Annette Vissing-Jorgensen's NBER Working Paper (No. 8896, April 2002) examines the elasticity of intertemporal substitution (EIS) using data from the U.S. Consumer Expenditure Survey (CEX). The study highlights that limited asset market participation significantly affects EIS estimates. The EIS measures how much consumers adjust their consumption over time in response to changes in asset returns. The paper finds that EIS estimates differ substantially between assetholders and non-assetholders, with stockholders showing EIS values of 0.3-0.4 and bondholders showing 0.8-1. These estimates are higher for households with larger asset holdings. Non-assetholders show smaller, often insignificant EIS estimates. The findings suggest that limited participation in asset markets is crucial for accurate EIS estimation, as including non-assetholders in Euler equation models leads to downward bias. The study also notes that EIS estimates for stockholders and bondholders are higher than for non-assetholders, supporting the limited participation theory. The paper emphasizes the importance of distinguishing between assetholders and non-assetholders in EIS estimation, as the Euler equation does not hold for non-assetholders. The results are consistent with earlier studies showing that EIS increases with consumption and is influenced by wealth and asset participation. The study also discusses the limitations of using the CEX data, including measurement errors and the need for careful handling of data frequency and instrument selection. Overall, the paper underscores the importance of considering limited asset market participation in estimating EIS and highlights the significance of distinguishing between assetholders and non-assetholders in consumption and asset return analyses.Annette Vissing-Jorgensen's NBER Working Paper (No. 8896, April 2002) examines the elasticity of intertemporal substitution (EIS) using data from the U.S. Consumer Expenditure Survey (CEX). The study highlights that limited asset market participation significantly affects EIS estimates. The EIS measures how much consumers adjust their consumption over time in response to changes in asset returns. The paper finds that EIS estimates differ substantially between assetholders and non-assetholders, with stockholders showing EIS values of 0.3-0.4 and bondholders showing 0.8-1. These estimates are higher for households with larger asset holdings. Non-assetholders show smaller, often insignificant EIS estimates. The findings suggest that limited participation in asset markets is crucial for accurate EIS estimation, as including non-assetholders in Euler equation models leads to downward bias. The study also notes that EIS estimates for stockholders and bondholders are higher than for non-assetholders, supporting the limited participation theory. The paper emphasizes the importance of distinguishing between assetholders and non-assetholders in EIS estimation, as the Euler equation does not hold for non-assetholders. The results are consistent with earlier studies showing that EIS increases with consumption and is influenced by wealth and asset participation. The study also discusses the limitations of using the CEX data, including measurement errors and the need for careful handling of data frequency and instrument selection. Overall, the paper underscores the importance of considering limited asset market participation in estimating EIS and highlights the significance of distinguishing between assetholders and non-assetholders in consumption and asset return analyses.