19/3/18: Drivers of the low labor force participation rate: U.S. data since 2001

Why doesn't U.S. economic expansion 'feel' like an economy is at full employment? Because of the low participation rate that has effectively reduced unemployment to superficially low levels without creating sufficient amount of quality jobs to offset the rise in working age population since the end of the Great Recession.

Here is a chart, via @ernietyedeschi, showing that the U.S. economic expansion has only recently started reducing the sticky non-participation drivers that remain in play since 2001:

The above is a fundamental problem for a range of advanced economies, not just the U.S. as I have noted in a number of previous posts, as well as a factor related to the secular stagnation thesis on both, demographic side (demand side secular stagnation) and technology / wages side (supply side secular stagnation).

27/7/17: The Gen-Lost is still lost...

Today, Marketwatch reported on a research note from Spencer Hill of Goldman Sachs Research claiming that the young workers cohorts in the U.S. have now caught up in terms of employment with older workers' cohorts.

Sadly, the argument is based on highly flawed analysis. The core data presented in support of this thesis is the unemployment rate, as shown in the chart below:

But official unemployment figures mask massive decline in younger cohorts' labor force participation rates, as evidence in this chart from Peterson Institute for International Economics:

In simple terms, when you reduce your employment base by moving people into 'out of workforce' category, you lower unemployment rate.  This is supported by other research, e.g. as reported here: http://trueeconomics.blogspot.com/2017/07/27717-work-or-play-snowflakes-or.html. Skewed, against the Millennials, workplace conditions are also to be blamed: http://www.epi.org/blog/young-workers-face-a-tougher-labor-market-even-as-the-economy-inches-towards-full-employment/. or as highlighted in these data:

Source: https://www.frbatlanta.org/chcs/labor-market-distributions.aspx?panel=1

So, no, beyond superficially deflated official unemployment metric, there is no evidence of the labor force conditions recovery for the younger workers. The Generation Lost is still lost. And that is before we consider the life cycle effects of the crisis.

17/7/17: New Study Confirms Parts of Secular Stagnation Thesis

For some years I have been writing about the phenomena of the twin secular stagnations (see here: http://trueeconomics.blogspot.com/2015/07/7615-secular-stagnation-double-threat.html). And just as long as I have been writing about it, there have been analysts disputing the view that the U.S. (and global) economy is in the midst of a structural growth slowdown.

A recent NBER paper (see here http://www.nber.org/papers/w23543) clearly confirms several sub-theses of the twin secular stagnations hypothesis, namely that the current slowdown is

1. Non-cyclical (extend to prior to the Global Financial Crisis);
2. Attributable to "the slow growth of total factor productivity" 
3. And also attributable to "the decline in labor force participation".

26/11/2014: QNHS Q3 2014: Participation and Unemployment Rates

Key summary of the two previous posts is:

1. Unemployment is falling across all durations and all demographic (age-defined) cohorts, but the pressure of long-term unemployment is rising in the cohort of older workers (40 years of age and older), and
2. Irish economy added 27,600 jobs in a year though Q3 2014 compared to Q3 2013. but only 17,300 of these jobs were private sector non-agricultural jobs. On longer-term trend: Non-agricultural Private Sector employment in Q3 2014 was 13.83 lower than 2008 average and Agricultural employment was 4.9% lower. In contrast, Public and State-controlled Sectors employment in Q3 2014 was 3.56% higher than 2008 average. 

Now, lets take a look at the Labour Force Participation Rate. CSO defines this as "The labour force participation rate is computed as an expression of the number of persons in the labour force as a percentage of the working age population. The labour force is the sum of the number of persons employed and of persons unemployed."

In other words, any serious improvement in employment conditions should have one of the two effects:
1) Unemployed moving into employment have zero effect on participation; and
2) New jobs added for new workers (including those previously in unemployment who have left labour force or have moved into training and are now gaining jobs or re-entering workforce) should increase participation.

What do we have in Q3 2014? Judging by the above stats on total employment and unemployment numbers, we should see an improvement in participation rates.

Based on CSO data, Q3 2014 participation rate stood at 60.4, down 0.3 points on Q3 2013. This marks an improvement on 0.5 points drop recorded in Q2 2013. Meanwhile, Unemployment rate dropped to 11.3% in Q3 2014 - a decline of 1.7 points on Q3 2013. The rate of unemployment decline slowed down in Q3 2014 compared to Q2 2014 when the rate dropped 2.1 points.

The above suggests that a significant share of the changes in unemployment is not related to jobs creation (something I will cover in a separate post).

For quarterly changes, consider seasonally-adjusted data. Seasonally-adjusted participation rate stood at 60.0 in Q3 2014, up marginally from 59.9 in Q2 2014 and marginally below 60.1 in Q1 2014. The historical average for the series is 60.7, which means we are still well below the average participation rate. Seasonally-adjusted unemployment rate posted 0.4 points drop to 11.1% in Q3 2014 compared to Q2 2014. Q2 2014 quarterly decline was 0.5 points, which suggests a slowdown in the rate of unemployment improvements in Q3 2014.

Rate of changes unemployment are shown in the chart below.

The key conclusion from the above data is that Ireland's participation rate remains below historical average and despite a slight improvement in Q3 2014 compared to Q2 2014, labour force participation rate remains lower than for the same period in 2013.

30/8/2014: Both Unemployment and Participation Rates Fell in Q2 2014

In the previous post, I covered duration of unemployment across age cohorts data for Q2 2014. This time around, let's take a look at labour force participation rates and unemployment rates.

As noted earlier, there are some good news in the latest QNHS data. And this theme continues with unemployment rate statistics.

Official unemployment rate has declined from 12% in Q1 2014 to 11.5% in Q2 2014 (seasonally-adjusted basis). Thus, Q/Q unemployment rate dropped by a substantial 0.5 percentage points, which is faster than the 0.2%points decline in Q1 2014 compared to Q4 2013. It is worth noting that in Q2 2013, Q/Q decline in seasonally-adjusted unemployment rate was shallower 0.1 percentage points. Overall, over H1 2014, unemployment rate declined by 0.7 percentage points compared to the end of 2013. Over the same period of 2013, decline was 0.6 percentage points.

Without seasonal adjustment, things are slightly different. Y/Y Q2 2014 unemployment rate is down 2.1 percentage points and in Q1 2014 the same decline was 1.7 percentage points. These rates are faster than the rates of y/y declines in unemployment for the same period of 2013.

All of which is good news as illustrated by the chart below:

Things are not as good on the participation rates front, however.

Non-seasonally-adjusted Participation Rate fell from 60.5 in Q2 2013 to 60.0 in Q2 2014. H1 2014 average participation rate was 59.85 against H1 2013 average of 60.0, H1 2012 average of 59.9 and H1 2011 average of 60.1, in H1 2010 the average was 60.8, in H1 2009 it was 62.2 and so on... Q2 2014 participation rate was lower than Q2 2011 reading of 60.5. By all metrics, participation rate has fallen and not just y/y, but also compared to other periods of the crisis.

On seasonally-adjusted basis, Participation Rate also fell, from 60.1 in Q1 2014 to 59.8 in Q2 2014. This marks the lowest Q2 participation rate since Q2 2000. Last time seasonally-adjusted participation rate expanded in Ireland was in Q2 2013.

Chart below to illustrate:

As chart above illustrates, the bad news is that participation rate is not growing and is basically flat since Q2 2011. Worse news is that it is now matching the crisis period low. Even worse things is that the participation rate is running below historical trend in every quarter since Q2 2010. Guess those higher taxes, as well as high cost of childcare for families, are keeping people out of the labour force (remember, labour force includes employed and unemployed workers who are searching for a job).

Or by definition: The labour force participation rate is computed as an expression of the number of persons in the labour force as a percentage of the working age population. The labour force is the sum of the number of persons employed and of persons unemployed.

So in basic terms, of every ten persons of working age, six are either employed or unemployed, and four are neither working nor seeking a job. At the peak of the Celtic Tiger, the ratio was 6.4 to 3.6.

5/3/2014: Broader Measures of Unemployment in Ireland: QNHS Q4 2013

Completing the coverage of Q4 2013 QNHS results for Ireland.

• First post covering detailed analysis of employment by sectors: http://trueeconomics.blogspot.ie/2014/02/2722014-employment-by-sectors-qnhs-q4.html
• Second post covering employment across broader sectors and categories: http://trueeconomics.blogspot.ie/2014/02/2822014-new-employment-across-broader.html
• Third post covering Participation and Unemployment Rates: http://trueeconomics.blogspot.ie/2014/02/2822014-participation-and-unemployment.html
• Fourth post covering duration of unemployment by age group: http://trueeconomics.blogspot.ie/2014/02/2822014-duration-of-unemployment-qnhs.html

Now, let's take a look at broader measures of unemployment.

Methodology note: CSO reports the following measures of broader unemployment:

• PLS1 indicator is unemployed persons plus discouraged workers as a percentage of the Labour Force plus discouraged workers. This indicator is broadly comparable to the previously published S1 indicator. In the nutshell, PLS1 = unemployed persons plus discouraged workers.
• PLS2 = PLS1 + Potential Additional Labour Force
• PLS3 = PLS2 + others who want a job, not available & not seeking for reasons other than being in Education and training.
• PLS4 = PLS3 + plus underemployed

In addition, I use CSO data from Live Register and emigration to add two more metrics:

• PLS4+STP: PLS4 + State Training Programmes Participant
• PLS4+STPE: PLS4 + State Training Programmes Participants + Emigration

So let's take a look first at labour force. The data is not seasonally-adjusted.

In Q4 2013 there were 2,163,000 people in labour force in Ireland, an increase of 19,600 year on year (+0.9%). A small increase, but a welcome one, suggesting that emigration is not offsetting demographic inflows of workers into the labour force. However, the level of labour force is still below Q4 2011 and is down 137,000 on pre-crisis peak. On average over the entire 2013, levels of labour force were 110,600 behind the pre-crisis period average levels.

As chart above shows, the greater challenge for us is the flat-trending labour force over the period of 2011-2013.

Table and chart below summarise changes in the broader measures of unemployment:

The key takeaway from the above charts and the table is shallower declines in the broadest measures of unemployment officially reported (PLS2-PLS4) compared to PLS1 and the adverse impact of 'sticky' State Training Programmes on the measure. It appears that these programmes are not moving workers out of unemployment fast enough.

Keeping in mind that Emigration is imputed only through April 2013 (we do not have official data beyond that), the PLS4+STP+E measure likely underestimates overall changes in broad unemployment.

Just how bad things are on overall unemployment front? With caveats to data and estimation errors, the above shows that in Q4 2013, 31.7% of Irish potential (including emigrants and state training schemes participants) labour force was either unemployed, underemployed, discouraged from seeking employment, not seeking employment for some reason other than being in education, toiling for free at Job Bridge and other fine 'activation' programmes or 'partying' abroad. Happy times...

That number, incidentally, is down from 32.5% in Q4 2012, but is still above 30.9 in Q4 2011.

28/2/2014: Duration of Unemployment: QNHS Q4 2013

Continuing with the coverage of Q4 2013 QNHS results for Ireland:
- First post covering detailed analysis of employment by sectors: http://trueeconomics.blogspot.ie/2014/02/2722014-employment-by-sectors-qnhs-q4.html
- Second post covering employment across broader sectors and categories: http://trueeconomics.blogspot.ie/2014/02/2822014-new-employment-across-broader.html
- Third post covering Participation and Unemployment Rates: http://trueeconomics.blogspot.ie/2014/02/2822014-participation-and-unemployment.html

In the current post let us consider changes in duration of unemployment.

The CSO reports two basic duration metrics:
- Less than 1 year,
- 1 year and over
And the data covers different age groups:
- All population 15 years and older
- Population 15-24 years of age,
- Population 25-44 years of age, and
- Population 45 years of age and older.

Here are the charts and quick commentary on these. The data is not seasonally-adjusted, so there is a lot of volatility and I am not going to do q/q analysis here.

For overall population 15 years of age and over:

• Q4 2013 total unemployment declined 14.05% (41,400 in level terms) compared to Q4 2012. In 2012, Q4 y/y decline was -6.1% (-19,300 in level terms). Thus, 2013 numbers are much better compered to 2012 numbers, as one should expect.
• Of these, unemployment numbers for duration less than 1 year have declined 18% (-20,900) in Q4 2013. Te good news - this reversed 2012 y/y rise of 2.0% (+2,300).
• Unemployment with duration over 1 year has fallen as well, in Q4 2013 this was down 11.8% (-20,900) compared to Q4 2012, which is a small gain on decline recorded in Q4 2012 (y/y -10% and -19,700).

So good news here is that numbers are declining for both long-term and short-term unemployed. However, while overall unemployment numbers have now fallen to the levels below those recorded in Q4 2009 (though ahead of those in Q4 2008), long-term unemployed numbers are down to the levels below those recorded in Q4 2010 and are way ahead of those in Q4 2009.

Since the current Government came to power (H1 2011), unemployed numbers for those over the age 15 are down 59,300, of which 36,050 declines came from the ranks of short-term unemployed and 22,050 declines came from the ranks of long-term unemployed.

Given the difficulty of reducing long-term unemployment compared to short-term unemployment, this is still a good record.

However, given that we do not know how many of long-term unemployed are gaining jobs vs how many are dropping out of the labour force (emigration or exits from workforce) we really have little to go in identifying how god the above aggregates really are.

Charts below plot unemployment by duration for different age groups.

Youth unemployment (15-24 years of age) is shrinking. Across total youth unemployment, in Q4 2013 numbers unemployed fell 17.3% y/y (down 10,200 in level terms), building on 12.9% (-8,700) decline in Q4 2012.

Compared to H1 2011, youth unemployment is down 19,950 (-29%) overall, with 11,000 of this decline coming from short-term unemployed figures and 8,450 from long-term unemployed ranks.

The problem with the above numbers is that we do not know the sources for these declines in youth unemployment. These, in addition to people gaining jobs, include demographic transition (entry of new young workers and exists of previously younger workers into the next category of 25-44 year olds), exits to and entries from education and training, including State Training programmes, emigration, including short-term migration on post-education visas and so on).

The weakest performance by age group is in the 45 and over category. Here, in Q4 2013 the numbers unemployed declined only 8.6% y/y (down 6,900 in level terms). The good news is that this reversed the rise in unemployment in this category recorded in 12 months through Q4 2012 (+3.0% and +2,300). Compared to H1 2011, numbers unemployed in this age group are still higher (+2.8% and +2,000). Numbers of long-term unemployed dipped in Q4 2013 only slightly (by 6.7% or -3,700) and compared to H1 2011 long-term unemployment in this age category is still up +11.4% or +5,250.

So overall, these are pretty solid numbers,with core reading showing that total number of unemployed for age 15 and over is currently at the lowest level for any Q4 period since (and including) Q4 2009.

Lastly, on severity of long-term unemployment, consider the chart plotting percentage of long-term unemployed in each age group total unemployment numbers:

This clearly shows that since around H2 2012, the positive trends in overall unemployment are broadly translating into symmetric reductions in unemployment for both short-term and long-term unemployed for all age groups. Again, this is a positive trend in the short run, as long-term unemployment is the hardest (or the stickiest) of all forms of unemployment and we can expect an upward trend in these charts. This was indeed the case in the period prior to H2 2012. Since then, however, we are seeing reductions in unemployment of relatively similar proportions for short-term and long-term unemployed.

28/2/2014: Participation and Unemployment Rates: QNHS Q4 2013

Having covered sectoral and broader aggregates QNHS results for employment levels (see post here for links: http://trueeconomics.blogspot.ie/2014/02/2822014-new-employment-across-broader.html), it is now time to take a look at participation rate and unemployment rate.

At the end of Q4 2013, seasonally-adjusted participation rate stood at 60.4%, up on 60.3% in Q3 2013 and 59.8% in Q4 2012. Compared to H1 2011 average (the period when the current Government came to power), the participation rate is up 0.3 percentage points.

Historical average for the seasonally-adjusted rate is 60.77% so we are just a whisker away from hitting this landmark. At the current rate of adjustment, we should be there around Q2-Q3 2014.

At the end of Q4 2013, official seasonally-adjusted unemployment rate fell to 12.1 from 12.7 in Q3 2013 (a very strong decline),  this marks the second sharpest correction downward in unemployment rate for the entire period of the crisis.

Over the year, seasonally adjusted unemployment rate has fallen full 2.1 percentage points from 14.2 to 12.1, improving on 2.0 percentage points decline y/y in Q3 2013. Compared to H1 2011, unemployment rate is now down 2.5 percentage points.

Q4 2013 marks the fifth consecutive quarter of declines in unemployment rate (q/q terms) and current rate is at the levels comparable with Q2 2009. All of which is very good news, despite all the possible caveats to data arising from potential impact of emigration and state training programmes participation (I will be covering these figures in my analysis of broader measures of unemployment as usual, next).

29/8/2013: Some positives v negatives from QNHS data: Q2 2013

Latest QNHS figures from Ireland are encouraging. Actually, given much of the tough news on the front of employment and jobs creation prior, these are heart warming. Here are the headlines:

Employment:

• "There was an annual increase in employment of 1.8% or 33,800 in the year to the second quarter of 2013, bringing total employment to 1,869,900. This compares with an annual increase in employment of 1.1% in the previous quarter and a decrease of 1.3% in the year to Q2 2012." This is good. Employment is up against adverse demographic effects, which is good, but it is also up due to superficial effects of reclassifications of some categories (see warning below).
• Even better news: "Full-time employment increased by 21,600 or 1.5% in the year to Q2 2013 while part-time employment increased by 12,100 or 2.8% over the year." So levels of increase in full-time employment are outstripping increases in part-time employment, implying that average jobs pool quality is not declining anymore.
• This marks third consecutive quarter of q/q increases in employment: "On a seasonally adjusted basis, employment increased by 9,600 (+0.5%) in the quarter." There was a seasonally adjusted increase in employment of 9,000 (+0.5%) in Q1 2013 and 12,100 (+0.7%) in Q4 2012.
• Employment increases and decreases composition are not sending a good signal, with higher value-added sub-categories of employment up: "Employment fell in five of the fourteen economic sectors over the year... The greatest rates of decline were recorded in the Administration and support service activities(-7.9% or -5,000), Transportation and storage (-5.4% or -4,900) and Public administration and defence; compulsory social security (-4.5% or -4,500) sectors. The largest rates of increase were recorded in the Agriculture, forestry and fishing (+18.7% or 16,300) and the Accommodation and food service activities(+8.0% or 9,600) sectors. 
• Here is a warning shot on the above figures: "In the case of the Agriculture, forestry and fishing sector it can be noted that estimates of employment in this sector have shown to be sensitive to sample changes over time." So, wait... +16,300 'new' jobs in Agriculture etc are really old jobs reclassified... or at least a large share of these are... Oops.. Note that this exactly matches decrease in the 'Not in the labour force' category (-16,300 y/y) and this knocks out quite a bit of wind out of the 'jobs creation' figures sails...

Unemployment:

• "The seasonally adjusted unemployment rate decreased from 13.8% to 13.7% over the quarter while the number of persons unemployed fell marginally by 500 persons, again on a seasonally adjusted basis." This is news in so far it is 'official' QNHS reading, but we knew 13.7% figure back in May when we had the standardised rate of unemployment estimate from Live Register.
• "Unemployment decreased by 22,200 (-6.9%) in the year to Q2 2013 bringing the total number of persons unemployed to 300,700. This is the fourth quarter in succession where unemployment has declined on an annual basis." Which is good news, indeed, except, wait... what about the 16,300 'new' jobs in Agriculture, Forestry & Fishing flagged above? Marginal decline of just 500 in terms of q/q seasonally-adjusted unemployment is a poor reading, to be honest. Better than an increase, but still, very weak. This weakness suggests that the bulk of 22,200 declines in unemployment rosters is due to exits and reclassifications of workers, not due to jobs creation.
• "The long-term unemployment rate decreased from 9.2% to 8.1% over the year to Q2 2013. Long-term unemployment accounted for 58.2% of total unemployment in Q2 2013 compared with 61.8% a year earlier and 56.1% in the second quarter of 2011." What we do not know here is whether this decrease was due to exits from benefits or entries into jobs or move to state-run training programmes. I will do analysis on these later, so stay tuned.

Labour force participation:

• Good news: "The total number of persons in the labour force in the second quarter of 2013 was 2,170,700, representing an increase of 11,500 (+0.5%) over the year. This compares with an annual labour force decrease of 19,600 (-0.9%) in Q2 2012." 
• The above is a good bit of news and it is made even better when we consider that increases in labour force were driven by increased participation rather than by demographic effects. In Q2 2013 there was a negative demographic effect cutting -16,300 from the overall labour force. This was more than offset by "a positive participation effect of 27,800 on the size of the labour force over the year.
• There was "an increase in the overall participation rate from 60.1% to 60.5% over the year to Q2 2013." Which is excellent news.
• "The number of persons not in the labour force in Q2 2013 was 1,415,600, a decrease of 16,300 (-1.1%) over the year." This seems to be related to reclassifications into Agriculture, etc. sector.

To summarise:

We have some positive news above, but overall, numbers remain obscured by reclassifications, changes in composition and lack of clarity on flows in- and out- of unemployment. 

Analysis of broader measures of unemployment, more indicative of underlying quality and nature of changes in the aggregate figures, is to follow, so stay tuned.