چکیده:
It has been suggested that existing estimates of the long-run impact of a surprise move in income may have a substantial upward bias due to the presence of a trend break in 1970s (1350s) and 1980s (1360s) gross domestic product (contained oil) data of Iran. This article shows that the statistical evidence does not warrant abandoning the no-trend-break null hypothesis at the 5% significance level. A key part of the argument is that conventionally computed p values overstate the likelihood of the trend-break alternative hypothesis. This is because they do not take into account that, in practice, the data is chosen based on pretest examination of the data.
خلاصه ماشینی:
"It shows that once pretest and small-sample distributional considerations are taken into account the F test reveals no evidence (at the 5% significance level) against the null hypothesis of no trend break in GDP (contained oil) data of Iran.
Although that test does not share the F test's small-sample distributional problem (Christiano 1992), we show it is still the case that, Once pretest considerations are taken into account, the p value is 98% and one can not reject null hypothesis of existing unit root for GDP (contained oil) of Iran.
This shows that, when the data are generated by the TS model, The break date is selected by the F Maxuntr method, and the conventional practice of using critical values from the F distribution is followed, then a test with nominal 5% size in fact has size 1.
For example, the F test based on Table 5: Size of Pretest Unadjusted Trend-Break Tests When Break Dates Are Selected Endogenously and the Data Are Generated by the DS Model and the TS Model F Max method Min Sig method Nominal sizea F(2,40) critical valueb Bootstrap critical valuec F(2,40) critical valued Bootstrap critical valuee (2) (3) (4) (5) (6) (7) (8) (9) A.
Conclution This article has tested the null hypothesis that the parameters of the time series model for GDP of Iran (contained oil) have been stable during the period 1959-2004 (1338-1383), against the alternative that there has been a one time break in trend."