Jump to main content.

Understanding Observed Differences in Time Preference Rates

Quick Links

Research Funding

Metadata


The primary objective of the proposed research is to improve the techniques used in the assessment and evaluation of project and program alternatives by investigating the conditions under which people discount future gains and future losses at different rates. Our experiments will examine the role and importance of the endowment effect and other context, specifically preference construction and affect or emotions, in explaining differences in intertemporal choices. We also will examine possible differences in time-preference rates for different types of outcomes (e.g., financial vs. environmental vs. health effects) and for near-term, as compared to far-term, consequences.

Approach: Responses to questionnaires will be analyzed, primarily using comparisons of the proportions of respondents answering questions that investigate (a) the conditions under which people discount future gains and losses at different rates, (b) the role and importance of context in explaining differences in intertemporal rates, (c) possible differences in intertemporal rates for different kinds of future outcomes, including a variety of financial, environmental policy, and health consequences, and (d) tests of different rates for near- and far-term outcomes in ways that take into account possible confounding by the endowment (gain/loss) effect. Both descriptive and inferential measures will be used; for example, comparisons of mean willingness-to-pay responses and comparisons of proportions in different cells (i.e., under different experimental conditions) will be analyzed using standard univariate and multivariate statistical tests to infer the reliability of the differences that we observe.

Expected Results: The expected results of this project will contribute to improvements in the methodology used in benefit-cost analyses of proposed changes to environmental quality and public health. A consistent difference in the observed rates among time-preference measures would call into question the conventional practice of using a single rate to discount all of the future outcomes attributable to a program or project. A disparity in rates used to discount future gains and losses also would cast doubt upon the common practice of calculating the net present value of a stream of future gains and losses as either (1) the sum of the present values of all the gains less the sum of the present value of all of the losses, or (2) the sum of the present values of the benefits minus the losses in each year. These would only be equivalent if the discount rate were the same for gains and for losses. In addition, the implications of the proposed studies extend beyond formal analyses of costs and benefits, present values and the like to include the way in which issues and problems involving time preferences are thought about and discussed -- the habit of mind that can influence social judgments and decisions regarding protection of the natural environment and public health.

Metadata

EPA/NSF ID:
R827931
Principal Investigators:
Gregory, Robin
Slovic, Paul
Technical Liaison:
Research Organization:
Decision Science Research Institute
Funding Agency/Program:
EPA/ORD/Valuation
Grant Year:
1999
Project Period:
November 30, 1999 to November 29, 2001
Cost to Funding Agency:
$228,463
Project Status Reports:
For the Year 2000

Objective: The primary project objective is to improve understanding of whether and why people may discount future gains or future losses, as well as different types of future outcomes, at different rates. Results of the project could influence the analysis of initiatives to protect the natural environment and public health and contribute to a reconsideration of the conventional discounting practices of economic benefit-cost and policy analysis.

Progress Summary: The first year of the research has focused on three questions. The first is to see if there are systematic differences in expressed time preferences depending on whether questions are framed as gains or as losses. In terms of the four-quadrant diagram we've been using (see below), this means a comparison between Quadrant I (the standard willingness to pay [WTP] quadrant) and Quadrant III (the standard willingness to accept [WTA] quadrant). Based on earlier experimental results that did not involve comparisons across time periods, our expectation was that discount rates for Quadrant I (gains) would be significantly higher than for Quadrant III (losses), and our initial results demonstrate this. Responses for Quadrants II and IV are, as expected, intermediate, but these quadrants are rarely appropriate for valuation purposes.

Quadrant I
(WTP)


Gain
Quadrant II
(choice of gains)
-Money
Quadrant IV
(choice of losses)


+Money
Quadrant III
(WTA)

Loss

Example of a four-quadrant diagram for displaying gains and losses over time.

The second research question initially asked whether there are systematic differences in expressed time preferences for different types of goods, specifically environmental, health, and financial goods. As the study evolved, we realized that something more fundamental than the type of good was at work. We believe this more fundamental difference could have to do with the emotional or affective response elicited by the item, although our current experiments also are considering the role of fairness and social/community interests. The more general point is that the context for the time preference task clearly matters, and ongoing experiments seek to understand more clearly the associated reasons and mechanisms.

The third research question involves taking these results into the domain of specific policy questions relevant to an agency such as EPA. This concern has guided our selection of examples and the experimental focus. A first round of experiments, conducted in March 2000, focused on pre-tests of questions comparing time preferences rates for changes coded as either gains or losses. The second, in November 2000, built on the lessons of these earlier tests and also asked questions about time preferences for different types of goods and in different contexts.

Future Activities: Our research continues to examine the same three questions outlined in the proposal. However, the original second research question, which focused on possible differences in time preferences for different types of goods, has evolved so that we now have become more interested in understanding the underlying contextual reasons?including affective and emotional responses as well as judgments of social norms and fairness?for subjects' expressed differences in rates of time preference.

We will continue to refine and expand our experiments over the coming 6 months, with the next round of experiments being run in Eugene, Oregon, in early March 2001. These experiments will clarify the relative contributions of gains and losses, and of differences in time preferences, to the observed results and also will include additional scales to further investigate the meaning of these results for conventional practices of benefit-cost analysis.

Book Chapter:

Knetsch J. Environmental valuations and standard theory: behavioral findings, context dependence and implications. In: Tietenberg T, Folmer H, eds. The International Yearbook of Environmental and Resource Economics 2000/2001, Edward Elgar, Cheltenham, UK, 2000.

Papers:

Knetsch J. A serious bias in policy choices and damage assessments.

Peters E, Knetsch J, Slovic P, Gregory R. Affect and the endowment effect.

Knetsch J, Gregory R. Discounting future gains and future losses: the context dependence of time preferences.

Gregory R, Knetsch J. Values and time: implications for deliberative environmental valuation.

Gregory R. Valuing risk management choices.

For the Year 2001:

Objective:

The main objective of this research project is to improve our understanding of the reasons why people may discount future gains or future losses, as well as different types of future outcomes, at different rates. Results could influence the analysis of initiatives to protect the natural environment and public health and contribute to a reconsideration of the conventional discounting practices of economic benefit-cost and policy analysis.

Progress Summary:

The second year of this research project focused on three guiding project questions. The first is to continue to test whether there are systematic differences in expressed time preferences, depending on whether questions are framed as gains or as losses. In terms of the four-quadrant diagram we have been using (see Figure 1), this means a comparison between Quadrant I (the standard willingness-to-pay quadrant) and Quadrant III (the standard willingness-to-accept quadrant). Our results demonstrate that discount rates for Quadrant I (gains) are significantly higher than for Quadrant III (losses). Responses for Quadrants II and IV are, as expected, intermediate but these quadrants are rarely appropriate for valuation purposes.

Quadrant 1
(WTP)
Future Gain
Quadrant II
(choice of gains)
-Money
Quadrant IV
(choice of losses)
+Money
Quadrant III
(WTA)

Future Loss


Figure 1. Four-Quadrant Diagram for Displaying Future Gains and Future Losses

The second research question asks whether there are systematic differences in expressed time preferences for different types of goods (e.g., environmental, health, financial). As the study evolved, we realized that differences in time preferences may reflect the emotional or affective response elicited by characteristics of the item, rather than solely by their type.

The third research question involves taking these results into the domain of specific policy questions relevant to an agency such as the U.S. Environmental Protection Agency. This concern has guided our selection of examples and the experimental focus.

Future Activities:

We will revise submitted papers to the extent suggested by reviewers and we will continue to work on the remaining draft papers so that they can be submitted for publication consideration. We may want to run one more small set of experiments to test out ideas that arose from the February and June 2001 rounds of experiments conducted in Eugene, OR.

Book Chapter:

Knetsch J. Environmental valuations and standard theory: behavioral findings, context dependence and implications. In: Tietenberg T, Folmer H, eds. The International Yearbook of Environmental and Resource Economics 2000/2001, Edward Elgar, Cheltenham, UK, 2000.

Journal Articles:

Peters E, Slovic P, Gregory R. Constructing prices: anchoring, adjustment, and affect. Quarterly Journal of Economics.

Knetsch J, Gregory R. Discounting future gains and future losses: further evidence of the context dependence of time preferences. Journal of Risk and Uncertainty.

Gregory R, Knetsch J, Arvai J, Burns K. Multiple discount rates: the influence of decision context on choices over time. Organizational Behavior and Human Decision Processes.

Finucane M, Gregory R, Knetsch J, Mertz CK. Reference points, affect, and sequential policy judgments. Risk Analysis.

Gregory R, Knetsch J, Mertz CK, Slovic P. Understanding differences in implied time preferences. Journal of Environmental Economics and Management.

Paper:

Knetsch J. A serious bias in policy choices and damage assessments.

Supplemental Keywords: time preferences, discounting, context, benefit-cost analysis.

Project Reports:

Local Navigation


Jump to main content.