Monday, 22 March 2010

Economically damaging and fiscally irrelevant

Michael Burke & Michael Taft: It is often stated that to reduce the fiscal deficit we must cut public spending. There is an assumption, never substantiated, that cuts equal savings. However, the evidence shows otherwise: public spending cuts will not significantly reduce the fiscal deficit and, in some scenarios, may actually increase it.

In April 2009 the ESRI assessed the economic impact of various fiscal measures (tax increases, spending cuts) on a number of variables over a six year period: GDP/GNP growth, consumption, employment, output, wages, borrowing, etc. On the basis of their estimates we will assess the impact of the Government’s current spending cuts on growth and the deficit up to 2014.

Impact on GDP

Since the 2009 budget the Government has announced slightly more than €6 billion in current spending cuts. These include wage cuts, employment reductions, cuts in purchases of goods and services, and social transfer cuts.

Public sector wage cuts: the ESRI estimates that the first year impact of public sector wage cuts on GDP was – 0.335, or -0.2 percent That is, for every €1 billion reduction in public sector wages, the GDP falls by €335 million. This is primarily due to reduced consumer spending (falling by 0.8 percent or approximately €700 million) with a knock-on effect on employment (a loss of 0.1 percent, or approximately 2,000 jobs).

The ESRI also projected that the deflationary effect accelerates – so that by the fifth year the impact on GDP is approximately -0.774, or -0.4 percent of GDP causing further loss in consumer spending and employment.

Other current spending cuts: The ESRI only provides a simulation for employment cuts. Here, they estimate a first year impact of -1.179, or -0.7 percent: for every €1 billion reduction through job losses, the GDP declines by €1,179 million. The driving force behind this impact is the loss of employment – 1 percent, or approximately 19,000 jobs in the first year. There is only a slight easing of the deflationary effect through the years. In the fifth year, the impact is estimated to be approximately -1.165 or -0.6 percent of GDP.

The ESRI does not provide simulations for reduction in government consumption or social transfers. Therefore, we will use the employment reduction multiplier as a proxy. This can be justified on the following grounds:

The fiscal shock from increasing government non-wage consumption was projected by Lane and Benetrix to have a first-year multiplier of in excess of 2.0. While it cannot be assumed the opposite will hold (an equivalent negative multiplier from a cut in consumption), it does show the Irish economy is very sensitive to this type of shock (given that Lane-Benetrix was measuring over a period of 20 years, this impact is likely to be higher during a recession).

While there are no Irish measurements for social transfers, in the US extension of unemployment benefits and transfers to food stamp recipients had higher multipliers (1.64 and 1.73 respectively) than other forms of spending increases or tax reductions. This should not be surprising. The size of the multipliers is directly related to the ‘propensity to consume’; i.e. what proportion of each additional € in income is consumed and what is saved. Transfers to the poor or low-paid are likely to have a stimulative effect since they are obliged to consume a greater proportion of their incomes. Again, while not assuming the opposite is true here, it is expected that such cuts will provoke a significant negative shock among groups with a high propensity to consumer.

Therefore, we find the following (assuming, for the purposes of this exercise, that the cuts of €6 billion took place in 2009 and taking the Government’s growth projection as the base-line):

The first year impact will result in a GDP decline of €4.9 billion, or -3 percent of GDP. In the long-term, the decline accelerates to a decline of €6.4 billion or -3.1 of GDP. These are significantly deflationary.

Impact on Borrowing Requirement

Turning to the impact on the borrowing requirement (EBR), the ESRI simulations estimate that:

(a) A reduction of €1 billion through public sector wage cuts, results in a reduction of 0.3 percent in the EBR in the first year. Because of the acceleration of the deflationary impact, this falls to 0.2 percent in the fifth year.

(b) A reduction of €1 billion through public sector employment losses results in a reduction in the EBR of 0.2 percent in the first year. By the second year, this is reduced to a mere 0.1 percent and continues at this level through to the fifth year.

Therefore, we find the impact on the EBR to be:

The Government’s current spending cuts reduce the EBR by 1.4 percent in the first year; declining to a mere 0.8 percent by the fifth year.

It is important to put this in perspective. The Government intends to reduce the General Government Balance by 8.7 percent of GDP between 2009 and 2014 (a reduction of 7.3 percent in the Exchequer balance).

Yet, the ESRI estimates that current spending cuts will, after factoring in the deflationary impact on the economy, make only the smallest of contributions to that reduction.

However, the ESRI simulations themselves may seriously under-estimate the debilitating impact on the economy and, therefore, over-estimate the reduction in borrowing.

• First, the ESRI simulations are based on long-run average behaviour, which includes both booms and busts. Fiscal tightening in a recession, when there is already spare capacity, will have a greater depressing effect than the same during a boom.

• Second, the ESRI model may assume certain behaviour – ‘crowding in’ of private investment, lower bond yields, increased economic activity, emigration levels – which may not occur. Factors such as the continuing credit crunch, household deleveraging, structural deficits in our infrastructure and indigenous enterprise base may overwhelm such theoretical assumptions.

• Third, the ESRI measures impact in tranches of €1 billion. However, when these tranches are multiplied (e.g. the impact of the April Budget tax/levy increases was €2.8 billion) the cumulative impact may be higher.

• Fourth, the ESRI simulations analyse fiscal measures in isolation. When combinations of these measures are introduced the cumulative impact may be higher.

• Fifth, when access to credit is constrained, the depressing impact on activity arising from fiscal contraction is also likely to be amplified.

In other words, such are the downside risks to the ESRI estimates, that we may experience perverse results: that the fiscal deficit burden may actually rise as a result of public spending cuts.


The Department of Finance is correct: quantifying impacts ‘requires a combination of econometric model simulations and judgement’. Judgement and experience tells us that cutting spending during a recession (a) reduces economic activity and (b) reduces tax revenue and increases unemployment costs. That this occurs is beyond doubt; what we need to do is find the extent.

We have shown that ‘savings’ are minimal and the impact on GDP severe. And such is the fractional impact on borrowing there is a distinct downside possibility that such cuts will increase the deficit burden. This is reinforced when we note the deflationary impact on the domestic economy is even more severe. Whereas GDP will decline by -3.1 percent by 2014 as a result of the current spending cuts, GNP will decline by -3.8 percent. This is what the TASC letter referred to as ‘a low-growth, high debt future’.

Cuts do not equal savings. Cuts degrade economic activity with only a marginal impact on borrowing. The next time a commentator says ‘we’re borrowing €400 million a week’ as a justification for more spending cuts, they can easily be answered: cutting spending won’t affect that ‘€400 million a week’ and it may only make things worse.

To bring the deficit under control we need another alternative – one based on growth and not deflation.


Rory O'Farrell said...

Very interesting.

Regarding the 3.1% lower GDP in 2014, is this the cumulative effect or the annual effect (with cumulative effect being approximately 15%)?

Proposition Joe said...

Interesting that you mention the propensity to consume in the same breath as food stamps in the US. Of course such in-kind benefits would have spending rates close to 100% because they simply cannot be hoarded. Would this be an argument for conversion of a portion of our welfare benefits from cash to some form of token with limited exchangability? (e.g. for food, but not alcohol etc) There could be an exchange rate higher than unity, versus the old cash benefit, funded by the guaranteed multiplier effect and the societal benefits of a better-fed population.

Obviously some would object to the implicit paternalism, but it mightn't be a bad idea. I for one wouldn't mind swapping some of my jobseekers for food stamps of higher value, if these were anything like the US version, which can now apparently be redeemed for all manner of artisanal and organic fare:

Proposition Joe said...

The link to the hipster foodies was chopped for some reason.

Slí Eile said...

@Michaels Thanks for the work on this. This is an important contribution to a debate that has hardly happened yet. Using the best available research and statistics to argue the case against deflation and for fiscal stimulus makes sense. I would welcome more discussion on this including a throughtful response from Ireland's well known macro-economists. Just how much modelling has been done by the EU and the Dept of Finance to allow for the impacts of spending cuts? I would welcome alternative views from what whatever side. Let the debate continue. The stakes are high and we need a healthy contestation. But, let it be informed by good data and evidence. We have far too much sweeping assertions about public spending, TINA, etc.

Michael Taft said...

Rory - the effect is in the year, not cumulative. Therefore, in year 5 (or 2014 in our baseline), GDP will be 3.1% lower, barring the many downside caveats.

I'd just like to point out that the multiplier estimates - in particular, the employment reduction proxy - is consistent with the Government's own estimates of the deflationary impact of their budgets (in April, they admitted GDP would be reduced by 1% on foot of the budget measures, while the December budget would reduce tax revenue by €880 million).

Proposition Joe said...

One wonders whether the inaccuracies of ESRI forecasts during the boom should cause us to question the veracity of their modelling during the bust.

Or are downturns just easier to simulate? I kinda doubt it, as there's an equal but opposite irrationality at play ... over-exuberance during the boom, replaced by an over-correction to doom and gloom during the bust.

Martin O'Dea said...

Michaels, excellent contribution and preaching to the converted here - however, the 'Greece' arguement will certainly be used to counter. The point then remains; is it a case of just borrowing to avoid having no money or,is it borrowing because we feel we can use it productivly. As I tried to suggest much earlier on this site the arguement as to our target when we re-adjust our economy i.e. spend and make €61 billion; €31 billion; or €1 billion is an essential point.

There are many incl dept of finance who point out, correctly, that the where we were €61 billion approx was unsustainable and artificially propped up by excessive activity in property and being overborowed. Of course, how you manage your readjustment is important and is what you are suggesting is not being managed well and is indeed in the throes of a disorderly cycle that is taking us down much too far (lack of investment ((public in place of private)) loss of employment, decreased spending/increased welfare etc.)

So the question remains if we feel that there is nothing to replace this construction boom then the govt are doing the right thing but in the wrong way i.e. as they often proport if we don't reduce now we are only borrowing more that we will need to pay back in the future. The progressive in progressive-economy here I feel we need to lay out strongly and clearly.

I am a relatviely poor man - but compare me to top 5% of earners or wealth holders in 1810 and in terms of life-expectancy, health, travel, leisure time pursuits, gadgetry, comforts etc. etc. I am much better off. That is economics as far as I can see. It is just stuff. I have more stuff than him. The financial economy largely functions as an enabler to the generation, adaption and usage of that stuff.
So, if we are now enetering into a technological 'dark ages' or a time when the chance to become 'wealthier' did not exist to a great level - I would agree with the government - - but boy oh boy are we not?

One doesn't have to be an enthusiast to appreciate the absolute stockpile of innovation that is awaiting our remedying our financial sector and investing intelligently.

Economic Models from a century ago, or even a number of decades ago, are just not that accurate now because technology progresses in a different way. It may take you a long time to programme one computer to be able to perform voice recognition on some ancient unused language; however, how long does it take for another computer to learn from it?

By the way where is Mary Hanafin I haven't seen her on these boards in a while!

David O'Donnell said...

Good progress guys ... Keep it up. The structual deficit does need to be tackled - questions are how and who?

Race to the bottom and potential destruction also needs to be tackled - and more than the neo-liberal leaning 'standard neo-classical' modellists ...... Judgement and Experience - How pragmatic of you. More please ...