Large estimates of the elasticity of intertemporal substitution: Is it the aggregate return series or the instrument list?

Since the 1980s, researchers have been puzzled by close to zero estimates of the elasticity of intertemporal substitution. Two possible causes are rates of return that are not representative of the agent's portfolio return and inconsistent estimates due to the weak instrument problem. We examine if the aggregate capital return series for the United States and several instrument sets can provide large estimates of this elasticity. Our findings indicate that this return series leads to large estimates of the elasticity using different instrument sets. An unusual set of instruments performed well and its use in consumption-model estimates seems promising.