Australian labour force data – weaker and slowly deteriorating

Today’s release of the – Labour Force data – for May 2014 by the Australian Bureau of Statistics confirms, once again, how weak the labour market is. All the talk recently from the financial markets and the government about how the economy has ‘turned the corner’ and the lean times are behind us are plainly wrong. Today’s data confirms the stagnant situation that we have been witnessing for the last few years. Employment growth was negative and unable to keep up with the underlying population growth and unemployment rose modestly as a result. The participation rate fell again and held the rise in unemployment down by around 0.2 points. Workers are exiting the labour force because there are not enough job opportunities available. As a result, hidden unemployment rises as well. underemployment has risen since February by 0.2 points and the broad labour underutilisation rate (sum of underemployment and unemployment) stands at 13.5 per cent or more than 1.65 million workers. If we add the workers who have exited the labour market due to the lack of job openings then the total labour wastage will be well above 15 per cent. The other on-going disaster is the teenage labour market and that group fell further behind this month. The policy settings are wrong and the politicians are moving in the opposite direction to what is needed.

The summary ABS Labour Force (seasonally adjusted) estimates for May 2014 are:

• Employment increased 4,800 (0.9 per cent) with full-time employment increasing by 22,200 and part-time employment falling by 27,000.
• Unemployment increased by 3,200 to 717,100 and it would have been worse if the participation rate had not fallen.
• The official unemployment rate was unchanged at 5.8 per cent.
• The participation rate fell by 0.1 percentage points to 64.6 maintaining a downward trend in recent months. It is still well below its November 2010 peak (recent) of 65.9 per cent.
• Aggregate monthly hours worked increased by 26.5 million hours (1.68 per cent).
• The quarterly ABS broad labour underutilisation estimate (the sum of unemployment and underemployment) was published this month and underemployment rose by 0.2 percentage points to 7.6 per cent and total labour underutilisation rate was 13.5 per cent. There were 936.59 thousand persons underemployed. The total workers unemployment or underemployed is now 1.653 million.

Employment growth – negative again

The too-good-to-be-true March 2014 data, has now clearly been shown to be an aberration driven by the rotation bias in the sample (see the March commentary for the explanation from the ABS).

The last three month’s data shows that the economy is back to the norm – stagnation – which has prevailed for the last 24 or so months. Total employment fell by 4,800 (0.9 per cent), that is, barely moving although, on the bright side, full-time employment rose by 22,100.

Over the last 24 months or so we have seen the labour market data switching back and forth regularly between negative employment growth and positive growth spikes. This month’s figure, however, is the strongest overall performance in that time.

The following graph shows the month by month growth in full-time (blue columns), part-time (grey columns) and total employment (green line) for the 24 months to May 2014 using seasonally adjusted data. It gives you a good impression of just how flat employment growth has been.

While full-time and part-time employment growth are fluctuating around the zero line, total employment growth is still well below the growth that was boosted by the fiscal-stimulus in the middle of 2010.

The following table provides an accounting summary of the labour market performance over the last six months. The monthly data is highly variable so this Table provides a longer view which allows for a better assessment of the trends. WAP is working age population (above 15 year olds).

The conclusion – overall there have 74.4 thousand jobs (net) added in Australia over the last six months. Full-time employment has risen by 70.4. thousand jobs (net) while part-time work has risen by only 4 thousand jobs.

The Working Age Population has risen by 162 thousand in the same period while the labour force has risen by 86.1 thousand. The participation rate has fallen substantially (0.1 percentage points) over the last six months.

This has had the effect of attenuating the rise in unemployment.

The weak employment growth has thus not been able to keep pace with the underlying population growth and unemployment has risen by 10 thousand as a result.

The rise in unemployment would have been worse had the participation rate not dropped (see below).

Today’s data is consistent with the very weak trends that have ruled for a long period now.

To put the recent data in perspective, the following graph shows the movement in the labour force and total employment since the low-point unemployment rate month in the last cycle (February 2008) to May 2014. The two series are indexed to 100 at that month. The green line (right-axis) is the gap (plotted against the right-axis) between the two aggregates and measures the change in the unemployment rate since the low-point of the last cycle (when it stood at 4 per cent).

You can see that the labour force index has largely levelled off yet the divergence between it and employment growth has risen sharply (in spurts) over the last several months.

The Gap series gives you a good impression of the asymmetry in unemployment rate responses even when the economy experiences a mild downturn (such as the case in Australia). The unemployment rate jumps quickly but declines slowly.

It also highlights the fact that the recovery has not strong enough to bring the unemployment rate back down to its pre-crisis low. You can see clearly that the unemployment rate fell in late 2009 and then has hovered at the same level for some months before rising again over the last several months.

The Gap shows that the labour market is now in worse shape than it was at the peak of the financial crisis in 2009. After the government prematurely terminated the fiscal stimulus the situation has progressively deteriorated.

In fact, in May 2014, the Gap (2.2 percentage points) exceeded the levels that appeared in May and June 2009 when the Australian economy was enduring the impact of the crisis. All the gains made since then have thus largely disappeared due to poorly crafted fiscal policy not responding appropriately to non-government spending changes.

There is a slight upward bias in the unemployment rate.

Aggregate participation rate fell by 0.1 points

The May 2014 participation rate fell by 0.1 points to 64.6 per cent after falling by 0.3 points over the last three months. It remains substantially down on the most recent peak in November 2010 of 65.9 per cent when the labour market was still recovering courtesy of the fiscal stimulus.

In May 2014, employment growth was below the population growth.

To understand this, one must appreciate that the labour force is a subset of the working-age population (those above 15 years old). The proportion of the working-age population that constitutes the labour force is called the labour force participation rate. Thus changes in the labour force can impact on the official unemployment rate, and, as a result, movements in the latter need to be interpreted carefully. A rising unemployment rate may not indicate a recessing economy.

The labour force can expand as a result of general population growth and/or increases in the labour force participation rates.

In the current month, the unemployment rate stayed unchanged at 5.8 per cent.

What would have the unemployment rate been had the participation rate not fallen by 0.1 points?

The following Table shows the breakdown in the changes to the main aggregates (Labour Force, Employment and Unemployment) and the impact of the fall in the participation rate.

In May 2014, employment fell by 4.8 thousand while the labour force fell by 1.6 thousand. As a result, unemployment rose by only 3.2 thousand.

The rising labour force in May 2014 was the outcome of two separate factors:

• The underlying population growth added 17.4 thousand persons to the labour force. The population growth impact on the labour force aggregate is relatively steady from month to month; and
• The fall in the participation rate meant that 19 thousand workers left the labour force (relative to what would have occurred had the participation rate remained unchanged).

So while employment growth was just below the underlying population growth, the falling participation took the pressure off somewhat as workers exited the labour force (19 thousand) and were taken out of the official unemployment count.

If the participation rate had not have fallen, total unemployment, at the current employment level, would have been 736.1 thousand rather than the official count of 717.1 thousand as recorded by the ABS – a difference of 19 thousand workers (the ‘participation effect’).

Thus, without the fall in the participation rate, the unemployment rate would have been around 6 per cent (rounded) rather than its current value of 5.8 per cent.

The conclusion is that hidden unemployment rose and this attenuated the rise in the official unemployment rise. In functional terms this signals a much worse deterioration in the conditions than signalled by the current official unemployment rate.

There is considerable monthly fluctuation in the participation rate but the current rate of 64.6 per cent is a long way below its most recent peak in November 2010 of 65.9 per cent.

What would the unemployment rate be if the participation rate was at that recent November 2010 peak level (65.9 per cent)?

The following graph tells us what would have happened if the participation rate had been constant over the period November 2010 to May 2014. The blue line is the official unemployment since its most recent low-point of 4 per cent in March 2008.

The red line starts at November 2010 (the peak participation month). It is computed by adding the workers that left the labour force as employment growth faltered (and the participation rate fell) back into the labour force and assuming they would have been unemployed. At present, this cohort is likely to comprise a component of the hidden unemployed (or discouraged workers).

Total official unemployment in May 2014 was estimated to be 717.1 thousand. However, if participation had not have fallen since November 2010, there would be 945.2 thousand workers unemployed given growth in population and employment since November 2010.

The unemployment rate would now be 7.6 per cent if the participation had not fallen below its November 2010 peak of 65.9 per cent.

The difference between the two numbers mostly reflects the change in hidden unemployment (discouraged workers) since November 2010. These workers would take a job immediately if offered one but have given up looking because there are not enough jobs and as a consequence the ABS classifies them as being Not in the Labour Force.

Note, the gap between the blue and red lines doesn’t sum to total hidden unemployment unless November 2010 was a full employment peak, which it clearly was not. The interpretation of the gap is that it shows the extra hidden unemployed since that time.

As the participation rate dropped over the period, the gap rose.

This is quite a different picture to that portrayed by the official summary statistics.

Full-time and Part-time employment in recovery

The following graph shows employment indexes for the last 3 recessions and allows us to see how the trajectory of total employment after each peak prior to the three major recessions in recent history: 1982, 1991 and 2009 (the latter to capture the current episode).

The peak is defined as the month of the low-point unemployment rate in the relevant cycle and total employment was indexed at 100 in each case and then indexed to that base for each of the months as the recession unfolded.

I have plotted the 3 episodes for 74 months after the low-point unemployment rate was reached in each cycle. The current episode is now in its 74th month.

The initial employment decline was similar for the 1982 and 1991 recessions but the 1991 recovery was delayed by many month and the return to growth much slower than the 1982 recession.

The current episode is distinguished by the lack of a major slump in total employment, which reflects the success of the large fiscal stimulus in 2008 and 2009.

However, the recovery spawned by the stimulus clearly dissipated once the fiscal position was reversed and the economy is now producing very subdued employment outcomes.

Moreover, the current episode is also different to the last two major recessions in the sense that the recovery is over and the economy is deteriorating again.

The next 3-panel graph decomposes the previous graph into full-time and part-time employment. The vertical scales are common to allow a comparison between the three episodes.

First, after the peak is reached, part-time employment continues to increase as firms convert full-time jobs into fractional jobs.

Second, recoveries are dominated by growth in part-time employment as firms are reluctant to commit to more permanent arrangements with workers while there is uncertainty of the future course in aggregate demand.

Third, the current recovery is clearly mediocre by comparison, with both very subdued growth in full-time and part-time work.

Teenage labour market – continues to deteriorate

Full-time employment for teenagers rose by 4.7 thousand in May 2014, while part-time employment fell by 7.1 thousand. Overall, teenage employment fell by 2.4 thousand in May 2014. So some good news in the full-time growth.

The following graph shows the distribution of net employment creation in the last month by full-time/part-time status and age/gender category (15-19 year olds and the rest)

If you take a longer view you see how poor the situation is.

Over the last 12 months, teenagers have lost 37.9 thousand jobs while the rest of the labour force have gained 136.7 thousand net jobs. Remember that the overall result represents a very poor annual growth in employment.

Even more disturbing is the attrition of full-time jobs among teenagers – losing 4.7 thousand over the last year.

The teenage segment of the labour market is being particularly dragged down by the sluggish employment growth, which is hardly surprising given that the least experienced and/or most disadvantaged (those with disabilities etc) are rationed to the back of the queue by the employers.

The following graph shows the change in aggregates over the last 12 months.

To further emphasise the plight of our teenagers, I compiled the following graph that extends the time period from the February 2008, which was the month when the unemployment rate was at its low point in the last cycle, to the present month (May 2014). So it includes the period of downturn and then the so-called “recovery” period. Note the change in vertical scale compared to the previous two graphs.

Since February 2008, there have been only 917 thousand (net) jobs added to the Australian economy but teenagers have lost a staggering 117.3 thousand over the same period. It is even more stark when you consider that 97.2 thousand full-time teenager jobs have been lost in net terms.

Even in the traditionally, concentrated teenage segment – part-time employment, has shed 20.1 thousand jobs (net) even though 521.3 thousand part-time jobs have been added overall.

Overall, the total employment increase is modest. Further, around 54 per cent of the total (net) jobs added since February 2008 have been part-time, which raises questions about the quality of work that is being generated overall.

To put the teenage employment situation in a scale context the following graph shows the Employment-Population ratios for males, females and total 15-19 year olds since March 2008 (the month which coincided with the low-point unemployment rate of the last cycle).

You can interpret this graph as depicting the loss of employment relative to the underlying population of each cohort. We would expect (at least) that this ratio should be constant if not rising somewhat (depending on school participation rates).

The facts are that the absolute loss of jobs reported above is depicting a disastrous situation for our teenagers. Males, in particular, have lost out severely as a result of the economy being deliberately stifled by austerity policy positions.

The male ratio has fallen by 11.3 percentage points since February 2008, the female by 7 percentage points and the overall teenage employment-population ratio has fallen by 9.2 percentage points.

Overall, the performance of the teenage labour market continues to be deeply disturbing. It doesn’t rate much priority in the policy debate, which is surprising given that this is our future workforce in an ageing population. Future productivity growth will determine whether the ageing population enjoys a higher standard of living than now or goes backwards.

The longer-run consequences of this teenage “lock out” will be very damaging.

The problem is that in the modest growth period that the Australian labour market enjoyed as a result of the fiscal stimulus and mining investment, teenagers failed to participate in the gains – they went backwards.

Now, with the economy entering a new period of slowdown, these losses will be added too given that teenagers are among the first in line to be shown the door by employers seeking to reduce staff levels in the face of declining aggregate sales.

The previous Government’s response was to push this cohort into endless training initiatives (supply-side approach) without significant benefits. The research shows overwhelmingly that job-specific skills development should be done within a paid-work environment.

I would recommend that the new Australian government immediately announce a major public sector job creation program aimed at employing all the unemployed 15-19 year olds, who are not in full-time education or a credible apprenticeship program.

It is clear that the Australian labour market continues to fail our 15-19 year olds. At a time when we keep emphasising the future challenges facing the nation in terms of an ageing population and rising dependency ratios the economy still fails to provide enough work (and on-the-job experience) for our teenagers who are our future workforce.

Unemployment – rises

The unemployment rate remained unchanged at 5.8 per cent in May 2014. Official unemployment rose by 3,200 to 717,100 and it would have been more had not the participation rate fell again (see analysis below).

Overall, the labour market still has significant excess capacity available in most areas and what growth there is is not making any major inroads into the idle pools of labour.

The following graph updates my 3-recessions graph which depicts how quickly the unemployment rose in Australia during each of the three major recessions in recent history: 1982, 1991 and 2009 (the latter to capture the 2008-2010 episode). The unemployment rate was indexed at 100 at its lowest rate before the recession in each case (January 1981; January 1989; May 2008, respectively) and then indexed to that base for each of the months as the recession unfolded.

I have plotted the 3 episodes for 76 months after the low-point unemployment rate was reached in each cycle. The current episode is now in its 76th month (0 being February 2008). For 1991, the peak unemployment which was achieved some 38 months after the downturn began and the resulting recovery was painfully slow. While the 1982 recession was severe the economy and the labour market was recovering by the 26th month. The pace of recovery for the 1982 once it began was faster than the recovery in the current period.

It is significant that the current situation while significantly less severe than the previous recessions is dragging on which is a reflection of the lack of private spending growth and declining public spending growth.

Moreover, the current episode is also different to the last two major recessions in the sense that the recovery is over and the economy is deteriorating again.

The graph provides a graphical depiction of the speed at which the recession unfolded (which tells you something about each episode) and the length of time that the labour market deteriorated (expressed in terms of the unemployment rate).

From the start of the downturn to the 76-month point (to May 2014), the official unemployment rate has risen from a base index value of 100 to a value 145. After falling steadily as the fiscal stimulus pushed growth along (it reached 122.5 after 35 months – in January 2010), it has been slowly trending up for some months now. Unlike the other episodes, the current trend, at this stage of the cycle, is upwards.

It is now above the peak that was reached just before the introduction of the fiscal stimulus. In other words, the gains that emerged in the recovery as a result of the fiscal stimulus in 2009-10 have now been lost.

At 76 months, 1982 index stood at 150 while the 1991 index was at 148.3. It is clear that at an equivalent point in the “recovery cycle” the current period is more sluggish than our recent two major downturns and trending upwards while the trend in the earlier episodes was moderately downwards.

It now appears that the recoveries have converging, which tells us that the current policy has failed to take advantage of the fact that the latest economic downturn was much more mild than the previous recessions. In other words, the policy failure is locking the economy into a higher unemployment rate than is desirable and otherwise attainable.

Note that these are index numbers and only tell us about the speed of decay rather than levels of unemployment. Clearly the 5.8 per cent at this stage of the downturn is lower that the unemployment rate was in the previous recessions at a comparable point in the cycle although we have to consider the broader measures of labour underutilisation (which include underemployment) before we draw any clear conclusions.

The notable aspect of the current situation is that the recovery is very slow.

Broad labour underutilisation

The ABS published its quarterly broad labour underutilisation measures in this data release.

In the May-quarter, total underemployment rose by 0.2 percentage points to 7.6 per cent and the ABS broad labour underutilisation rate remained at 13.5 per cent (the sum of unemployment and underemployment).

There were 936.5 thousand persons underemployed. Overall, there are 1.653 million workers either unemployed or underemployed.

The following graph plots the history of underemployment in Australia since February 1978.

If hidden unemployment is added to the broad ABS figure the best-case (conservative) scenario would see a underutilisation rate well above 15 per cent at present. Please read my blog – Australian labour underutilisation rate is at least 13.4 per cent – for more discussion on this point.

The following graph shows the same type of indexes as the previous graph except it uses the ABS broad labour underutilisation rate (unemployment plus underemployment). It also is in terms of quarters rather than months.

We also show the full evolution of the the 1982 and 1991 episodes from the low-point (= 100) through the peak and back to the next low-point. In the case of the 1982 recession the index had risen to a peak of 172.8 in May 1983 (a broad underutilisation rate of 14.4 per cent) and then fell back to 9.8 per cent by November 1989 (index value 117.4).

At that point, the cycle turned down again signalling the beginning of the 1991 recession. That cycle reached a peak of 185 (or 18.1 per cent in November 1992) and it took until February 2008 for it to reach the low-point of 9.9 percent (an index value of 100.9). That point marked the beginning of the next cycle.

That should tell you how severe the 1991 recession was and how asymmetric the labour market response is on either side of the cycle. From its start in November 1989 it took 74 quarters (18.5 years) to return to more or less that level.

In terms of the three recession comparison, at the same period in the recovery (using quarterly data), the broad labour underutilisation rate (unemployment plus underemployment) had an index value of 147 in the 1982 recession (absolute value of 12.2 per cent); an index value of 154.1 in the 1991 recession (absolute value of 15.1 per cent); and an index value of 136.4 in the current period (absolute value of 13.5 per cent).

So while the level of unemployment is much lower now than in the 1982 recession (at a comparable stage), underemployment is now much higher and so the total labour underutilisation rates is higher. Further, the 1982 recovery in broad underutilisation terms was more robust than the current stagnating situation.

Commentators who think of the 1982 recession as severe, rarely see it in these terms. Joblessness is probably worse than underemployment but both mean that labour is wasted and income earning opportunities are being foregone. For a worker with extensive nominal commitments, the loss of income when hours are rationed may be no less severe than the loss of hours involved in unemployment, if the threshold of solvency is breached.

The next update will be in the August Labour Force data release.

Hours worked – rose in May 2014

Aggregate monthly hours worked rose by 26.5 million hours (1.68 per cent) in seasonally adjusted terms, which somewhat reversed the large fall last month

The following graph shows the trend and seasonally adjusted aggregate hours worked indexed to 100 at the peak in February 2008 (which was the low-point unemployment rate in the previous cycle).

The next graph shows the monthly growth (in per cent) over the last 24 months. The dark linear line is a simple regression trend of the monthly change – which depicts the shallowest upward trend you will ever see. You can see the pattern of the change in working hours is also portrayed in the employment graph – zig-zagging across the zero growth line.

Conclusion

In general, we always have to be careful interpreting month to month movements given the way the Labour Force Survey is constructed and implemented. But today’s data release is consistent with the message that has been generated for at least the last 2 years – the stagnant state of the labour market is on-going.

In May 2014, employment growth was negative repeating the pattern of zig-zagging across the zero growth line that has been evident for long period now. Overall employment growth is not even keeping pace with the underlying population growth and the result has been a further rise in unemployment (albeit modest).

The rise in unemployment would have been worse if the participation rate had not fallen again. The participation rate effect reduced the unemployment rate by 0.2 points on what would have been the case if the 19 thousand workers hadn’t dropped out of the labour force, presumably due to the lack of job opportunities.

Further, underemployment has risen since February by 0.2 points and the broad labour underutilisation rate (sum of underemployment and unemployment) stands at 13.5 per cent or more than 1.65 million workers. If we add the workers who have exited the labour market due to the lack of job openings then the total labour wastage will be well above 15 per cent.

What cannot be understated or uncertain is the continuing appalling performance of the teenage labour market. Employment has collapsed for that cohort since 2008. I consider it a matter of policy urgency for the Government to introduce an employment guarantee to ensure we do not continue undermining our potential workforce.

This month, teenagers lost further ground in the labour market.

The neglect of our teenagers will have a very long memory indeed and the negative consequences will be stronger given the ageing population.

The data certainly shows that the government has to reverse the contractionary bias or else conditions will condition to slowly deteriorate.

It is time for the conservatives to abandon their free market/anti-government biases and do some public sector job creation, which will arrest the current decline and stimulate private sector activity and employment!

That is enough for today!

(c) Copyright 2014 Bill Mitchell. All Rights Reserved.