The NAIRU should have been buried decades ago

In 1983, I started a PhD at the University of Manchester working within the Phillips curve framework. At the time, all the talk was Monetarist – eschewing the use of fiscal policy to reduce unemployment. Unemployment was high after the OPEC oil shocks and governments were abandoning their responsibilities to reduce it because they had drunk the Monetarist Kool-aide. The Monetarists invented a concept – the Non-Accelerating Inflation Rate of Unemployment (NAIRU) or the ‘natural rate of unemployment’, which became part of the dominant macroeconomic approach and influenced policy makers to pursue microeconomic reform (deregulation, privatisation, outsourcing etc) and obsessing about fiscal surpluses. My work was an attempt to show this shift in thinking – away from a commitment to full employment was based on a lie. The whole NAIRU story was a fraud. I was largely ignored along with other progressive economists who were also producing credible research that refuted the main propositions. Some 40 years later, the ECB has produced a research paper which now supports the position I took back then. Millions of jobless people later!

When I returned to Australia and took up a position at the Flinders University in South Australia (Adelaide) one of my first publications on that work was – The NAIRU, Structural Imbalance and the Macroeconomic Equilibrium Unemployment Rate – which was published in the Australian Economic Papers, June 1987.

It had been in the refereeing process for 2 odd years.

If the referees had have been more efficient, then my paper would have beaten the paper by Summers and Blanchard to press. Their paper is the one everyone references.

Regular readers will know that I have written about the NAIRU before and have done years of work on the topic:

1. My PhD thesis included a lot of technical work (theoretical and econometric) on the topic – beginning in the mid-1980s, when I was just starting out.

2. In my 2008 book with Joan Muysken – Full Employment abandoned – we analysed the technical aspects of the NAIRU in detail.

3. Many refereed academic papers.

4. The following blog posts

(a) The NAIRU/Output gap scam reprise (February 27, 2019).

(b) The NAIRU/Output gap scam (February 26, 2019).

(c) No coherent evidence of a rising US NAIRU (December 10, 2013).

(d) Why we have to learn about the NAIRU (and reject it) (November 19, 2013).

(e) Why did unemployment and inflation fall in the 1990s? (October 3, 2013).

(f) NAIRU mantra prevents good macroeconomic policy (November 19, 2010).

(g) The dreaded NAIRU is still about! (April 6, 2009).

At the time I began this research, the debates were about whether inflationary expectations drove the inflation process independent of the state of the labour market.

The traditional Phillips curve was based on the notion that there was a trade-off between inflation and unemployment, such that when unemployment was low, wage pressures would push up costs and firms, exercising price-setting power through mark-ups, would pass the rising costs on as price rises.

To reduce inflationary pressures, unemployment had to rise.

So governments had to choose between the ‘twin evils’ and a lot of work was done in government departments trying to work out the socially optimal trade-off.

There was never any doubt that governments could reduce unemployment to some irreducible minimum. The only question was how much inflation had to be tolerated.

Then the Monetarists came along and claimed that there was no trade-off really because inflationary expectations drove inflation quite apart from whether unemployment was low or high.

They said that if governments tried to drive unemployment down, there would be some inflationary pressures, which would then become self-fullfilling because wage and price setters would expect higher future inflation and build those expectations into their behaviour.

As a result, the only unemployment rate that was consistent with stable inflation was the so-called ‘natural rate’, later called the NAIRU and governments could not influence that rate using aggregate fiscal and monetary policy.

The ‘market’, so the story ran, would keep the economy at that level of activity and the government should just butt out!

This marked the attack of fiscal intervention and began the era that obsessed with running fiscal surpluses etc.

And so we endured elevated levels of labour underutilisation.

The mainstream economists said that the only way the nation could reduce its ‘natural rate’ of unemployment would be through structural reforms and so we had the array of neoliberal policies emerging – attacks on pensions, minimum wages, unemployment benefits etc and the introduction of harsh work tests – all designed to divert attention away from the fact that the government had abandoned its macroeconomic responsibility to create enough jobs.

At the time, all the econometric work was targetting the sensitivity of inflation to inflationary expectations and the mainstream claimed that a one per cent rise in expected inflation would be fully passed through to the inflation rate, which seemed to support the Monetarist assertion.

Heterodox Keynesian economists tried to estimate a sensitivity of below 100 per cent.

All sorts of econometric papers were published debating this issue.

For me, the issue was unresolvable within that framework.

As a young researcher, I came up with the idea that it was better to target the relationship with the Phillips curve between the actual unemployment and the so-called steady-state rate (which they called the NAIRU).

The point of my work was that even if expectations were passed on fully into the inflation rate, there could still be a trade-off, if the steady-state unemployment rate was related to the actual unemployment rate.

My 1987 paper showed that demand deficient unemployment occurs when the number of people wanting gainful employment exceeds the number of vacancies being offered.

The composition of the unemployed relative to the skills demanded is not the binding constraint.

Structuralists suggest that structural imbalances can originate from both the demand and supply sides of the economy.

Technological changes, changes in the pattern of consumption, compositional movements in the labour force and welfare programme distortions are among the pot-pourri of influences listed as promoting the structural shifts.

The distinction between demand deficient and structural unemployment is usually considered important at the policy level.

Macro policy will alleviate demand deficient unemployment, while micro policies are needed to redress the demand and supply mismatching characteristic of structural unemployment.

In the latter case, macro expansion may be futile and inflationary.

But my PhD work was part of a new research agenda that was able to show that structural changes were in fact cyclical in nature – this was called the hysteresis effect.

Where the economy is at today, is a function of where it has been in the past!

Accordingly, a prolonged recession may create conditions in the labour market which mimic structural imbalance but which can be redressed through aggregate policy without fuelling inflation.

I produced a theoretical model which showed that any structural constraints that emerge during a large recession can be wound back by strong fiscal policy stimulation. This was my main PhD contention. I also produced several empirical articles during that period to verify the claims.

In other words, recessions cause unemployment to rise and due to their prolonged nature the short-term joblessness becomes entrenched long-term unemployment.

The unemployment rate behaves asymmetrically with respect to the economic cycle which means that it jumps up quickly but takes a long time to fall again.

But this behaviour has to be seen in the context of the policy position that the national government takes at the time of the recession and the early recovery period.

It is true that once unemployment reaches high levels it takes a long time to eat into it again because labour force growth is on-going and labour productivity picks up in the recovery phase.

You need to run GDP growth very strongly at first to absorb the pool of idle labour created during the recession unless you provide a strong public employment capacity that is accessible to the most disadvantaged (for example, this is what the Job Guarantee is about!).

It is also the case that if GDP growth remains deficient then the idle labour queue will remain long and employers will use all sorts of screening devices to shuffle the workers in the queue.

They increase hiring standards and engage in petty prejudice.

A common screen is called statistical discrimination whereby the firms will conclude, for example, that because on average a particular demographic cohort is unreliable, every person from that group must therefore be unreliable.

So gender, age, race and other forms of discrimination are used to shuffle the disadvantaged from the top of the queue.

Enter the hysteresis effect.

The hysteresis effect describes the interaction between the actual and equilibrium unemployment rates.

The significance of hysteresis is that the unemployment rate associated with stable prices, at any point in time should not be conceived of as a rigid non-inflationary constraint on expansionary macro policy.

The mainstream theories of the natural rate (NAIRU) all began by assuming that the rate was invariant to the state of the economy.

That only microeconomic reform could reduce it.

But this was really just a ruse to argue against macroeconomic policy interventions (and the use of fiscal deficits) and to justify the raft of neoliberal policies that gathered pace in the 1980s – privatisation, outsourcing, deregulation, attacks on unions, attacks on income support schemes, and the rest.

If hysteresis is present, then policies that reduce the actual unemployment rate, also can be shown to reduce the equilibrium rate itself – which means in theoretical terms, the so-called inflation constraint is pushed away by the expansionary fiscal policy.

That was what my PhD work was about.

The idea is that structural imbalance increases in a recession due to the cyclical labour market adjustments commonly observed in downturns, and decreases at higher levels of demand as the adjustments are reserved.

Structural imbalance refers to the inability of the actual unemployed to present themselves as an effective excess supply.

I documented the non-wage labour market adjustments that accompany a low-pressure economy, which could lead to hysteresis.

For example, training opportunities are provided with entry-level jobs and so the (average) skill of the labour force declines as vacancies fall.

New entrants are denied relevant skills (and socialisation associated with stable work patterns) and redundant workers face skill obsolescence. Both groups need jobs in order to update and/or acquire relevant skills.

Skill (experience) upgrading also occurs through mobility, which is restricted during a downturn.

If wage demands are inversely related to the actual number of unemployed who are potential substitutes for those currently employed, then increasing structural imbalance (via cyclical non-wage labour market adjustment noted above) drives a wedge between potential and actual excess labour supply, and to some degree, insulates the wage demands of the employed from the cycle.

The more rapid the cyclical adjustment, the higher is the unemployment rate associated with price stability.

The importance of hysteresis is that stimulating job growth can decrease the wedge because the unemployed develop new and relevant skills and experience.

These upgrading effects provide an opportunity for real growth to occur as the cycle reduces the unemployment rate that might be associated with stable inflation.

In my PhD thesis I called this the Macroeconomic Equilibrium Unemployment Rate to distinguish it from the conservative NAIRU concept (which at the time was claimed to be cyclically invariant).

Why will firms employ those without skills?

An important reason is that hiring standards drop as the upturn begins.

The faster is the recovery the more quickly the hiring screens are taken away as firms are forced to compete for available labour resources.

Rather than disturb wage structures firms offer entry-level jobs as training positions.

So the imbalances that arise as the economy goes under are stripped away in the new growth period. New jobs, new technologies emerge but a strongly growing economy forces firms to offer training and capacity building. If the recovery is tepid, then these upgrading effects are lost or a very slow to emerge.

What this research showed was that the whole NAIRU industry, which had been used to argue against the use of fiscal policy to reduce unemployment, was based on a lie.

That was the point of my work then.

The research, however, disappointingly, was ignored as the mainstream became obsessed with the NAIRU.

And so the dominance of Monetarism, which morphed finally into the current dominant New Keynesian macroeconomic paradigm, eschewed the use of fiscal policy, prioritised the ‘inflation-first’ monetary policy dominance, and tolerated elevated levels of unemployment and underemployment as a result.

Billions of dollars of lost income has been the result.

The lives of individuals and their families who have endured unemployment have been significantly compromised as a result.

If I was wrong, okay.

But it seems I was right all along.

In the last week (December , 2021), the ECB has put out Working Paper 2625 – Hysteresis in unemployment: evidence from OECD estimates of the natural rate (it was actually published as an NBER working paper in October 2021).

You might reasonably ask what is this research being published for now, when we knew all the results back in the 1980s when I was one of the leading researchers in this field and publishing my work.

I even discussed these concepts with one of the authors of the ECB Working Paper (Laurence Ball) at a European Commission workshop in Florence in 1996, when I was invited to present the sceptic viewpoint regarding the creation of the Eurozone.

But that is the nature of paradigm dominance – suppress conflicting research until it is convenient to claim it as your own.

The ECB Working Paper uses OECD data for 29 countries to “the dynamics of unemployment (u) and its natural rate (u*)”.

u* is, of course, unobservable, being a theoretical construct in mainstream macroeconomics.

So it has to be proxied – estimated – from real world data using a range of estimation techniques, all of them deeply flawed.

The ECB paper ostensibly studies the concept of hysteresis – so “short-run fluctuations cause the natural rate to change”.

Take a look at this graph for Italy.

I could produce similar graphs for any number of countries.

It shows the European Commission estimate of the NAIRU and the actual unemployment rate.

See the point?

The NAIRU estimate really is just a filtered replication of the actual unemployment rate.

It provides no independent information.

Given the way the estimates are achieved, the NAIRU time series will always just track the unemployment rate up and down (with some lags)

It is meaningless.

It was obvious that the NAIRU concept that has dominated economic policy debates since the mid-1970s and persists today was a fraud.

The ECB paper now find that:

1. Aggregate demand shocks (rises and falls in spending) have “permanent effects” – so when the economy goes into a slump and the unemployment rate rises, so does the estimate of the natural rate.

2. These hysteresis effects are symmetric – which is what I found in 1987.

In other words, expansionary fiscal policy drives the unemployment rate down and pushes any inflation constraint arising from structural impediments out.

3. The paper “finds strong evidence of hysteresis …”

Which means the mainstream are finally catching up.

These findings reject core New Keynesian propositions.

They should be abandoned.

But then the whole house of cards would fall.

We are waiting for that.

The decision in 2020 by the Federal Reserve Bank to abandon the NAIRU approach, which other central banks have followed really marks a decision by policy makers to move on after they have seen the adherence to the NAIRU nonsense has left millions of people without work – quite unnecessarily.

Well, it served the purpose of redistributing income to profits by suppressing wages growth.

But society is becoming fed up with that corruption and the role economists have played in perpetuating it.

So the central bankers have resisted buying into the inflation mania because they know that the current price pressures are temporary and they will only worsen the unemployment if they acted as they did in the past.

This shift away from a ‘forward looking’ anti-inflation stance where they would hike interest rates if someone ‘thought’ there were future inflationary pressures is causing havoc for the speculators who are demanding interest rates rise (as they would have in the past).

Why these demands? Because the gamblers have bets on that the rates will rise and they lose money if they don’t.

When commentators say that rate rises are already ‘priced into the market’, all they are saying is that the financial market speculators have made these bets.

The real economy will not be affected one way or another if they lose their cash.

And central bankers are now wiser to that con.

As an aside, another interesting aspect of the ECB paper, which is common in this ‘we knew it all along’ world as the mainstream economists work hard to salvage their reputations, is that the ECB paper doesn’t reference any of my work, even though I was one of the earliest researchers to develop the concept of hysteresis and present econometric modelling to examine the validity of the concept.

There is a sort of club – that gets the platform. It becomes a self-reinforcing cycle – to never reference work that is awkward so as to control the debate and make it look as if the mainstream were on top of all this all along.

They weren’t.


So when you hear an economist rave on about the NAIRU and the need for fiscal constraint – disregard them.

That is enough for today!

(c) Copyright 2021 William Mitchell. All Rights Reserved.