You are here: Home Projects C: Financial market regulation C01: Financial decisions over the life cycle

C01: Financial decisions over the life cycle

In this project, we take a novel perspective on financial decision-making. We enhance classical economic models of the life cycle with features such as non-standard expectation formation, behavioral biases in decision-making, and peer effects. We develop estimable models and collect novel longitudinal data that allow us to identify and structurally estimate the parameters of the enhanced models. Our results generate new insights into the design of institutions such as social insurance schemes.


Project members (C01)

Policy outreach (C01)


Publications (C01)


Discussion papers (C01)



For important decisions taken irregularly and with long-term consequences (such as financial decisions for retirement savings) learning from experience can be difficult. Thus, traditional assumptions about individuals maximising utility under perfect knowledge of the institutional environment and the distributions of future shocks might not be realistic. The empirical evidence suggests that a significant fraction of individuals (a) have beliefs which are non-standard, heterogeneous, or not even well-defined; and/or (b) employ ad-hoc decision rules such as rules of thumb. We propose to collect new data and to develop and estimate rich models in order to explore the consequences of such behavior over the life-cycle.

Policy relevance

Quantitative insights into the design of welfare systems when a richer set of motives in life-cycle decision-making is considered, including distributional consequences of various actual and counterfactual policies. Analysis of policies that are irrelevant in standard models, e.g., nudging, information interventions, financial literacy education. Provide a microfoundation for analyzing financial market stability if not all households adhere to the rational-expectations paradigm.

Project plan

Work package 1 - Quantifying the effects of persistent heterogeneous survival beliefs in a life-cycle model

  • Add an individual-level time-constant degree of optimism or pessimism in survival beliefs to a state-of-the-art estimated life-cycle model.
  • Use a framework of ambiguity since no objective probabilities are available.
  • Learn about the effects of survival beliefs on savings and retirement.

Work package 2 - Developing a life-cycle model with heterogeneous learning about stock market returns

  • Develop a process for the formation of stock market expectations that can be taken to the data within a life-cycle model.
  • Use a framework of ambiguity with generalized Bayesian learning.
  • Calibrate the model, generate insights into the mechanics of life-cycle models and learning processes.

Work package 3 - Collecting novel data on beliefs, expectations formation, and preferences

  • Collect nine waves of data using incentivized experiments on the CentERpanel.
  • Bi-annual frequency: Stock market expectations.
  • Annual frequency: Preferences, qualitative questions on decision-making process.

Work package 4 - Analyzing the interplay between expectations, non-standard decision-making, and the social environment

  • Analyze the first waves of data collected in the CentERpanel in work package 3.
  • Relate the precision of expectations and the data on preferences, choice rules, and background characteristics.
  • Learn about how to identify individuals whose behavior is best described by nonstandard decision rules and which departures from a rational model are empirically important.

Work package 5 - Estimating life-cycle models with heterogeneous beliefs, learning, and alternative decision rules

  • Estimate comprehensive life-cycle models of consumption, savings, and portfolio choices featuring heterogeneity in preferences and beliefs about survival and stock market returns.
  • Model heterogeneity in decision-rules using discrete types.

Document Actions