I develop an adaptive learning model to study the welfare effects of Social Security policy uncertainty in an aging economy. Agents combine full knowledge of the political process (which Social Security reforms are possible and when they could occur) with limited knowledge about the structure of the economy. The adaptive learning amplifies cyclical dynamics along the transition path to the new steady state. This magnifies the welfare effects of policy uncertainty, compared to a standard rational expectations model. The ex ante consumption equivalent variation that equates the expected utility of consumption (with policy uncertainty) to the utility of expected consumption (across the possible policies without uncertainty) ranges between −0.29% and 0.21% of lifetime consumption in the adaptive learning model compared to −0.012% to −0.018% in the standard model. The welfare cost to future generations is also larger in the adaptive learning model compared to the rational model.

中文翻译：

---

适应性学习，社会保障改革和政策不确定性

我开发了一种自适应学习模型，以研究老龄化经济中社会保障政策不确定性的福利影响。代理人将对政治过程的全面知识（可能进行社会保障改革以及何时进行社会保障）与对经济结构的有限知识相结合。自适应学习将沿着过渡路径的循环动力学放大到新的稳态。与标准的理性预期模型相比，这放大了政策不确定性的福利影响。在事前在自适应学习模型中，等效消费量变化将预期消费效用（带有策略不确定性）等同于预期消费效用（跨可能的策略，没有不确定性），介于终身学习的−0.29％和0.21％之间，而−0.012％在标准模型中为-0.018％。与理性模型相比，在自适应学习模型中，对子孙后代的福利成本也更大。