The Participation Puzzle. The regions being hit harder by an aging population

Over the past few years there has been much discussion about the impact that demographic changes (particularly an aging population) are having on the labour market. In particular, focus has been on the effects of varying participation rates within different age groups and the impact these are having on the level of employment.
First, some definitions (simplified) as used by the ABS:

  • Employed:- anyone over the age of 15 who worked for pay at least one hour in the reference week, or people who had a job but were away from paid work for less than 4 weeks, or were on strike or workers compensation etc.
  • Unemployed:- anyone over the age of 15 who does not have a job but who has actively looked for a job at any point during the previous 4 weeks period.
  • Labour Force:- the sum of those employed and unemployed.
  • Unemployment Rate:- the percentage of the Labour Force who are unemployed.
  • Participation Rate:- the percentage of the civilian population aged 15 and over who are in the Labour Force.

One effect of those definitions is that as the Participation Rate (PR) falls (i.e. a smaller percentage of the population are in the Labour Force) so the lower will be the unemployment rate for the same level of those employed. The inverse is also true.

Therefore, if we see significant changes to PR levels then it may be that the “headline” unemployment rate is no longer giving us a true and fair representation of what is happening in the labour market. For example, if the unemployment rate has risen only slightly, and yet the PR has fallen sharply, then the labour market is almost certainly far weaker than the headline unemployment rate would suggest.
Classically one sees the PR fall during periods when the labour market is weak. The reason for this can be easily understood; with high unemployment many workers will remove themselves from the Labour Force. They will become discouraged and stop looking for work. This results in the PR falling. Once the labour market returns to strength these people may well return to the Labour Force again and thus the PR moves back up.
However, we need to bear in mind that there may be reasons behind a change in the PR that have little to do with the actual health of the labour market. Rather they are due to changes in the demographic make-up of the population. To understand why this might be we need only consider a couple of examples. When we consider the age group of 15 to 24 year olds it would come as little surprise to discover that the PR will be depressed by a large number within this group being in tertiary education, and therefore not in the Labour Force. Alternatively, if we consider the “peak” age group of workers (i.e. 25 to 44 year olds) the PR will be high since the vast bulk of this group are likely to be in the Labour Force (either in work or looking for work). In particular, if we consider the older demographics (45+ years) it should be no surprise that the over 65’s in particular tend to have a lower PR; few of them are still working and those that aren’t are not looking for work (i.e. they are not “unemployed” and therefore not in the Labour Force).
What becomes clear therefore is that a change in PR will be due to the impacts of two factors;

  • the change of PR within age groups (the effects of discouragement etc.)
  • the change in the proportion of those age groups amongst the total population (the effects of demographic change)

As we can see from the graph below, the PR has shifted significantly throughout Queensland since 1998 with particular falls witnessed since the impact of the GFC. But how much of that change in the PR is due to demographic shifts and how much to weakness in the labour market?


Fortunately there is a way of splitting these two effects out from the data. The method was described and detailed in 2009 in a paper entitled Decomposing Changes in the Aggregate Labor Force Participation Rate written for the Federal Reserve Bank of Atlanta by Julie L. Hotchkiss. The precise method employed is not relevant for this discussion (although keen readers will be able to access the original paper here) but suffice to say that it involves weighting changes of the relative percentage of total population in each age group by the PR rate of that age group at the start of the comparison period. At the same time we weight the change in PR within each age group by the percentage of total population in that age group at the end of the comparison period. The sum of these two changes will total to the change of the total PR across all age groups between the comparison dates. We have used this process together with our own Conus Trend series for the three age groups (Youth 15-24, Middle 25-44 and Older 45+ yrs) for all the SA4 regions within Queensland. We have considered demographic changes since October 1998 to January 2017.
The chart below shows the original Conus PR along with the PR adjusted to remove the impact of the change in demographic make-up for both Rest of Queensland and Greater Brisbane. Therefore, what the dotted lines shows us is what the PR rate would have been had there been no change in the demographic make-up since October 1998.


In makes it clear that had we not seen these demographic changes the underlying PR’s across the state would have been somewhat higher. What it also demonstrates is that the impact of the changes (largely an aging population) has had a far greater effect outside of Greater Brisbane. Quite simply, over the period the Older cohort in Greater Brisbane has increased from 41% in 1998 to 45% now; in the Rest of Queensland it’s gone from 44.5% to almost 53% over the same period.
Since the end of 1998 Trend PR in Greater Brisbane has increased by 0.5%; however, demographic changes have decreased PR by 1.2%. The impact of increases in PR rates within the age groups has more than outweighed that decline to result in the net increase. In the Rest of Queensland over the period Trend PR has declined by 2.3% with demographic changes knocking 2.6% off PR with actual PR changes within the age groups adding only slightly.
If we consider the situation for our own region we see a similar pattern. Trend PR in Cairns has fallen by almost 8% since 1998 with demographic changes causing 2.5% of that decline (very similar to the impact across the Rest of Queensland as a whole) with the remainder of the fall coming from PR changes within the age groups.


The story to our south in Townsville is somewhat different. Here Trend PR has also fallen by 7.5% since 1998 but only a much smaller part of that decline (1.6%) can be laid at the door of demographic change; the rest is down to a decline in the PR within age groups.

7 replies
    • Pete Faulkner
      Pete Faulkner says:

      Thanks Gene. It’s an interesting area of study and while I don’t think this methodology necessarily captures all the elements, it does go some way to understanding the impact that demographic changes are having on our participation…and the disproportionate effect being felt in the regions. Thanks for the comment.

      Reply
  1. Glen
    Glen says:

    Pete, interested in how you think the large number of defence personnel in Townsville has affected this, in particular against Cairns which would be a more comparable analysis. The 7000 or so defence personnel in Townsville would be overwhelmingly younger people, as would their households which are not included in the ABS data for employment stats, but I presume they would be counted as an overall in any demographic stats.

    Reply
    • Pete Faulkner
      Pete Faulkner says:

      Glen, that’s a really good question and something I’ll need to think about (and perhaps ask the ABS how they handle). I think you’re right to assume that the defence personnel and families would be included in demographic stats, and we know they’re excluded from the labour force survey. How this is handled by the ABS, and what kind of impact it would have on comparative data…I’m not sure. I’ll try and get an answer and let you (and others) know what I can find out.
      Thanks for the comment, Pete

      Reply
    • Pete Faulkner
      Pete Faulkner says:

      Glen, having looked into this question in a little more detail it becomes clear that defence personnel are excluded from all the calculations. Working population measures only include civilians over the age of 15, so the PR rates provided by the ABS will relate only to the civilian population. How the ABS handle the fact that other members of defence personnel households (who could well be civilians, and therefore included in working population calculations) are not included in the labour force survey I am not yet clear on.

      Reply
      • Pete Faulkner
        Pete Faulkner says:

        I’ve crunched some numbers (just for sake of argument) using your 7,000 defence personnel figure. If we included those 7,000 people into the labour force as employed then on my calculations the Trend unemployment rate in Townsville would drop from 11.8% to 11.5%. In addition, if we include those 7,000 in the working population then the Trend PR rises from 59.7 to 61.2. As you point out, many of these 7,000 are likely to be in the youth cohort so the impacts could well be more dramatic here than overall. If we assume that civilian members of defence personnel households are, in general, no different to any other civilian cohort then their exclusion from the labour force survey shouldn’t have any material effect. Cheers, Pete

        Reply
    • Pete Faulkner
      Pete Faulkner says:

      Glen,
      I’ve checked with the ABS and can confirm that defence force personnel are excluded from the survey (and all series derived from that, which would include PR). Therefore, the impact on the data of excluding the defence force would be as per my response from the other day; it would reduce the unemployment rate by some 0.3% and increase PR by 0.5%. This impact is likely to be greater in the youth cohort but without data of the breakdown of the defence force personnel I cannot estimate any further. Clearly the relative impact (say to Cairns) would be of this scale only if Cairns has NO defence force personnel itself. Thanks for asking an interesting question and I hope I’ve gone some way to answering it. Cheers, Pete

      Reply

Leave a Reply

Want to join the discussion?
Feel free to contribute!

Leave a Reply

Your email address will not be published. Required fields are marked *