Tuesday, 6 September 2011

This time, it's different

Michael Burke: The ESRI’s latest research paper, ’Irish Government Debt and Implied Debt Dynamics: 2011-2015’ has received much coverage. Commentary has focused on the projected stabilisation of the overall debt in relation to the economy; the debt/GDP ratio or the debt/GNP ratio. From this, some more excitable commentary has suggested that this is a vindication of the broad-based attack on living standards, increased taxes on working people (but not on corporates) and public spending cuts euphemistically known as ‘austerity’ policies.

However, this last contention, that ‘austerity works’, is not supported by the authors’ findings themselves. To quote their own summary of their paper, ‘[Our projected debt ratios are] much lower than had been projected in official figures earlier in the year, partly because the cost of the bank recapitalisation was much lower than anticipated and also because of the reduction in EU interest rates’.

So, the lower projections are based on the lower level of bank recapitalisation and lower rates for bailout funds for those recapitalisations. According to the authors’ abstract of their own findings, none of the projected improvement is attributable to ‘austerity’ measures.

Quality of Forecasts

But how ‘good’ are these forecasts, in both senses? First, what is the quantum of improvement that is being forecast by the ESRI? Secondly, how likely is it that these forecasts will prove correct? The debt ratios forecast by the ESRI are set out in the table below:

Debt Ratios, % GDP

Source: ESRI, Eurostat, IMF databank, Euro Area Spring Forecasts

The first point to note is that the ESRI is not projecting any reduction in the level of government debt over the period 2010 to 2015. Over that period, debt will rise from 94.9% of GDP to a projected 106.2%. It is in effect projecting a deterioration in the debt level to next year, then stabilisation and then a reduction. Both the IMF and Eurostat forecasts are much worse.

This disparity is a function of two factors. First, the Eurostat and IMF forecasts were made before the reduction in interest rates and before the lower projections for the level of the bank bailout were made. Secondly, the ESRI has stronger real growth projections than either Eurostat or the IMF.

Before dealing with the substantial point of how large the bailout and interest rate savings are, it is worth highlighting the main source of the discrepancy in growth forecasts. In effect, ESRI has a much larger growth forecast of +1.8% real GDP in 2010 than either Eurostat (+0.6%) or the IMF (+0.5%). The difference arises because whereas the IMF and Eurostat both have prices rising by 0.6% in 2010 to reduce real GDP by that degree, the ESRI projects falling prices of 1.1% (implied from the gap between real and nominal GDP (Table 7). Given that deflation both reduces the nominal level of taxation revenues while also increasing the ratio of existing debt to nominal GDP, there can be little argument that this ‘stronger’ growth forecast is responsible for the ESRI’s more optimistic debt/GDP forecasts.

Instead, it is the combination of lower interest rates and a lower projected bank bailout cost which is responsible for the projections of a substantially improved debt outlook. On the former, the consensus appears to be that the annual saving will be in the order of €1bn per annum, perhaps slightly more. Implicitly the ESRI authors assume a saving of €1.125bn per annum, on the basis of a former interest rate of 6% (p.20, point 2.). Compounded, this saving over a 5 year period amounts to €8.5bn or approximately 4.6% of the GDP level projected for 2015.

On the lower bank bailout costs, the international bailout of creditors to Irish banks in November 2010 included €10bn of immediate bank recapitalisation plus another contingency amount of €20bn. To date, of this a total of €17bn has been provided by the State (banks funding €7bn themselves in the financial markets for a total of €24bn). The ESRI authors expect €3bn to be repaid to the State by 2014. This ‘saving’ of €13bn also incurs interest. However, the authors now argue that funding from the Troika will be needed in 2014, even though the terms of the original bailout were that the government would return to the financial markets in 2013. Therefore, there will be no net interest saving, based on the authors’ projections. Instead, there will be an additional cost of approximately €0.5bn (based on the 3.5% interest rate, rather than a projected interest rate of 6% in the financial market borrowings that are also assumed in the ESRI paper). As a result, the projected saving from a less onerous bank bailout is a net €12.5bn.

Taken together the actual interest rate saving of €8.5bn and projected saving on the bank bailout of €12.5bn combine for a total €21bn. This is equivalent to 11.4% of projected 2015 GDP.

Without these actual and projected windfalls, the ESRI forecast would otherwise have been 117.6% of GDP in 2015. This compares to a debt/GDP ratio of 94.9% in 2010.

Debt Dynamics

The idea that there will be no more bank bailouts has firmly taken hold and is largely responsible for the specific rally in Irish government debt in recent weeks. This is despite the fact that the EBA’s stress tests were widely discredited by the failure of two small Spanish banks shortly after publication, with total losses exceeding the EBA’s estimate of EU-wide recapitalisation requirements.

It may be the case that further losses do not require further recapitalisations. But the key exposure of the Irish banks is to the domestic economy, which continues to deteriorate on all forecasts, including those of the ESRI. This has an impact on the banking sector. Currently, this is most evident in the rise in the rate of mortgage defaults.

This highlights a key misconception regarding the relationship between the banking sector, government finances and the real economy. It is assumed that, if the banking sector is stabilised to the extent that it requires no further taxpayer funds, this will restore government finances to health, as long as public spending is reduced towards the level of taxation revenues (sufficient to provide a ‘primary surplus’, that is before interest payments are included). It is argued that, if all three occur, bank stabilisation, no more bailouts, a swing toward a primary surplus, then the crisis ‘will be over in 3 years’.

This is the premise that underlies a section of the ESRI paper dealing with debt dynamics. It is not denied that significant remedial action was required to resolve the crisis in the banking sector, even while many argue that a bailout of all their creditors was one of the least effective means of doing so. But, ever-greater contraction of the domestic economy can only be fatal to any ambitions to remove the banking sector from the life support it has been given by taxpayers. In effect, the ESRI and many others look through the world through the wrong end of the telescope. Neither banks’ balance sheets nor government finances can be restored to health until and unless there is an economic recovery.

The authors argue that, absent any further negative ‘shocks’, the debt/GDP ratio will begin to fall from 2013 onwards. The horizon for an imminent improvement never seems to alter. It is always 18 months hence. But the Irish economy received no external shocks in the way that economists use the term. The recession began here nearly a year before it began in the world economy, the slump in investment also a year earlier (which preceded the recession in both cases).

Three years ago in the Autumn 2008 Quarterly Economic Commentary the ESRI was forecasting a general government debt of 47.5% of GDP in 2009 and was fully supportive of government efforts to cut the deficit, in particular urging cuts in public sector pay. But this contractionary fiscal policy was a shock to the economy and the effect was slower growth, rising unemployment and falling tax revenues.

In the event, the debt/GDP ratio was 65.6% of GDP, not 47.5% in 2009, even while the government implemented ‘austerity’ measures equivalent to 9.1% of GDP. In effect the ESRI was forecasting a near-term deterioration in the debt level followed by stabilisation and then reduction, based on ‘austerity’. In fact a trawl through the QECs since 2008 shows that this debt profile is what the ESRI has been forecasting since 2008.

The authors clearly haven’t asked themselves the key question, so we must: Why will it be different this time?


Paul Hunt said...

An impressive report which you may have spotted:

Particularly good on the stupidity of (1) focusing structural reforms on labour markets (and not on the real structural problems) to suppress real wage growth below productivity growth and contribute to continued secular decline in labour share of national income, (2) failing to restructure banks and to re-regulate the international financial system effectively, (3) being guided by the views of financial market participants on economic policy, (4) relying on the private sector to replace reduced public investment without providing a long term contractual basis for such investment and (5) regressive taxation policies.

Seamus Coffey said...

Hi Michael,

It is clear there are differences between the IMF's debt projections and those produced by the ESRI but I don't think we can ascribe them to the reasons you provide.

It would be great if the interest rate reduction was going to reduce our interest bill by €8.5bn between now at the end of 2015. This is not going to happen.

First, the rate reduction wasn't agreed until July 21 of this year so it cannot be applied fully for 2011. Initially the rate reduction will apply only to money borrowed from the EFSF (and some bilateral loans). It is expected that a lower rate will also apply to money borrowed from the EFSM but this will only be discussed later this month.

Second, the €1.1bn annual saving only arises when the full amount has actually been drawn down. So far we taken €3.6bn from the EFSF and €11.3bn from the EFSM. This will limit the projected interest reductions for 2011, 2012 and 2013 until the full €45bn has actually been accessed.

The full €1bn of interest reduction is possible in 2014 and 2015 but for 2011-2013 it is likely to be no more than €1.5bn in total. This gives a total interest reduction of €3.5bn rather than €8.5bn, and even that is likely to be at the upper end of the range.

There are also some issues with the suggested savings from the bank recapitalisation. The IMF 2015 projection is that the government debt will be €216.6bn. This is based on a projected bank recapitalisation cost of €19bn in 2011. The final cost to the State has ended up being €17bn.

There is only a difference of €2bn between the IMF's bank recap figure and the actual figure used by the ESRI.

Unlike the IMF, the ESRI have predicted that the State will get back €3bn of the "contingent capital" provided to the banks this year.

All told the reduced interest rates and the lower bank recapitalisation costs will reduce the debt by no more than €8.5bn rather than the suggested €21bn given above.

How then do we explain the €23bn difference between the IMF (€217bn) and the ESRI (€194bn)?
We have explained €9bn with the lower interest rates and bank recapitalisation costs so there is still €14bn to go.

Once we account for the reduced interest costs there is not much difference between the ESRI's and the IMF's projections of the annual deficits up to 2014 and there is around €3bn of a difference between them across the four years.

However, 2015 is somewhat different. The ESRI project a deficit of €3.2bn. The IMF project a deficit of €8.1bn. Although their projections are closely aligned for 2011 to 2014 they diverge greatly in 2015.
These differences in the deficits account for around €7bn of the difference between the two, with most of this largely unexplained in 2015.

So we still have around €7bn to reconcile between the two figures. Thankfully this is relatively easy to do.

At the end of 2010 the government had €16bn of cash in various accounts. The IMF assume that this remains intact. The ESRI recommend that €7bn of the cash be drawn down to fund expenditure rather than look to new borrowings. The IMF leave us with €16bn of cash at end-2015; the ESRI leave us €9bn.

So there you have it. Easy as that!

There is a €23bn gap between the ESRI and the IMF when it comes to our 2015 debt.
- €9bn is due to the recent developments in interest rates and bank costs.
- €7bn is due to differences in the annual deficits over the five years (with most of that in 2015)
- €7bn is due to the fact that the ESRI assume we spend €7bn of our cash reserves.

Once the IMF account for the recent changes, their debt projections will be very close to those of the ESRI, and if they revise the deficit for 2015 they will be almost identical!

Apologies for the length but we have made a right mess of handling our public finances which makes everything much more complicated than it should be.

Michael Burke said...

Hi Seamus,

no dispute on the mess in government finances. Aside from policy choices, a contributory factor is the lack of transparency in official policy documents.

Unfortunately, ESRI tends to reproduce this. So, as far as I am aware it hasn't made any projections of its own for the deficits out to 2015. As stated in the Kearney/Fitzgerald document (p.14) the projections of stabilisation are actually based on the April 2011 Stability Programe Update. That is, they are government/Troika projections.

No apology needed for the length of contribution. But I think it misses the point. The purpose of my piece was to highlight the source(s) of the projected stabilisation of the debt level, which, as I point out is not actually an improvement at all, just less disastrous than previously forecast. These are the sources of the authors' projecte improvement, not mine.

The authors do not, unlike your analysis, attribute any of the improvement to lower deficits, if by that you mean the primary deficit. Perhaps this is because they haven't made any indedependent forecasts of their own.

Instead, they explicitly state that that they are using SPU projections (p.14) and that these are for the primary balance (before interest payments) and are unchanged as between the SPU and their own 'debt dynamics', set out in Tables 6 & 8 respectively.

The fact that these primary balances are unchanged between the SPU and the authors' projections is also set out, in terms (p.21, 4.2, 2nd para.),

"In the top panel [Table 8, 'Debt Dynamics] the primary blance is as set out in the Department of Finance's April estimates (see Table 6)".

It is also clear from the authors' own abstract that the entire projected stabilsation is due to lower interest payments and a projected reduction in the cost of the bank recapitalisation.

The debt stock is also reduced by drawing down assets. But this is just gaming the silly rules of the EU/IMF who insist on accounting only liabilities in gross general government debt and not taking account of assets in a measure of net debt.

We are are repeatedly told that there is no money for stimulus, 'we are broke'. Yet there were €19.9bn in liquid assets in 2010 and will still be €8.6bn even under the authors'. Any adviser who couldn't come up with a steream of projects yielding mre than 3% should be sacked. And, as this is about market confidence rather than Maastricht limits, ond markets understand the NPV of cashfows, even if the EU doesn't.

Seamus Coffey said...

Hi Michael,

The ESRI does have "stronger real growth projections than the IMF", but not by much.

2011 0.6% 1.8%
2012 1.9% 2.3%
2013 2.4% 3.0%
2014 2.9% 3.0%
2015 3.3% 3.0%

The IMF are projecting the real GDP in 2015 will be 111.6% larger than the actual 2010 level. The ESRI forecast that it will be 113.8% larger.

The IMF used their growth forecasts back in May (before the interest rate and bank recap changes) and projected that the debt would begin to "stabilise" in 2013, with a peak of 120% of GDP before dropping to 117% of GDP by 2016.

The IMF do not have the ESRI's growth forecasts and have not included the recent changes and they also have the debt stabilising.

It is not new that the ESRI are forecasting that the debt will stabilise. The stabilisation has very little to do with the interest rate and bank recapitalisation changes. The debt was forecast to stabilise even before they came to pass.

All the ESRI have forecast is that the stabilisation will come at a lower level. I think a lot of this is artificial as it is based on drawing down €7 billion of our cash resources and the assumption that the banks will be in a position to return €3 billion of "contingent capital" in 2014. Neither of these are an improvement.

The only improvement relative to the IMF forecast is the €3 billion reduction in our EU interest bill and the fact that the bank recapitalisation bill for 2011 was €2 billion lower than expected. This €5 billion brings merely brings down the projected debt level (but not by much!).

You ascribe the ESRI's debt stabilisation projection to stronger growth forecasts and the recent changes in interest rates and bank recapitalisation costs. I just think that this is not a complete assessment of the ESRI's projections as the effect of them is smaller than you indicate.

All the ESRI have done is taken the forecast debt stabilisation of the DoF/EU/IMF and revised it down to a lower level based on the developments over the summer.

Michael Burke said...

Hi Seamus

we seem to be violently agreeing.

"All the ESRI have done is taken the forecast debt stabilisation of the DoF/EU/IMF and revised it down to a lower level based on the developments over the summer." Absolutely correct.

The only things I would stress is that these are all merely forecasts or in the ESRI's case, forecasts based on forecasts. Oh, and maybe put stabilisation in inverted commas.

Seamus Coffey said...

I hope we are agreeing! I just feel that the ESRI can't be accused of "some outstanding optimism" as Paul Sweeney put it. I think that are showing the same optimism as every other body so it is not that outstanding at all.