I am taking a brief rest from the Eurozone crisis – which will probably blow up again in the coming weeks as the Spanish austerity drives the bond markets in the opposite direction than was intended by the Troika – and the latter call for even harsher cuts to unemployment benefits at the same time as their austerity policies force unemployment to continue its inexorable rise upwards. Today, I have been reading the latest OECD Report (April 12, 2012) which is attracting attention – Fiscal Consolidation: How much, how fast and by what means?, which is part of their Economic Outlook series. It is really a disgraceful piece of work but will give succour to those politicians who are intent of vandalising their economies and making the disadvantaged pay more and more for the folly of the elites. It is an amazing situation at present. I am also reading a book – Pity the Billionaire – which I will review in the coming week. It examines how it is that the the popular response to the crisis which was caused by an excess of “free markets” is to attack government regulation and intervention and demand even freer markets. The OECD are part of the battery of institutions that fuel this crazy right-wing conservative response (the “unlikely comeback” in Pity the Billionaire terms) to the crisis through their highly tainted publications.
The OECD is still trying to pretend it has something to offer the public debate. Like the IMF, the OECD has moved well beyond its original charter (which was as a progressive force on the World stage), and is now a major part of the problem.
Their latest report is 5.9 mbs of nonsense.
Normally astute UK Guardian journalist – Larry Elliot – wrote (April 12, 2012) that the – OECD report will please George Osborne. The article went like this:
If you want to understand why George Osborne is so keen on austerity, you could do worse than dip into a report released on Thursday by the Organisation for Economic Co-operation and Development, which details how much pain countries in the rich west are going to have to endure in order to knock their public finances into shape. Plenty, is the short answer … Osborne will like this study, since it provides intellectual cover for his deficit-reduction programme.
Larry Elliot provides three qualifications to the OECD analysis, which I will discuss in more detail soon:
1. Debt levels are not historically high in many nations – they were much higher at the end of the Second World War.
2. The 50 per cent debt ceiling set by the OECD as being prudent is “a bit arbitrary”.
3. “it may be that we are getting things back to front” because “low growth leads to high debt levels, rather than high debt levels leading to lower growth”.
But then, in closing he gives the game away by saying that:
Osborne is right to be worried about debt … This, though, is not really about ends but about means – how much austerity the economy can bear before it kills off the growth that has to be a big component of any deficit reduction plan, and the right mix between fiscal and monetary policy. At present, monetary policy is doing all the heavy lifting, but quantitative easing contains risks of its own, not least the serious repercussions if financial markets suspected for a moment that central banks were secretly seeking to “monetise” (inflate away) the debt.
At which point we are back into a neo-liberal framework which has very little predictive capacity. The central banks could quantitatively ease all the outstanding public debt if they chose at present and there would be no inflation.
And if banks did suddenly find enough credit-worthy customers and dramatically increased their loans, and those customers suddenly withdrew all the credit created and spent up big on goods and services then we would observe three things immediately: (a) real GDP growth would rise – in the World and in the nation undergoing to credit expansion; (b) employment would rise and unemployment would fall; and (c) budget deficits would decline (and public debt issuance under current arrangements would decline).
It would be highly unlikely that inflation would be a concern given the amount of spare capacity that is around at present.
The overall problem I have with articles such as that above is that they are totally uncritical of the initial premises. He takes as given (other than saying things are a “bit arbitrary”) the key propositions that the OECD base its conclusions on. Closer examination shows that the analysis is more than a bit arbitrary. On key propositions it is totally unreliable and should not be used as the basis of policy analysis.
Just using OECD-style terminology – such as “how much pain countries in the rich west are going to have to endure in order to knock their public finances into shape” – is loaded
Upon what basis does one conclude that the rich western nations have public finances that are out of shape? What is the benchmark? Do we conclude that because there has been a serious collapse in real GDP growth in recent years and tax revenue has shrunk that the cyclical impact of the budget outcome means the latter is “out of shape”?
The more reasonable conclusion – once we fully appreciate what a budget outcome actually is and what role fiscal policy plays in an economy – is that the real economy is out of shape and that our efforts should be focused on restoring growth and reducing unemployment. The budget outcome reflects the malaise but isn’t the cause of the malaise.
The OECD and IMF and the rest of these out-of-date organisations have created the perception that the budget outcome is the problem and so policy has to be focused on “fixing” that problem. In doing so, they are making the actual problem worse.
The budget outcome is, in fact, a rather irrelevant accounting number and should never be the focus of policy.
Take for example the question of debt limits – which Larry Elliot described as being a bit arbitray. This is what the OECD said in their “Key Policy Messages” in the latest publication:
Many countries face enormous fiscal consolidation challenges. Even if debt-to-GDP ratios stabilise over the medium-term, they would remain at dangerous levels.
Countries should reduce debt levels to around 50 per cent of GDP or lower to provide a safety margin against future adverse shocks.
Dangerous in relation to what? We have a nation that for many years has had public finance parameters well beyond these limits – the second largest economy in the World (Japan). So what danger has it faced?
Solvency? – it issues its own currency and can do that whenever it wants to meet any yen-based obligations.
Interest Rates? – it has maintained near zero short- and long-term rates for two decades.
Inflation? – it fights against deflation.
More recently, only the Eurozone nations which clearly face insolvency risk because they do not issue their own currency have seen any upward movement in bond yields as their deficits have risen. That is entirely predictable given the flawed monetary system. In the US, for example, we have not observed that at all.
The Report nor the supporting papers used as authority also fail to establish that nations are in “danger” if they fail to reduce their deficits.
In the main body of the text the OECD say:
A natural question is then what reduction in debt would be warranted. While different debt targets will be appropriate for different countries, a target of around 50 per cent of GDP can nonetheless be supported by some arguments. For example, empirical estimates suggest that the performance of the economy weakens in various respects around debt levels of 70%-80% of GDP: interest rate effects of debt seem to be more pronounced (Égert, 2010), offsetting saving responses to discretionary policy changes become more powerful (Rohn, 2010), and trend growth seems to suffer …. Building in a safety margin to avoide exceeding the 70%-80% levels in a downturn suggests aiming at a target of 50 per cent or even lower during normal times.
The two papers cited above are – Égert, B. (2010) ‘Fiscal Policy Reaction to the Cycle in the OECD: Pro- or Counter-cyclical?’, OECD Economics Department Working Papers, No. 763 (access) and Röhn, O. (2010) ‘New Evidence on the Private Saving Offset and Ricardian Equivalence’, OECD Economics Department Working Papers, No. 762 (access).
The relevant sections that the latest OECD Report referred to above calls upon in the Égert paper are paragraphs 49 to 53. In relation to debt thresholds we read:
50. A similar non-linear effect is detected for public debt … For public debt levels above 89% to GDP, fiscal policy is pro-cyclical. At intermediate levels of between 30% and 89%, the fiscal policy reaction is either neutral or mildly pro-cyclical and becomes counter-cyclical below 30%. This result can however be obtained only if the cycle is captured by the output gap as no non-linear effects can be established, if the cycle is measured by GDP growth …
52. Panel threshold models were used to analyse a possible non-linear relationship between public debt and the difference between short-term and long-term interest rates in OECD countries. While it is difficult to establish strong non-linear effects for all OECD countries for 1995 to 2010, long-term interest rates appear to be a non-linear function of public debt for the G-7 countries (excluding Japan) in recent years (2007:q1-2009:q4). The estimation results indicate a 4 basis point increase in long-term rates relative to short-term rates if public debt exceeds 76% of GDP. The effect of public debt on long-term interest rates is not statistically significant below this threshold.
We thus learn that the key result is highly sensitive to specification and variable construction. The basic result can only be obtained if they use their output gap measure. This variable, in turn, is generated from other modelling and is based on flawed notions of the NAIRU (which in turn is based on highly constrained specifications with particular “long-run neutrality” properties asserted).
In this regard, the author probably uses what we term to be a “generated regressor” (for the output gap) and unless that is accounted for bias is likely in the results (bias means in this context that the point estimates will be reliable estimates of the true mean. For those that are interested in this sort of issue read the original article – Murphy and Topel (1985) ‘Estimation and Inference in Two Step Econometric Models’, JBES, 1985, pp 370-379. There is nothing in the paper to explain how the author handled this issue.
When the more obvious and more precisely measured – real GDP growth variable – to reflect the swings in growth, their non-linear results disappear.
Further, “it is difficult to establish strong non-linear effects for all OECD countries”. But when the author does squeeze out something that suits the ideological bias of the framework used (and the OECD in general) all they can come up with is that the spread between long-term and short-term interest rates rise by 4 basis points when public debt exceeds 76 per cent of GDP.
Need I remind you that a basis point is equal to 1/100th of 1 per cent.
That is a frightening hike in long-term interest rates that the OECD has discovered – even if we take their highly contentious econometric modelling as reasonable.
The paper uses the Comparative Political Data Set compiled by the Department of Political Science at the University of Bern. Among the variables the author uses, there are “Three types of political economy variables are added as controls: a) the strength of the government, b) the background of the head of government and c) the timing of general elections … The first variable measures the strength of the left-wing government and the second variable measures the strength of the right-wing government”.
I have seen this dataset used before and each time I shake my head wondering how it distorts the econometric results derived. Here is what that variable (in raw form) says about Australia from 1960 to 2009. Australian readers will join with me in laughter.
The variable in the Bern dataset prorates 100 per cent for each year across the style of government in power. So if the same party is in power an entire year their “style” gets 100 per cent. The left-hand axis is thus in percentage terms.
Apparently, Australia has either right-wing conservative government (blue bars) or left-wing governments (green bars) and nothing in between. Apparently, Australia is now being governed by a left-wing government – hence the laughter.
This variable is meant to be picking up the economic approach to Government with the naive underlying view that right-wing governments are more likely to engage in “fiscal discipline” relative to left-wing governments. The term “discipline” is, of-course, loaded.
There is no way that any reasonable commentator would call the present Australian government left-wing. It is firmly right of centre and has moved more to the right in recent years.
Take the following snapshot of policies:
1. It imprisons refugees from who land on our coastline having escaped desperate circumstances (some of which our own military has created – Iraq, Afghanistan).
2. It continues to fund elite schools and ensure that public schools are starved for funds. When confronted with a recent report (the Gonski Report) which demonstrated that our public school system is falling dramatically behind world standards (literacy, numeracy etc) and need $A5 billion injected immediately, the Government responded saying its priority was to achieve a budget surplus.
3. It refuses to increase the unemployment benefit which is now well below acceptable measures of the poverty line even though their own budget forecasts predict their policies will increase unemployment.
4. It continues to “occupy” indigenous communities with its punitive “intervention” policy.
5. It is obsessed with achieving a budget surplus even though it is killing economic growth. It says it has to do this to placate the financial markets and preserve its credit rating.
None of these policies, which are centrepieces of the current government approach are remotely “left-wing”. They all define what is reasonably called a “right-wing” approach.
Further, the key econometric results appear to be highly sensitive to specification (the model design) and the data used to represent the variables (the so-called transformations).
For example, the author admits that:
… the results for the quarterly fiscal policy reaction functions show that whether or not fiscal policy is pro-cyclical depends crucially on how the business cycle is measured: different measures can result in contradictory results for a country using the same specification and estimation method. Third, some of the controls (e.g. the political economy variable) are available only at annual frequency, which decreases the degrees of freedom rapidly for a data set that typically spans over 20 years …
The general point is that the key results are not robust across specifications and transformations.
In the Röhn paper, we learn that:
… there was considerable heterogeneity in the results … [of the preliminary unit root tests which are crucial to model specification] … Despite some of the tests being inconclusive, one cannot exclude the possibility that there exists a co-integrating relationship between the I(1) variables … the insignificance of the error correction term points to possible misspecifications in some of the countries …
So like all these OECD studies there is a lot of manipulation of the data and econometric specifications to get the results that finally appear. I also would note that the datasets combine economies with quite different monetary systems and exchange rate setting modes which are not recognised nor controlled for. That alone renders much of this sort of panel-set modelling difficult to interpret.
For example, it is clear that at present the dynamics with respect to the interaction of government and, say, bond markets is quite different in the Eurozone nations than elsewhere. In many cases, the Eurozone deficits are lower (as a percent of GDP) than is the case in other nations.
Further, when you dig a little deeper (for example, footnote 9) you read that endogeneity is a problem. We read that the:
The MG estimator does not explicitly control for endogeneity issues. Therefore also IV estimators were used (difference and system GMM). These estimators, however, are generally designed for large N small T panels, which does not apply to the sample used. In addition these estimators rely on a homogeneous panel. As an intermediate approach we also applied the Pooled Mean Group (PMG) estimator, where long-run coefficients are restricted to be homogeneous whereas short-run parameters are unrestricted. While in all cases the parameter estimates of the fiscal variables were in general comparable, estimates for the controls varied considerably.
So we cannot be sure what variable is driving the show. The model assumes that the right-hand variables are independent of each other and explanatory of the left-hand variable – private saving as a percent of GDP.
There is no formal analysis of the likely interdependence between the government deficit and real GDP growth for example (in the short-run model). I could go on.
I am well-trained in these techniques and issues. The modelling and results presented did not leave me confident that the results would stand up to scrutiny. It is very likely that I could get very different results by more formally modelling the systems-nature of the data (for example).
But even if we suspend our disbelief somewhat the author finds that:
… all else equal a fiscal stimulus of for example 5% would lead to an immediate increase in private saving of 2% … This may limit the ability of discretionary fiscal policy even more as an indirect offset may occur to the extent that government actions put upward pressure on bond rates and thus borrowing costs.
Note the “may limit” terminology. The modelling does not examine the link between fiscal deficits and interest rates. The suggestion in this conclusion is pure assertion and reflects the ideological bias of the author and the framework used. How does the author explain Japan, the US – for example. Relatively high budget deficits and zero interest rates and no suggestion that interest rates will rise. In Japan’s case, long-term interest rates have been around 1 per cent for two decades. How do they explain that?
But the first finding (in the quote above) is entirely what Modern Monetary Theory (MMT) would suggest (in direction). The private saving is the ratio to GDP. The authors find strong evidence that a fiscal stimulus stimulates real GDP growth (in their dynamic model component) and that real GDP growth stimulates the private saving ratio.
A central plank of the attack that Keynes led on the mainstream approach of his day was that the latter was compromised by its propensity to enge in compositional fallacy. While these type of logical errors pervade mainstream macroeconomic thinking, there are two famous fallacies of composition in macroeconomics: (a) the paradox of thrift; and (b) the wage cutting solution to unemployment.
In first semester of a credible macroeconomics course, students learn about the paradox of thrift – where individual virtue can be public vice. So when consumers en masse try to save more and nothing else replaces the spending loss, everyone suffers because national income falls (as production levels react to the lower spending) and unemployment rises.
The paradox of thrift tells us that what applies at a micro level (ability to increase saving if one is disciplined enough) does not apply at the macro level (if everyone saves aggregate demand and, hence, output and income falls without government intervention).
So if an individual tried to increase his/her individual saving (and saving ratio) they would probably succeed if they were disciplined enough. But if all individuals tried to do this en masse, and nothing else replaces the spending loss, then everyone suffers because national income falls (as production levels react to the lower spending) and unemployment rises. The impact of lost consumption on aggregate demand (spending) would be such that the economy would plunge into a recession.
As a result, incomes would fall and individuals would be thwarted in their attempts to increase their savings in total (because saving was a function of income). So what works for one will not work for all. This was overlooked by the mainstream.
The causality reflects the basic understanding that output and income are functions of aggregate spending (demand) and adjustments in the latter will drive changes in the former. It is even possible that total savings will decline in absolute terms when individuals all try to save more because the income adjustments are so harsh.
Keynes and others considered fallacies of composition such as the paradox of thrift to provide a prima facie case for considering the study of macroeconomics as a separate discipline. Up until then, the neo-classical (the modern mainstream) had ignored the particular characteristics of the macro economy that require it to be studies separately.
They assumed you could just add up the microeconomic relations (individual consumers add to market segment add to industry add to economy). So the representative firm or consumer or industry exhibited the same behaviour and faced the same constraints as the individual sub-units. But Keynes and others showed that the mainstream had no aggregate theory because they could not resolve the fallacy of composition.
So they just assumed that what held for an individual would hold for all individuals. This led the mainstream opponents to expose the most important error of the mainstream reasoning – their attempts to move from specific to general failed as a result of the different constraints that the macroeconomic level of analysis invoked.
In the context of today’s discussion, it is thus entirely reasonable that there is a private saving offset to fiscal stimulus.
The OECD paper by Röhn thinks this means that
… forward looking agents may anticipate that given a constant government spending path, current increases in budget deficits will have to be financed through higher taxes in the future (Ricardian equivalence).
So the author concludes that that “(o)verall the results provide evidence against a strict version of the Ricardian equivalence hypothesis in the long term”.
However, there is nothing in the analysis that models the behaviour of the saving ratio (that is, the motivation). The study is at the aggregate level and nothing can be deduced as to why private saving as a percentage of GDP has risen when there is growth or higher deficits (the two effects not capable of being disentangled in the modelling framework used).
The other point is that the saving offset in no way reduces the “effectiveness” of fiscal stimulus. It just means that to drive the economy to full employment, the structural budget balance has to be larger the higher is the private saving offset. The OECD want us to believe that higher deficits are “dangerous” although they don’t articulate exactly why.
It is all smoke and mirrors sort of stuff – the occasional veiled reference to “financial markets losing confidence” and the like.
There is nothing in any of the work cited that convincingly shows that a public debt ratio of 50 per cent is optimal or even desirable.
There is nothing in the OECD work that explains why Japan continues to issue debt at low yields and every time it expands its deficit in a discretionary manner the economy grows.
There is nothing in the OECD work that documents why fiat-currency issuing nations will encounter “dangerous” waters if they don’t cut their deficits in the short-run. It is all smoke and mirrors inference. Ideology rather than logic or fact.
I could provide more detailed analysis of the OECD Report and the supporting econometric studies that they draw upon for authority but I have run out of time. I have to catch a flight.
The overall conclusion is that the OECD is part of the problem not part of the solution. The sooner the national government withdraw their funding from institutions such as this the better we will all be.
That is enough for today!