Tag Archives: unemployment

Premier Malinauskas’ “Full Employment” – a dangerous definition

Yesterday (21 February 2024), on ABC Radio National, the South Australian Premier Peter Malinauskas told the country that we have full employment in the state. It is a claim I have heard him make before. Given that I think that full employment is a historical myth, at best an unstable temporary state, and probably unattainable in the long term, I did a quick fact check on the Premier’s statement.

Head shot of Peter Malinauskas

The latest data from the ABS shows that in January 2024, the unemployment rate in South Australia was 3.9%, equating to around 40,000 unemployed people. The underemployment rate was 7.6% of the labour force, which means that there were an additional 75,000 people who were less than fully employed. Further, the proportion of the state’s population who were employed was significantly below the national average, so we probably have an unknown number of unregistered unemployed people.

So, on its face, the Premier’s claim is simply wrong. There are clearly unemployed and underemployed people in SA looking for jobs. We do not have full employment.

At this point a populist would bemoan “lying politicians”, but the Premier was probably relying on the infamous NAIRU, the Non-Accelerating Inflation Rate of Unemployment, to define full employment. The NAIRU is basically the minimum level of unemployment that is not expected to trigger supply constraints and inflation. It is a definition of “desirable” rather than “full” employment, but Malinauskas is not alone in using this as the definition full employment. In fact, such definitions are crucial to current federal policy debates.

The latest review of the Reserve Bank was celebrated for giving the goal of full employment equal status with its inflation management goal – a radical departure from its inflation-focused recent history. However, the RBA’s December 2023 Statement on the Conduct of Monetary Policy makes clear that this full employment is the “maximum level of employment that is consistent with low and stable inflation” (the NAIRU).

As Mike Beggs argued in a recent JAPE article, when the RBA defines full employment as the NAIRU, then the change is essentially meaningless. Economic growth and the existing inflation target will deliver the NAIRU “full employment” (almost by definition) – while actual unemployment will continue.

The Federal Treasurer Jim Chalmers also notes the problem of using the NAIRU as a measure of full employment, saying of the NAIRU:

“While it’s a useful measure, it doesn’t capture the full potential of our workforce and it shouldn’t – and doesn’t – limit the Government’s ambitions for getting more Australians into work.”

Chalmers’ White Paper on future work defines full employment as where anyone who wants a job can get a job without searching for too long.

This is a more common-sense definition, but it is hard to see how we will achieve this full employment if State Premiers and economic institutions like the Reserve Bank use a language which renders unemployed people invisible. Unemployed people simply cease to exist when the NAIRU is translated as full employment.

As a general rule, if you actually want to get unemployed people into work, you probably need first to acknowledge that they exist.

Alternatively, if the reality is that you don’t want or can’t have full employment (in the sense of everyone being able to get a good job), then you might want to find other ways of ensuring unemployed people have access to a dignified life (rather than pretending they don’t exist, or trying to starve them into work with conditional welfare at below-poverty-line levels).

Either way, pretending that unemployment does not exist by adopting a NAIRU definition of full employment is dangerous. It misdirects political debate and shuts down policy options that would deliver better outcomes for unemployed people and the community more broadly.

Did We Ever Have Full Employment?

Did we ever have full employment? It seems like a simple question – just a matter of checking the historical data and answering yes or no. But strangely, it is not so simple. And nor it is just a curiosity of history. It is a question loaded with contemporary significance.

The Long Boom

The golden era of full employment was the Long Boom (or the Trente Glorieuses in France – because everything sounds better in French!) which followed the Second World War and lasted until the 1970s. In Australia, with post-war construction and Keynesian economic strategies dominating economic management, the unemployment rate averaged 1.9% from 1941 to 1974 – with most of that accounted for by frictional unemployment as people changed jobs. Further, the employment model was largely full-time, under-pinned by a centralised wage-fixing system and industries protected by significant tariff barriers (as well as the expense of pre-containerised shipping – which also meant mass employment on the waterfront).

So, question answered. We did, for all practical purposes, have full-employment in that long boom. Since then, the world has changed. Unemployment has generally hovered between 5% and 8%, peaking in December 1992 at 11.2%, and a different policy mix is clearly required to re-establish full-employment in the modern economy. But the data clearly suggests that it is possible to have full employment in a capitalist economy.

Hmmm – not so quick. Firstly, given the historic record before and after the Long Boom, that period may have simply been an aberration – a one-off event caused by a unique coincidence of factors which can’t be repeated. More importantly, it should be remembered that these very low unemployment rates were obtained in part by culturally or legally excluding parts of the workforce.

Women who had been employed through the Second World War were forced out by cultural norms, or social pressures to give up their jobs for returning heroes or later for men who had families to feed, or women were simply barred from work. Indeed, the marriage bar in the Commonwealth public service (and in other major employers like banks) lasted officially until 1966, and probably informally in some instances for much longer. Women like my mother simply lost their jobs on marriage and were out of the workforce for years, but they were not “unemployed”.

Similarly, given the long history of Aboriginal exclusion, wage theft and the ambiguous labour force status of station workers and domestic servants, it is likely that the unemployment data did not capture the reality of Aboriginal employment and unemployment.

And many people with disabilities who are now in the labour force (but not necessarily employed) were kept at home and either not expected to work, or placed in sheltered workshop “employment”, both potentially hidden forms of unemployment.

So, on closer inspection the golden age of full employment was not – at best it was full male employment with caveats. To refer unproblematically to the full employment of the long boom is to take a gender-blind and privileged view of history.

Participation Rates and Full Employment

The crucial factor here is the labour force participation rate, that is, the proportion of the adult population in the workforce. The graph below shows the significant rise in participation rates since that golden age, from 60.2% of adults in the workforce in 1966 rising to 66.5% in January this year. This growth was driven largely by the increasing proportion of women entering the workforce, from 36.6% of adult women in the workforce in 1966 to the current 62.1%.

Line graph of labour force participation rates from August 1966 to January 2023, showing rate increasing the through the 1970s, dropping back to around 60% in the late 1970s early 1980s and then a long term secular growth since then.
Source: ABS Labour Force, Historical Charts.

This increasing participation rate is important because, if today’s participation rates applied in 1966, then there would have been 515,000 more people in the workforce (my calculation from ABS Labour Force data). If there were same number of jobs at that time, then the unemployment rate would have been 11.1%, not the 1.8% in the ABS data. Again, not exactly the “full employment” of legend.

Of course these counter-factual statistics are flawed because if the labour force participation rates were higher and those hypothetically unemployed people were actually “in the workforce” then the price of labour may have been lower (especially as this was pre-“equal pay” so women were cheaper to employ). Further, expenditure patterns would have been different if some of those people found employment (and the non-market production of cooking, housework and childcare may have been monetised sooner creating more jobs). All of which may have created more employment.

The counter-factual data therefore does not definitively give us a negative answer to the question of whether we ever had full employment. Rather, it reminds us that the statistics are not neutral. The statistic data, like the very notion full employment, is in part socially and politically constructed by the social patterns and expectations of the time.

Theory and Policy

This social construction immediately makes me suspicious of any economic theory that posits employment dynamics (full or un/under-employment) as being driven solely by market dynamics, for example aggregate supply and demand.

At one level, this is a critique of most macroeconomic theories, but the proposition that we possibly never really had full employment is particularly problematic for Keynesian and post-Keynesian theories. As noted in my previous post, such theories posit full employment as not just possible, but as the aim of macroeconomic policy. It is also those theories which tend to hark back most to the golden era of “full employment” for real-world legitimacy.

Of course, the historical record does not definitively rule out full employment (even if we did not have it in the past, it may still be possible in a theoretical future), but the long-term record is not promising.

Given the enormous social and economic costs of unemployment, the idea that full employment may not be possible is hard to stomach. But simply wanting it does not make it possible.

By contrast, both the Monetarist-neoclassical and Marxian-classical schools of thought posit some level of unemployment as being inherent in the system – the former with a NAIRU, the later with a reserve army of unemployed. (Again, see previous post).

What is at Stake?

If the mainstream theory or the Marxists are right, then the post-Keynesian full-employment approaches create false hope and may be economically counter-productive, but of course if full employment is possible, then both mainstream economics and Marxist theories may be selling the unemployed short. That is what is at stake in the curious debate over historical full employment.

That said, I do not think that accepting that full employment may not be possible is giving up on the unemployed. Certainly, as I suggested previously, the idea that there is a natural level of unemployment has been deployed by neoliberalism to undermine progressive interventions in the economy (unions, minimum wages, welfare) and limit government spending.

However, the idea that unemployment is inherent in the economic system can equally be deployed to undermine the stigma of being unemployed (because someone has to be unemployed) and to create arguments for minimum income guarantees not attached to workforce participation.

The policy approaches here and are as debateable as the political economic theories which underpin them. I am simply suggesting that, at a minimum, we might want to have those debates without reference to an idealised era of full employment which was socially constructed at best, and fictitious at worst.

What Level of Unemployment is Acceptable? Theory and Policy

In a previous post I highlighted unemployment data which showed that, despite all the business and economic talk of labour shortages, the rates of effective unemployment are higher than in the official ABS headline data and that there remains a problem of entrenched long-term unemployment. However, in policy terms, what we make of the level of unemployment – and what to do about it – is shaped by one’s theoretical starting point (with the following discussion of theory drawn primarily from my reading of Anwar Shaihk’s Capitalism, Competition, Conflict and Crisis).

Theories of Unemployment

Neoclassical

In the neoclassical theoretical model there should be no unemployment because in a perfect market wages would be at the price where the market cleared – that is, the demand for labour matched its supply. Any unemployment is therefore a product of market imperfections (lack of full market knowledge or mobility of resources, or pesky things like minimum wages, unions, government regulation, or welfare payments). While unemployment obviously exists in the real world, this is a “natural” consequence of the departure from the perfect competition and the setting of wages above the level of market clearance (zero unemployment). Unemployment then becomes “voluntary” – an option chosen by society and by individuals where welfare payments allow people not to work at the market equilibrium rate.

For Friedman and others, this “natural rate of unemployment” morphed into the NAIRU – the Non-Accelerating Inflation Rate of Unemployment. In theory, once you approach full-employment, there is increasing competition for labour and the price of labour increases, which drives inflation. But with a certain level of voluntary unemployment, that inflation tendency kicks in below the level below full employment – hence, the effective rate of full employment (or the desired level of unemployment) is the point just before employment “accelerates inflation”.

The graph below shows the Australian Treasury estimates of the rate of NAIRU over the last 40 years, with the pre-pandemic estimate just under 5%. While there are endless arguments about the estimates and the theoretical relationships underlying it, the NAIRU and the need to contain inflation has long been a part of the dominant rationale for maintaining some level of unemployment.

Line graph showing the estimated NAIRU declining from around 7% in 1980 to just under 5% in 2019, and the much more volatile official unemployment rate above the NAIRU in the 1980s, and again in the 1990s, falling to below the NAIRU in the early 2000s then rising after the Global Financial Crisis.

Keynesian

By contrast, those coming from Keynesian and post-Keynesian traditions assume an imperfect market and have a theory that demand drives supply and economic growth. Given that the imperfect market will be below full employment, the government can (and should) via stimulus spending boost aggregate demand up to the point of full employment. Modern Monetary Theory (MMT) fits within this tradition, but adds a jobs guarantee as an additional “automatic” stabiliser (see for instance, Kelton, pg 65-66).

In this context, the persistence of unemployment simply shows that neoliberal governments worried about inflation have not provided enough stimulus to reach the levels of full employment.

Classical

Like many classical (and neoclassical) political economists, Anwar Shaikh believes that such Keynesian stimulus measures can lead to short-term economic boosts, but in the longer term unemployment will return as a natural part of capitalist competition. Marx referred to this as the “reserve army of the unemployed”, but it is not a conspiratorial “disciplining of labour”. Rather, in Shaikh’s view ongoing unemployment is a function of the logical response of profit-driven capital to full employment.

Supply and demand processes suggest that as unemployment falls, there is upward pressure on wages. This negatively impacts on profit rate, which in turn, leads to lower investment which decreases the demand for labour. Alternatively, capital can look to counter this increasing cost pressure by increasing labour supply (by migration or capital flight to source labour overseas), and this increased labour supply will lead to unemployment. Or, in response to rising wages, capital can invest to increase productivity, which displaces labour and creates unemployment. There is no pre-determined path, but ultimately capitalist production is driven by profit not demand, and the self-governing mechanisms of profit-seeking ensure some level of unemployment.

Implications of the Different Theories

There are of course lots of variations within the three broad approaches described above, and I am not going to adjudicate these long-standing theoretical differences here. However, it is interesting to look at the implications of the recent employment data (in my previous post) for different aspects of these theories.

The existence of the sustained and significant unemployment highlighted in my previous post is obviously not consistent with the neoclassical paradigm of a perfect market, but it is compatible with the revised Friedmanite version where less than actual full employment is functionally full employment. Since many of those in long-term unemployment are never going to get jobs (as new labour market entrants out-compete them), the relevant inflationary impact will be at some level of unemployment below 0%.

That said, it is not immediately clear how the official unemployment rate relates to the NAIRU and inflationary pressures if, as we saw in my previous post, 72% of the newly-employed do not come from those who are counted as unemployed. The unemployment rate is simply not measuring the potential labour supply which is purportedly driving wages and prices.

The persistent unemployment data is also consistent with Marx’s notion of a reserve army of the unemployed, but again, the data suggests that most of this reserve army is located outside of the ranks of the unemployed. Shaikh’s analysis of capital’s response to near full-employment may be valid, but it relates more to a labour supply not-in-the-labour-force than to the long-term unemployed, who are more an outside sub-class than part of the reserve army.

But perhaps the biggest challenges of persistent long-term unemployment are posed for post-Keynesian theory and the attempts at government intervention to ensure full employment. At the macroeconomic level, having a cohort of semi-permanently excluded job seekers points to the limits of generic demand-led stimulus strategies. It is not just that there has not been enough stimulus. The barriers faced by many unemployed people may prevent them from winning most of the newly-created jobs (as jobs are filled from outside the labour force). More demand management will not (or not easily) create full employment.

A Job Guarantee

This leads many, particularly in MMT, to argue for more direct government job creation through a job guarantee (with the government or government-funded organisations “employers of last resort”). In theory, this should be good for those entrenched in long-term unemployment. Direct job creation can target disadvantaged groups, and there would be a guarantee of a job that they would probably not get in a competitive market. However, here we are not necessarily talking about people who freely move in and out of jobs as the labour market changes. The jobs need to be created and supported in a way to address the specific needs of those with barriers to employment – which is a difficult and resource intensive task.

This is recognised at least in passing by some proponents of a job guarantee, but much of their description of how such a program might work seems to suggest a workforce moving flexibly in and out of the guarantee program. Moreover, even if people with employment barriers flock to a job guarantee, if sufficient numbers of them can’t go back to the private workforce when demand for labour increases, then the macroeconomic “automatic stabiliser” can’t do its job in adjusting labour supply.

These problems may or may not be solvable, but they do mean that the Keynesian-inspired solutions are not as easy as they are often portrayed.

Other Possibilities

There are of course dangers in focusing on labour “supply-side” issues like the barriers faced by long-term unemployed people. This can and has led to reactionary “victim-blaming” policies, and to surveillance and back-to-work programs which pretend that the problem is a deficit in the unemployed person rather than the labour market.

However, if we take the persistence of long-term unemployment as evidence of systemic unemployment within capitalism, as per the classical (and even the neoclassical) theories, then perhaps we can stop demonising and starving those who are unemployed. We might accept that the labour market can’t provide income for everyone, and actually build support and value them outside the labour market.

This could be an argument for a Universal Basic Income, but it could also simply be an argument for the removal at some point of “job seeking” pretences of current social security payments. This already happens (in a minimum and controlled way) for people over 55 where mutual obligation requirements change after 12 months to basically accept volunteer work instead of employment. But the payment is still called “JobSeeker” rather than, say, a Community Participation Allowance, and in practice it may be closer to a job guarantee than a UBI.

Whatever, any income support (universal or not) needs to be at a level which provides for a minimum healthy standard of living – which clearly is not the case currently. This need to increase payments is not just for humanitarian reasons, but a recognition that the unemployed are the collateral damage of capitalist production (the classical theory) or that they are bearing the cost on behalf of the community of fighting inflation (the neoclassical model).

The acceptance of the inevitability of some level of unemployment may not initially sit well with those interested in equality and the rights of those who are struggling, but it can still have progressive possibilities. And it may be better than simply hoping for a full employment which may not be possible.

Of course, if full employment is possible, accepting less is selling out. That is why the theory matters!

Unusual Unemployment Statistics: Highlighting Long-term Unemployment

With businesses complaining about labour shortages, and the official unemployment rate at a long-time low, this should be an optimistic time for unemployed people – a job offer should be just around the corner! As the ABS Labour Force data in the graph below shows, unemployment is trending down strongly after the spike caused by COVID lockdowns, and it is significantly below the pre-pandemic levels. Even long-term unemployment (more than 12 months duration) is decreasing the in face of economic growth and a tight labour market.

Line graph showing ABS data on the unemployment rate from 2018 to 2022 - sitting around 5% pre-COVID, spiking at over 75 in 2020, and a long decline after that.

However, this is overly-simplistic and other data shows a pattern of higher unemployment and sustained, long-term unemployment. The ABS Labour Force data defines a person as employed if they have any employment at all (even 1 hour per week), which means that when people are minimally employed, or pick up the odd shift or week’s work they don’t show up as unemployed and the clock is reset on their duration of unemployment.

A very different story

What is actually happening for people on the margins of the labour market is much better captured by the Department of Social Security (DSS) data on those receiving unemployment payments: JobSeeker and Youth Allowance (Other). The 2022 September Quarter data shows that there were 839,000 people in receipt of JobSeeker and Youth Allowance, far more than the 504,000 unemployed in the ABS Labour Force data.

Approximately 183,000 (22%) of those receiving JobSeeker/YA had some form of earnings for the fortnight, and so would be excluded from the ABS definition of unemployed. However, the remaining 656,000 unemployed people in the DSS data is still a much higher number than the ABS estimate of unemployment, and in contrast to the ABS unemployment estimate, the total numbers of people receiving JobSeeker/YA has not yet returned to pre-pandemic levels. Further, as the graph below shows, the gap between the ABS employment data and the numbers of people receiving income support payments has widened substantially since COVID – probably reflecting a more fractured or precarious post-pandemic labour market.

Line graph comparison of DSS income-support data to ABS unemployment data. Both lines track closely from 2018 until COVID, then the number of income support recipients spikes more than ABS unemployment data and remains higher.

The DSS data not only shows higher levels of effective unemployment, but also of much longer duration. The ABS labour force data suggests around 25% of unemployed people are out of work for 12 months or longer, while the DSS data shows that some 71% of those receiving JobSeeker have been in receipt of the payment for more than a year – and 80% have been in receipt of JobSeeker or other income support payments for more than a year.

According to the DSS data, in September 2022 there were a massive 665,000 people on JobSeeker/YA who had been receiving income supports payments for more than a year. That is higher than the total ABS unemployment figure, and still higher now than prior to the pandemic. As the graph below shows, the numbers have reduced since the height of the pandemic, but it is not the rosy story painted by the ABS data. There remains a problem of entrenched and long-term unemployment.

Line graph comparing data on long-term unemployed where DSS data shows significantly higher rates of long-term unemployment.

Why are people unemployed in the long-term if there is work available?

One reason for the ongoing long-term unemployment is that, while macro-economic models posit a smooth relationship between economic growth and jobs growth leading to reduced unemployment, for individuals looking for work the process is anything but smooth. There is not a queue where you get to the font of and get a job when it is your turn – it is a competition for each job. And if you have a disability and only a partial capacity to work (or need additional supports) or if you have caring responsibilities, employers may simply prefer someone else (that they see as less complicated).

That is important because the DSS data shows:

  • Over 40% of unemployed people receiving JobSeeker are assessed as having only partial capacity to work only (i.e. they have some form of disability which prevents them working more than 30 hours a week)
  • Around 13% are the principal carers for children, with 80% of those people not having a partner and so being the principal carer on their own.

These people are at a considerable disadvantage in each individual employment competition, and someone still ends up in long-term unemployment if they finish second in every employment process!

The extent to which people who are unemployed are constantly beaten to jobs is shown in ABS data on changes in labour force status. This tracks the flows of people who are employed, unemployed or not-in-the-labour-force (NILF) in one month to see if their status changed in the next month. From this, we can track where people who get jobs are coming from.

From November 2021 to November 2022, the vast majority of jobs that went to people who were not already employed did not go to people who were unemployed. 72% of such jobs went to those who were previously not in the labour force, with only 28% going to those who were unemployed. The result was that on average only a quarter of those that were unemployed one month were employed by the next month, just over a half were still unemployed and the rest had dropped out of the labour force.

So, while unemployment levels did drop last year (in both ABS and DSS data sets), the labour market expansion did not simply soak up the unemployed. Rather, we saw:

  • extra hours going to workers who already had jobs (with the numbers of people employed but looking for more hours falling by about 20%),
  • people entering (or re-entering) the workforce getting the majority of jobs that did not go to those already in work, and
  • the long term unemployed remaining unemployed.

Or put more broadly, we can see that entrenched long-term unemployment is compatible with (and will not be fixed by) jobs growth when the unemployed are constantly or disproportionately beaten to jobs.

Implications of long-term unemployment

The existence of significant levels of entrenched long-term unemployment is important because the longer someone is unemployed the more likely savings will run out, housing and other basics become more difficult to secure, personal confidence declines, motivation and mental health challenges become more acute, skills and networks degrade, and the harder it becomes to get a job. In a society where identity and access to resources is fundamentally determined via the labour market, being unemployed sucks. And being long-term unemployed is damaging.

In a future post I will consider the implications of these significant levels of entrenched long-term unemployment for political-economic theory and policy, but finish here simply with a table summarising the September 2022 data which shows the extent of the problem.

 ABS Labour Force DataDSS Payment
Demographic Data
Unemployment (U)504,200839,421
Less than 12 months377,930174,724
More than 12 months126,300664,697
Unemployment Rate (% of LF)3.6%5.4%
Long Term U % of Total U25%79.2%
Median Duration of U/Payments13 weeks315 weeks*
* JobSeeker only