Essays in the Development and Costing of Income Contingent Loans

Abstract

(Note: Abridged abstract) Income contingent loans (ICL) are policy instruments that can enable participation in activities that yield both public and private benefits, with repayment conditions that promote equity, remove the risk of loan default, smooth consumption, and thus increase utility for loan recipients. Three essays are presented on the development and costing of ICLs. The first two essays consider the application of ICLs to paid parental leave and to student income support, and they are approached from the perspective of financial product design: context and motivation are described; product features are chosen to balance consumer needs and affordability with provider costs; models are developed and populated with assumptions and parameters; and the models are used to undertake risk assessment and costing. In the first essay, an ICL is developed as an extension to the recently introduced Australian statutory paid parental leave scheme. Design features are proposed to mitigate adverse selection and moral hazard, and it is shown that the ICL could be a cost effective and equitable means of providing parents with the necessary leave so as to optimise both private and public returns. In the second essay, an ICL is motivated and developed for the shortfall in student income support for higher education. The consequences of different loan indexation arrangements to expected taxpayer costs, loan recipient outlays, cross-subsidisation, and participation rates are discussed. A novel ICL sourced from superannuation is proposed separately as a means of financing income support for mature aged students. The viability of ICLs will depend, in part, on aggregate taxpayer costs. In the final essay, labour force and earnings models are developed to explore how ICL cost estimates are affected by model structures and assumptions. Nested bivariate logistic models for labour force transitions are developed that incorporate lagged labour states. Hourly wage is modelled, and residuals from the mean fit are partitioned into permanent and transitory components, incorporating serial dependency and non-normal shocks. A non-parametric model for weekly hours worked is developed that incorporates conditional transition probabilities. Monte Carlo simulation is used to estimate ICL debt, repayments and subsidies under the fitted models. It is found that under a wide range of conditions, dynamic stochastic earnings models lead to greater repayments, lower projected debt, and significantly lower aggregate taxpayer subsidies when compared with models that ignore earnings mobility.