Background: The gamma‐aminobutyric acid (GABA) hypothesis in essential tremor (ET) implies a disturbance of the GABAergic system, especially involving the cerebellum. This review examines the evidence of the GABA hypothesis. Methods: The review is based on published data about GABA dysfunction in ET, taking into account studies on cerebrospinal fluid, pathology, electrophysiology, genetics, neuroimaging, experimental animal models, and human drug therapies. Results: Findings from several studies support the GABA hypothesis in ET. The hypothesis follows four steps: 1) cerebellar neurodegeneration with Purkinje cell loss; 2) a decrease in GABA system activity in deep cerebellar neurons; 3) disinhibition in output deep cerebellar neurons with pacemaker activity; and 4) an increase in rhythmic activity of the thalamus and thalamo‐cortical circuit, contributing to the generation of tremor. Doubts have been cast on this hypothesis, however, by the fact that it is based on relatively few works, controversial post‐mortem findings, and negative genetic studies on the GABA system. Furthermore, GABAergic drug efficacy is low and some GABAergic drugs do not have antitremoric efficacy. Discussion: The GABA hypothesis continues to be the most robust pathophysiological hypothesis to explain ET. There is light in all GABA hypothesis steps, but a number of shadows cannot be overlooked. We need more studies to clarify the neurodegenerative nature of the disease, to confirm the decrease of GABA activity in the cerebellum, and to test more therapies that enhance the GABA transmission specifically in the cerebellum area.