Structural Break

1. This Data has 1 structural break or 2?
Bring The Clouds‎ made a query as such > My data is logarithm of Electrical
Consumptions. I tested it with Dickey-Fuller and the result is I(2). Then, I continue to run it
with Clemente-Motanes-Reyes for 2 structural break and 1 structural break (Clemao2 and
clemao1), respectively. The result is shown in this picture. Could anyone help me explain
the result, This Data has 1 structural break or 2?   (Dec 3, 2016)

Muili Adebayo Hamid commented> The Du1 (1996) and Du2 (2004) ,shows the dummies
for the period of structural breaks as suggested by the test. checking the significance of
the pvalues of the break , it is apparent that there's a break during this period and such
break takes occurs rapidly judging by the AO (additive outlier ) approach used. the null
hypothesis for the test is = there's unit root while the alternative is that there's no unit root.
the min -t ( rho -1) is 2.789 and the critical value at 5% level of significance for 2 periods is
5.490. thus the min t calculated < min t critical. thus, the null hypothesis of unit root cannot
be rejected. hence, we conclude that. in the presence of rapid structural breaks in the two
periods, there's still unit root. which means the series is still nonstationary. in the second
result shows that du1 the dummy for 1999 structural break is significant at the
conventional level of sig, and also exhibits a rapid structural break due to the AO
approach, . the null hypothesis of unit root cannot be rejected as the rho -1 (1.931) is <
The critical value of 1 structural break of 3.560. thus, we conclude that even with the
structural break, the series is still non stationary. you can check related stata journal on
this test or see the work of Isabelle (2009) " test and structural break in Sweden electricity
Meo School of Research
Shishir Shakya
Noman Arshed
Structural Break
Hossain Academy Note
Univariate Models
Multivariate Models
Panel Data Model