Tuesday, 19 May 2009

Will a Domestic Stimulus work in a Small Open Economy like Ireland?

Sli Eile: Taoiseach Brian Cowen commented over the last weekend that the main opposition parties wants the Government to cut more while the other (minor) opposition party wants the Government to cut less. Hard to disagree with this observation. Draw your own conclusions. But, who is for not cutting at all and instead spending more? Yes, spend more but very differently and under very different arrangements with regard to how banking, public finances and corporations are run in Ireland and Europe.

Turning to political economy, the trouble with many economists is that they can only look back – to old theories, old evidence and old empirical models based on what is measurable and what was given in terms of the external environment. On the other hand, it is hard to look into the future without regard to what has happened in the past and why it happened. The 1960s were the heyday of econometric modelling as economists discovered new data sources and put all their quantitative prowess on display through the science of multi-variate statistical modelling. This was, also, the time of ‘manpower planning’ and Economic Programmes (Whitaker closer to home). The first ‘Oil Shock’ of 1973, and following it the slowdown in economic growth in the early 1980s, disturbed many of the stable empirical relationships.

The new orthodoxy was monetarism – strong medicine for a new world – allied to ever more complex modelling of micro-economic behaviour and macro-economic impacts.

We are in a muddle again as the old world dissolved in 2008. In particular, the emergence of a very different profile of industrial output, labour market structure and public-private balance emerged in Ireland in the 1990s and the present decade.

The bottom line is that it is hard to model an economic ‘readjustment’, let alone a recovery, when you in the midst of an economic tsunami. The ‘old reliables’ are gone in more ways than one.

A future blog will comment on the latest ESRI publication Recovery Scenarios for Ireland (published last week). This blog goes back to an earlier, less publicised, document that sought to quantify the impact of various policy shifts in regard to expenditure, taxation, employment as well as ‘exogenous’ shifts in competitiveness and world trade.

In a heroic attempt to model the impact of various (simplistic) adjustments to taxes, public spending and nominal wages, the authors of the ESRI paper entitled ‘The Behaviour of the Irish Economy: Insights from the HERMES macro-economic model’, (Adele Bergin, Thomas Conefrey, John Fitzgerald and Ide Kearney) have done some service in assessing the impacts of various changes on key economic outcomes such as GNP, GDP, Unemployment, Government Borrowing and price inflation. Using historical data and based on a complex forecasting model (HERMES) drawing data from another global era, they have modelled forward the projected or estimated impacts of a number of ‘shocks’ or adjustments including the following:

  • 5% cut in nominal wages

  • Cuts in public spending

  • 1% Increases in world growth

  • 1% Improvements in competitiveness

A number of salient points are in order:

  • Economics is not a perfect science – the questions you choose to ask and explore empirically are a function of your values and those underlying assumptions and interests that you hold dear;

  • Instability, uncertainty, conditionality and the impact of ‘exogenous’ variables renders standardised econometrics in the league of heroic simplicity, in spite of all its finesse and seeming complexity;

  • The past is a different place and not necessarily a sound guide to the future or present; and

  • No policy response is ideologically, politically or morally neutral.

That said, fair play to the ESRI for using the only empirical data available to assess various possible outcomes. But, empirics can never tell the full story, neither can they tell you what to do. The ESRI authors fully acknowledge the limitations (‘expectations in the model are backward looking’, p7 and ‘the unquantifiable effect on confidence’ is omitted).

The doctrine of ‘expansionary fiscal contraction’ (public spending cuts fuelling recovery of private consumption and investment) is firmly rebuked in the paper, as it had been already by Bradley and Whelan in a 1997 paper. The problem identified in many ‘growth studies’ over recent decades is that it is fiercely difficult to account for factors such as new technologies, the impact of political changes and swings in business mood. Some economists such as Harberger (1998) have distinguished between “yeast” and “mushroom” effects in explaining economic growth.

Factors such as knowledge and human capital act like yeast to increase productivity relatively evenly across the economy, while other factors such as a technological breakthrough or discovery suddenly mushroom to increase productivity more dramatically in some sectors than others. The ‘X’ factor is a lot bigger than people imagine. During the heady days of the Celtic Tiger, some growth studies identified a very large ‘unexplained’ residual in the case of Ireland, suggesting productivity increases well above what could be accounted for by standard input measures. Buried in the plot was, no doubt, the impact of temporary foreign direct investment, price transferring and international spillover effects.

The number-crunching (based on ceteris paribus on the explanatory side – all else constant while one variable is shifted but allowing for interactions in all the outcome variables) on various scenarios is summarised in the ESRI paper as follows:

Gross National Product and Gross Domestic Product would fall initially but recover in the medium- to long-term as a result of a 5% cut in nominal wages (details are summarised on page 3 of the paper).

GNP/GDP would fall initially as well as in the medium-term (to 2013) as a result of a one-off hike of €1billion in any of the following: income tax, property tax, public sector pay cuts, employment cuts and Government investment reductions. The extent of the impact varies with larger negative impacts in the case of cuts in employment, public sector pay and income tax. Carbon taxes would be mildly expansionary in the case of the GNP measure due to a reduction in profit repatriations by the manufacturing sector (but not GDP).

GNP/GDP would be higher in the medium-term as a result of either a 1% increase in world growth or a 1% improvement in competitiveness.

One of the major sources of instability is migration. Unemployment peaked at 17% in the late 1980s but would have gone much higher were it not for the huge level of outward migration at the time reflecting job opportunities in the UK and other destinations. Clearly, the same does not hold now. A safe bet is that labour supply will remain fairly ‘inelastic’, at least until green shoots of recovery appear in the UK labour market. The implications of this – not spelt out in the ESRI paper – is that any deflationary shock (recall that the ESRI scenarios entailed a €1billion shock, which is a small compared to what Government is promising us for the next 4 years) will have large and difficult-to-predict impacts on unemployment. Although the ESRI didn’t model the impacts on poverty, health and well-being it is safe to assume that these will be substantial – in the absence of any reversal of current economic policy.

Still, while modelling for a fall in employment in the education and health sectors (p20) the authors bank on ‘extensive emigration’ so that the unemployment rate would initially rise by 0.9 per cent points and fall back to 0.2 by 2015 (for a reduction of around 17,000 in the numbers employed in health and education in 2009). They acknowledge the uncertainty in the labour market situation internationally. Nobody has provided solid evidence, yet, that we are looking at return to net outward migration, and certainly nothing of the order last seen in the late-1980s (when unemployment peaked at 17% and net outward migration at 44,000 in 1989). Put another way, it is not obvious that unemployed teacher graduates or nurses can readily find employment in the UK and further afield. But, it may come to that if labour markets pick up elsewhere before the Irish labour market. And, it would seem that domestic policy is, implicitly aiming for this outcome.

Instructively, the ESRI conclude that pay cuts in the public sector have bigger bucks than employment cuts. Hence, a cut of 17,000 jobs in 2009 would save ‘only’ €500 – much less than half of what could be saved from a cut of 5% in public sector pay. The lesson they seem to be strongly hinting at is cut pay before you cut jobs.

Very crudely, if Government is promising an ‘adjustment’ of some €4 billion for each of the coming 3 years on top of the full-year adjustment of €5 billion, this year then we are looking at (very crudely) something possibly like €17 billion in total cumulative terms. In other words, a one-off impact of a €1 billion multiplied 17 times gives a downward long-run adjustment of 7 % in GNP – other things equal. Who is to know the dynamic effect of such an adjustment if it further depresses demand and undermine confidence?

Perhaps the one of the most intriguing aspects of this paper is the estimated impact of cuts in Government investment (p22). They write:

we consider the impact of a €1 billion reduction in expenditure on public investment under the National Development Plan. These results only take account of the demand side impact of the change in investment. They take no account of the longer-term supply side impact reducing national output and productivity as a result of the reduced stock of infrastructure.

Then they spell this out:

Thus the longer-term impact of this cut on output and employment would be substantially greater than shown here.

The impact on public finances is large (reducing borrowing) for a cut of €1billion in the public capital programme. The net impact on national output is very small in the medium-term (to 2015) allowing for some positive impact on private manufacturing and services in the ESRI model. However, the long-term impact on the ‘supply-side’ is unknown and unquantifiable. Put another way, cuts in the PCP (like in the Ireland of the 1950s), along with cuts in public services such as health and education, will have lasting effects and these effects will interact with the rest of the economy and society. Have we not learned the lessons from the lasting impact of health cuts in the late-1980s?

The ESRI paper standardises all the impacts into a monetary-based multiplier Table 9 (p25). The biggest long-term negative impact is -1.35 in respect of a €1b value cut in public sector employment and the strongest positive impact is 0.15 from a carbon tax hike of €1b. Underlying the ESRI analysis is:

  • An inherent assumption that deflation is a necessary part of the medicine to get back on track – (real) pay cuts, a mix of tax increases and some pruning of public investment is assumed appropriate in the circumstances; and

  • A hope that improved trade conditions in conjunction with a moderately conservative domestic fiscal stance will lift the Irish boat – in 2011 if thing go well and later if not – eventually the storm will subside and how quickly we bounce back depends on things outside our control and things inside our control.

The Paper does not deal with issues around supply-side initiatives such as training and labour market flexibility – but this can be factored into a package as well.

A problem with the ESRI one-off ‘ceteris paribus’ shocks is that in the real world everything is changing and interacting and a policy stimulus to lower unemployment and improve competitiveness and reduce borrowing in the medium-term and raise national output needs to combine a range of measures into a coherent package. However, the ESRI exercise is useful at least analytically in quantifying the separate effects of one-off shocks on the assumption that everything else on the ‘policy instrument’ side is held constant.

In conclusion one could ask why the ESRI authors chose to try some particular set of scenarios and not others. They probed the impact of cutting expenditure and wages as well as raising taxes. And, they probed the impact of improved international trade and Irish competitiveness on global markets. But, they did not probe the impact of a fiscal stimulus and still less a forensic one targeted at particular sectors and spenders within the economy. Moreover, nobody can quite model the impact of a political stimulus based on a new leadership, new hope and reform of democracy and governance in the corporate and political worlds. It could surprise everyone – even for a small open economy like Ireland with some leverage in Europe and the wider world.

As always, comments, corrections, disagreements, suggestions from the blogosphere on the above welcome.

No comments: