Tuesday, 27 April 2010

Contractionary measures should fall on the revenue side

Tom McDonnell: Last week’s Eurostat figures officially confirmed that Ireland has the largest general government deficit as a percentage of GDP of all the EU member countries and this news will increase the already very strong likelihood that the next Budget will contain both substantial cuts in expenditure and increases in taxation.

So what are the likely effects of this contraction? An ESRI working paper presented by Thomas Conefrey at the Irish Economic Association’s Annual Conference this weekend in Belfast gives us some idea (The Behaviour of the Irish Economy: Insights from the HERMES Macro-economic Model). The ESRI paper uses a medium-term macro-economic model called HERMES to forecast the effects caused by a shock (such as an expenditure cut or a tax increase) on economic aggregates such as employment, GDP and government borrowing. As the authors acknowledge, the HERMES model does not handle how households’ expectations are formed and does not take into account how concerted policy changes alter perceptions about the Irish economy. A further caveat is that the paper is a little out of date in so far as it forecasts the effect of a shock occurring in 2009. Nonetheless the findings are still of more than academic interest.

So what are the medium term effects of a menu of possible shocks? The ESRI calibrated property, income and carbon taxes to raise €1 billion in tax revenue in 2009 and the forecast effects generated for year 6 are shown below (I use year 6 here because it is the longest range forecast). If the shock occurs in 2010 then year 6 refers to 2016.

Table 1: Year 6 effects of a €1 billion increase in tax revenue
(% change relative to benchmark)

Based purely on these forecasts it would appear that both the property tax and the carbon tax on the non-tradable sector are superior alternatives to an increase in the personal tax rate. Raising income tax appears to have the most damaging impact on employment; the most damaging impact on GNP and the smallest positive effect on the general government balance. The choice between a property tax and a carbon tax is less clear cut; however what is clear is that either option will negatively impact both growth and employment. The carbon tax will be more successful in minimising the damage to economic growth (an impact of -0.1 on GNP compared to an impact of -0.4 from the introduction of a property tax). Thus if output is the policy goal then a carbon tax seems preferable. However, pursuing a carbon tax in lieu of a property tax will come at the cost of extra jobs lost. The latest Quarterly National Household Survey estimates that there are currently 1,887,400 people in employment. Thus a decline of just 0.1% in employment is equivalent to a decline of almost 1,900 jobs.

Table 2: Year 6 effects of a €1 billion cut in public expenditure
(% change relative to benchmark)

 Table 2 shows the various impacts of a €1 billion cut in public expenditure. Unfortunately, as the authors are quick to highlight, the HERMES model’s estimated impacts of a cut in investment do not take account of the long-term supply side impact of the cut on output and productivity caused by the reduced stock of infrastructure. Thus the figures shown for the impact of a cut in public investment are not comparable to the rest of the results and are excluded from the rest of this discussion. What is clear from Table 2 however is that focussing cuts on reducing the level of wages in the public sector is a superior policy option (in terms of growth; employment and managing the deficit) to a strategy of reducing the deficit through cutting the number of public sector employees.

Tables 3, 4 and 5 summarise the year 6 effects of the five comparable policy options.

Table 3: Year 6 Effect on GNP – Ranking the Policies
% Change in GNP relative to benchmark

Table 4: Year 6 Effect on Employment – Ranking the Policies
% Change in Employment relative to benchmark

Table 5: Year 6 Effect on the Government Balance – Ranking the Policies
% Change in the Balance relative to benchmark

In a different paper presented by Vincent Hogan at the IEA conference (Expansionary Fiscal Contraction: Deja Vu All Over Again?), the author found that there is no real evidence for the historical existence of expansionary fiscal contraction. Certainly the year 6 estimates generated by the ESRI’s HERMES model bare that conclusion out. The conclusion is that contractionary policies will have negative effects on output and employment. If contraction is the policy then the goal becomes one of minimising this damage. What the rankings in Tables 3, 4 and 5 show is that the damage can be minimised by focussing fiscal rectitude on the revenue side i.e., the least damaging long term economic impacts will be obtained through tax increases rather than through expenditure cuts.

The carbon tax and property tax options appear to be the most compelling options. However it is important to also consider the effects on the real economy. For example a carbon tax (depending on how it is designed) can have adverse distributional effects. In a recent study by Callan et al., (2009) published in the Journal Energy Policy; it was found that the distributional impact of a carbon tax in Ireland is indeed regressive. From a social policy perspective carbon and property taxes should be calibrated in such a way as to avoid any regressive effects occurring. If incorrectly designed property and carbon taxes have the potential to greatly increase poverty levels and to magnify Ireland’s already high level of inequality. Thus it isn’t enough simply to decide that a carbon/property tax is the best option. There are numerous ways that €1 billion can be generated through a property/carbon tax and careful consideration must be given at the design stage to minimise any regressive impact and to ensure that such a tax does not adversely impact the poorest in society.


Anonymous said...

I guess you didn't get the memo: the public sector is the enemy and must be destroyed. It doesn't matter if it makes any sense, for the great Sindo has spoken!

Michael Burke said...

I think it is very useful to use the research from the ESRI and others, which tends to undermine the entire basis of the discussion on cuts.

In official and media circles, the effect of all cuts are consequence-free and have no impact on the real economy whatsoever. From that entirely mistaken supposition, it is then asserted that cuts = savings.

Yet, as the ESRI and others remind us, contractionary fiscal policies have exactly that effect, they contribute to economic contraction.

But it is also worth noting that the ESRI and other models are based on long-run average experience, both booms and busts. In a bust, as now, the negative impact of withdrawn government spendng is increased relative to the average, ie the 'multipliers' increase both positively and negatively as the case may be in terms of fiscal stimulus or contraction. This is because the private sector is otherwise incapable/unwilling to fill the breach left by withdrawn public spending during a recession, and, conversely, is much more inclined to invest on its own account where it sees demand risng as a result of government spending.

It might be useful to examine the actual course of events in the recession to date, and the impact on growth and government finances arising from government policy.

I hope we can publish something on that here shortly.

Slí Eile said...

@Tom Very useful data updating on previous econometric work by ESRI in May 2009. Do you have a link to the IEA paper in Belfast? I can't find on a quick search.