Most stringent test of null of cointegration: a Monte Carlo comparison

Abstract

To test for the existence of long run relationship, a variety of null of cointegration tests have been developed in literature. This study is aimed at comparing these tests on basis of size and power using stringency criterion: a robust technique for comparison of tests as it provides with a single number representing the maximum difference between a test's power and maximum possible power in the entire parameter space. It is found that in general, asymptotic critical values tends to produce size distortion and size of test is controlled when simulated critical values are used. The simple LM test based on KPSS statistic is the most stringent test at all sample sizes for all three specifications of deterministic component, as it has the maximum difference approaching to zero and lesser than 20% for the entire parameter space.

Monetary policy and nonperforming loan ratios in a monetary union; a counterfactual study

“…The stationarity testing found nonperforming loan ratios, the Monetary policy rate of Ghana and the Monetary policy ratio to be nonstationary whiles the GDP growth and the exponential of the GDP growth (ex gdp) were found to be stationary. Owing to the mixed order of integration of our variables, the estimated ARDL model was tested for cointegration using the Pesaran et al (2001) ARDL bounds testing cointegration methodology, the most theoretically sound cointegration test (Khan et al, 2019) and reported in Table 4.…”

“…Owing to the mixed order of integration of our variables, the estimated ARDL model was tested for cointegration using the Pesaran et al. (2001) ARDL bounds testing cointegration methodology, the most theoretically sound cointegration test (Khan et al. , 2019) and reported in Table 4.…”

The probabilities of type I and II error of null of cointegration tests: A Monte Carlo comparison

“…However, there are fewer real data based comparative studies of cointegration tests ([ 16 , 17 ]). MCS have been frequently used for such comparative studies ([ 18 – 23 ]). These studies ([ 14 , 15 , 24 – 26 ] and many more), using MCS were assessing and evaluating the performance of tests based on two properties, the size: “the probability of rejection of null hypothesis when actually it is true” and the power: “the probability of rejection of null hypothesis when actually it is false”.…”