Sunday, 10 January 2010

Mirror, mirror on the wall: who is the most deflationary of them all?

Slí Eile: The ESRI authors of their latest Quarterly Economic Commentary make the following two claims:
‘In spite of popular perceptions, analysis contained in the Commentary shows that Budget 2010 was not the most contractionary of modern times..’
‘while Budget 2010 was clearly regressive, the combination of Budgets 2009 and
2010 placed most of the burden of fiscal adjustment on higher earners’
Predictably, the media picked up these claims and they have entered the lexicon of canonical and uncontested truths along with all the others with regard to pay, social welfare, debt … In this blog I survey the first claim. Next week, I will explore the second of the two claims.
As pointed out in a previous blog, it is difficult to assess these claims without access to the full and complete publication of the Quarterly Economic Commentary (QEC). However, having seen the same I am struck just how thin the evidence is. The Winter 2009 QEC is 76 pages in length. The claim with regard to the scale of contractionary impact is discussed in Box 1 on three pages only: 21-23. Three figures are contained in this Box (A, B and C). ‘Own Estimates’ is mentioned at the foot of each Figure (A, B an C) in Box 1. There is no citation of working papers or other published research apart from a paper written in 2000 (Assessing the Stance of Irish Fiscal Policy)

In the decomposition of fiscal stance (Figures A and B) no precise figures are shown. Visually, Figure A suggests a combined contractionary effect of around 2.8% of GNP (and not GDP) in the 2009 Budget (presumably October 2008 and April 2009 Budgets combined?). The hit in 2010 is in the order of 1.5% of GDP. These estimates are hard to accept on first sight. The scale of adjustment, to date, in combination with the negative multiplier impact of lost jobs and income on household consumption is likely to be greater. The ESRI estimate – however it was derived and whatever assumptions were made in its generation via the Hermes model is uncanningly the same as the Department of Finance (DOF) estimate of the contractionary impact of last month’s budget.
In the background documentation, DOF (The document can be downloaded here) has estimated that the combined fiscal adjustment in Budget 2010 was €4bn = 2.5% of GDP (section 3.1). The document states that: 'Table 8 below sets out the estimated loss in tax revenue of €897 million associated with the introduction of the budgetary package in 2010.' Adding these two elements together gives an estimate closer to 3% of GDP and not 2.5%. Once again, the full set of assumptions behind these estimates is not provided. Going by the collapse in consumer spending in 2009, one suspects that this is not unrelated in a significant way to cuts in public spending (plus a host of other factors including fear about what lies around the corner by way of further income cuts or redundancy).
A significant (majority) component of the fiscal adjustment in 2009 and, again, in 2010 was on the capital side. In that regard, it is worth referring to earlier empirical work by the ESRI in April measuring the impact on GDP – in the long-term – from cuts in public spending. For example:

…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.

What is striking about the comparisons between the peak fiscal constractions of 1976, 1983, 1989 and 2009 (refer to Figure B) is the following:
The 1976 and 1983 contractions are estimated by the ESRI at 6% and 3.5% of GNP, respectively, (and not GDP);
The 2009 contraction impact, the lowest of the four peaks, is about 3.3% (going by the naked eye and no precise figures in the article) of GNP; and

The composition of contractionary effects differs as follows:
The bulk of adjustment in 1976, 1983 and 2009 was by way of the Public Capital Programme (schools, hospitals, transport) while the bulk of the adjustment in 1989 was via current expenditure (and with an estimated expansionary impact from lower taxes at the same time in the 1989 Budget).
Peering into the past may be deemed a purely academic pursuit. However, it has implications for informing political economy debate in 2010. Recall that politics is about who gets what and economics is also about who gets what – in the midst of a depression it is comforting to be told that:
The whole world is in recession
We must devalue across the board (well for most groups..)
Things aren’t as bad as they were in the 1970s and 1980s on a particular measure of ‘fiscal stance’
Prices are falling so the apparent cut in wages and living standards is not as much as it seems (if at all…); and
There is no alternative anyway (refer to the una voce media in Ireland)

It should not escape our attention that:
All the major fiscal contractions of the past (the 1923 cut in the OAP, the 1930’s cuts in public sector pay, the 1950s cuts in capital expenditure, the 1976 cuts in capital spending, the 1989 cuts in health spending and the 2009/10 cuts in Social Welfare, public sector pay, capital funding and other programmes were all PRO-CYCLICAL. In other words, the country under successive Governments of tweedledum and tweedledee have been following the McCreevey school of economics: when I have the money, I spend it; when I don't, I don't. Either they never heard of Keynes or Keynes doesn’t apply here (small open economy, unique position vis-à-vis the international financial markets, national sovereignty as risk blah blah)

However, the key to the ESRI claim that Budget 2009 or 2010 were not as contractionary as the 1983, 19988 or 1989 budgets is the impact of price inflation. With consumer prices falling, the cuts in public spending are not as deflationary as they appear. With prices increasing in the 1980s, restraints in public spending implied real cuts.
However, its seems to me that a stronger contradiction emerges in Figure C on page 23. The graph (Measure of Fiscal Stance and GDP Growth Rate) shows, on two axis, (i) GBRR (Government Borrowing Requirement) as a % of GDP and (ii) GDP growth. The impact on GDP – as distinct from the General Borrowing Requirement is much more severe in 2009-2010 than it was in the 1970s or 1980s 94% in 2009-10 compared to 2.5% in 1987-89. How can this be reconciled with the information provided in Figures A and B (which are in turn difficult to reconcile with each other)? Has there been a mistake somewhere? No explanation or elaboration is given in the QEC.

Caution is struck on Page 21 ‘There is, however, no universally accepted indicator or methodology for assessing fiscal stance.’
The ESRI work is based on modelling using the ESRI Hermes Model and long historical time series. It is just impossible to know exactly how the results were derived and what particular assumptions were used in running the simulation.

In a telling comment in the 2000 paper referred to here the following is provided in the summary:
Fiscal stance is a measure of the discretionary changes in budgetary policy, though there is no universal acceptance on its measurement.
The appropriate stance of budgetary policy needs to take account of a number of factors such as the state of the public finances, the stage of the economic cycle and the growth prospects for the economy reflecting its stage of development. These three intertwined considerations are crucial in interpreting what fiscal stance should be.

In the QEC, the ESRI authors acknowledge that ‘there are a number of difficulties in interpreting the structural budget balance as an indicator of fiscal stance’. These refer to the measure of ‘capacity output’ which, in turn, drives the estimate of ‘structural’ and ‘cyclical’ components of the overall fiscal deficit and – by deduction – the net impact on GDP. The QEC authors state that: ‘Such difficulties can be avoided by basing the measure of fiscal stance on the change in discretionary policy relative to the previous year’s budget.’
I suggest that:
The inner workings of the HERMES be made public and available to other researchers to explore different scenarios and possible model specifications;
A full set of working papers be provided on the ESRI website including the work underlying pages 21-26 of the Winter 2009 QEC.
More caution be adopted in regard to any big claims about the scale of fiscal stance and its estimated impact on GDP, unemployment and borrowing.

Lets say that forecasting in any domain of macroeconomics has taken a severe battering at home and abroad in the last decade.
In one very telling comments (footnote 16 on page 23) the ESRI authors state:
‘However, this is not a fully valid comparison since 2011 is also expected to include a
contractionary budget which is not included here.’
This follows their comment:
Figure C shows that between 1999 and 2002 there was a cumulative
giveaway equivalent to far more than the “spendthrift years” of the late 1970s. It also suggests that the current period of retrenchment is less severe in impact than in the period 1987-1989.
Which, as I have pointed out, does not match the story in Figure C.

Aside from the above, it is worth noting that, in Table B in the Summary, some stark figures for the likely drop in income in 2009 are provided. It shows, for example, that:
Income in Agriculture will drop by about 25% in 2009 compared to 2008 (on top of a fall of 11% in 2008 on the previous year) signalling a cumulative fall of 36% of very roughly 30% over two years in volume terms (adjusting for price falls). This is a mighty fall in income spread unevenly across a diverse sector.
GNP is estimated to have dropped by 14% in 2009 (or 10% in volume terms when the impact of price deflation is factored in)
A fall of around 9% in wage income in 2009 (and 6% in 2010) – signalling somewhat smaller per capita real falls when numbers in employment and price deflation are taken into account.
A projected (rather than forecasted) hike in interest rates including real mortgage rates in Ireland (page 8)
‘This means that we expect a modest increase in private consumption spending in the final quarter of 2009. However, we do expect a further fall in consumption in 2010, in response to the contractionary budget’ (p11)
‘We expect the General Government Deficit in 2010 to remain essentially unchanged in, at 11½ per cent’ (p12)
A continuing sharp fall in capital investment (down by almost a third in 2009 and forecasted to fall by one sixth in 2010).

Fiscal policy is fuelling the depression. The main difference between now and 1976 or 1988 is that the world is in a much bigger recession now than it was then - which only reinforces the impact of fiscal contraction.

No comments: