Scottish-born economist - Angus Deaton - recently published his new book - An Immigrant Economist…
The economics journalists were out in force again today in Australia after being fed their latest copy from the neo-liberal propaganda machine. In this case, the propaganda was in the form of the first report published yesterday (May 22, 2013) from the newly established Parliamentary Budget Office – Estimates of the structural budget balance of the Australian Government 2001-02 to 2016-17. The Report estimates that huge unsustainable budget deficits and has led to a flurry of media activity all just repeating what the PBO told them was the message. I wonder if any of the journalists have actually read the report in detail particularly the Appendix where the technicalities are exposed. Technicalities is too strong a word because it suggests there is something robust going on. Nothing could be further from the truth. This is another shoddy attempt to bias the public perception towards thinking the current (pitifully small relative to the scale of the problem) budget deficit is problematic.
The Parliamentary Budget Office (PBO) is one of those neo-liberal constructs where the government establishes a so-called non-partisan, independent budget watchdog to advise it on fiscal policy. The body is unelected and unaccountable to the people.
Yet, everytime it makes a noise the lazy economics media jumps to attention, gets their copy for the day from the PBO press release and the elected government is immediately under pressure because “the deficit is too big”; “we have a structural blow-out” or all the usual mis-guided headlines that appear the next day.
These organisations are designed to reduce the discretion of governments to use fiscal policy in the interests of the well-being of the people. They bias fiscal outcomes to being too restrictive and maintain much higher levels of labour underutilisation that is necessary.
They perpetuate the worst lies that my profession push into the public debate – all the usual deficit and budget myths. They are essentially anti-democratic and anti-prosperity.
The PBO claim that the Report:
… is the first in a series of reports that the PBO proposes to issue to help explain how underlying budgetary trends and discretionary fiscal policy decisions impact on the government’s fiscal position.
They will excuse me if I roll off my chair laughing. Explanation is about provided connected knowledge. This Report is a fudge around a miss-mash of orthodox theory which provides very little credible information to inform fiscal policy decisions.
It supports the mainstream view that structural budget deficits are undesirable.
The purpose of the Report is to provide estimates of the structural budget balance (SBB) which they claim is:
… a partial measure of the sustainability of the budget
Whatever that means. What they mean by this is meaningless in the case of a currency-issuer such as the Australian government. They have some insolvency model underpinning their view that budgets should generally be in surplus.
There is never a solvency issue with respect to the Australian government budget. The only thing the federal government should worry about it whether they are providing enough spending support to the economy to ensure there is no output gap and everyone who wants a job can find one.
That is the sustainable budget position.
Please read my blog – The full employment budget deficit condition – for more discussion on this point.
Also please read the following introductory suite of blogs – Fiscal sustainability 101 – Part 1 – Fiscal sustainability 101 – Part 2 – Fiscal sustainability 101 – Part 3 – to learn how Modern Monetary Theory (MMT) constructs the concept of fiscal sustainability.
After the press release yesterday by the PBO, there was a flurry of news reports. For example, in this morning’s Melbourne Age there was an article – Telstra cuts jobs amid bleak outlook.
… the new Parliamentary Budget Office blamed both sides for Australia’s slide into a structural budget deficit – a deficit Treasury warns is now likely to remain for another six years … In its debut research paper, the budget office estimates Australia will still be in significant budget deficit in 2016-17, even though the budget papers forecast a $6.6 billion surplus by then.
That is, it accepted without question the PBO estimates released yesterday.
Weren’t the economic journalists the slightest bit curious to see whether the estimates could be trusted as reliable indicators of the actual position?
Further, they reinforced the current fears that this “slide into a structural budget deficit” is something that we should be worried about.
Our National Broadcaster, the ABC ran a story late yesterday after the PBO release – oe Hockey defends Howard government tax cuts after reports say they contributed to structural deficit.
The author once again reinforced the myths and chose to quote the Access economics director (who is used all the time by the ABC) who said the following:
The usefulness of the figures out of both the Parliamentary Budget Office and the Treasury is it’s underscoring the fact that the budget is in more trouble than it looks … We overdid it. We’ve overdone it for a decade.
Australia did over do for a decade or more – they ran excessive surpluses. That is why there were still 9.9 per cent of people unemployed or underemployed at the peak of the last cycle.
The Access commentator claimed that the release of the PBO report was “great day for Australia” and:
The Parliamentary Budget Office has that greater degree of independence and Australians need to hear the message that the politicians on both sides aren’t telling us.
And … not one journalist during yesterday or overnight has questioned these PBO estimates. They are now part of the bank of “facts” that deficit frenzy in the press draws upon.
Our fourth estate is largely a failure. They are mostly mouthpieces, waiting for the next press release, which they faithfully reproduce. I asked on journalist recently why their story was so unbalanced. I was told that there wasn’t enough time between the filing deadline and the release of the embargoed report.
The question then is why go along with the press manipulation by the organisation releasing the report. If they manage the schedule so that they cannot be attacked by those critical then the press should just tell them to take a long walk of a very short pier.
But no, the article was uncritical and created an aura of respectability to a piece of analysis that was poor at best.
In compiling their estimates of the SBB they say:
The estimates have been prepared using internationally accepted methodologies.
Which, of-course, doesn’t make them right. The whole clutch of neo-liberal organisations use methods which deny reality – in particular they assume economies are much closer to full capacity than any reasonable estimate of the output gap would conclude.
They produce the following graph (their Figure 1).
The supporting text is fall of words like “a sharp improvement” if the SBB fell and “recovering” when it fell further. They say that “Based on the latest budget estimates further improvement is expected from 2013-14”. So you immediately grasp the ideological position the PBO adopts – a rising deficit is a deterioration and vice-versa.
On what grounds? Simply because they think surpluses are good and deficits are bad.
Note the estimated “further improvement” coincides with estimates in the 2013-14 Budget of fall real GDP growth (further off trend) and rising unemployment rates.
In this 2010 interview with the ABC current affairs program PM (May 13, 2010) – Modern Labor’s full employment – half a million out of work – the Treasurer was asked to define full employment.
Best of all, the unemployment rate is expected to fall further from 5.3 per cent today, to four-and-three-quarter per cent by mid 2012; around a level consistent with full employment
We can dispute the full employment benchmark defined here but he was wrong. Unemployment has risen since and is forecast in the 2013-14 Budget to rise further to be 5.75 per cent in 2013-14 fiscal year. That is an optimistic forecast. I suspect it will go higher.
But the point is that the PBO is now pumping out so-called independent research analysis which defines a “further improvement” in the fiscal position of the government as being associated with a rise in the unemployment rate, which even on the Treasurer’s own assessment is 1 per cent higher than his definition of full employment. That is 125 thousand people on current labour force size.
But that doesn’t take into account the lower participation rate that will result (pushing up hidden unemployment) and the rising underemployment.
Why didn’t any of the economics journalists point that out today?
But have a look at the graph. The broad SBB line reflects their margin of error and define what they claim is a “a plausible band”. Okay so lets look at 2011-12. Based on this reasoning, there was zero cyclical component to the budget deficit – that is, the SBB was larger than the actual budget balance.
Remember that the SBB is “the underlying position of the budget after adjusting the actual budget balance for the impacts of major cyclical and temporary factors”.
In other words, once you take out the cyclical and “temporary factors”, the PBO is claiming the Australian economy was at over-full employment – beyond its capacity.
The following Table provides a comparison between the peak at the end of the last cycle and the position over 2011-12 – the PBO’s so-called over-full capacity year where there were no cyclical factors impacting on the budget outcome.
For the labour force aggregates I have used February 2008, which was the low-point unemployment rate; whereas for the national accounts aggregates I have used the fiscal year. There is not much hinging on that choice.
I have run econometric and other types of statistical models all my career. I have supervised several successful PhD students who have also estimated things. The rule of thumb is when you estimate something you check the forecasts/projections against reality. If the estimates seem far-fetched when confronted with what you actually know then you can be sure something is wrong with the methodology being used to generate the estimates.
Make up your own mind but if the left-hand column was the best we could do out of the last cycle then what assessment are you going to make about 2011-12 outcomes.
Any reasonable person would conclude that things have got worse. That is couldn’t possibly make any sense to construct a situation where if the economy had have maintained its trend growth (between 2000-2008) of 3.05 per cent after 2007-08, then it would be some $A59,960 million or 3.8 per cent larger than it was in 2011-12.
Or that a rise in total labour underutilisation from 9.9 per cent to 12.5 per cent is a deterioration and represents extra productive capacity not being utilised.
The participation rate in 2011-12 was also 0.4 points below the peak of November 2010, which means that an extra 102 thousand workers who wanted to work were officially classified as being discouraged when compared to the 2010 peak.
The PBO know that the average person won’t be able to debate these technical matters and so they just lie and rely on the clutch of economics journalists then to spread the lie as widely as they have readers and listeners.
Why didn’t any of the economics journalists question that aspect of the PBO modelling today? They were too busy trying to make some political points out of the PBO estimates rather than working out whether the estimates made any sense in the first place.
You can see that in the 2012-13 fiscal year the gap between the actual budget balance and the PBO estimate of the SBB widens. In other words, even though unemployment and broader labour underutilisation have increased and real GDP fallen further below trend output, the PBO thinks the economy has gone further into an over-full capacity state.
If a first-year student produced this sort of analysis I would fail them immediately. The PBO work is very shoddy.
I won’t get into the subsequent analysis presented by the PBO – just a waste of time really. But it is educative to examine the information presented in the Appendix because it is there you can work out how they came up with their SBB estimates.
The steps in the calculations are straightforward:
1. “the cyclically-adjusted budget balance (CAB) is calculated … [by removing] … the cyclical component of the actual budget balance … To do this an estimate is made of the extent to which the actual output of the economy is either above or below the economy’s potential output, that is output produced when the economy is operating at full capacity, consistent with stable inflation.”
2. “Having derived an estimate of the output gap it is then necessary to determine the impact of the output gap on the budget balance by estimating the sensitivity of tax revenue and government expenditure to this gap”.
How did they estimate potential GDP?
To estimate potential production they use a – Cobb-Douglas production function – which is one of the arbitrary mathematical constructs that the mainstream pull out because under the assumption of perfect competition it presents “desirable” properties – that is, results that match their theory.
Thus, it embodies the “law of diminishing returns”, fixed factor shares (under perfect competition) and typically, constant returns to scale.
So it imposes a particular world view on empirical analysis which constrains (biases) the analysis from the outset.
For those interested in history, the Cobb-Douglas production function first was really launched in an article published by Paul Douglas in 1948 (it was his presidential address to the 1947 American Economic Association). Much earlier (1928) he had published an article with Charles Cobb which defined the function.
Douglas, by the way, was one of several authors of the 1939 paper – A_Program_for_Monetary_Reform – (aka The Chicago Plan) which has enjoyed somewhat of a revival in the literature in the last year. It is a flawed plan though which I might write about sometime in the future.
[References: Cobb, C.W. and Douglas, P.H. (1928) ‘A Theory of Production’, American Economic Review, 18 (Supplement), 139-165; and Douglas, P.H. (1948) ‘Are There Laws of Production?’, American Economic Review, 38(1), 1-41]
These constructs were theoretically demolished (by the Cambridge Critique) and then proponents claimed that it didn’t matter that we could not derive the function in a way consistent with the underlying theory (that is, it did not have solid theoretical foundations) as long as the empirical results were consistent with that neo-classical theory.
It was an extraordinary example of the way the mainstream enters denial when their approach is shown without doubt to be flawed. This was not just a matter of opinion. It was a case of their theory being internally inconsistent even on their own grounds.
There reaction – never mind … students, turn to page 20, where we see the aggregate Cobb-Douglas production function … let’s proceed.
Somewhat later, New School economist Anwar Shaikh established that the production function had no empirical relevance (he called it a “humbug function”).
[Reference: Shaikh, A. (1974) ‘Laws of Production and Laws of Algebra: The Humbug Production Function’, The Review of Economics and Statistics, 56(1), 115-120]
That devastating critique washed of the mainstream’s back like the water from a duck’s back. Denial again. Just ignore it and it will go away.
The PBO also seem to have constrained the function’s coefficient to 0.6 (hours worked – labour input) and 0.4 (capital input). This implies a wage share in factor shares of 60 per cent. It is currently around 52 per cent, which means they are over-weighting the employment input. The first area of bias.
The inputs to the Cobb-Douglas production function are trends in employment by hours, capital stock and total factor productivity.
How did they compute the trend in hours worked?
They used the formula:
Trend hours worked = (working age population x trend participation rate x (1 – trend unemployment rate) x trend average weekly hours worked) x 52
Note: at the outset the stated object of the exercise was to estimate potential output – which by any meaning is the maximum the economy can produce if all its productive resources are being deployed to their own potential.
But you will see that without explanation, the PBO slips trend in for potential. The two concepts are different. Trend is the underlying (non-cyclical) pattern of the time series.
But imagine an economy in chronic recession, the trend measure of, say unemployment would bear no relationship to a reasonable definition of full employment.
It is true that a persistently weak trend will eventually cause damage to the potential path. But it is unreasonable to replace trend and call it potential. At present the economy is growing somewhere around 2.5 per cent annum well below the 3.05 trend that it achieved up to the crisis.
Unless there has been major destruction of productive capital and fundamental shifts in worker preferences to retirement then 3.05 is about what we should expect the trend growth to be at present. So 2.5 is below trend which means even by that benchmark there is excess capacity. But then that trend was still associated with persistently high labour underutilisation which makes the excess capacity at present even worse.
Anyway, the PBO then proceeded to compute trends for each of these components (other than working age population) using a filtering technique – Hodrick-Prescott filter. All this does is smooths out a time series. It tells you nothing about the relationship between that time series and, for example, its potential maximum or minimum value.
So if an unemployment rate has been very high for say a decade, the HP filter will produce a smooth series of that very high unemployment rate. To infer that was a full capacity component is ridiculous in the extreme.
Also the HP technique, while useful to see smoothed trends, has serious flaws. As it is a purely backward looking measure it is known to cause misleading estimates of future values because, unlike a moving average filter, the HP estimate of the past state of the series is sensitive to each future value.
This relates to the so-called end-point problem that all stochastic filters encounter. I will spare you the technicalities but essentially you can only filter up to a certain point and then the filter can change significantly as you add observations irrespective of whether the added observations are in this context cyclical or trend.
The HP filter is particularly affected by this problem (unlike other time-series filters).
The way to overcome the “end-point” problem is to extend the time series and then apply the filter over the extended horizon. This can clearly introduce bias if the “made up” observations do not turn out to be accurate when time catches up with them.
The HP filter approach cannot deal with these sorts of errors and so are prone to producing biased trend estimates in extended samples.
A way of minimising this problem is to extend the time-series based on a model which is close to the true data generating process (that is, having a reliable forecasting model).
I could go further but then I would have to introduce very technical matters that do not belong on my blog (given the readership).
The question is how did the PBO choose to extend the time series to allow it to use HP filters to generate future trends?
Not much imagination was used I am afraid.
The unemployment rate and the participation rate utilise the 2013-14 Budget medium-term projections to address the end-point problem of the filter.
It is almost laughable – are these characters kidding us with this type of stuff? The problem is as a comedy it is great but in reality the consequences of this folly are tragically large – hundreds of thousands of jobs large.
The result is that trend unemployment hovers around 5.6 per cent out to 2016-17, participation falls about 0.75 per cent (note it is already well below its recent peak) and average hours worked is totally flat – that is, no change expected in the period out to 2016-17.
The result is of-course predictable. The output gap they produce is miniscule for the current period and beyond, which means that the measure of the structural deficit they come up with is much larger (in terms of its proportion of the actual budget outcome) because they are grossly underestimating the cyclical component in the actual budget outcome.
This is classic IMF/OECD tactics and then leads them to advocate fiscal austerity even though the degree of fiscal support for aggregate demand is much lower than they claim.
Then fiscal policy becomes pro-cyclical and the austerity bias sets in.
The PBO produced this graph of the output gap – comparing the OECD, IMF and their own measures.
I don’t have their data and I haven’t time to reverse-engineer their estimates from the graph. But you will recall I went through this exercise a few weeks ago.
Please read my blog – Australia output gap – not close to full capacity – for more discussion on this point.
My method was less open to bias. The RBA and Treasury have from time-to-time indicated they considered the trend growth rate to be between 3.1 and 3.5 per cent.
With that in mind, I did a regression analysis of trend. The method was standard. Real GDP was decomposed into its trend and cyclical components and for the period 2000Q1 to 2008Q1.
The estimated model was then forecast out-of-sample (dynamic) for the period 2008Q2 to the most recent observation 2012Q4. The forecasted trend series can be interpreted as telling us what would have happened to the evolution of real GDP if the GFC had not have occurred and the Government had not pursued its surplus obsession.
This model is entirely plausible and fits well with other data indicators that are available which I will consider in further background papers in the coming days.
The trend rate of real GDP growth for the period was 3.05 per cent. So it is entirely within the ballpark that the RBA and Treasury use (in fact just below their oft-quoted estimates). So there is no bias entering there.
I then computed the gap between trend output and actual output for the forecast period. Note that the gap between real GDP and trend GDP is not strictly an output gap because that would require the additional assumption that trend growth always defines the potential growth of the economy. An economy may be held in a state of austerity for years (as we are seeing in Europe at present) and its trend will be much lower than its potential – given that tens of thousands to millions of people might be persistently unemployed or underemployed.
So we should more accurately call the gap between real GDP and trend GDP to be an incremental output gap because it does not presume that the trend output upon which the forecast was based coincided with potential output or full employment.
In the context of my measure, the incremental output gap depicts the increase in whatever gap existed at the time the model forecast began (March-quarter 2013).
The following graph compares the IMF and OECD estimates with my “incremental output gap” measure. You can then more easily compare my measure with the PBO measure using the IMF/OECD relativities as guide posts.
So while the PBO is claiming there is an output gap of less than 1 per cent at present and close to zero, the OECD estimates it to be around 2 per cent in 2012 and my estimate is closer to 4 per cent.
My method is so transparent and so insensitive to technical manipulation that what the PBO and all the journalists who choose to advertise their work have to be able to demonstrate is why they assume the output gaps are close to zero when it is more likely closer to at least $60 billion.
What they have to be able to show is that the trend rate of growth of 3.05 is a massive over-estimate of the what the economy is capable of achieving in the next few years.
What has changed which might justify such a bizarre conclusion? Nothing – labour hasn’t migrated (like in Ireland, Latvia, Spain etc for example).
Why don’t all the economics journalists demand that sort of answer from the PBO? It is a simple question – Do you think trend real GDP growth is around 3.1 per cent? Yes or No?
If Yes, then there is no way the PBO estimates (nor the IMF estimates) are remotely true.
If that is the case, then their estimates of the SBB are useless and the whole department should be closed down. It is an ideological beast with no place in a democracy.
Another day of angst dealing with this stuff.
At some point in future history we are going to be exposed for what we are – stupid, dim-witted idiots – who soak up bulls***$ every day from the elites and their sycophants (including the fourth estate), whose only objective is to screw more of the real income out of our share.
I need some fresh air.
That is enough for today!
(c) Copyright 2013 Bill Mitchell. All Rights Reserved.