Investigators: Pamela Giustinelli
Funding: Michigan Center on the Demography of Aging, 2015-2017
During year 1, the project will analyze expectations data in the 2008 HRS, where respondents were asked follow-up questions to the standard percent chance point-expectations about their labor supply, survival, and performance of the stock market to better understand the nature of respondents' use of focal values (0, 50, 100) or rounded/approximated figures (e.g., multiple of 10 or 5 percent). The primary aims of this analysis will be: (1) to describe patterns of
rounding/approximation across respondents and questions; (2) to explore whether the rounding/approximation follow-up questions can be used to empirically separate rounding that is used to simplify communication from rounding that reflects ambiguous beliefs point expectations questions do not enable respondents to express; (3) to assess the consequences of rounding and/or ambiguous beliefs on inference when subjective expectations data are used on the left hand side, and to validate existing methodologies for handling rounded expectations data. During year 2, the project will analyze the data collected in an experimental module of the 2016 HRS on long-term care preference, expectations, and sources of uncertainty/ambiguity. The overarching aim of the project will be to identify and study those determinants of individuals' longterm
care expectations that have the potential to inform the discussion about, and design of, longterm care policies, that is: (1) major sources of uncertainty, which might make it hard for people to
form precise beliefs about those plans (e.g., personal health or survival—preliminarily analyzed in year 1—or surrounding family support); (2) characteristics of long-term care arrangements or
insurance plans which are currently unavailable to, or unattainable by, individuals, but which would fulfill individuals' preferences and well-being better than existing options. From a methodological
and modeling perspective, this project will shed light on how new interval and conditional measures of subjective expectations may be used to empirically model and analyze individuals' stated preferences for long-term care options.