Personal Security Accounts and Mandatory Annuitization in a Dynastic Framework

Abstract

The aging of the populations in the OECD countries has prompted various calls for reforming the existing pay-as-you-go (PAYG) pension systems. Currently, there is renewed discussion in the United States about partial privatization where a fraction of the social security payroll tax would be diverted to Personal Security Accounts. In this paper, we quantitatively evaluate\nthe welfare effects of reforming social security by introducing a PSA with and without mandatory annuitization in an economic environment with bequests and borrowing constraints. Our setup allows us to assess whether mandatory saving or mandatory annuitization of accumulated PSA wealth at retirement is welfare enhancing, and if so, for what type of individuals. Our setup follows Fuster, Imrohoroglu, and Imrohoroglu (2003) and studies various pension schemes in a two-sided altruistic framework where social security provides insurance against individual income and lifespan uncertainty. This framework is well\nsuited to consider the annuity role of social security for single individuals versus for households where families also provide annuity insurance to their members. Our main findings can be summarized as follows:\n? A majority of households prefer a PSA reform (with or without mandatory annuitization) over the current PAYG pension system. Aggregate capital, output, and consumption, as well as individuals' lifetime welfare, are higher in the reformed pension system. \n? Mandatory annuitization benefits most households.\nIn light of these findings, structuring the social security reform along a two-tiered system with a safety net for low income households that do not have access to family insurance, and allowing all households to accumulate retirement wealth faster through PSAs, and finally, requiring some level of annuitization of this wealth appear welfare improving for a large fraction of households.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 16 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.