The inexact science of calibrating fiscal policy

In the showdown between France and the European Commission last week, France clearly is the winner on points, which is not surprising given the impossibility of the task the Commission had set it in meeting the Excessive Deficit Procedure (EDP) rules and the danger to the latter if France was to openly defy it. We have a sort of stand-off between the surrender monkeys – France is going along with the rules sort of and the Commission is bending the rules to save face. It is 2003 all over again. The public might actually think this EDP process is based on a fairly definite science with respect to measuring fiscal policy positions which provide unambiguous statements of deficits. The public would be very wrong if they did adopt that conclusion. In general, the applied work associated with informing the EDP process is very inexact. But, moreover, it is ideologically tainted which makes the process very damaging for any notion of prosperity. All applied work has measurement and other technical issues, which means it is always just an approximation. But when those errors are overlaid by a systematic bias against government net spending and therefore full employment, then the exercise becomes a scandal.

To get started, the so-called – New Pact for Europe project released its second report – Towards a New Pact for Europe – last month (November 13, 2014).

On Page 27, you learn why the whole exercise will fail:

Measures to strengthen the social dimension of EMU by identifying and agreeing on binding common social convergence benchmarks/criteria matching the quality of the convergence criteria laid down in the Stability and Growth Pact or the fiscal criteria included in the Stability Treaty. Introducing such benchmarks within the framework of a social compact (if necessary via an intergovernmental treaty/agreement) would guarantee mini- mum social standards and guide EU spending priorities. To make the social dimension an integral part of economic and fiscal governance, these criteria could also be used as the basis for mechanisms to trigger preventive and corrective actions if social indicator targets are not met (‘excessive social imbalances’), similar to the excessive deficit procedure in the Stability and Growth Pact.

Are they kidding? The convergence criteria laid down in the Stability and Growth Pact were farcical, biased to stagnation, caused unemployment to be unnecessarily high, and were violated by most (if not all) the nations in the lead up to Stage III of the monetary union.

Many of the nations tricked the Commission with various statistical dodges. The Commission turned a blind eye such was the political necessity for all nations to be allowed to enter the union.

Italy should never have been allowed in. Belgium categorically failed to satisfy the criteria. Greece a little later hired Goldman Sachs to defraud the Commission with some tricky derivatives deal Germany, itself, did not make the criteria and in 2003, the SGP was changed because France and Germany still couldn’t come within the rules.

Which brings me to today’s topic.

The EU Observer article (November 28, 2014) – France gets three months to tweak budget – said that:

EU economics commissioners … defended “political” reasons for giving France more time to fix its budget deficit …

The EDP is full of ‘let outs’ for the Commission if they choose to use them, which means they could declare the whole of the Eurozone as an ‘exceptional’ situation and allow fiscal deficits to go to whatever level was required to support large job creation programs and infrastructucture investments without violating the current Treaty.

The ECB could then announce they were beginning to buy the debt of any Member State government and the bond yields would remain low and the bond markets would be powerless.

The Eurozone crisis would be resolved very quickly then. That is how easy it would be if the political elites and their bureaucrat economists could let go off their ideological attraction to austerity and escape the neo-liberal Groupthink that has caused millions to lose their jobs and poverty rates to rise.

There is no shortage of idle real resources in Europe. They are idle as a result of some arbitrary and ideologically motivated ‘fiscal rules’ that the elites insist on maintaining.

Well, maintaining at least until a powerful nation such as France consistently violates them and has no hope in the foreseeable future of operating within them.

Then the slippage comes and the latest justification is couched as “political” reasons – meaning that the Commission is too scared to take France on at this stage, given the instability in French politics and the impossibility of meeting the rules anyway in the current economic climate.

The Commission also went limp on Italy and Belgium – extending thier time lines for adherence to the rules. Italy will still not meet the revised deadlines – it is mired in recession and things are getting worse.

The problem is that in playing this on-going game with the EDP, austerity is still imposed and growth damaged. It would be better, as noted above, if the slippage, was growth oriented.

Consistent with form, and so as not to disappoint us, the inevitable “accounting tricks” enter centre stage as the next way out for the Commission.

The media is reporting that:

… an accounting trick might get France off the hook.

The trick relates to how the “structural deficit” is calculated – either “bottom up” or “top-down”.

The calculation will form the core of the assessment in March on whether France is complying with its obligations.

Accordingly, the Commission wants France to carve around 1 per cent of its structural deficit to GDP ratio by the end of 2015 to bring it in line with the 3 per cent rule under the SGP.

Using so-called ‘bottom-up’ measurement methodology, France will only achieve a cut of 0.3 per cent. But if a ‘top-down’ approach is taken, then France will achieve a cut of 0.9 per cent on current estimates. QED. EDP ends and the Commission economists and officials retire to have yet another long lunch with plenty of wine.

The unemployment rate worsens or stays around the same. Europe continues in its absurd approach to socio-economic organisation.

So apropos of yesterday’s blog – The Australian economics media guilty of false reporting – which discussed structural deficits, I thought it might be useful to explain what these methodological ruses are all about and to demonstrate how flaky they can be.

If you ever thought all these fiscal rules and the supporting databases to implement them were in anyway scientific or rigourous, think again. Arbitrary is a euphemistic description for them!

So we will keep this as simple as possible.

For the technically-minded, this Chapter from a European Commission publication – Part III Measuring the fiscal effort – will be useful.

Other background material can be found in these blogs – Structural deficits – the great con job! and Structural deficits and automatic stabilisers.

Recalling yesterday’s blog (which summarised the earlier two blogs cited above), the conceptual difference between the cyclical and the structural components of the published fiscal balance is as follows:

• The published fiscal balance is in deficit if the government spends more than it takes in via revenue (acknowledging that ‘revenue’ may be a tainted term in the case of a currency-issuing government that faces no funding constraints). A fiscal surplus occurs in the opposite case.
• A deficit means the government sector is contributing to total spending in net terms and assisting economic growth whereas a surplus means the opposite. Changes in the deficit/surplus indicate the direction of the impact. So if the deficit is falling, the net public spending is lower and that might undermine the current growth rate depending on the response of the non-government spending.
• However, the total tax intake and welfare spending for given policy settings (rates etc) are sensitive to how strong the economic cycle is. When the real GDP and employment growth are strong, the tax intake is higher and the welfare payments lower than otherwise.
• Thus, a component of the published fiscal balance is sensitive to the economic cycle and the outcome thus varies without any discretionary change in government policy settings.
• It is thus difficult to know whether the government is choosing to stimulate or contract the economy with its fiscal policy settings by just seeing consulting the stated fiscal data. A rising deficit might indicate discretionary government stimulus but then it might indicate declining private spending and an ailing overall growth rate. Or a mixture of both!
• To provide a clearer picture of the intentions of fiscal policy, economists devised a way to net out the cyclical effect (called the automatic stabiliser component) from the discretionary component of the public fiscal balance. The discretionary effect, that component which reflects policy choices, is called the structural deficit or cyclically-adjusted fiscal balance.
• The structural balance used to be called the Full Employment Balance, because it was measured by asking what would the fiscal balance be if there was no cyclical impact on the published figure? In other words, given the current tax and spending decisions in place, what would the fiscal balance be at full employment when the cyclical component is by definition zero?

So that is what a structural balance is. The top-down approach, which is the long-standing way of calculating the balance involves a series of steps.

First, the full employment level of activity – or potential capacity – has to be estimated as the benchmark, against which variations in the cycle are judged. At potential capacity, the economy cycle is considered to be at its peak and neutral to the fiscal balance.

Deviations from the peak thus invoke non-neutral cyclical impacts on the fiscal outcome.

The calculation of the structural deficit spawned a bit of an industry in the past with lots of complex issues relating to adjustments for inflation, terms of trade effects, changes in interest rates and more.

Much of the debate centred on how to compute the unobserved full employment point in the economy. There were a plethora of methods used in the period of true full employment in the 1960s. All of them had issues but like all empirical work – it was a dirty science – relying on assumptions and simplifications. But that is the nature of the applied economist’s life.

In the 1970s, a new ideological component entered these calculations. At the time that governments abandoned their commitment to full employment (as unemployment rise), the concept of the Non-Accelerating Inflation Rate of Unemployment (the NAIRU) entered the debate – see my blogs – The dreaded NAIRU is still about and Redefing full employment … again!.

The NAIRU became a central plank in the front-line attack on the use of discretionary fiscal policy by governments. It was argued, erroneously, that full employment did not mean the state where there were enough jobs to satisfy the preferences of the available workforce. Instead full employment occurred when the unemployment rate was at the level where inflation was stable.

NAIRU theorists then invented a number of spurious reasons (all empirically unsound) to justify steadily ratcheting the estimate of this (unobservable) inflation-stable unemployment rate upwards. So in the late 1980s, economists were claiming it was around 8 per cent. Now they claim it is around 5 per cent. In Spain the Commission estimates recently have seen the NAIRU rise from below 10 per cent to 23 or 24 per cent. A similar (ridiculous) scenario is estimated for Greece.

The NAIRU has been severely discredited as an operational concept but it still exerts a very powerful influence on the policy debate.

It is widely used to define full capacity utilisation. An economy that is operating at the so-called NAIRU is said to be operating at full capacity.

Of-course, the economists who use these concepts kept changing their estimates of the NAIRU which were in turn accompanied by huge standard errors. These error bands in the estimates meant their calculated NAIRUs might vary between 3 and 13 per cent in some studies which made the concept useless for policy purposes.

The reason this concept is used is not because it is statistically sound but because it carries the ideological weight justifying the neo-liberal attack on government intervention.

The statistical models are rather spurious in fact and because they inflate the level of unemployment that would be consistent with full employment, they, in turn, severely underestimates the tax revenue and overestimates the spending that would be consistent with full employment.

In other words, they allow policy makers to conclude that the structural balance is more in deficit (less in surplus) than it actually is.

They thus systematically understate the degree of discretionary contraction coming from fiscal policy.

That is the first major issye with ‘top-down’ techniques.

Second, once the benchmark full employment level is agreed, the statisticians estimate how sensitive tax collections and welfare spending are to variations in real GDP from the full capacity level.

Economists estimate how much tax revenue will fall/rise for every percentage variation in real GDP (down/up). They call that response an elasticity.

If for example, the current tax intake equals 100 and the intake varies by 0.5 per cent every 1 per cent variation in real GDP (as a result of variations in income).

Then if real GDP growth consistent with full employment is 3 per cent. The fiscal balance calculated using the tax intake and the welfare payments forthcoming at level of activity would be the structural balance.

What happens if real GDP growth falls to 2 per cent? We would expect the current tax intake to fall (to 99.5) even if the current tax policies (rates, structure, concessional allowances etc) are unchanged. The decline has nothing to do with a change in policy – it is all down to the sensitivity of the intake to real GDP growth.

If nothing else changed on the spending side, then the fiscal balance will change by 0.5 towards deficit or into a larger deficit. We would typically express that change as a percent of nominal GDP which means the 0.5 would be a different number.

A similar exercise is undertake on the spending side to work out how much welfare payments rise for each percentage fall in real GDP growth (and vice versa).

The sum of the tax declines and the spending increases driven by the deviations from full employment growth rates are the measure of the cyclical component of the fiscal balance.

So if the government starts with a small (responsible) fiscal deficit at a time of full employment and the economy goes into recession, then the deficit will rise, even if the government changes no policy parameter.

This methodology is the ‘top-down’ approach. The European Commission says that this approach produces the “Cyclically-adjusted Balance (CAB)” which:

… consists of removing from headline balances the impact of the business cycle, based on standard methodologies … [it] … is easily interpreted as the balance that would prevail if GDP was at potential … Moreover, the change in the CAB measures the fiscal stance, i.e. the change in the fiscal balance that is not driven by the automatic reaction of the balance to the business cycle. This provides a gauge of the non-automatic impulse from the fiscal balance on the economy. An increase in the cyclically adjusted deficit provides an expansive impulse on the economy.

There are several problems with this approach which are in addition to the ideological bias used to estimate full employment. That bias is unsurprisingly not acknowledged by the European Commission.

What measured changes in the structural balance or in Commission-speak the CAB try to depict is the change in the fiscal position “due to clearly identified government actions”.

But there are some factors which will change the fiscal balance outside of the cyclical impact, which are not necessarily attributable to discretionary government policy changes. The factors influence the accuracy of the estimates of the responsiveness of taxation revenue and welfare spending to real GDP growth fluctuations.

For example:

The best-known factor is the presence of windfall/shortfall in revenues or unemployment expenditures, loosely correlated with the evolution of GDP but not taken into account in the cyclical correction because of the decupling between the evolution of the tax base and GDP. Fluctuations in asset or housing markets, are known to generate non-permanent but long-lasting shifts in revenues that are not captured by the CAB …

So a nation might have a real estate boom which drives stamp duties and property tax intakes through the roof. The property bubble is an abnormal event but could allow the government to run surpluses or increase spending and run small deficits.

If economists then estimate the responsiveness of tax revenue to real GDP growth without discounting the ephemeral nature of the property bubble and its impacts on tax takes, it is likely they will overestimate the cyclical sensitivity.

So then if real GDP falls, the cyclical component of the fiscal balance will be overestimated, which means the structural component will underestimated.

The conceptual problem is that the increase in spending may be structural whereas the tax surge is ephemeral. The actual ‘structural’ position of the fiscal balance will indicate a short-run discretionary surplus using the top-down approach, whereas, in reality, given the transient nature of the tax boost, the structural balance may be in signficant deficit.

Then, according to the Commission:

… to improve the structural balance the government will have to put in place new measures large enough to more than offset underlying negative trend.

The alternative might be where the tax responsiveness starts to decline relative to the level of activity but the estimates retain the prior measures of responsiveness. In that case, even if the economy remains at potential activity levels, the true structural deficit estimates will steadily rise and “signal a loosening in fiscal policy, despite no action having been taken in this sense by the government”.

Then the net spending cuts will have to be larger to achieve some smaller structural deficit.

To overcome these problems, economists have come up with an alternative way to estimate the so-called ‘fiscal effort’ which has been called the “‘narrative approach’ or the ‘bottom-up approach'”.

In brief:

Fiscal effort is measured as the sum of the value that government authorities have attributed to the measures in their budget at the time of adoption. Consolidation periods are then defined as periods in which the fiscal effort is above a given threshold.

So this avoids having to calibrate a full employment level. It just builds incrementally on the announced policy shifts and the treasury estimates of the impact on the fiscal balance.

The benchmark in the ‘bottom-up’ approach is called the “unchanged policy” scenario – “what would have happened in absence of government intervention”.

You will immediately see the shortcomings of that approach. It requires forward-looking estimates of the counter-factual – and the official estimates derived from OECD, IMF, EC models are typically disastrous – biased to predicting too much growth when there is austerity cuts and too little when there is fiscal stimulus.

Identifying the underlying revenue responsiveness (that is, netting out ephemeral and other revenue impacts) evades accurate estimation anyway. This is a problem common to both methodologies.

Further, trying to distinguish changes in spending which are controllable by government and those which are triggered by the strength of non-government spending is also difficult.

The other major problem with the ‘bottom-up’ approach is that:

… the methodologies underlying the quantification of the measures are neither transparent nor replicable, differ across countries and in time within each country, are influenced by the cyclical position of the country and can be affected by the scope and the aim of the assessment and by political decisions of the governments.

In other words, they do not constitute a very sound basis for cross-country comparisons, which is an essential requirement of the EDP within the Eurozone.

There is no certainty using the ‘bottom-up’ technique that a given fiscal balance in in France would signify a similar fiscal effort to the same balance in Italy.

For these reasons, economists have come up with hybrid approaches, which I will leave for another time.

The point is that the French government is currently claiming via the Commission that its projected fiscal position by the end of 2015 is likely to deliver close to the adjustment required by the Commission relative to if it didn’t change policy at all. That is the ‘bottoms-up’ assessment.

From a ‘top-down’ assessment, the level of adjustment is much lower than required because the deviation from potential output remains huge.

All of this presumes that the decompositions of the fiscal balance matter. Do they?

While Modern Monetary Theory (MMT) considers the fiscal balance to be a relatively meaningless artefact in isolation from the real state of the economy – that is, it should never be a policy target in itself, one should not get the impression that the distinction between the cyclical and structural components of the fiscal balance is irrelevant for those who consider MMT to be applicable.

Why? Policy interventions have to be consistent with the aim of reducing idle real resources without promoting accelerating inflation.

In that sense, if a government is facing a recessed economy and desires to stimulate total spending with a discretionary net spending increase, then it has to be able to, more or less, work out how much of a spending injection will be required to generate the required change in total spending.

In that sense, a fairly reliable set of estimates of the impact of spending and tax changes on total spending and the impact of a change in net public spending on total spending (the so-called expenditure multuplier) are required.

Please read my blog – Spending multipliers – for more discussion on the expenditure multiplier.

The governments seeks to know how much the initial net injection should be given that the fiscal balance will move towards surplus when the recessed economy resumes full employment.

So estimates are required and the discussion about methodology essential for a coherent interventionist strategy. This is why the tools of econometrics and mathematical statistics are important to acquire if one wants to be a professional economist.

Many progressive economists eschew the study and use of econometrics, claiming issues regarding formalism, realism etc. But then they are left floundering when they simultaneously advocate an interventionist approach.

They cannot tell you how much intervention should be introduced! Their voices are thus hollow in the policy debate.

Conclusion

When millions of jobs are at stake you might consider these methodologies were more accurate and sound than they are.

The worst part of the way in which organisations such as the European Commission operate however are not related to the difficulties of estimate tax and spending elasticities.

Rather it is the ideologically constructed measures of full employment/capacity that allows them to then justify huge structural shifts in government net spending when, in fact, the effects are largely cyclical, and due to a massively recessed economy.

The recent cases of Spain and Greece, where unexplained but more than double-digit shifts over short time horizons (two to three years) in the estimates of full employment unemploymenr rates have been provided by the Commission and then used to justify harsh cuts in public net spending.

At that point, you know the economists have stepped across the line of empirical difficulty, which bugs all applied work in social sciences, and entered, instead, the world of paid gunslinger the neo-liberal ideologues who occupy the elite positions in the policy making structures.

That is enough for today!

(c) Copyright 2014 William Mitchell. All Rights Reserved.