6 Conclusions
This paper has applied a framework for policy making under uncertainty to select a small number of policy options as the focus for future quantitative analysis. The paper is based on the view that detailed qualitative analysis assessing the relative losses and risks of alternative policy options is worthwhile prior to empirical estimation and numerical simulation, which typically are more resource intensive.
Three policies were selected:
- a low risk policy that places emphasis on time-consistency and agency cost issues while down-weighting the significance of distortionary taxation;
- a medium risk policy that applies a balanced weighting to the three issues; and
- a high risk policy that places emphasis on distortionary taxation while down-weighting time-consistency and agency cost.
The detailed policy targets that would apply for each candidate policy have been specified in terms of Crown net worth, overall risk/return properties of the Crown balance sheet, and the level and structure of financial assets and public debt.
The three policies are potential candidates for eventually putting to the government for consideration. The intention is that the government would select a preferred policy by comparing the incremental benefit versus risk of each policy relative to the status quo and other policy options.
A significant program of empirical estimation and numerical simulation analysis will need to be completed before any such recommendations can be made. Work will be required to formulate the loss functions, assign numerical values to the policy targets, and identify the variables that drive the sensitivity of the results. While initial rounds are likely to be rudimentary, as modelling progresses it will be important to place emphasis on comparative institutional analysis. This will require that, within reason, the policy targets and policy reaction functions incorporated in modelling work reflect how policy would actually be implemented based on the incentives, information and other constraints that would apply under the proposed institutional arrangement.
Inevitably, further analysis will identify significant knowledge gaps regarding the choice between alternative policies. Decisions will be required on which gaps can be reduced significantly through further empirical work and where judgemental assessments will be required. In respect of the latter, a small-scale survey of informed people could be desirable (e.g. for assessing ‘prior beliefs’). Again, attention will be required to ensure any judgements made are consistent with the incentives, information and other constraints under the proposed institutional arrangement.
