The optimal policy in a wide array of areas turns on how one weighs benefits to earlier and later generations or cohorts. Using an incentivized survey experiment, we infer respondents' intergenerational social discount rates and compare them to how respondents make decisions for themselves. Respondents allocate resources over time to people in different age cohorts. Respondents choose how many meals to the homeless and how much clean water will be donated now or in twenty years. We use these choices to estimate a social discount rate consistent with respondents' choices. Using a series of further choices, we study whether respondents' social preferences are consistent with the sort of social welfare functions typically used in economic analysis.
Intervention
Respondents decide how to allocate goods over time for varied goods and settings.
Experimental design
We will present respondents with a series of convex time budget (CTB) decisions, where they allocate some good over time. We vary who receives the good and what the good is. We also introduce some decisions involving multiple people or goods. We carry out each decision for one survey respondent.
Design details
Randomization method
Random number generator on Qualtrics platform.
Randomization unit
Respondent
Clusters
596 Social Planners
Observations
596 Social Planners; 20 "Citizens" affected by their choices.
Arms
298 Social Planners make choices about audio and video; 298 make choices about money and either audio or video (50/50). The order of the three settings (self, within generations, across generations) is evenly randomized across Social Planners, with the additional settings specific to different treatment arms coming immediately after the across-generation choices.
Primary outcomes
• The share of a convex budget respondents allocate to the later period for four different settings and five different goods. The settings are long-term, across generations, within generations, and self, and the goods are meals for the homeless, clean water, and cash transfers to low-income Americans for the long-term setting, and money, audio-related work, and video-related work for the other three settings.
• The log-ratio of how much respondents allocate sooner compared to later in each of the four settings and five goods.
• The share of respondents' budgets they allocate to the later period for two different goods, audio and video work tasks, when they simultaneously divide budgets of each good across members of two generations.
• The log-ratio of how much respondents allocate sooner compared to later for two goods, audio and video work tasks, when they simultaneously divide budgets of each good across members of two generations.
• The log-ratio of how much money respondents allocate to a college junior compared to a college senior at the same point in time.
• The log-ratio of how much money respondents allocate sooner compared to later across generations when the older generation will also receive a transfer in the early period.
Secondary outcomes
• Respondents' policy attitudes and donation choices for a set of questions involving intergenerational policies and charities.
• The log-ratio of how much money respondents allocate sooner compared to later within generations when a separate, additional person will get a transfer in the early period.
• The share of a convex budget of money that respondents allocate to a senior now, a junior now, or that junior in a year (when they are a senior), given the choice of how to allocate a budget among all three.
Cited by
No citation data available for this registration yet.