Retail companies can employ a variety of return policies ranging from a no-questions-asked return policy to specifying a duration in which products can be returned. This duration is often termed as return period. The impact of return period length on product return volumes and their variability is not understood. This paper shows that increasing return period length results in reducing variability of product returns-- commonly known as risk pooling. Risk pooling in product returns by increasing return period length is structurally different from existing works on risk pooling. Commonly discussed risk pooling examples such as demand substitution, are accrual in nature where the behavior of underlying stochastic system does not alter. In the case of product returns, risk pooling occurs by changing customer behavior, which is a behavioral change in the underlying stochastic system. Additionally, the necessary conditions for behavioral change to result in additional risk pooling benefits are derived.