There can be a trade-off between output and unemployment which is formally called in literature as Okun's Law for G-20 countries. The objective of this research is to reconsider Okun's Law for G-20 countries with time-series data over the period 1991-2014. First order difference and "gap" specification methods are used for testing the relationship. Three different filtering techniques; Hodrick-Prescott (HP), Chiristiano-Fitzgerald (CF) and Butterworth (BW) are conducted for the gap specification method. For all methods, the results validate the existence of inverse relation between output and unemployment in most of the countries. An interesting finding is China, Indonesia, Saudi Arabia and, Turkey do not satisfy the Okun's coefficient with at least one of these filtering techniques. It is understood that, Group of G-20 countries display different Okun's coefficients depending on their development structures and productivity heterogeneities.