Background paper

Background Paper for the 2021 Statement on the Long-term Fiscal Position: Shocks and Scenarios Analysis Using a Stochastic Neoclassical Growth Model

Acknowledgments

I thank Jean-Pierre Andre, Matthew Appleby, Matthew Bell, Steve Cantwell, Karsten Chipeniuk, Shane Domican, Craig Fookes, Margaret Galt, Peter Gardiner, John Janssen, Anita King, Sam King, Natalie Labuschagne, Udayan Mukherjee, Murat Özbilgin, Oscar Parkyn, Hemant Passi, Bettina Schaer, Nathan Spence and Niven Winchester for their help and their useful comments. All mistakes and errors are my own.

Accessible HTML version

Only the Executive Summary of this paper has been prepared in HTML. If you require a full HTML version, please contact [email protected] and cite Background Paper for the 2021 Statement on the Long-term Fiscal Position:  Shocks and Scenarios Analysis Using a Stochastic Neoclassical Growth Model as a reference.

Executive Summary#

At least once every four years the New Zealand Treasury is required, under the Public Finance Act 1989, to produce a long-term fiscal statement, outlining how the Crown’s fiscal position could evolve over the next 40 years. With the onset of COVID the publication of the 2020 long-term fiscal statement was delayed. The updated 2021 long-term fiscal statement has been combined with the Treasury’s inaugural long-term insights briefing. The long-term insights briefing, a new requirement under the Public Service Act 2020, outlines future fiscal challenges and provides analysis and policy options government could consider when responding to these challenges. The largest fiscal pressures, likely to affect revenue, expenditure and net government debt, are the ageing population, increasing health expenditure and climate change. While the fiscal impacts of an ageing population and increasing health expenditure have been extensively covered in previous long-term fiscal statements, this is the first time the fiscal implications of climate change are considered. This is also the first time shocks and scenarios analysis is produced using a general equilibrium model (in this case, a stochastic neoclassical growth model), to support the long-term fiscal statement, which I document in this paper.

Analysis from the long-term fiscal model (LTFM), a spreadsheet-style accounting model, has formed the basis of the modelling work for previous long-term fiscal statements. The LTFM is an extremely useful tool, providing detailed projections of government expenditure, revenue and net government debt, under the assumption that policy does not respond and there is no feedback to the rest of the economy. However, the LTFM (like all models) has its limitations. In particular, it does not capture behavioural or feedback responses, and it is deterministic, which rules out the investigation of uncertainty. I address these issues in this paper by abandoning spreadsheet-style accounting models in favour of a general equilibrium model. More specifically I develop a stochastic neoclassical growth model (NCGM), with a government sector, to carry out analysis investigating the fiscal impacts of an ageing population, increased health spending and some aspects related to climate change. The model I develop is plain vanilla in many regards, sharing a number of features with the textbook growth model. Government plays a key role stabilising net debt around its long-run target. The NCGM provides complementary analysis to the LTFM.

In the baseline ageing population scenario key spending tracks from the NCGM are matched with their counterparts in the LTFM and net debt as a share of GDP is kept reasonably stable around its target over the 40 year reporting period. However, this requires substantial increases in tax rates, with the average tax rate on labour increasing by 6.4 percentage points, the average tax rate on capital income increasing by 10 percentage points, and the average tax rate on consumption expenditure increasing by 5 percentage points, resulting in a GDP path that is 3.7 percent lower than what have prevailed in the absence of these spending pressures in 2061.

I produce a range of alternative scenarios where some of these assumptions are altered or the model economy is subjected to economic and physical shocks. The main results from these scenarios can be summarised as follows:

  • The recessions modelled raise net debt to GDP by between 11 and 13 percentage points, and require reasonably aggressive tax responses from government to return net debt to target before the next recession.
  • A large earthquake raises net debt to GDP by 12 percentage points, similar to the recessions scenario. The overall fall in GDP is smaller compared with the recessions scenario, due to the faster decline and recovery in trend total factor productivity and the rebuild which factors in a healthy degree of “building back better”.
  • Fast fiscal consolidations are more costly (in a GDP sense) in the short run, but generate more benefit in the long run as debt servicing costs are reduced. Slow fiscal consolidations spread the economic costs over a longer period of time, resulting in smaller costs per period. However, elevated debt in the slow consolidation scenario translates into higher tax rates and lower spending over a longer period of time and a larger cumulative GDP cost.
  • Delayed fiscal responses to increasing debt generate benefits in the short run, as additional government spending increases aggregate demand without the costs of higher taxation. However, when government decides to stabilise debt at a higher level, or consolidate, taxes need to be raised rapidly to fund higher debt servicing costs and the higher primary surpluses required to bring debt back to target. These higher taxes reduce economic activity and the level of GDP. In most scenarios investigated the net gains from delaying are outweighed by the costs of stabilising or consolidating debt.

I produce separate stochastic scenarios that investigate the fiscal and economic implications of changes in the frequency of droughts and the intensity of storms due to the effects of climate change. In the droughts scenario, the increase in drought frequency could lower GDP by 0.5 percent relative to trend, on average, by 2061, while net debt to GDP could be 1.2 percentage points higher on average. In the storms scenario, the increase in storm intensity suggests that GDP could be 0.7 percent lower than trend, on average, by 2061 and net debt to GDP could be 3 percentage points higher, on average. I note that droughts and storms add a measurable amount of volatility to both the net debt and GDP projections.