Although this blog post considers some very technical material its message is simple. Mainstream macroeconomic models that are used to determine policy choices by governments are deeply flawed and the evidence strongly supports a central thrust of Modern Monetary Theory (MMT) – that fiscal policy is powerful and that austerity will kill growth. In that sense, it helps us understand why various nations and blocs (such as the Eurozone) struggled after the onset of the GFC. It also explains why the deliberate attack on Greek prosperity by the Troika was so successful in demolishing any prospect of growth – an outcome that the official dogma resolutely denied as they constructed one vicious bailout after another. It also explains why New Keynesian approaches to macroeconomics are flawed and should be ignored. I was reminded this week by a research paper I had read last year (thanks Adam for the reminder) which presents a devastating critique (though muted in central bank speak) of the mainstream approach to macroeconomic modelling. A research paper from the ECB (May 2017, No 2058) – On the sources of business cycles: implications for DSGE models – provides a categorical critique of DSGE models and a range of other stunts that mainstream economists have tried to introduce to get away from the obvious – economic cycles are demand driven.
The paper is highly technical and I will avoid discussing those aspects, even though they are very interesting in their own right.
The mainstream macroeconomics profession relies heavily on so-called Dynamic Stochastic General Equilibrium (DSGE) models.
Central banks and other forecasting agencies deploy to make statements about the effectiveness of fiscal and monetary policy.
I considered this type of model in several blog posts, including:
1. Mainstream macroeconomic fads – just a waste of time (September 18, 2009).
2. The myth of rational expectations (July 21, 2010).
3. Fiscal austerity damages real growth and prolongs the financial downturn (June 21, 2012).
4. Mainstream macroeconomics in a state of ‘intellectual regress’ (January 3, 2017).
5. Austerity is the problem for Britain not Brexit (January 9, 2017).
This is not meagre academic to-and-fro. The use of these models is prominent in current debates about Brexit.
For example, the National Institute of Economic and Social Research (NIESR) in the UK, which has been pumping out ridiculous forecasts of doom should Britain leave the EU uses as a core approach to its modelling a New Keynesian style, Dynamic Stochastic General Equilibrium (DSGE) model.
Its main model NIGEM has in their own words “has many of the characteristics of a … DSGE model” including the use of rational expectations, which means their long-run solution is supply-determined.
But you need to understand the ‘supply-determined’ bit means that there is a dichotomy between the real and nominal worlds or in simpler language – money doesn’t matter in a long-run general equilibrium.
That is standard nonsense and the ECB paper provides withering evidence to support my assessment.
This is one of the reasons the early forecasts of the impacts of Brexit have proven to be so wrong, a topic we considered in our recent Jacobin article (April 29, 2018) – Why the Left Should Embrace Brexit.
On March 9, 2009, the hardly radical economist Willem Buiter wrote in a Financial Times article (now deleted but available here) – The unfortunate uselessness of most ‘state of the art’ academic monetary economics – that:
Most mainstream macroeconomic theoretical innovations since the 1970s (the New Classical rational expectations revolution … and the New Keynesian theorizing) … have turned out to be self-referential, inward-looking distractions at best. Research tended to be motivated by the internal logic, intellectual sunk capital and esthetic puzzles of established research programmes rather than by a powerful desire to understand how the economy works – let alone how the economy works during times of stress and financial instability. So the economics profession was caught unprepared when the crisis struck … the Dynamic Stochastic General Equilibrium approach which for a while was the staple of central banks’ internal modelling … excludes everything relevant to the pursuit of financial stability.
The ECB paper provides powerful support to that conclusion.
The question and conclusions
The ECB seek to:
… investigate sources of economic fluctuations … We consider real GDP, real consumption, real investment, real exports, real imports, the unemployment rate, and core inflation
They consider the US and other advanced countries.
The importance of this enquiry is that the mainstream models of the business cycle emphasise supply-side motivations and claim that demand shocks are temporary at best.
This is crucial because if that was true then the claim that fiscal policy cannot permanently alter the course of an economy would be valid.
The ECB paper though refutes the mainstream belief.
Their overall conclusions:
1. They find “great regularities in business cycle co-movements of key macroeconomic variables across multiple economies”.
They find “one dominant source of real co-movements … business cycle dynamics of key macroeconomic data can be largely, .. explained by a single source of variation”.
2. The one dominance source of co-movement is a “demand factor” – that is, variations in aggregate demand (spending).
3. “any structural economic or econometric model of business cycles must be able to generate the principal component structure that we present” – that is the dominance of aggregate demand.
“We argue that the recent vintage of structural economic models fails this test—these models cannot explain business cycle dynamics.”
The authors find that:
… most prominent DSGE models today are not compatible with our empirical findings on the number of factors and the nature of co-movement in the macroeconomic data.
There it is!
The ECB authors deploy Dynamic principal component analysis, which is a relatively recent derivative of – Principal component analysis – (PCA). PCA was introduced as a statistical technique by Karl Pearson in 1901.
The statistical procedure manipulates data in such a way that principle component are generated which are ordered by the extent to which they account for the variability in the data (largest possible variance down). Once the first component is selected, the next component has the large variance subject to the constraint that it is “orthogonal” (totally uncorrelated) with the first component.
And so all the variance in the data is eventually accounted for by the set of orthogonal components.
PCA is limited by the assumption that the data set is time independent (that is, it is unsuitable for time series data where relationships are intrinsically correlated across time – so-called “serial correlation”).
Dynamic principal component analysis extends PCA to account for these lag structures within time series data. It adds lags to the data and then extracts the principle components once a specific lag structure is determined.
There are all sorts of options and techniques available to statisticians to determine these things and they are too complicated for me to deal with in this setting.
The point being that the technique is now well established, having been introduced in the early 1980s, and is an interesting way to examine dynamics in economic processes such as income and inflation generation.
The implications of the results can be summarised fairly simply.
The results point to only “a single dominant dynamic principal component” – aggregate demand.
The authors conclude that any model:
… for instance a DSGE model—that does not feature a structural shock that dominates the cyclical frequencies of consumption, investment, output, hours worked, and inflation is very likely to be misspecified.
They also say that “most DSGE models in the literature would struggle to explain the principal-component space of the data in a satisfactory manner”.
The authors find that:
… it simply does not happen that investment plummets while private consumption remains resilient or rallies during a recession. Again, it does not happen that the unemployment rate drops when output slumps. It just does not happen, except in many DSGE models.
This has implications for the ‘fiscal contraction expansion’ or the ‘growth friendly austerity’ that the IMF and other bodies were pushing during the GFC as they coerced governments to cut their deficits using fiscal austerity but claimed that the public spending cuts would spawn new private sector growth.
It was never going to happen because that is not the way economies work.
The ECB research confirms the remarkable stability of the co-movements between the aggregate variables.
If investment falls, or for that matter public spending falls, private consumption falls. Why?
Because the declining investment (or spending in general) creates an output gap and firms lay off workers. National income drops and household consumption and saving falls.
Standard observations that Marx, Keynes, Kalecki etc all understood intrinsically. And, a core component of what we now call MMT.
In simple terms, their results means that most of the macroeconomic models that are used by policy agencies, New Keynesian researchers etc ignore the dominance of aggregate demand in explaining growth and inflation dynamics and will therefore perform poorly and generate misleading policy inferences.
This “misspecification will cause remaining structural shocks or measurement errors to be correlated”, which means the models fail to say anything about what is going on in the real world.
The authors suggest that some of the mainstream models are more obviously flawed than others.
For example, they point to the Real Business Cycle theory literature (the so-called ‘new classical macroeconomics’) which emerged in the early to mid-1970s (mainly out of University of Chicago and Minnesota) to counter the Keynesian orthodoxy.
RBC theory thinks that economic cycles – boom and bust are driven by ‘productivity shocks’ (supply-side shocks), so that when new technologies arrive, raw material prices change, or new production methods are introduced – booms and busts occur depending on whether the shock is good or bad.
Adjusting aggregate demand has no impact – the so-called ‘policy ineffectiveness’ postulate that allows RBC theorists to eschew government attempts to manipulate the economy in any systematic way using fiscal policy.
So it challenged the keystone of Keynesian thinking.
The ECB authors show that in one recent RBC case (a 2010 paper they cite), the ‘supply shock’ “cannot explain more than 10% of the consumption dynamics”, which is due to them ignoring the key demand component.
Similar flaws can be found in most of the major mainstream approaches to explaining business cycle dynamics.
The ECB authors consider that “any structural economic model must at least be able to generate the principal component structure of the data it is supposed to represent”, which is why the mainstream macroeconomic models fail.
They argue that many mainstream models capture, say, the dynamics of perhaps one or two of the aggregate variables, but fail badly to capture the dynamics of other variables.
In other words, they have been unable to fully embrace what is going on in the economy in general.
The other important result is that:
… the analysis of both real variables and inflation reveals their tight co-movement — often doubted in the literature …
Thus real output and inflation move together as a result of demand dynamics.
In technical terms, this means that the vast majority of macroeconomic models (such as NIGEM, above) which exploit what is known as the ‘classical dichotomy’ – that money is neutral in the long-run, which is determined purely by supply-side forces – are misspecified and fail to appreciate these robust co-movements between real and monetary aggregates.
Which means that when a New Keynesian starts waxing lyrical about the ineffectiveness of fiscal policy in determining long run trajectories for the economy, you should turn the volume down or turn the page.
The authors are clear:
1. “We reach a conclusion that the business cycle dynamics of key macroeconomic data can be largely explained by a single source of variation. Since this dominant unobserved principal component behind the business cycle explains the positive co-movement of the output cycle and inflation, we label the principal component as a ‘demand factor.’”
2. The demand factor – the “dominant component” – captures the “positive co-movement of output and inflation”.
This is a very important finding and supports the core structure of Modern Monetary Theory (MMT) and its emphasis on the primacy of fiscal policy.
That is enough for today!
(c) Copyright 2018 William Mitchell. All Rights Reserved.