Wharton Pension Research Council Working Papers

Document Type

Working Paper

Date of this Version



Public employee pension systems have traditionally been of the pay-as-you-go defined benefit (DB) variety, where retiree payments are financed by taxes (contributions) levied on the working generation. The same holds for Germany, where civil servants are promised a (mostly) unfunded, noncontributory, tax-sponsored DB pension, representing substantial liabilities currently not recognized as explicit obligations to the public sector. This paper analyzes the risks and rewards of moving to a (partially) prefunded pension system for most civil servants in the German federal state of Hesse. First, we conduct an actuarial valuation of pension promises to retired and active civil servants, which we conservatively put at € 44 billion in present value, or about 150 percent of explicit state debt. Second, we project 50 years into the future and estimate the payroll-related contribution rate sufficient to fund the civil servant pension obligation. Next, using a Monte Carlo framework and a stochastic present value approach, and a Conditional Value at Risk measure, we identify an asset allocation for plan assets that minimizes worst-case pension costs. Prefunding the pension with a tax worth about 20 percent of payroll, and investing the assets 30 percent in equities and 70 percent in bonds, substantially reduces expected and minimizes worst-case pension costs. Finally, we illustrate contribution rates and asset allocation when the plan sponsor is limited to a particular risk budget. In one interesting case, current taxpayers are asked to pay additional regular contributions of only 15 percent while the portfolio is held 43 percent in equities. This mix allows future generations to benefit from possible contribution holidays and withdrawals, while providing an acceptable level of risk of supplementary contributions resulting from underfunding.


public sector pensions, defined benefit pensions, risk

Working Paper Number


Copyright/Permission Statement

Opinions and errors are solely those of the authors and not of the institutions with whom the authors are affiliated. © 2007 Maurer, Mitchell, Rogalla. All rights reserved. © 2007 Pension Research Council. All rights reserved.


This research was conducted with support from the TransCoop Program of the Alexander von Humboldt Foundation. Additional research support was provided by the Pension Research Council (PRC) at The Wharton School of the University of Pennsylvania. We are grateful for useful comments from Peter König and for data support by the Hessian Statistical Office.

Included in

Economics Commons



Date Posted: 17 December 2019