Economic debate in Australia today reminds me of one of the great lines in The Rocky Horror Show. As he invites Brad and Janet to meet Rocky, Frank-N-Furter taunts Janet with the lines:
I would like–if I may–to take you on a strange journey…
“I see you shiver with anticipation!
But maybe the rain is really to blame
So I’ll remove the cause,…
But not the symptom!”
It seems that in Australia, the reverse can apply: the symptom can be removed without eliminating the cause. We can avoid a serious recession while doing nothing to reduce private debt.
Excessive private debt caused the Global Financial Crisis. The expansion of private debt caused a false bubble prosperity for the last one and a half decades. Now that the growth in private debt has ceased, economies worldwide are going into a severe downturn driven by deleveraging: consumers in particular are spending far less as they try to reduce their debt levels; output is plummeting, and unemployment is rising.
Australia was as caught up in this process as anywhere else in the OECD. And yet somehow, we aren’t going to suffer much in the way of ill consequences? It seems heroic to assume so, and yet this was the basis of the recent Budget. Discussion and criticism of it by the Opposition and many economic commentators likewise focussed far less on the causes and expected severity of the downturn, and far more on the government’s projections of how rapidly it would return the federal budget to surplus.
Treasury Secretary Ken Henry’s speech to Australian Business Economists (Contemporary challenges in Fiscal Policy, Sydney, 19 May 2009) provided the most detailed defence of the Budget estimates that the recession would be over by 2010-11 and be followed by a boom, do no worse than reduce annual GDP by half a per cent in 2009-10, and see net government debt peak at 14 per cent of GDP in 2013–14 (Budget Paper No. 1, p. 3–9). In this Debtwatch I want to take a more critical look at this document than our economic journalists have done to date, and contrast the assumptions behind it, and its modelling, with my approach.
For God’s sake keep a grip on yourself Janet. I’m here — there’s nothing to worry about.
The Treasury’s Approach
I’ll quote the key economic argument in Henry’s speech in its entirety first, and interpret it later:
So let me tell you what we actually did. It’ s a bit complicated. But not too complicated, I’ m sure, for this audience.
I’ ve already outlined some of the context for this work. But a quick re-cap: We developed the medium-term scenario to provide parameters for the medium-term projections of the budget balance. The latter are required to span the gap between the four-year forward estimates period and the 40 year projections contained in the intergenerational reports.
In spanning the gap we also wanted to reconcile the short and medium-term GDP trajectory with the long-term projections contained in our IGR [“Inter-Generational Report”] modelling. Call us fastidious if you like, but we don’ t like discontinuities in our economic projections. We wanted to be sure that we were describing a medium-term scenario that is consistent not only with the short-term forecasts, but also with the long-term IGR projections.
The approach is predicated on a gradual recovery in aggregate demand in the final forecast year (2010–11), after which the supply-side drivers of the economy take over.
This is how we went about it. First, recall that we can describe GDP growth using a number of different decompositions. Usually, we employ the components of aggregate demand to tell the story. But increasingly we have used the supply-side decomposition that isolates changes in the following key variables:
(1) the population aged 15-plus;
(2) the participation rate;
(3) the employment rate;
(4) average hours worked; and
(5) labour productivity.
All of these things are cyclically sensitive, including the first because of the cyclical sensitivity of immigration.
We can obtain an index of real GDP simply by multiplying together these five things.
The weak GDP growth rate forecast for 2010-11 can be explained as follows. Despite relatively strong productivity growth of 2½ per cent, and the population aged 15-plus growing by 1¾ per cent, GDP growth is only 2¼ per cent. Weak demand sees both a fall in the participation rate (which subtracts about 1 percentage point), and an increase in the unemployment rate from 7½ per cent to 8½ per cent (which subtracts another percentage point).
As we move into 2011-12, however, growth strengthens sharply: productivity growth is weaker, but still a healthy 2 per cent; the growth in population aged 15-plus is also weaker, but still a healthy 1½ per cent; the participation rate is unchanged; and the unemployment rate falls by one percentage point. You might be interested to know that in coming out of the recessions of the early 1980s and 1990s the unemployment rate fell by about a percentage point a year for the first two years. Taken together, these factors produce a GDP growth rate of 4½ per cent.
That pattern is repeated in the following year, 2012–13. Then in each of the following four years the unemployment rate falls by only one-half of a percentage point a year; productivity growth slows to only 1½ per cent; and the participation rate adds half a percentage point a year. These factors produce annual growth of 4 per cent.
The productivity growth rate of 1½ per cent is less than that used in the two IGRs published to date. It is the actual backward-looking 30-year average. The profiles of the unemployment and participation rates over the period 2013–14 to 2016–17 weren’ t plucked out of the air either. They ensure that by the end of that period, potential GDP is back on the long-term trajectory contained in our IGR [Inter-Generational Report] modelling, having spent nine years from 2008-09 below that trajectory. By 2016–17 the output gap has closed, with an unemployment rate of 5 per cent and a participation rate consistent with the labour force models used in the IGR projections. (Henry 2009, pp 14–16)
Deconstructing Treasury I: The context
Three contextual points stand out in this statement.
Firstly, there is no discussion of what actually caused the GFC, here or anywhere else in Henry’s speech. The fact that it originated in the financial system is noted, but why it originated there, and what caused it, is not considered.
Secondly, “the long run” in the Budget papers is determined by assumptions the Treasury made about the economy’s future in its Intergenerational Report, which was published in 2007. Since those assumptions were made prior to the Global Financial Crisis, that means that the Treasury is assuming that the GFC will have no long term effect on the economy: it will suppress output and employment for a couple of years, but after that everything will return to how it was before the GFC came along.
Thirdly, the Treasury produced its Budget estimates for GDP growth by decomposing growth into 5 factors, making assumptions about those factors over time, and adding them up to produce estimates for 2010–2017. Four of the five numbers used–and the Treasury’s expectations for inflation as well–are shown below (the Treasury didn’t provide its estimates of average hours worked, so that factor is presumably collapsed into the employment rate; and their table [Table One on page 15] says “Unemployment Rate” where I believe they meant “Employment Rate”).
Note that Henry describes this decomposition as a “supply side decomposition”. There’s no argument that population is a “supply side” issue–not withstanding Peter Costello’s “Baby Bonus”, it’s fair to assume that the households decisions about whether to have children are not determined by economic conditions. Ditto productivity to some degree: higher economic growth should lead to higher productivity, but there will be productivity growth even when the economy is stagnant, so at a first pass one can treat productivity growth as independent of aggregate demand.
But are the participation rate and the employment rate “supply side” factors–set by household decisions alone rather than influenced by the demand that currently exists for workers? Henry also notes that all these factors “are cyclically sensitive”, which implies they are affected by demand conditions. So why call them “supply side” factors?
Because if you follow neoclassical economic theory, the rate of growth “in the long run” is determined by the labour supply decisions of households and the rate of productivity growth. While “in the short run”, neoclassical economists will concede that there can be insufficient aggregate demand to employ all workers who wish to work, in the long run they assume that everyone who wants a job can get one, so that “in the long run” the unemployment rate is determined by households, based on their preferences for income and leisure.
Credit issues–even ones as severe as the GFC–don’t factor into the Treasury’s modelling because, following neoclassical economics, they assume that money and credit don’t have any long term impact. Monetary factors like credit and debt are not included in Treasury’s macroeconomic model (TRYM) in any way. Ultimately, their model assumes that the economy will settle down to a long run equilibrium rate of growth determined solely by household labour supply decisions and the rate of technological progress.
Deconstructing Treasury II: Alternative Long Run Assumptions
Now let’s consider the Treasury’s approach in detail, starting with the the framing of the Budget’s assumptions so that they were consistent with the outcomes of the IGR. Henry’s defence that “Call us fastidious if you like, but we don’ t like discontinuities in our economic projections” doesn’t mean that the approach it took here was the only one available. It could also have revised the IGR projections in the light of the GFC, and then made the Budget assumptions consistent with these lowered assumptions about the future.
And there are severalways it could have done that. It could have:
- stuck with the IGR’s assumptions about long run rates of growth, but treated the GFC as a “one-off” event that took a permanent chunk out of economic performance for a few years; or
- treated the GFC as something that permanently altered the economy’s long term growth rate so that the economy was on a permanently different path; or
- some combination of the two.
Because it did none of the above, the Treasury was forced to assume that the medium term impact of the GFC would be to accelerate Australia’s economic growth: since their long run scenario assumed that the GFC altered neither the level of output in 2040 nor the rate of growth then, and yet in the short run (the next two years) it would reduce growth, then in the medium term growth had to accelerate to catch up with the IGR forecasts.
Therefore the Budget’s projections come down to hope. IF the GFC has no impact on the economy in the long run, AND the economy necessarily settles down to an equilibrium rate of growth independent of financial factors, THEN the Budget’s forecasts will be correct.
If on the other hand the GFC does have a long term impact–either by taking pushing the economy down but not affecting the long run rate of growth, or by taking a chunk out of the economy now and altering the rate of growth–and the economy’s equilibrium isn’t independent of financial factors, then the Budget’s forecasts will be wildly wrong.
Deconstructing Treasury III: Alternative Short Run Assumptions
Equally the short run assumptions–that the GFC will cause output to fall by 0.5 percent this financial year and grow by just 2.25 percent next year–can be challenged. The Treasury’s Budget Papers abound with statements like “The 2009-10 Budget has been framed against the backdrop of the deepest global recession since the Great Depression” (Statement 1, p. 1–1). And yet this recession is expected to be shallower than those of 90–92 and 82–84, when GDP fell by 1.25 percent and 2.5 percent respectively?
What if “the deepest global recession since the Great Depression” has an outcome like its predecessor, when output fell by as much as 10 percent a year?
This might sound extreme, but it’s in the ballpark of what’s happening in the rest of the OECD, where US output is falling at 6 percent and Japanese at over 10 percent. Certainly it would have been more realistic to assume that “the deepest global recession since the Great Depression” would have a greater impact than any post-WWII recession, especially since the figures coming out of the rest of the OECD imply that, in the words of Rocky Horror’s Narrator, this recession will be “no picnic”.
What if Treasury had assumed that something half that bad–something like the current downturn in the USA of a 5 percent fall in real GDP in a year–as the immediate impact of the GFC, rather than a mere 0.5 percent?
Though Treasury had to use a single forecast to frame the Budget, it could also have considered a range of scenarios, rather than the single incredibly upbeat scenario it presented and now defends.
On its record with picking the GFC to date, the Treasury is in no position to rest on its forecasting laurels. The comparison of its expectations in 2008 for the most recent financial year with reality makes for interesting reading:
Deconstructing Treasury IV: Doing the Time Warp Again
The real problem with the Budget is that it is based on the same approach to economic forecasting that Keynes lampooned in his often cited but rarely appreciated statement on the long run. It was not a quip about mortality, but a justified criticism of the manner in which the economists of his day assumed that the economy would always return to equilibrium after any disturbance:
“ But this long run is a misleading guide to current affairs. In the long run we are all dead. Economists set themselves too easy, too useless a task if in tempestuous seasons they can only tell us that when the storm is long past the ocean is flat again.” (Keynes, A Tract on Monetary Reform, 1923)
Though one might hope that economics had learnt from the past and improved since Keynes’s day, the reality is that it has instead reconstructed the same manner of thought that existed before Keynes and the Great Depression. The vast majority of models used by economists to predict the future course of the economy today are subject to Keynes’s criticism of 1923. The Treasury’s TRYM model has these same characteristics–though now expressed in mathematical models rather than impenetrable prose:
There are those who say life is an illusion — And reality as we know it, is merely a figment of our imaginations.
- A short term that into which “shocks” can be fed like a change in the growth rate, which then affects all related variables;
- A long run which is in equilibrium with assumptions made about productivity, labour supply, and population growth that generate growth projections which match recent experience, altered only by anticipated changes in demographics; and
- Assumptions about how quickly the equilibrium will reassert itself that generate a cyclical convergence from any short term disturbance.
So if an enormous financial shock–whose causes are not understood–has only a minor impact on the Australian economy, after which we will return to an equilibrium path that reflects recent history, then the Budget will be accurate. And if not?…
It’s just a jump to the Left…
There is a minority of economists who completely reject this approach to economics. But when the economy itself appeared to be booming, that minority was ignored by the majority of “neoclassical” economists, and regarded as “left wing critics of capitalism” by the public.
Now that the GFC is afoot, it is finally possible to get across the fact that the criticism was directed not so much at capitalism itself, as at woolly and delusional thinking about capitalism masquerading as economic logic.
Much of this minority’s time was taken up with pointing out why neoclassical economics was delusional (see my Debunking Economics for a collation of those critiques). Only part of it was devoted to developing alternative theories of how the economy operated, with the outstanding contribution there being Hyman Minsky’s Financial Instability Hypothesis. The core aspects of this approach to the economy are that:
And then a step to the right…
- The economy is inherently cyclical and will never be in equilibrium, either now or in the future;
- Credit and money are critical and impact on real economic activity in “the long run” as well as in the short, and must be included in any macroeconomic model; and
- Asset markets matter because debt-financed speculation on them can both drive economic activity and lead to financial crises.
These insights lead non-neoclassical economists to focus on issues that neoclassical economists ignore–such as the level and rate of growth of private debt–and require a different, truly dynamic approach to modelling the economy, as opposed to the equilibrium-obsessed approach that dominates neoclassical economics.
I’ve recently published two academic papers based on this credit-driven, non-equilibrium approach to economics (one is already available in Economic Analysis and Policy, the other will be published in the Australian Economic Review in September). They’ll be combined into one approach in the book I’m writing on the financial crisis for Edward Elgar Publishers (Finance and Economic Breakdown, estimated date of publication December 2011). Here I’ll show how both of them predict a rather different outcome from the GFC to the “return to business as usual” hopes of the Treasury.
These models are still very basic and incomplete–in particular, I have not attempted to fit them to the Australian data. But they indicate the essential differences between the modelling approach the Treasury takes and true dynamic modelling.
What is a Credit Crunch?
The phrase “credit crunch” has entered the vernacular, but what in fact is it? The easiest way to describe it is as a sudden reduction in people’s willingness to take on debt, and banks’ willingness to extend credit. This phenomenon can be captured by the model I outlined in the Roving Cavaliers of Credit by changing three key parameters:
It’s something we ourselves have been working on. But it seems our friend here has found a way of perfecting it.
- The average time horizon over which borrowers aim to repay their loans drops;
- The rate of recirculation of existing bank reserves falls; and
- The rate of creation of new credit money drops.
The changes to financial parameters are shown in the table below. The impact of these three variables on economic activity in this model of a credit-driven economy are dramatic.
The model itself is quite simple, but it differs from TRYM and its neoclassical cousins in one vital way: it is fundamentally dynamic. As noted above, TRYM’s “dynamics” simply result from the model’s key variables converging to some assumed “long run” values that are arbitrarily imposed on the model. For example, the graph below from TRYM’s documentation shows what TRYM assumes will happen to unemployment “in the long run”: it will converge to a steady state level that in 1995 they assumed was 7%.
The fact that unemployment behaved very differently between 1996–when this document was written–forced the Treasury to change the 7% figure to 5.25%. But it it still imposed on the model–it is not something that results from the interaction of unemployment with other variables in the model.
This is why the Treasury can assert that the GFC will have no long term impact on the unemployment rate: because its model simply assumes so.
Things are rather different in the real world, and a truly dynamic model goes at least part of the way to capturing that. Below are four simulations of the impact of a permanent credit crunch–a permanent shift in financial parameters–in the Roving Cavaliers model. Not only does the size of the shock have a dramatic impact on unemployment in the short term–causing everything from a short recession with a mild shock to a long Depression with a large one–but the equilibrium rate of unemployment changes permanently as well.
The model generates an acceleration of growth after a credit-crunch-induced recession, rather like the one that Treasury is assuming will happen–but again, in contrast to the Treasury’s model, this acceleration is not merely an assumption, but a product of the model’s many feedbacks.
Deleveraging and Economic Breakdown
The Roving Cavaliers model above simulates a crisis of liquidity, something which can easily be overcome if firms’ willingness to take on debt, and banks’ willingness to lend, can be restored. If this were the nature of the crisis we are experiencing, then it could be reversed simply by increasing confidence–which is the one thing that can be said in favour of an official forecast that the recession will only last a year and result in unemployment no worse than 8.5 per cent.
But it’s the pelvic thrust That really drives you insane…
But what if the crisis is one of solvency instead? What if the real cause of the crisis is not merely a sudden drop in confidence resulting in lower rates of creation and circulation of credit, but too much debt altogether?
That possibility is captured in my Minsky model, in which a series of booms and busts leads to one final bust where the accumulated debt is so great that the economy can no longer service it. Output and employment collapse, and the only way out is to deliberately reduce the debt.
This model’s dynamics are generated by four interacting factors:
- Wage demands by workers based on the rate of employment
- Investment decisions by firms based on the rate of profit
- Speculative borrowing by firms based on the rate of economic growth; and
- Lending by banks
Its equations can be summarised as saying that:
- Workers share of output will rise if wage demands exceed the growth in productivity
- The employment rate will rise if economic growth exceeds the sum of productivity and population growth
- Banks lend to finance investment and speculation and charge interest on the outstanding debt; and
- Speculation rises as the rate of economic growth increases.
Those four factual statements result in a model that generates a series of trade cycles, each apparently like the previous one, but with the level of debt altering over time.
A crucial factor is the distribution of debt between productive and unproductive purposes–between genuine investment and mere speculation. The former builds additional productive capacity that can be used to meet financial commitments in the future, even if today excessive increases in the debt burden leads to a recession. But borrowing that merely finances speculation on the stock exchange or the housing market about the prices of shares and houses adds to debt now without increasing our capacity to service it in the future.
The “No Speculation” simulation shown in the following graphs has only productive borrowing; the “Ponzi Finance” simulation has borrowing to speculate on asset prices as well as productive borrowing.
It is possible to generate a crisis in the “No Speculation” case by choosing a set of initial conditions that result in more extreme cycles–it is a “chaotic” model that has “sensitive dependence on initial conditions” in the technical vernacular–but generally speaking it is a stable system that won’t lead to a debt crisis (in fact in the simulation here, debt to GDP ratios are negative, which means that firms accumulate positive bank balances).
The Ponzi Finance system however is inherently unstable: the growth of unproductive debt during a boom–when people borrow money to speculate on rising asset prices–adds so much to debt that the amount accumulated in the previous boom is never completely repaid before the next boom takes off. The debt to GDP ratio therefore ratchets up over time, until ultimately, so much debt is taken on that the economy experiences not merely a recession but a Depression.
These models are far from “the ant’s pants” in terms of what truly dynamic economic modelling could be. But they are far in advance of models like Treasury’s TRYM that predict the future simply by assuming that it will be pleasant.
END OF COMMENTARY
Comments on the Data
The latest credit figures indicate that, were it not for the First Home Vendors Boost (let’s call it what is really is), private debt in Australia would now be falling. The increase in debt for the month of April was a mere A$776 million, with a A$6.5 billion increase in mortgage debt almost offset by a A$4.8 billion fall in business debt and a A$1 billion fall in personal debt. The change in private debt is therefore on the cusp of reducing aggregate demand, whereas for the past 17 years, it has been adding to it.
Table One
Table Two
Even though I expect that the government’s “Keynesian pump priming” will not prevent a Depression, it is still instructive to compare the Government debt levels, which are the focus of Canberra’s current hysterical debate over the Budget, with private debt levels.
It is an indictment of economic and political thinking that our last forty years of economic debate have been dominated by a song and dance about the Government’s debt levels when out of control private debt levels have been virtually ignored.
The run up in private debt since the recession of 1990 has been so great that changes in government spending are simply too small to neutralise the impact of de-leveraging by the private sector. The next graph shows the contribution that change in debt makes to aggregate demand–defined as the sum of GDP plus the change in debt. Deleveraging has only just begun in Australia, and yet already the reduction in the rate of growth of private debt has sliced about 8 percent off aggregate demand. The government’s attempts to counter this, though huge by historical standards, are trivial compared to the scale of private sector de-leveraging.
