Essays on public pension reform

Thumbnail Image
Journal Title
Journal ISSN
Volume Title
University of Alabama Libraries

In my first essay, we examine the financial impact of states’ legislative response regarding pension reforms since 2008 recession. Since 2001 the combined funded ratio of the 115 state level public systems in our data set has fallen from being fully funded in 2001 to roughly 72% as of 2013. In absolute dollars this translates to a shortfall of almost $1 trillion. We found that 49 of 50 states passed pension reform legislation during this period; however, the impact of these reforms on funded ratios has been minimal due to the inability of most states to apply reforms to existing employees. In short, we don’t expect funded ratios to improve without significant increases in plan contributions. This result is likely due to the issue mentioned above regarding existing employees, amortization elections previously made by plans and unrealistic investment return assumptions. In my second essay, we use asset allocations and investment returns from the same data set to address the issue of pension plans taking on additional investment risk for the purpose of reducing annual plan contributions. We examine changes in portfolio allocations over a fourteen-year period paying special attention to responses to market downturns. We then examine investment returns on an actual and risk-adjusted basis. We find no evidence of moral hazard related to the asset allocations of public pension plans. In fact, we find that public pension plans outperformed the markets on an actual and risk-adjusted basis in many cases. We confirmed our findings described above using several models including the Capital Asset Pricing Model, Fama French three-factor and five-factor models and the Carhart four-factor model. We regress our equal-weighted and value-weighted pension investment return data against each of the models. All of our regressions produced positive and significant alphas. We then perform panel fixed effects regression on the same set of models. We find positive and significant alphas using all five models confirming the previous results. Finally, we employ a bootstrapping procedure to confirm the alphas in our time series regressions were positive and significant. We find that state level pension plans do generate positive alphas.

Electronic Thesis or Dissertation