13/7/19: Russian v European Dependency Ratios: 1950-2100

Doing some numbers crunching on a different project, I just came across this interesting database from the UN showing population projections through 2100. One interesting aspect of this data is the forecasts/projections for the dependency ratio - basically, a number of working age population per 100 people of non-working age.

There are caveats attached to the analysis of this data, including the changes in the duration of the working age (over the years, younger age dependency has moved toward 24 years from 19 years due to extended period spent in education, while for older age dependency, the mark has been moving from 64 years to 69 years as the last year in working age group). These caveats aside, here is a really eye-opening chart:

We consistently hear about the demographic catastrophe that has visited Russia since 1990-1991 collapse of the USSR. We are also constantly hearing the claim that the Russian society is demographically so challenged, it is running out of people. The chart above shows that, actually, that is not exactly true. Russia has been showing pretty decent readings on population dependency ratio compared to its peers ever since the mid-1970s. More so, through 2020, the estimates from the UN suggest that Russia is performing better than its peers in Europe in terms of overall dependency. This is expected to change - to the detriment of the Russian society and economy - in 2030-2040, but thereafter, Russia is expected to once again perform better than overall Europe.

Similar picture arises when one looks at more modern definition of dependency age ranges:

This data suggests that the popular narrative about the relative decline of Russian population dynamics compared to other European states is at least highly imperfect.

7/7/19: Employment to Population Rate in the U.S.: General Labor Force vs African Americans

With 'booming' and 'tight' labor markets, the White House is only happy to argue these days that we are in a Golden Era of employment/unemployment for all, including the African Americans. Is this, in fact, the case?

Firstly, I am not too keen on the arguments that any President in office should get the credit for jobs creation. At the very best, Presidential decisions simply support jobs creation by the private sector, and can be instrumental in creating jobs (albeit less in sustaining them) in the public sector. Secondly, jobs are just numbers, unless they are distinguished by their quality - something that is hard to do.

But the White House claims are usually about the aggregate jobs numbers / statistics, as opposed to the more granular analysis. So it might be worth taking them to the test.

One comparable - across different cohorts and time periods, as well as business cycles - metric is that of employment to population ratio. It takes total number in employment and divides it into the total population of working age for a specific group. Here is the chart (data from FRED database with calculations performed by myself):

Across the entire workforce, E-to-P ratio is sitting at 60.6%, statistically indistinguishable from the historical average of 60.64%, and 4.1 percentage points below all time high of 64.7%. For the category 'Black or African American', the E-to-P ratio is currently at 58.2%, which is above 55.13% historical average and is 3.2 percentage points below all time high of 61.4%.

Which means that current reading for African Americans population in terms of employment-to-population ratio is better, relative to their own historical trends than for the overall population. But, the ratio is still lower for African Americans than for the overall population in level terms.

What about the historical positioning of the gap between the overall population E-to-P ratios and that for the African Americans?

The gap between the employment-to-population ratio for the African Americans and that for the overall population is around the lowest levels it has ever been and is well below the historical average.

So, yes, the claims that employment has been relatively strong for both the general population and for the African Americans pans out to be true in at least this metric, which is - as noted above - by far not the only metric that matters.

30/11/18: Ireland’s Dependency Ratio Problem?

Ireland seems to have a twin dependency. or rather a triple dependency problem:

• Younger population means larger share of population is either below the working age or in education;
• Older population largely working less in their post-retirement age due to a number of factors, such as family/household work (‘grandparents duties’ in absence of functional childcare and early education systems), and tax effects (low thresholds for the upper marginal tax rate application act as disincentive to supply surplus labor over and above retirement income), plus the workplace practices and regulations that restrict post-retirement age work; and
• Working-age adults in large numbers drawing various forms of allowances (labor force participation rate being low for Ireland despite a relatively benign unemployment statistics).

All of which means that the aggregate (and very broad) dependency ratio for Ireland is yet to recover from the decade-old crisis, and is below that for other small, open economies, for example, Iceland:

The latter observation was true before the crisis, but the onset of the GFC and the Great Recession have pushed Ireland’s employment to population ratio to such dire lows that the country is yet to recover from its woes. Iceland recovered its pre-crisis levels of employment to population ratio back in 2016. It also endured much less pronounced impact of the crisis in terms of ratio decline (peak to trough) and duration of the peak-to-peak cycle. Ireland is still climbing out of the mess, and the rate of recovery is expected to slow down dramatically in 2018 (based on the IMF data).

While many observers and analysts are quick to discount this ratio, the reality is that economy’s resilience to shocks, its productive capacity today (and, via on-the-job training, learning by doing and other forms of career-linked investments in productivity growth, its future capacity) are determined by how many people work in the economy per capita of population. The lower the ratio, the less income producing capacity the economy has, the lower the absorption capacity of the economy in the face of adverse shocks.

30/7/18: Broader Unemployment vs Official Unemployment: Ireland

In the first post (see above) looking at the broader measures of unemployment and dependency ratios, recall that CSO publishes several extended series for broader unemployment rates. 

The official unemployment rate at 1Q 2018 stood at 6.4 percent (well within the pre-crisis historical range of the average of 5.31 percent and the 99% confidence interval of (3.70%, 6.92%). In more simple terms, statistically, the current official unemployment rate is indistinguishable from the average rate prevailing in 1Q 1998 - 4Q 2007. Which is the good thing, implying that in official terms, Irelands economy has recovered from the crisis at last. In fact, the recovery in official terms has been attained in 4Q 2017.

However, the CSO also reports the PLS2 measure of broader unemployment. The Above analysis was based on reported PLS1 data, covering unemployed plus discouraged workers, as a percentage of the labour force. Adding to the PLS1 those in potential additional labour force (basically able bodied adults who are neither employed nor unemployed, nor discouraged, and are not in studies or formal training), the CSO gets PLS2 measure of broader unemployment. In 1Q 208 this number read 10.2% of the Labour Force, plus Potential Additional Labour Force, which was statistically higher than the pre-Crisis average of 6.1% (falling into the 99% confidence interval range of (4.39%, 7.81%). In other words, the economy has not yet recovered from the Crisis based on PLS2 broader unemployment measure.

Extending PLS2 to cover all unemployed, plus those who want a job and not seeking for reasons other than being in education or training, in 1Q 2018 the broader PLS3 unemployment measure stood at 14.2 percent, unchanged on 4Q 2017. As with PLS2, the 1Q 2018 reading for PLS3 falls well beyond the range of the pre-crisis historical average of 8.36% (with 99% confidence interval of (6.52%, 10.20%).

As noted above, by two broader unemployment measures: PS2 and PLS3, Irish economy has not recovered from the crisis, even if we take a relatively benign recovery measure of the economy reaching the pre-crisis 1Q 1998 - 4Q 2007 average rate of unemployment. 

Worse, taking 4 year moving average and a 4 year rolling standard deviation in PS3 rates, 1Q 2018 PLS3 rate of 14.2% is closer to the upper margin of the 99% confidence interval for 1Q 2018 based on prior 4 years of data (the CI is given by (9.81%, 15.63%) range). Which means that 1Q 2018 data shows no statistically significant break-out from the PLS3 broader unemployment dynamics of the past 4 years. The same holds for the 5 years MA and rolling STDEV. 

So while the official unemployment readings are showing a very robust recovery, broader measures of unemployment continue to trend in line with the economy still carrying the hefty legacy of the recent crises. 

27/11/2014: QNHS Q3 2014: Employed, Unemployed & Retired

Key summary of the 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. 
3. 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. 

In this post, lets take a look at the QNHS breakdown by principal status.

Total population over 15 years of age stood at 3,595,600 in Q3 2014, which marks an increase of 0.08% y/y - a significant drop from 0.21% rise in Q2 2014. Compared to Q1 2014, total population over 15 years of age is down 0.1% (-3,500). In level terms, total population of age over 15 has increased 2,800 in Q3 2014 compared to Q3 2013 while in Q2 2014 y/y increase was 7,600.

Population at work stood at 1,859,500, close to Q4 2009 levels of 1,859,200 and up 1.7% y/y (+31,000). However, this marks a slowdown in the rate of growth in numbers at work from 2.17% y/y growth in Q2 2014 (+39,100). Since Q1 2011 some 58,500 more people are at work, although this reflects seasonal variations.

Unemployed numbers fell to 294,800 in Q3 2014, down 9.76% y/y which is a faster rate of decline than in Q2 2014 when numbers unemployed declined 9.67%. In level terms, the rate of decline y/y in Q2 and Q3 2014 remained identical at 31,900.

Student numbers rose 0.5% to 401,000 in Q3 2014 compared to the same period of 2013. This compares to a 0.89% decline y/y in Q2 2014. Currently, number of students is running at 0.5% below Q1 2011 levels.

Numbers engaged on home duties dropped to 470,300 - a decline of 1.63% y/y in Q3 2014, having previously dropped 1.77% in Q2 2014. Overall, there has been a dramatic drop in numbers of those engaged on home duties compared to Q1 2011 - down 11.3% or 59,900.

Numbers of those retired from employment rose once again to 416,700 in Q3 2014 - a rate of y/y increase of 2.76% (+11,200) y/y. This marks acceleration in the y/y increases compared to Q2 2014 when numbers retired grew at an annual rate of 1.61%. Since Q1 2011, numbers retired are up dramatically - rising by 68,400 or +19.64%.

Overall, numbers of unemployed, retired and other (including in state training programmes) stood at 864,800 in Q3 2014 - a decline of 2.58% y/y, but down only 100 (-0.01%) on Q1 2011.


As the result of the above changes, Ireland's dependency ratio improved slightly in Q3 2014, as shown in the chart below:

Overall, Q3 2014 dependency ratio was 40.34 individuals at work to 59.66 individuals not working for various reasons. This represents a slight improvement on H1 2014 ratio of 39.97 : 60.03. Nonetheless, dependency ratio remains higher than historical average of 42.58:57.42.

31/8/2014: Irish Dependency Ratio Rises in H1 2014

Previous posts covering QNHS release for Q2 2014 provided analysis of

• Breakdown of population over 15 years of age by their principal economic status: employed, unemployed, in retirement, students, engaged at home, and 'others': http://trueeconomics.blogspot.ie/2014/08/3182014-qnhs-by-principal-economic.html
• Unemployment and Participation Rates: http://trueeconomics.blogspot.ie/2014/08/3082014-both-unemployment-and.html
• Duration of Unemployment, including distribution by Age Cohorts: http://trueeconomics.blogspot.ie/2014/08/3082014-irish-unemployment-plight-of.html

In the present post, a quick summary of changes in overall dependency ratio. I define a ratio of those at work to total population. This is imprecise as we only have estimated population figures for the year, and year coverage is for 12 months through April each year. The data is directionally-indicative more than actual levels, so many caveats here. However it does show where this economy is heading in terms of how many people support via work the rest of the population.

Here are the trends, set against 5 year averages:

In H1 2014, the ratio of those at work to total population declined to 0.3997 from 0.4026 in H2 2013 and up on 0.3927 in H2 2013. This is less material than the fact that current average (from H1 2013 through H1 2014) is running at 0.3983 - the lowest for any 5 year period on record.

In simple terms, there are fewer people supporting through work larger number of those who do not work for any reason.

6/2/2014: Dependency Ratios Out to 2050

You've heard many analysts talking about the 'demographic dividend' for Ireland as if young people (we still have more of them than other countries in Europe) can be locked safely within the shores of this island. But supply of emigrants from this (and any other country) is driven, in part, by the demand for younger workers in the labour forces around the world. And here things are rather testy:

So that story of the 'demographic dividend'... do we have a policy plan for protecting or securing it here in Ireland? one can only keep hoping... 

7/6/2012: QNHS Q1 2012: Sectoral Decomposition

In the previous post I covered the top-of-the-line data on QNHS for Q1 2012. This time, lets take a look at some sub-trends by occupation and public v private sector numbers.

A handy summary table to outline changes by occupation:

Few surprises in the above table are:

• Twin (q/q and y/y) rises in Wholesale & Retail Trade category, 
• Y/y rise in Accommodation and food service activities with a level increase of 8,600. This appears to confirm the Government claims on the sectoral jobs creation on the foot of jobs stimulus. The problem with comparatives is that the y/y increase comes on foot of a sudden decline in employment in the sub-sector in Q1 2011 when it fell to surprisingly low, seasonally-unjustified level of 102,900. Sub-sector employment remains down on Q1 2010 when it stood at 123,700 or 12,100 ahead of Q1 2012, and it is down on Q4 2011 when it was at 113,400 against Q1 2012 at 111,600. The core factor in Q1 2012 differential on Q1 2011 might have been not so much jobs creation as the increased expense of jobs reductions under Budget 2012. This, however, is speculative argument at best. My suggestion would be to wait and see how the numbers employed in the sector pan out in Q2 2012.
• Another surprising thing is that in the category of skills closely aligned with Accommodation and food service activities there was a decrease, not an increase, y/y in terms of employment. Caring, leisure and other service category of workers saw employment drop from 142,300 in Q1 2011 to 141,500 in Q1 2012. Something is not adding up, unless the jobs created in the sub-sector were managerial and/or associate professional and technical.
• Not surprisingly, ICT sub-sector grew employment y/y with 6.81% increase on Q1 2011 - the only private sector sub-sector that posted an increase in jobs on 2007 levels (+5.31%), with only other two sectors adding jobs on 2007 levels being Education (+4.64%) and Human Health and Social Activities (+6.57%).
• For all the claims of MNCs employment gains, the core sub-sector of Professional, Scientific and Technical Activities has seen employment shrinking, not rising in Q1 2012 relative to Q4 2011 (-0.53%), to Q1 2011 (-7.56%) and on 2007 (-15.44%). Striking feature of these changes is that this sector was the hardest hit in Q1 2012 of all sub-sectors listed by CSO, amidst the robust IDA and Government claims that jobs creation in MNCs is ongoing and that R&D and innovation activities are booming.

In the core series for sub-sectors:

• There was a recorded rise in Education sub-sector. Employment in education stood at 144,200 in Q1 2012 - up 2.2% (or 3,100) on Q4 2011 and down 2.2% (-3,200) on Q1 2011. Since Q1 2007, employment in the sector grew by 4.64% or +6,400.
• Employment levels in Health and social work activities fell q/q by 1.96% (-2,000) but are up on Q1 2011 by 1.72% (+4,000). Compared to Q1 2007, Q1 2012 employment in the sector is up 6.57% (+14,600). 
• The two sectors above represent front-line services in their definition.  Between them, during the austerity period the two sub-sectors added 29,700 new jobs.

And lastly, two charts on dependencies ratios. Without any comment.