) ^{-1}\). Define the \(m \times (m+1)\) matrix \(R\) with \(R \, \widehat{\beta }=( \widehat{\psi }_1,\ldots ,\widehat{\psi }_m)^\prime \). The \(F\) statistic becomes
and its limiting distribution follows the usual way. \(\square \)
Rights and permissions
About this article
Cite this article
Hassler, U., Meller, B. Detecting multiple breaks in long memory the case of U.S. inflation. Empir Econ 46, 653–680 (2014). https://doi.org/10.1007/s00181-013-0691-8
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00181-013-0691-8