Normal arm swing plays a role in decreasing the cost of transport during walking. However, whether excessive arm swing can reduce the cost of transport even further is unknown. Therefore, we tested the effects of normal and exaggerated arm swing on the cost of transport in the current study. Healthy participants (n=12) walked on a treadmill (1.25 m/s) in seven trials with different arm swing amplitudes (in-phase, passive restricted, active restricted, normal, three gradations of extra arm swing), while metabolic energy cost and the vertical angular momentum (VAM) and ground reaction moment (GRM) were measured. In general, VAM and GRM decreased as arm swing amplitude was increased, except for in the largest arm swing amplitude condition. The decreases in VAM and GRM were accompanied by a decrease in cost of transport from in-phase walking (negative amplitude) up to a slightly increased arm swing (non-significant difference compared to normal arm swing). The most excessive arm swings led to an increase in the cost of transport, most likely due to the cost of swinging the arms. In conclusion, increasing arm swing amplitude leads to a reduction in VAM and GRM, but it does not lead to a reduction in cost of transport for the most excessive arm swing amplitudes. Normal or slightly increased arm swing amplitude appear to be optimal in terms of cost of transport in young and healthy individuals. This article has an associated First Person interview with the first author of the paper.