Despite its importance for the analysis of life-cycle behavior and, in particular, retirement planning, stock ownership by private households is poorly understood. For instance, in contrast to the predictions of standard models of life-cycle behavior and portfolio choice, relatively few households hold stocks while those who do hold stocks often trade too often and too much.
Among other approaches to investigate these puzzles, recent research has started to elicit private households’ expectations of stock market returns (Dominitz and Manski, 2005). Whereas variations in households expectations about stock market returns elicited at one point may help predict variations in stock ownership, such an approach is only a first step in understanding stock ownership and trading. A natural extension is to test how changes in financial market expectations and changes in portfolio composition are related. Households’ beliefs about future events play a central role in forward-looking models of decision-making. Examples of probability beliefs that may affect individual decisions abound. They include beliefs about future labor market experiences, the future value of retirement portfolios of stocks, bonds, and social security benefits, and beliefs about receiving or leaving bequests, and health and mortality risks. Obtaining reliable measures of households’ beliefs with respect to future events has been at the center of much research in survey design and analysis over the past decades (see Manski, 2004, for an overview of the literature).
There is now a broad consensus that data about households’ beliefs should be obtained using probability formats (rather than using discrete response alternatives and verbal descriptors such as “very likely”, “likely”, and “somewhat unlikely”). The idea that probabilistic elicitation of expectations might improve on the traditional qualitative approaches of attitudinal research appears to have originated with Juster (1966). After some history in market research, probabilistic expectations questions have been used successfully in economic surveys since the early 1990s (Dominitz and Manski, 1997, 2004).
In the United States, the Health and Retirement Study (HRS) has pioneered asking questions about subjective probability beliefs on a wide variety of topics, including general events (e.g., economic depression, stock market prices, weather); events with personal information (e.g., survival to a given age, entry into a nursing home), events with personal control (e.g., retirement, bequests). Recent research by Dominitz and Manski (1997, 2004), Gan, Hurd, and McFadden (2003), and others shows that responses to probabilistic expectations questions are predictive for behavior. For example, responses to questions about subjective mortality risk are generally predictive for subsequent mortality experience (Hurd and McGarry, 1995, 2002; Smith, Taylor, and Sloan, 2001) and more predictive for savings behavior than objective life table hazard rates (Hurd, McFadden, and Gan, 1998). Manski (2004) provides an extensive discussion of research using survey questions on subjective probabilities.
Download
Short-run variations in households’ financial market expectations
