One of the extraordinary things that arose in a recent discussion about whether employment guarantees are better than leaving workers unemployed was the assumption that the costs of unemployment are relatively low compared to having workers engaged in activities of varying degrees of productivity. Some of the discussion suggested that there were “microeconomic” costs involved in having to manage employment guarantee programs (bureaucracy, supervision, etc) which would negate the value of any such program. The implicit assumption was that the unemployed will generate zero productivity if they are engaged in employment programs. There has been a long debate in the economics about the relative costs of microeconomic inefficiency compared to macroeconomic inefficiency. The simple fact is that the losses arising from unemployment dwarfed by a considerable margin any microeconomic losses that might arise from inefficient use of resources. in this blog, I discuss some of those issues.
I last visited this issue in some detail in this blog – The daily losses from unemployment. In terms of estimates, not much has changed in the US economy over the last 2 years. The daily losses in income alone are enormous.
One of the strong empirical results that emerge from the Great Depression is that the job relief programs that the various governments implemented to try to attenuate the massive rise in unemployment were very beneficial. At that time, it was realised that having workers locked out of the production process because there were not enough private jobs being generated was not only irrational in terms of lost income but also caused society additional problems, such as rising crime rates.
Direct job creation was a very effective way of attenuating these costs while the private sector regained its optimism. In fact, it took about 50 years or so for governments to abandon this way of thinking. Now we tolerate high levels of unemployment without a clear understanding of the magnitude of costs that that policy position imposes on specific individuals and society in general.
The single most rational thing a government could do was to ensure that there were enough jobs to match the available labour force.
In the growth period before the current crisis few countries had returned to the full employment states that they achieved in the Post World War 2 period up until the mid-1970s when the OPEC oil shocks led to a paradigm change in macroeconomic policy setting. The neo-liberal approach emphasised fiscal austerity to support an increasing reliance on monetary policy for counter-stabilisation.
So already in this period of relatively better economic growth there were substantial losses being recorded both in terms of lost GDP (production and income foregone) and the additional personal and social costs that accompany persistent unemployment.
It is well documented that sustained unemployment imposes significant economic, personal and social costs that include:
- loss of current output;
- social exclusion and the loss of freedom;
- skill loss;
- psychological harm, including increased suicide rate (which I will return to later);
- ill health and reduced life expectancy;
- loss of motivation;
- the undermining of human relations and family life;
- racial and gender inequality; and
- loss of social values and responsibility.
Many of these “costs” are difficult to quantify but clearly are substantial given qualitative evidence.
In the past I have done work on this with a colleague Martin Watts and you can see an early working paper (subsequently published but the working paper is free), which outlined the sort of method we used to compute the “costs of unemployment”.
There is a lovely quote from 1977 (Page 468) by the late James Tobin, which economists who read the blog will relate to:
Any economics student can expatiate on the inequities, distortions, and allocation of inefficiencies of controls or guideposts or tax rewards and penalties. But just consider the alternative. The microeconomic distortions of incomes policies would be trivial compared to the macroeconomic costs of prolonged underemployment of labor and capital. It takes a heap of Harberger triangles to fill an Okun Gap.
[Reference: Tobin, J. (1977) ‘How Dead is Keynes?’, Economic Inquiry, 15(4), 459-68].
Arnold Harberger was a University of Chicago economist/econometrician who coined the term “Hargberger Triangle” to denote the graphical area in a demand and supply graph that measures the deadweight loss arising from taxation. The triangle became an oft-used graphical device to portray microeconomic or so-called welfare losses in a variety of “markets”.
An Okun’s Gap (named after the US economist Arthur Okun) measures the loss in real output (income) arising from unemployment when an economy is operating below full employment because of demand constraints.
Harberger said that:
Tobin was implicitly attributing as a social cost the entire gap between actual and potential GDP, and by innuendo criticising the approach of measuring costs solely in terms of the types of distortions usually considered in applied welfare economics. In effect, he was saying that the distortions approach misses the point, that it fails to capture the lion’s share of the real costs of a shortfall of output.
The simplest calculation reveals that the daily income losses alone of having that many people idle dwarf any reasonable estimate of microeconomic losses arising from the so-called “structural inefficiencies” or microeconomic rigidities (a favourite of the IMF) that have dominated public debate over the neo-liberal era.
It is just plain madness to ignore huge costs and then go about pursuing small costs (if they exist). One of the problems is that in pursuing these micro costs the government almost always will increase the macroeconomic costs.
We know that the losses encountered during a prolonged recession reverberate into tortured recoveries and that the damage that unemployment causes spans the generations.
Even before the crisis hit, these costs in most countries were huge as policy makers began using unemployment as a policy tool rather than a policy target as the obsession with inflation-targetting took hold.
Most people do not consider the irretrievable nature of these losses. Every day that unemployment remains above the full employment level (allowing for a small unemployment rate arising from frictions – people moving in-between jobs) the economy is foregoing billions in lost output and national income that is never recovered.
The magnitude of these losses and the fact that most commentators and policy makers prefer unemployment to direct job creation, shows the powerful hold that neo-liberal thinking has had on policy makers. How is it rational to tolerate these massive losses which span generations?
As noted, to some extent these losses are a mystery to society in general. While the unemployed and their families are certainly aware of them, the remainder of the society are less aware. For example, we might notice rising crime rates in our neighbourhoods but do not associate it with unemployment.
Neo-liberalism has also changed the way we think about unemployment. In the past we understood clearly that it arose as a result of a shortage of jobs.
However, in recent decades, we have been conditioned by a relentless (lying) press and government statements to perceive unemployment as an individual problem. So the unemployed are type-cast as being lazy; having poor work attitudes; refusing to invest in appropriate skills; and subject to disincentives arising from misguided government welfare support, and all the rest of the arguments that mainstream uses to obfuscate the social problem.
The focus in the public debate is to “blame the victim” and suggest that most are unemployable and prefer to live on welfare, where that support is available.
The overwhelming evidence from the informed research literature is that almost all the unemployed (when surveyed) prefer to work and are willing to take work if offered.
The overwhelming evidence from studies in most countries suggests that the unemployed are highly motivated to find work and are victims of a shortage of jobs rather than personal/individual deficiencies.
The dominance of the neo-liberal ideology led governments in most countries to have eschew the adoption of policies of direct job creation to reduce the rate of unemployment and to minimise these massive costs. Fiscal policy has became geared to the achievement of budget surpluses as some sort of token of prudent financial management.
After the release of the OECD Jobs Study, employment policy shifted from a demand-side emphasis (ensuring there were enough jobs) to a supply-side focus (active labour market policy) which targetted individuals with futile training programs divorced from a paid-work context and pernicious welfare-to-work rules.
While the governments were pursuing labour market deregulation and attempting to retrench the Welfare State in the face of persistent unemployment they were also deregulating financial markets. The lack of oversight of the latter is the fundamental reason we are now enduring the worst crisis in more than 80 years.
While the previous work I have done on this topic has sought to compute the broadest possible measure of costs, in this blog I just focus on lost GDP income.
In focusing on the foregone output resulting from unemployment and underemployment you have to decide a benchmark to measure the current situation against.
In the work that follows, I used the low-point unemployment rate prior to the crisis – for the US 4.45 per cent in December 2006 and the high-point participation rate prior to the crisis – 66.3 per cent for the US.
The high-point participation rate allows us to adjust the labour force to eliminate the cyclical decline that occurs in a crisis as the discouraged workers who stop looking for work are not counted by the statistician as being unemployed. Economists call these people the hidden unemployed – they would work immediately if a job was offered to them.
The use of the low-point unemployment prior to the crisis is just a convenience and does not presupposes that this level was a full employment state.
The question being asked is: What are the daily GDP losses arising from the increases in the unemployment rate as a result of the crisis.
Some facts – the peak labour force participation rate in the US occurred in December 2006 (66.3 per cent). The low-point unemployment rate was achieved in the same month (4.45 per cent).
The real GDP peaked accrued in December 2007. The average gross in real GDP between December 2003 and 2007 was 2.8% per annum. We use this period as a representative period of stable growth.
Note, the calculations assume that the full employment unemployment rate is 4.45 per cent, which may be a conservative estimate. It certainly doesn’t bias the results upwards.
First, I computed potential employment – which is the employment level that would have existed if the participation rate had have remained at its peak value given the actual underlying population growth that has occurred.
So we’re correcting for the cyclical decrease in participation rates that has occurred as a result of the crisis.
The unemployment rate used here was the low-point unemployment rate that the US economy sustained in December 2006.
Second, I computed potential GDP using the potential employment levels multiplied by actual labour productivity (real GDP per person employed). So this series tells us what real GDP would have been if employment was at its potential level (defined above) and those workers were producing the average GDP per unit. Note that the concept of potential in this case does not mean the maximum GDP that could be produced.
Remember the selective way I am defining potential employment – as a departure from the maximum achieved prior to the crisis.
Further, some will ask – why use average productivity when it is obvious that the unemployed are typically drawn from the lowest productivity pool? Surely putting these people back to work will not generate average productivity per person. That is true but not relevant in this case. By using actual labour productivity series I am probably understating the production gains that would occur given that labour productivity slumps during a downturn.
But to allay fears, I also computed a 2nd potential GDP series, which assumes that all the unemployed above the 4.45% level, would achieve a productivity level 50% below the normal productivity levels achieved in the US economy.
It is of course impossible to determine the productivity profile of the unemployed. Some of the unemployed will have productivity close to the economy wide average, while others will have substantially lower productivity.
The calculations abstract from these differences by assuming that all the unemployed can only function at half the rate of the existing employed. I judge that this is a conservative assumption, which will not bias the results upwards.
Third, I calculated the daily real GDP losses in billions of dollars per day for the US by dividing the gap between actual and potential real GDP by the number of days in each quarter.
This approximates the daily loss in real income that is permanently foregone by allowing unemployment to remain either above the level that was achieved immediately prior to the crisis.
The following graph compares the actual real GDP series since the March-quarter 2006 (to September-quarter 2011) with the simulated real GDP series for the US based the assumption that real GDP continued to grow at the average rate of real GDP between December 2003 and the had held for the 8 quarters prior to the peak (that is, 2.8 per cent per annum).
The scale of the current crisis and lost income is clear from that graph.
The next graph shows the calculations of the daily losses in real GDP (national income) for the two different productivity assumptions employed.
The daily GDP losses that the US economy is enduring as a result of the decline in economy activity below it previous peak are shown in the following graph. You can see that in the September-quarter 2011 these stood at $US9.7 billion per day.
Just say it to yourself – every day the US government is allowing $US9.7 billion to go down the drain in lost income just because they are too stupid to implement sensible direct job creation strategies.
Even with the conservative productivity assumption the daily losses are of the order of $US6 billion per day.
There has been a marked unwillingness by the US government to engage in direct job creation. How can these deadweight daily losses be justified? The policy inaction is culpable in the extreme in this regard.
For the reasons I have explained in the introduction, these estimates are just the starting point – they only cover real income losses arising from foregone production.
While they are bad enough, once you start adding in the other losses that have itemised above – the social, health, crime, etc costs – then you have a national catastrophe on your hands.
Suicides and the business cycle
I read an interesting academic paper the other day which contributed to the large literature that finds that the rate of suicides rise when unemployment rises and vice versa.
The study – Impact of Business Cycles on US Suicide Rates, 1928–2007 – was published in American Journal of Public Health in June 20011 and written by researchers from the US Centers for Disease Control and Prevention,
The publication was accompanied by a CDC Press Release.
The study “examined the associations of overall and age-specific suicide rates with business cycles from 1928 to 2007 in the United States”.
They defined a crisis as “the state of affairs provoked by a sudden and severe economic recession” and pointed out that “(m)ost recessions in history have been mild”.
They suggest that the:
… association between suicide and unemployment can be traced back to Durkheim,15 who stated that unemployment weakens a person’s social integration and increases suicide risk.
In addition, they discussed the so-called vulnerability model “which suggests that unemployment may result in limited access to supportive re- sources, thereby increasing the impact of stressful events and then suicide risk.”
… indirect causative model indicates that unemployment may bring about relationship difficulties or financial problems that may lead to events precipitating suicide.
They do not dismiss the possibility that both suicide and unemployment are actually driven by a 3rd factor. The research aims to disentangle these influences.
I won’t discuss the statistical methodology here because, if you’re interested, you can read the paper yourself.
Overall, they found:
1. “the overall suicide rate generally increased in recessions, especially in severe recessions that lasted longer than 1 year”.
2. “The largest increase in the overall suicide rate occurred during the Great Depression (1929-1933), when it surged from 18.0 in 1928 to 22.1 (the all-time high) in 1932, the last full year of the Great Depression.”
3. “Not only did the overall suicide rate generally rise during recessions; it also mostly fell during expansions.”
4. “the overall suicide rate rose relatively in 11 of 13 recessions, suggesting the countercyclical nature of the overall suicide rate”.
In terms of policy recommendations the study provides three strategies:
– Providing social support and counseling services to those who lose jobs or homes
– Promoting individual, family, and community connectedness, ie greater degrees of social integration (e.g., number of friends, high frequency of social contact, low levels of social isolation or loneliness); positive attachments to community organizations like schools and churches; and formal relationships between support services and referring organizations help ensure services are actually delivered and promote a clients’ well-being (as in the case of the primary care system and the mental health system) all serve as protective factors against suicidal thoughts and behaviors.
– Increasing the accessibility of prevention services (e.g., crisis centers and other community services).
I found these recommendations to be useful but disappointing nonetheless.
If there is a strong relationship between the suicide rate and unemployment in the most effective response from government should be to reduce unemployment. Mass unemployment can only be reduced by job creation.
When private spending is weak, public spending has to fill the gap. The best way to attenuate the massive costs – both direct and indirect – that arise from unemployment, is for that public spending to be concentrated on job creation.
These insights bear on the debate about the relative costs and benefits of using unemployment versus employment buffer stocks to maintain price stability.
Please read my recent blogs – Whatever – its either employment or unemployment buffer stocks and MMT is biased towards anti-crony – for more discussion on this point.
In 1978 while still a student at the University of Melbourne and in recognition of the then rising economic and social costs of unemployment, I formulated an early model of what I now call the Job Guarantee approach to full employment.
It wasn’t a new thought – public sector job creation as an antidote to cyclical unemployment was the norm in those days. Others have since independently proposed large-scale job creation.
I considered that the JG would work under the principles of a buffer stock mechanism (which were common in agricultural policy during that period) and the jobs would be designed to increase per capita social welfare by satisfying social needs that are not met by the private sector in areas including environmental services, community and social services, health and education.
Thus this increase in public sector employment would contribute to the reduction in the negative externalities that tend to increase with increasing levels of production by increasing the share of final output that is associated with green, public sector employment.
But the buffer stock would also ensure price stability, a point that my friend Warren Mosler independently came up in the early 1990s. That recognition helped consolidate our working relationship since then that has evolved into Modern Monetary Theory with the input of several others (Randy, Stephanie, Pavlina, Mat, Scott to name the early contributors).
The current crisis should make it obvious that unemployment arises when there are too few jobs being created (net) relative to the available labour force.
But despite the transparency of the causes of mass unemployment governments have failed to use direct job creation as a way of dealing with the rising costs associated with the joblessness. The vestiges of neo-liberal ideology are proving hard to shake.
If governments had have had a Job Guarantee in place there would be hardly any discernible rise in official unemployment and the costs of the crisis would have been significantly reduced. Yes there would have been some skills-based structural underemployment (high skills workers taking Job Guarantee jobs) and some obvious loss of income.
But overall these losses would have been attenuated and some of the other costs would have been mostly negated. For example, most (if not all) workers would not have been in danger of losing their homes. The contagion factors in this crisis would have been limited.
Even under conservative assumptions, the economic and social costs of sustained high unemployment are extremely high. The inability of unemployed individuals and their families to function in the market economy gives rise to many forms of social dysfunction, in addition to output loss.
The apparent failure of neo-liberal supply side policies to reduce unemployment prior to the crisis is now highlighted during the crisis.
There is now an urgent need to address the large pools of unemployment in world economies.
The daily income losses alone are enormous and overwhelm other inefficiencies notwithstanding the productivity heterogeneity that exists across the workforce.
There is no financial reason why the government should not deal with this problem directly by introducing a Job Guarantee as a starting point. Then broader investment in public infrastructure could follow according to political preferences. If the Government had the political will, it could readily overcome the problem of persistently high unemployment.
The Saturday Quiz will be available sometime tomorrow. I hope everyone gets 5/5 so I can stop writing them.
That is enough for today!