Public Policy and What's Missing in Action?
Our deputy chair, Karen Chester, says evidence based research is a critically endangered beast in her speech to the CSIRO-Monash Superannuation Research Cluster Conference 2016 in Melbourne on 6 December 2016.
Download the speech
Read the speech
Good morning. And welcome to the 2016 CSIRO-Monash Superannuation Conference.
Today, you will hear and learn of the endeavors of the super research cluster. And research dividends in a field of public policy endeavour that matters. Today the $2 trillion behemoth meant to deliver better retirement incomes for working Australians. And some would argue it’s a field of much policy tinkering, but perhaps less well informed public policy design and development.
So when Deborah Ralston asked me to open the conference — my response was a resounding yes. For the PC shares much in common with the cluster. My Commissioner colleague, Angela McRae who is also here today, represents the PC on the cluster’s steering committee. And Angela and I, along with our PC team, have been beavering away in the retirement incomes space over the past 18 months.
And will continue to do so for the foreseeable future thanks to David Murray. And our work has benefitted from the insights Deborah and her cluster colleagues have shared with us along the way. But this good old fashioned evidence-based research you will hear about today has much in common with the black rhino.
A critically endangered beast — seldom seen and rarely funded. For today we find ourselves struggling to deal with what should not be intractable policy issues. And why intractable? I fear it is because three simple but essential prerequisites to good public policy are sorely missing in action.
Now the first prerequisite missing in action is evidence-based policy research and analysis. Something core to the cluster and the Productivity Commission’s DNA. The second essential prerequisite and very much missing in action, is policy and program evaluation. The third, and arguably the twin sibling of evaluation, is experimentation or piloting. And for the PC, these are the essential E’s of good public policy.
So this morning I would like to outline the case for today’s public policy imperative following a decade of the ’nothing era’. And then share two streams of recent endeavors of the Commission. For these streams — one directly linked to the work of the cluster (ageing and retirement incomes) and one indirectly linked to the cluster (overcoming Indigenous disadvantage) — illuminate the policy design benefit of the three E’s.
But perhaps more importantly, they expose the high price of the counterfactual. Policy made in their absence. And this is stark for the latter stream — overcoming Indigenous disadvantage.
But first indulge me in sharing why I began worshipping at the altar of public policy back in the 80’s. And why after a decade of non-worship I returned to it three years ago.
So allow me to rewind the clock to 1972 Brisbane. The then daggy suburb of Salisbury. Working parents. Three kids. My mother had been squirrelling away money for two years to afford a return flight to Perth to visit her mother. The price of domestic air travel at the time was in real terms over four-fold what it is today.
The price of clothing (with TCF tariffs north of 40 per cent — think President Elect Trump) was also more than three-fold higher in real terms than today. And three kids (each 5 years apart in age) experienced a simultaneous exponential growth spurt. A perfect storm for my mother who ended up raiding the squirrel tin to re-clothe us. No flight to Perth. She never saw her mother again. Now fast forward the clock to 1983, and the middle child from Salisbury is studying economics at the University of Queensland.
Why economics? Her entrance score was just short of what she needed to get into law. So the plan B was to get a HD average in economics and switch across to law at the end of 1st year. But then 2nd semester microeconomics happened — here she learned that TCF tariffs and the two airline policy were why her mother didn’t make it back to Perth. So Plan B became Plan A. And she got the HD average.
Now fast forward the clock another 4 years and the kid from Salisbury is an economics graduate at the Department of Prime Minister & Cabinet, where she luckily joined what John Durie has coined the ‘Rat Pack’ – Rod Sims, Peter Harris, Phil Gaetjens and Daryl Quinlivan. And it’s not long before the graduate from Salisbury is sitting in the then Prime Minister’s Office (as humble, silent note taker), as Rod Sims and Malcolm Gray (her bosses) discuss with Prime Minister Hawke how they can convince the ACTU to embrace tariff reform.
The note taker mumbled (and was heard) ‘because tariffs screw workers’. A plan was hatched. A model built to estimate the effective indirect tax incidence of tariffs on the price of goods. And why? To ask and answer how much the lower quartile are getting screwed. A model was quickly built. An evidence base was formed. A case made. And good public policy happened.
And the girl from Salisbury was hooked. But it wasn’t, to quote Peter Harris, Camelot, as some romantically recall. It wasn’t even really West Wing. It was still a dirt under the fingernails clamour for good policy, but it was informed by evidence-based analysis. And that was what ultimately won the day.
Now fast forward the clock to late 2012. Sydney. And a decade away from the public policy altar. Now a partner in Mercer’s investment business running global infrastructure. Having been recruited a few years earlier by ex-Treasury Secretary, Tony Cole, a chat with Peter Crone at the BCA resulted in Tony Cole and I teaming up to author a report for the BCA’s budget submission — grandly titled Back to Policy Fundamentals. Nothing to do with our day jobs at Mercer. But the BCA was prepared to pay us well, and Tony and I quickly found ourselves back at our favoured altar. And with some help from David Knox.
We reframed and redesigned the fiscal policy rules, and hollered a warning cry that in the absence of immediate and disciplined fiscal repair, our fiscal flex would be lost for a good 20 years and an intractable fiscal wedge created. And it wasn’t just a simple case of returning to surplus in the foreseeable future — it was an immediate imperative and needed more than a balanced budget.
And why? Fiscal policy needs to be cast against three policy objectives.
First — the size of government as a discipline for Australia’s international competitiveness— so tax as a percentage of GDP of around 23.5 per cent (30 year average).
Second — countercyclical effectiveness. Need to have fuel in the tank for countercyclical fiscal policy and this needed us to accumulate a surplus of 3 per cent of GDP by 2024 — which required an immediate return to surplus in 2013.
Third — intergenerational equity to deliver our readiness for future fiscal gaps. Now we had the Future Fund delivering on one sizeable unfunded liability. But nothing preserving for the biggest unfunded liability — our ageing population. So we proposed an intergenerational reserve for ageing.
And much of this was drawn upon and revisited in the Commission of Audit, but alas little acted upon.
We also lamented that our review of Australia’s major economic policy developments over the last 50 years led us to the view that Australia suffers from ’vintage’ concentration in Government policy efforts. As can be seen on this — my all-time favourite story telling figure. And what it shows is that reforms were concentrated in the 80’s, fell away in the 90’s and today are sparse.
Now our report — Back to Policy Fundamentals — was released on the day Rudd re-tilted for the then Gillard leadership. The report sank with little public trace.
But my temporary return to the altar soon became a permanent one when I was easily persuaded to join Peter Harris’ merry band at the PC.
So while this may all appear an indulgent personal reflection on modern economic history, it actually reveals much of how we’ve ended up with today’s policy imperative and the ‘nothing era’.
So why does good public policy matter so much now for Australians’ prosperity? And perhaps our narrative should be the prosperity imperative — as eyes often glaze over with utterances of our productivity imperative.
Now the single best indicator of economic prosperity is the so‑called ‘real net national disposable income’ per capita because it represents the income available for consumption by Australians. And Australians enjoyed strong growth in average incomes through the 1990s and 2000s, but the source of this growth has shifted over time.
It’s largely accepted that growth in the 1990s was due in large part to the impact on productivity growth of successive structural reform efforts. So the improvement to people’s material living standards was not by chance, but by design. But in the last seven years, average annual growth was below 0.5 per cent. Indeed, there have been four successive reductions in real net national disposable income per capita from 2012-13 to 2015-16 (the only time a sequence of this kind has been experienced in close to the last six decades). Now this outcome was strongly associated with the precipitate and (over this period) unparalleled fall in Australia’s terms of trade, abetted by weak productivity growth.
Now in the 2000s, the impact of the resources boom on the terms of trade (in glaring red) fortuitously offset generally poorer productivity growth. And in doing so, it ’masked’ the home truth that this lower productivity growth was, at least in part, because of the slowdown in reform efforts by governments (recall my earlier favourite figure on vintage concentration). And perhaps it is that higher growth (thanks to resources, albeit temporarily so) perversely diluted the incentive for concerted reform efforts.
Whichever way you cut it, our recent income growth has been the opposite to the 1990s — it was less by design and more by chance.
Of course, the benefit from the terms of trade was only ever going to be ephemeral, and we are now seeing the decline in the terms of trade drag significantly on income growth. This is the reality we confront — real incomes, on average, are falling.
And while labour productivity growth has returned to average, income growth has declined significantly since the peak of the terms of trade.
Also, shifts toward part-time hours of work combined with an ageing population may imply less of a contribution from labour utilisation in future.
Over the last 40 years, aggregate labour productivity growth in the market sector (real output per hour) has stayed mainly in the band between 2 and 2.5 per cent per year over the various business cycles.
The data give the beguiling sense that not much is wrong with Australian productivity trends. The only marked exceptions to generally stable productivity growth over these decades were a period of stagnation in the mid-1980s and exceptional growth in the mid-1990s. The longer historical story however shows virtually zero growth in GDP per capita in the nearly four decades from federation until the mid-1930s — proof that it is possible to have protracted periods of sluggish growth. And history can repeat itself, unless people and policy settings ensure it does not.
Moreover, the ‘good’ labour productivity outcomes of recent years have almost entirely reflected the contribution of more physical capital, rather than any underlying improvement in the capacity to ‘get more out of all inputs’. That capacity — measured by multifactor productivity growth rates (MFP) — languished from 2003-04, creating what has been referred to as the ‘nothing era’.
This means, at least in aggregate, that Australians have become no better or smarter in producing what we produce for over a decade, implying slow (or no) underlying technological progress. The slowdown in Australia’s ability to do more with the same is puzzling because, at the same time, we have seen rapid technological growth penetrating many aspects of consumer and business activity.
Now mismeasurement has been cited as a reason to worry less about this trend. And there are difficulties in measuring productivity. While some of the productivity slowdown could be mismeasurement, the timing and scale of the effects suggest this is unlikely to be the sole culprit of the recent observed productivity outcomes. It may be that the productivity returns from the new technologies will emerge after a lag. Regardless, even if mismeasurement suggested that the slowdown was more apparent than real, the policy issue is always ‘can we do better?’
And it’s in the how we do better that our herd of black rhinos need careful preservation.
So with such a prosperity imperative, our priority should obviously be making policy issues less intractable than they need be. And this is where the three E’s really matter.
And it is the Commission’s recent work in two policy streams — retirement incomes and overcoming Indigenous disadvantage — that illustrates the policy design benefit of the three E’s. But perhaps more importantly illuminate the high price of the counterfactual. Policy made in their absence.
So let me canter through those recent endeavours which I hope will illuminate why the black rhinos of public policy (the three E’s) are critical to any semblance of good public policy. And I should mention that all but one are self-initiated projects:
- An Ageing Australia
- Superannuation policy for post-retirement
- Housing decisions of older Australians
- Indigenous Primary School Achievement and
- Overcoming Indigenous Disadvantage.
So turning first to some fundamental evidence-based research the Commission self-initiated in the ageing and retirement incomes space. And you may ask why we self-initiated these streams of work? The practical reason — last year, the Government had left us with a not so full inquiry dance card and we like to be productive at the Productivity Commission. And perhaps the more important reason for us today, was simply we wanted to lay the groundwork.
In 2013, we released the first flagship research paper on ageing and its impacts on economic growth and government budgets.
To put it bluntly, we are experiencing a demographic seismic shift. Our projections suggest the share of the population over 65 will increase from 1 in 7 today to 1 in 4 by 2055. So our cohort of taxpaying and super accumulating workers is quickly shrinking relative to the cohort of longer living Age Pension recipients and decumulating super retirees. And this is primarily because of increasing longevity — which is a good thing.
Now we’ve all heard the adage that older Australians on the whole are income poor and asset rich — and importantly this spread has widened over the past decade. This is primarily as a result of high rates of home ownership for older Australians, coupled with recent strong house price growth (5.5 per cent on average each year over the past 10 years).
So current older households (the orange line) are a lot wealthier than earlier cohorts of the same age (the green line).
So while the fall out of our ageing population has been examined extensively by us and others — the focus has been around wellbeing and fiscal sustainability.
We found that only limited work had been done on how well placed the retirement income system is to deal with this demographic shift especially in terms of policies and products for the decumulation phase of super. And this is what really prompted our two self-initiated projects on super and housing for older Australians last year.
So let me share how I view the retirement income system.
First when you think of retirement incomes policy, the most important thing to have front of mind is that it’s a wheel of policy fortune. And each component part needs not only to be well understood, but also with the recognition that they collectively and never in isolation impact behaviour. Pull one lever without understanding the collective interaction, and you’ll just end up with poor public policy.
So with our not so full dance card, and prompted by our ageing work, we started to work our way around the retirement incomes policy wheel. So turning to our first cab off the rank.
So we turned first to the relative policy orphan of super in retirement. And arguably orphan in both the super industry and public policy development — albeit some helpful advances and policy prods from David Murray on impediments to pooled retirement income products. But we wanted to understand what policy levers might matter most.
So we set out to ask and answer: what would happen if we lifted the preservation age of super? To do this, we had to better understand when, how and why do people access their super approaching and in retirement. And to do this, we had to build from the ground up a behavioural micro sim model. Why? We viewed it as a must have (not a nice to have or good to have, but a must have) to form an understanding and evidence base for retirement incomes policy.
So turning to the policy context.
Now the figure on the left shows the preservation age is currently 55 and being raised to 60 by 2025 — so reducing the gap between the preservation age and the Age Pension age from 10 to 7 years. But the mooted increase in the Age Pension age to 70 would simply restore the gap to 10 years.
So as part of our Super report, we modelled (after building our model) what would happen if you raised the preservation age to 65 — the figure on the right.
So let me cover off our key findings — because they reveal much about the greater than anticipated diversity of folk approaching and entering retirement.
The modelling suggested that a gradual increase in the preservation age to 65 years would see by 2055:
- older worker participation would increase, but only by a modest 2 percentage points
- those who delay retirement would stay in the labour force for 2 years longer, giving them a 10 per cent boost to their super balances at retirement.
And in doing this modelling it also meant we better understood things like:
- Most people have low or no super. Many people have very low levels of super or simply none at all. 15 per cent of people aged 44–55 had no superannuation whatsoever, and only 10 per cent had super of more than $230,000.
- Diversity matters. The average super balance of a 65–74 year old is over $300,000 … But the median is zero. So retirement income policies and products based on the ‘average household’ are unlikely to match most people’s circumstances. So put simply — CIPR (comprehensive income product for retirement) ain’t simple.
- Not everyone chooses when they retire. While there’s a perception that retirement happens at 65, for many people it’s quite a bit earlier. Almost one half of men and around one third of women who retire between the ages of 60 and 64 do so involuntarily. Ill health and redundancy are the main reasons.
- Most super is not taken in lump sums. And interestingly, when retirees take their super as a lump sum, it’s typically those with modest balances (makes perfect sense).
- And those lump sums are often used to invest … or pay down debt. So folk with modest balances are most rationally getting ready for retirement and a lower income stream.
Then in the second half of last year, we analysed the housing decisions of older Australians. Another part of the wheel of fortune. And perhaps the biggest policy sacred cow of all. But given its pre-eminence on the balance sheets of older Australians and how those housing decisions impact their wellbeing — we set forth.
Now getting solid evidence quickly became our priority. First, in addition to our bread and butter consultation and research, we commissioned a national survey of older people’s attitudes to the key issues for our study. Because when looking at housing decisions of older Australians, we identified a paucity of data on the behavioural drivers of decisions (needed to inform good policy). So we set about filling this information gap on older people’s attitudes and preferences for housing. To do this, we surveyed about 1500 Australians aged 60 years and over. And in doing so, we uncovered older people’s attitudes and views on a range of housing and retirement issues, like:
- planning for retirement
- attitudes towards the family home
- debt and bequests
- housing preferences and awareness of residential aged care, and
- attitudes towards home equity release products.
We also did some pretty neat quant work and modelling to simulate the impact for asset modest and income poor older Australians of drawing down on the family home equity.
Firstly, we looked at the scope for ‘asset rich, income poor’ older Australians to boost retirement living standards by using their home equity. Specifically, how would such borrowing affect the balance of home equity at the end of life, in the context of an uncertain life span. And this required some funky Monte Carlo modelling.
The height of the overall bars on this chart shows how many people would need to draw on their home equity to reach the ASFA (Association of Superannuation Funds of Australia) modest and comfortable standards.
So based on real people in 2010 (thanks to HILDA), around 50 per cent of older singles and 30 per cent of older couples would need to draw down on the value of the family home to meet ASFA’s modest retirement living standard. The green portion of the bars shows the proportion of people drawing down who would still have some positive home equity left over at the end of their lives.
As you can see, when pensioners draw down on the value of the family home to reach the ASFA modest standard, over 90 per cent retain some home equity at the end of their lives. But once they start to borrow to attain a comfortable standard, a much larger proportion exhaust their home equity entirely (the black shaded part of the bar).
So here we have a few insights from our report, including from our survey results of older Australians and how they view the family home. Again, an important part of the evidence puzzle is understanding what behavioural responses may be to policy changes. Or what nudges may be needed from Government. And from the 10 key take-outs, perhaps the most policy relevant are:
- some older Australian are even net savers — and perversely those in the lowest quartile
- older households are slow to draw on housing wealth — and while the biggest part of their balance sheet, it remains an untapped source of retirement income
- and while people stay in the family home until late in life, few understand the cost or likelihood of entering residential aged care — and those odds have lengthened considerably (fewer older Australians entering residential aged care and then much later in life), but folk are still preserving like they haven’t changed.
So turning to the second, and perhaps Australia’s most intractable policy issue — overcoming Indigenous disadvantage. And here at least 2 of the E’s — evidence-based analysis and evaluation — are missing in action. So perhaps it is here that the price being paid — and by Indigenous Australians — for the absence of the E’s is so starkly illuminated.
While there is extensive reporting on the extent of Indigenous disadvantage, there is a lack of evidence about what works (and what does not) in bridging outcome gaps. And while evaluating the impacts of policies on Indigenous outcomes can be challenging, the challenges should not be seen as insurmountable.
So we recently undertook a self-initiated research project looking at Indigenous primary school achievement. Why education? When you think of perhaps the two most important and linked drivers of Indigenous disadvantage, they are poor educational outcomes and low economic participation.
And we sought to better understand the factors that contribute to the literacy and numeracy achievement of Indigenous students. Now this was an evidence base first. It represented the first effort to statistically analyse the factors contributing to Indigenous student achievement at a national level. And we did this by using a newly available dataset developed by ACARA (the Australian Curriculum, Assessment & Reporting Authority) that links NAPLAN results to student demographics and school characteristics.
And it is revealing that our work was novel.
Now the underlying question and genesis for this project was why, after decades of good intentions and much government expenditure, do large gaps stubbornly remain between the literacy and numeracy achievement of Indigenous and non-Indigenous primary school students?
In 2014, across Australia, 73 per cent of Indigenous students in year 5 were at or above the national minimum standard for reading, compared with 95 per cent of non-Indigenous students. This equates to a national gap in achievement of 22 percentage points. Similar gaps exist for writing and numeracy.
So despite decades of policy attention, there has been no consistent improvement in the literacy and numeracy achievement of Indigenous primary school students over at least the last 16 years. This can only but suggest that current policies are not working, and that we need a stronger evidence base about what might work to improve Indigenous education achievement.
And we found this is a policy issue where gaps abound.
First, our analysis could only take us so far because there are large gaps in the data available at the national level. Only a subset of the characteristics thought to be associated with education achievement are observed in the data. Yet with the cooperation of state and territory governments, whom we know have much of this data (the middle grey column), our analysis could be expanded to include drivers that really matter, such as teacher characteristics and student attendance rates.
And partly as a consequence of the data gaps, we found that the available data on school and student characteristics only explain about a quarter of variation in Indigenous student achievement. Three quarters of the variation in achievement remains unexplained.
Something else we discovered was that geographic context really matters. We asked and answered the question: where do Indigenous students attend school?
Why? Because policies to lift Indigenous achievement need to be informed by the distribution of Indigenous students across schools. And from this analysis, policy effort is clearly needed across all geographies to reduce the gap.
Indigenous students in remote areas are often the focus of public and therefore policy attention, and it’s true that gaps in achievement between Indigenous and non-Indigenous students are largest in remote areas. But this is not where most Indigenous students go to school. Most Indigenous students (80 per cent) attend school in metropolitan and regional Australia.
And we felt this finding was pretty key to policy thinking. So we also decomposed the gap in achievement between Indigenous and non-Indigenous students by geography. And found that while all geographies contribute to the national gap, the majority of the gap is not attributable to students in remote and very remote communities.
So improving Indigenous education achievement needs policy effort across Australia, and not just in remote and very remote schools.
We also found that there are potential policy insights to be gleaned from looking at outlier schools. Outlier schools ‘punch above their weight’. And they exist and can be identified — not by us, but by state and territory governments. These are schools where Indigenous students perform considerably better than you’d expect from analysis of their student demographics and school characteristics.
The Commission identified (not by name) more than a dozen schools where Indigenous students perform much better than expected on both reading and numeracy tests. So it remains a glaringly obvious thing to do — for state and territory governments to identify and systematically evaluate these high achieving schools. For lifting the bonnet on these outliers could shed light on what works best to lift the achievement of Indigenous students.
Turning to the final chapter in our case for why the three E’s matter. The Overcoming Indigenous Disadvantage report is the single most comprehensive report card on the wellbeing of Indigenous Australians. This biennial report card (released 2 weeks ago) reveals a very mixed performance.
This figure shows that outcomes (green shading) have improved in a number of areas, including some COAG targets like mortality rates for children, school completion rates, and economic participation. However, there has been little or no change in other areas (in yellow) like rates of family and community violence and risky long‑term alcohol use.
Outcomes have alarmingly worsened (in red) — in areas such as high levels of psychological distress and hospitalisations for self‑harm. And most alarmingly, adult imprisonment rates have increased by 77 per cent over the past 15 years.
But data alone cannot tell the complete story about the wellbeing of Indigenous Australians, nor can it fully tell us why outcomes improve (or not) in different areas. To support the indicator reporting, case studies of ‘things that work’ are included in this report.
And here lies the rub — these are more than sparse. To put it bluntly, the overwhelming absence of meaningful evaluation of Indigenous policies and programs beggars belief.
We know that annual government expenditure on Indigenous Australians is just over $30 billion. And of that $5.9 billion on Indigenous specific programs. And we know from some recent research the $5.9 billion is allocated across just over 1000 Indigenous specific programs.
But for the 2016 Overcoming Indigenous Disadvantage report, we could only find 34 case studies of ‘things that work’ out of that 1000. And it wasn't for want of trying to find more; and for having the full involvement of federal, state and territory governments; and Indigenous consultation.
And it is nothing more than common sense that to know whether things are making a difference on the ground, you need evaluations. So of the 34 ‘things that work’, only 24 were what we could view as soundly evaluated (so meeting basic criteria for a solid evaluation). The remaining 10 are promising programs, but yet to be evaluated— so the jury is still out.
And anecdotally, we heard of potentially good programs being discontinued and ineffective ones persisting all in an evidence vacuum.
So we should not feign surprise that we are failing to make substantive progress on overcoming indigenous disadvantage. And it’s not just short changing the taxpayers underwriting these programs. For more importantly, it is ultimately short-changing Indigenous Australians. And it poses the question: will overcoming Indigenous disadvantage remain elusive if we continue to allow the systemic evaluation of Indigenous programs to remain missing in action?
So what does all this mean for our ‘where to from here’? And perhaps the answer lies in the three E’s.
Stop looking for silver bullets and policy sound bites. And just get back to the dirt under the fingernail work of building evidence-based policy and building a much stronger evaluation culture (we need to know more about what works and why).
For the Commission, we now have a full dance card which we professionally relish. Perhaps the most encouraging sign of an appetite for good old fashioned public policy and the three E’s is the Australian Government commissioning us to undertake the 5 year Productivity Review. And we are up and running in doing so. Looking for the new and novel, not old tired lists of the past.
And on our super to do list — our recently released study on the assessment framework to assess the competitiveness and efficiency of the super system next year — reveals our DNA. We have developed a robust assessment framework, which will be informed by a comprehensive evidence base that we have set out in detail. And this will be used to ask and answer the $2 trillion question sometime next year. And that question is: after a couple of decades of our super system, is it today competitive and efficient?
And now you will hear and learn from the endeavours of the CSIRO-Monash super research cluster. And it’s a great program ahead today. Analysis on super and the broader economy, super member behaviour and outcomes (including for Indigenous Australians), and labour market and health in retirement.
So enjoy today’s veritable herd of black rhinos.