One weird trick that proves the IGR is nonsense

I have a piece in today’s Guardian, written before the release of the Intergenerational Report and making the case that the intergenerational equity problem, as it was conceived in the 1980s and 1990s has already been resolved. Key quote

The resolution of the intergenerational fiscal problem was a major public policy achievement of the reform era of the 1980s and 1990s. But a political class still fixated on the most ideological version of the reform agenda, in which cutting public spending is desirable in and out of season has refused to drop the club of intergenerational equity. The idea that (very modest) budget deficits and public debt levels constitute “robbing our children” remains a staple in calls for “reform”.

Having seen the IGR, there’s a single statistical choice that shows the entire exercise to be worthless. The key issue in all this is whether changes in our demographic structure will create an unreasonable fiscal burden. The Report chooses to summarise this by reference to what it calls the “dependency ratio”, defined as the ratio of people aged over 65 to those aged 15-64.

In what kind of world would this make sense? Essentially, one in which
* Children aged 14 and under cost nothing to raise and required no public expenditure on schools, daycare etc
* Children leave school at 15. After this, they not only support themselves, but contribute to the support of those over 65
* People retire become eligible for age pensions at 65

The first of these assumptions is obviously silly, but it has become more so over time as the cost of raising children and of school education has risen steadily, outstripping growth in average incomes. As regards schools, back in the 1960s the standard class size was 40 (at least in SA where I went to school), and there were hardly any ancillary staff. Since then, the staff-student ratio would have doubled

The second assumption was broadly accurate in the 1960s, when measures like this were introduced. But it’s totally out date today, when high school completion is the norm and some post-secondary education is expected

The third assumption is particularly striking. The IGR is all about the supposed burden of age pensions, but they neglect to mention that the pension age will be 67 in 2023, and will almost certainly increase further after that.

The reason for this silliness is obvious. If children and increased retirement ages are taken into account, the dependency ratio in 2050 will be very similar to that in the 1960s, which we managed with ease.

I’ve accused the IGR of reflecting the dead hand of the reform ideology of the 1980s, but it seems as if their thinking really belongs in the 1960s, when Abbot was in school and Hockey in nappies. That’s one kind of intergenerationalism, I guess.

37 thoughts on “One weird trick that proves the IGR is nonsense

  1. If the “working age definition” was shifted to 20-70 rather than 15-65 it would accomodate JQ’s points, but probably not really change the numbers very much. The key statistics are longevity and birth and immigration rates.

  2. The working age against which you count the dependents needs to be better: but that wasn’t JQ’s only, or dominant, point. He said you have to count the dependent young as well as the dependent old – and if you do so there turns out not to be a huge swing.
    Terje has acknowledged this; Nevil Kingston-Brown should do the same.
    Of course it really changes the numbers. And it really changes them very much.
    Then there’s the unmentioned problem – how many actually work depends on ending the nostrum of ‘necessary unemployment’ that isn’t a spur to growth, just the mechanism for cutting the wage share and funding transfer of power and wealth to the already powerful and already wealthy.

  3. Terje has acknowledged this; Nevil Kingston-Brown should do the same.

    Close but not quite. I’ve said the report includes conflicting definitions and the problem may just be a clerical error in one of the footnotes. But it could also be that the report actually does have the flaw as described by JQ. At this point I’m still on the fence as to which class of problem this is.

  4. So many factors could change between now and the time when the IGR projects out to, 2055, that all the projections are meaningless anyway even if the parameters assumed now did seem well defined and reasonable (which they don’t anyway).

    1. What will climate change do to the world by 2055?
    2. What will Limits to Growth do to the world by 2055?
    3. What will geopolitical events and wars etc. do to the world by 2055?
    4. Will a nation to north of us invade us by 2055?
    5. Will limited or general nuclear war have happened by 2055?
    6. Will a major flu pandemic comparable to Spanish Flu 1918 happen before 2055?
    7. Will some other great disease pandemic happen before 2055?
    8. Will a major world wide economic depression occur before 2055?

    Although it is hard to say what major negative black swan event will happen by 2055, the chances seem high (I would say 50% at least) that some major negative black swan event will happen before 2055. It is even possible, though less likely I think, that a major positive black swan event could happen before 2055. Projecting current trends out that far is “beyond ridiculous” as the saying goes.

  5. Correction, I should have said:

    “Projecting current demographic and economic trends out that far is “beyond ridiculous”.

  6. Donald Oats said it all. Just a giant Ponzi Scheme, to echo down through the ages, like rattling chains on a cold Elsinor night.

    Use and abuse of economics.. a cause for the most myopic, arcane, self serving and least visonary thinking of all; no wonder it is called the “Dismal Science

  7. Ikonoclast :

    At the other end, with the economy in the toilet, your chances of being employed over 60 are not good.

    Presumably that will change as the population ages.

  8. @Ron E Joggles
    I’ve already used a big texta to write “THIS SPOT RESERVED FOR MR D OATS FROM 2025 ONWARDS” on a big cardboard box, and put it under the Torrens River bridge: that is an investment in my future.

    Cardboard City: building better futures.

  9. It seems to me that the question of gov’t deficits is secondary at best to this question. Suppose, for example, that all of the retirees are rich. If there are not enough workers to supply their needs, what good do their riches do? And if there are enough workers to supply the needs of the retirees and everyone else, then why condemn the elderly to poverty?

    And what about productivity? If productivity increases as much as the dependency ratio, what is the problem?

  10. I’m in the middle of reading The Imaginary Time Bomb: Why An Ageing Population Is Not A Social Problem, by Phil Mullan, published in 2000.

    Most of the points made in the book have already been made here (and most of the points made here are also made in the book), so I won’t repeat them, but as those points are supported with more extensive analysis and documentation in the book I thought it might interest some people.

  11. TerjeP (and others) – re the ‘conflicting definitions’:

    It is perfectly routine to have multiple different definitions of dependency ratios, just like their are different definitions of life expectency (period vs. cohort as the most obvious example). Traditionally, definitions have been as follows – same as on Wikipedia:

    Total Dependency Ratio: young (0-14) + old (65+) / working age (15-64)
    Aged Dependency Ratio: old / working age
    Child Dependency Ratio: young / working age

    As I understand it, John’s point is two-fold (as ChrisH indicates). First, the IGR focuses almost exclusively on the aged dependency ratio, when the Total Dependency Ratio is more realistic and important in terms of ‘fiscal pressures’.

    Second, the 15-64 range is arguably old fashioned. I think the argument to increase the bottom to 20 is stronger than for moving the top to 70 (do I hear 75?), but this is a judgment call.

    As for JQ’s claim about what that would mean for total dependency ratios compared with the 1960s, using ABS data, a demographic model (MoDEM 2.0) and assumptions/data as consistent as I could* with IGR:


    1960s (avg): 1.62
    2015: 1.96
    2055: 1.50

    Custom (20-69 / (0-19 + 70+)) TDR

    1960s (avg): 1.29
    2015: 1.85
    2055: 1.47

    So total dependency ratios will be slightly ‘worse’ than the 60s using the traditional definition, but no way near as dramatic as the typical ‘4.7 to 2.7’ spin story.

    Using the custom definition, things will be ‘better’ than the 1960s. Why? The Child Dependency Ratio bottomed out in the 60s with the boomer babies. (It’s always the Boomers’ fault.) However, it will be worse than at any time since the 1980s, and a roughly 20 per cent fall from today.

    There is no doubt there will be impacts from demographic change – but it is far from a ‘crisis’.

    * I used IGR assumptions for NOM, TFR and life expectancy. I couldn’t find its data on age profiles for NOM and TFR, but based these on PC’s 2013 report (updated for more recent NOM data), which Treasury might have done too. These slight potential differences would have minimal effect.

  12. My thoughts are that extrapolating current demographic and economic trends out to 2055 for Australia or any other country is an exercise fraught with many pitfalls. The notion that smooth growth and smooth development will continue from 2015 to 2055 is a very heroic assumption. This is the case when we face the limits to growth on multiple fronts;

    (1) Overpopulation (exceeding sustainable ecological footprint).
    (2) Climate change.
    (3) Loss of Biodiversity.
    (4) Disruption of the Phosphorus and Nitrogen Cycles.
    (5) Inadequate access to clean fresh water (exponential increase in this problem).
    (6) Ocean acidification and dying oceans.
    (7) Pollution in all its forms.
    (8) Ozone depletion (this problem might be arrested for now but ozone levels are still low)
    (9) Over-fishing.
    (10) De-forestation.

    There are many more problems which could be added to this list. There are plenty of signs that the entire ecological web of life on the planet is becoming seriously de-stabilised. It is also clear that weather patterns are becoming seriously altered for the worse (more extremes of drought and flood). Other possibilities for rupture or tipping-points exist from changes in ocean currents to methane clathrate releases (already commencing the tundra of Russia and Canada).

    Instead of worrying about hypothetical demographic trends which might well be catastrophically interrupted and altered, we need to worry about the potential catastrophic biosphere events which science advises us are not just probable but highly probable with many intial effects due well before 2055. Indeed, many effects are already occuring now.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s