A sequential procedure to determine the number of breaks in trend with an integrated or stationary noise component

Abstract

Perron and Yabu (2009a) consider the problem of testing for a break occurring at an unknown date in the trend function of a univariate time series when the noise component can be either stationary or integrated. This article extends their work by proposing a sequential test that allows one to test the null hypothesis of, say, l breaks versus the alternative hypothesis of (l + 1) breaks. The test enables consistent estimation of the number of breaks. In both stationary and integrated cases, it is shown that asymptotic critical values can be obtained from the relevant quantiles of the limit distribution of the test for a single break. Monte Carlo simulations suggest that the procedure works well in finite samples. Copyright Copyright 2010 Blackwell Publishing Ltd

Trends of U.S. Emissions of Nitrogen Oxides and Volatile Organic Compounds

“…Instead it treats the change in the intercept, if it occurred, as an outlier. This problem can be overcome by using an extension of the Perron and Yabu (2007) test in Kejriwal and Perron (2009) to allow for multiple breaks.…”

Revisiting Sulfur Kuznets Curves with Endogenous Breaks Modeling: Substantial Evidence of Inverted-Us/Vs for Individual OECD Countries

“…, 1930-1939). Yet, allowing for endogenous breaks involves an information trade-off; indeed, Harvey et al (2013), Kejriwal and Perron (2010), and Kejriwal and Lopez (2013) recommended allowing for a maximum of two structural breaks (and considered over 100 time observations). But if we restrict our analysis to allow for no more than two endogenous breaks, such breaks likely would/may be calculated to occur before 1950 for most countries.…”

Revisiting sulfur Kuznets curves with endogenous breaks modeling: Substantial evidence of inverted-Us/Vs for individual OECD countries

“…5 The figures clearly indicate why the consideration of breaks is important: for all countries the emissions series display breaks around the two World Wars (e.g, 1914Wars (e.g, -1921Wars (e.g, and 1943Wars (e.g, -1945; in addition to breaks during those two periods, all countries display a substantial break in GDP per capita around the Great Depression (e.g., 1930Depression (e.g., -1939). Yet, allowing for endogenous breaks involves an information trade-off; indeed, Harvey et al (2013), Kejriwal and Perron (2010), and Kejriwal and Lopez (2013) recommended allowing for a maximum of two structural breaks (and considered over 100 time observations). But if we restrict our analysis to allow for no more than two endogenous breaks, such breaks likely would/may be calculated to occur before 1950 for most countries.…”

“…If cointegration is confirmed, the different regimes are estimated similarly by assuming the previously determined number and timing of breaks. Table 1 Again, to determine the number and timing of breaks, we consider two information/decision criteria, i.e., the sequential method of Kejriwal and Perron (2010) and the LWZ criterion. If the sequential method did not determine a break, we went with the number of breaks determined by the LWZ (as in Kejriwal 2008).…”

“…Indeed, echoing previous work, we uncovered a prevalence of shared timing in regards to breaks/transitions as well as somewhat similar levels of development. Notes: S=sequential procedure (as described in Kejriwal and Perron 2010). LWZ=Schwarz criterion.…”