Modelling non-stationary is one of the most debated problems in time series econometrics. Several kinds of methodologies are developed in the literature to examine the relationship between time series variables. These tests are compared on the basis of their size and power characteristics using Monte Carlo experiments. In the literature, variety of comparison of tests were carried out on the basis of size and power properties. No previous studies have compared co-integration and co-breaking in term of size and power using real economic data. In this study, comparison of three different econometric techniques i.e. Co-integration, co-breaking and Modified R are carried out while applying to real data. This study also evaluates the performance of these three techniques to distinguish between genuine and spurious relationship taking consumption and income real data of 44 countries. Moreover, we calculate the size and power because size portrays the probability of spurious relationship between consumption and income of two different countries While calculation of power denotes the probability of genuine relationship between income and consumption of same countries. Finally, the estimated results are interpreted in three different scenarios. Firstly, if our focus is to minimize the size distortion then Engle and Granger co-integration is best one as it shows 7.88% size distortion. Secondly, the Size distortion of Johansen and Juselius and MR are approximately equal and comparable So when we compared these two tests MR show better performance as their Power is greater than JJ. Thirdly, we calculated the operational power for each test. On the basis of operational power, we concluded that MR is best tests Which show better performance in our analysis as compare to remaining tests Supervisor:- Dr. Atiq-ur-Rehman