A way to expand the GST and make it less regressive.

A fair way to broaden the GST, collect more revenue, and avoid increasing the burden of the less advantaged would be to extend it to education.

educ spending

Education spending shows more of an association with income than even alcohol, clothing, vehicles or holiday accommodation. 

Screen Shot 2014-10-06 at 1.09.29 pm
Screen Shot 2014-10-06 at 1.19.07 pm
vehicle purchase spend
Accommodation spend

Taxing education has the potential to make Australia’s GST fairer. Perceptions that it is unfair are a major impediment to reforming the GST.

The GST is regressive. (nb. I had to figure this out the old-fashioned way after being unable to find a single publication that contained this data.) Lower income households pay more of their (equivalised disposable household) income in GST than higher income households, according to my calculations.

GST current burden
Calculated using ABS 6530.0 Household Expenditure Survey, Australia: Detailed Expenditure Items, 2009-10; ATO GST rulings; 6523.0 Household Income and Income Distribution, Australia, 2009-10

Lowest income households pay 11.5 per cent of their income in GST (higher than 10 per cent because some spend more than they earn). Highest income households pay 8.7 per cent.

In absolute terms, this means the highest income households pay four times as much GST ($148/week) as the lowest income households ($36/week). They can do so because they have 5.5 times the disposable income ($1704 vs $314).

This is not quite as regressive a tax as I had assumed. The exemptions agreed by the Democrats and the Coalition in 1999 make a difference. Without exemptions, the GST would be even more regressive.GST share without exemptionsThe definition of a progressive tax is one where the rich pay more as a percentage of their income. Levying GST on education will not turn the GST into a progressive tax. But it moves it in the right direction.

Screen Shot 2015-04-09 at 3.32.33 pm
Those are percentage points on the vertical axis.

You can see the comparison a different way in this next graph.

Screen Shot 2015-04-09 at 3.32.46 pm

The estimated revenue gain of taxing primary, secondary and tertiary school fees would be around $1.2 billion.

Two caveats remain.

One is pure politics. Education is seen as A Good Thing. Even though Australia would continue to provide free and universal school education, a decision to tax education expenditure would cause confusion and dismay. I wrote just yesterday about how insufficiently distinguishing sin taxes and revenue taxes causes trouble, and this would fall right into that trap. This government would also face the problem that its supporter base are more inclined to be patrons of private education. (I should note here that I am not biased against private schools, having attended one myself.)

The second comes to the exact question of fairness that prompted my investigation into how GST could simultaneously be extended and made less regressive.

The association between education expenditure and household earnings is not simply a matter of wealthy people sending their children to private schools. The period of peak earnings in a person’s life is also commonly in their 40s – the period where they have children in secondary school. In other words, the statistics insistence that education spending is a luxury good could be more of a life-cycle artefact.

Nevertheless, the same is true of many types of expenditure, and an opportunity to plug a hole in the GST – and do so fairly – is rare enough to be worth grasping.

How Exams Maybe Ruined Your Whole Life

Your time starts… Now. 

The exam hall fills with an updraft as a thousand students flip over the front page of the exam booklet. Some are nervous. Others feel they are in their element.

I was always in the latter group. The pressure of having two hours to show everything you know is just like the pressure of producing a newspaper story right on deadline. The marks I got on exams were almost always better than the marks I got on assignments.

Whether I studied hard or crammed like a boss didn’t matter. I generally knew the answers or a decent way to bluff, and my biggest problem would be getting cramp in the hand that wrote the essays and coloured in the little circles.

Today is the day of the Victorian Certificate of Education Maths Methods exam – one of the biggest and most important subjects. I remember that exam clearly. I had prepared more for it than any other. I spent hours doing every practice problem in the book so on the day I was the captain of calculus and I nailed an A+.

Many students will not be so calm. Many will be in a state of nervousness that will actually hurt their score.

The bad news is this will ruin their life.

New research out of the US National Bureau of Economic Research argues exam performance has long-term consequences for schooling attainment and labour market outcomes.

The way they ran their test was by finding an external factor that could plausibly disrupt student performance. They settled on air pollution, which has been shown to hurt human productivity on a range of tasks.

Students who did their exams on days with bad pollution went on to have lower scores, reduced years at university, and lower monthly earnings. (The test in question is Israel’s version of the SAT test, which is heavily weighted in university admissions).

To get the data they used air quality testing sites within 2.5 km of the schools where the tests were being run.

An increase in particulates by one standard deviation on testing day reduced test scores by 0.65 of a point (a tenth of one standard deviation), with the result mostly driven by strongly worse performances on strongly polluted days.

When the air clears, the effect on cognition also clears. But the effect of the low score plays out throughout students lives.

An increase in particulates on exam day by just half of one standard deviation reduces monthly adult earnings by $29.

“Our analysis highlights a major drawback of using high-stakes examinations to rank students. If completely random variation in scores can still matter ten years after a student completes high school, this suggests that placing too much weight on high-stakes exams … may not be consistent with meritocratic principles.”

Exams are not good preparation for real life any more. No serious job forces people to solve problems without reference to the internet, conferring with colleagues, or time to think (airline pilots dealing with a nosedive and firefighters maybe excluded).

Exams discriminate in favour of people who can go in there calm, and hurt people who didn’t get a good night’s sleep a good breakfast, who have a headache or a cold, or – for whatever reason – are wigging out.

These are likely to be talented people, who may miss out on the course of their dreams, and a subsequent job that they would have excelled at. Exams seem fair but they are not. It’s not just random bias as described above, but systematic bias too. We should call “pens down” on the whole exam system.

I went to an expensive private school. What was the benefit?

I spent six years of my life attending an independent private school – Melbourne Grammar School. It was expensive. I remember fees being around $10,000 a year.

Was it worth it?

There is an article in The Conversation today about private schooling, entitled Private schooling has little long term payoff.

It cites research that shows “after controlling for tertiary entrance score, university students from government schools outperformed students from private schools.”

Which is interesting. There is certainly an argument that private school kids are “spoon-fed” and can’t handle university thereafter. (I also seem to remember research that students who scored less than the cut-off but got into their course tended to do better in first year, so there could be confounding effort effects.)

The new research looks further along the lives of these students and finds “former independent school students were no more likely to be employed full-time than those who attended a government school after controlling for the effects of level of education, sex and age.”

The study also finds that graduates of independent schools are not likely to have higher earnings, nor more “prestigious” employment, after controlling for education.

In a comment thread on the topic, a user called Gabrielle calls out the methodological elephant in the room.

“Yep, ‘controlling’ variables everywhere! E.g. controlling for whether or not people attend a Go8 university when they look at occupational prestige, or controlling for field of study when looking at earnings…what the?! I mean sure, that’s interesting, but it totally bypasses questions about whether your schooling or university help you get into high status or high earning careers.”

If a private school helps you get into university (independent school students had an 80 per cent higher chance of graduating from a G8 university, according to the research), and university helps you get a job, then controlling for university when trying to measure the effect of private school is distracting. Silly even.

It takes just a glance at year 12 results to see that private schools dominate the top scores. I think we can conclude that going to a private school has a real payoff.

But what form does it take?

Second Row, fourth from the left.
Dug this out. I’m second row, fourth from the left.

I’m still close friends with some people I went to school with. But I have other friends who went to other schools, public and private.

As far as networking goes, I’m not aware of any benefit I’ve got from “the old school tie” and quite acutely aware of the benefits that have come from people I’ve met in my professional life. Could seeing my school on my CV have helped me in job applications? It’s not impossible. 

I’ve never worked at Goldman Sachs or a law firm, so maybe it’s different there, but I’ve rarely been aware of the schools my colleagues have gone to. When I have known, they’ve mostly been graduates of government schools. I don’t think my employers have been deeply biased to private schools.

So what was the benefit of my private school education?

You can never know the counter-factual but I think I fit the data perfectly – I suspect got a better year 12 result at the school I went to than at another school chosen at random.

Did the teachers “spoon-feed” me? I don’t know. They taught me, definitely. They were (mostly) highly qualified, diligent and happy to answer questions.

I also worked my arse off in year 12, despite having been rather disinterested in years 9, 10 and 11. By working hard I think I maximised my potential. I got into a good uni course that eventually helped me get a good job.

I wanted to beat the other kids by having a better score, and I knew they wanted to beat me. Even then, I was aware that the level of competition was peculiar.

Would that competitive aspect have been there at another school? It could have been. It might have been stronger at a selective entry school. But the skew of high scores at my school suggests it was acutely competitive.

So I’m not here to say, “I worked hard so I deserve everything I get!”

I’m trying to refute the crazy notion that private schooling is not advantageous, and answer the question of where and how that advantage accrues.

I know the world is unfairly skewed toward people who have the privileges I have. I’d like to see the opportunities of really motivating and rewarding educational environments shared really widely. Denying that private schools have benefits seems unhelpful in pursuing that goal.

a+b=c. Algebra can save a poor black kid’s life.

Doubling the amount of time poor kids in Chicago spent studying algebra in grade nine led to an 8 per cent increase in their high school graduation rate, and an 11 per cent increase in college enrolment.

That’s the finding of a terrific new paper out of the US National Bureau of Economic Research.

Given that US high school dropouts die earlier than graduates by 3-5 years and make up 7 out of 10 prisoners in the US, it is fair to argue that for thousands of these kids, algebra has saved their life.

The paper exploits a natural experiment whereby students in the Chicago Public School System were placed into double dose algebra classes (largely replacing music or art classes) if they scored below the national median on an eighth-grade math test. The researchers compare kids just above and below the cut-off. 


Barack Obama in Chicago in 1995. Photo By Mark Pokempner

The numbers of students in the algebra courses is substantial, because maths skills are thin on the ground in the 73 public high schools studied. In the US, just 20 per cent of hispanic students, 13 percent of black students, and 17 per cent of students poor enough to qualify for free lunches are rated proficient in maths. In the Chicago Public School System 90 percent of students are black or Hispanic.

“In the Chicago Public Schools (CPS), the focus of this study, roughly half of high school freshmen fail at least one course, with the highest failure rates in math courses”

Schools assigned weak students to a back-to-back lesson in Algebra with the regular algebra teacher, and gave the teachers professional development, permitting them to use different instruction methods, including working in small groups, solving problems verbally, and having students set problems for each other.


The impact on test scores was modest. But the impact on the students lives was not.

“…the test score impacts of this policy dramatically understate its long-run benefits as measured by educational attainment.”

The effect of double dosing in the first year of high school remained over time. Students who were in the program were 9.3 per cent more likely to pass algebra in that year. But they are also 7 per cent more likely to still be in school in the fourth and final year of high school.

The students who got the most out of the instruction were ones with poor reading skills. They end up passing three more subjects in high school on average, after taking the double dose, creating a 13 per cent increase in the number of poor readers who complete enough subjects to pass high school. The researchers suggest that the focus on using verbal methods to solve maths problems may explain this result.

Black students who got the double dose  also saw a dramatic effect, with a 15 per cent increase in college enrolment. For the double-dose population at large, the increase was 11 per cent. The effect can be seen in this chart:


For some of these student, the impact of these classes – which their 14 and 15 year old selves no doubt dreaded – will be dramatic over their life times.

Should you do a PhD?

In order to answer this question, I’ve surveyed six people who started PhDs – friends of mine – from the following fields: physics, biology, computer science, chemistry, economics and history.

Five emerged from their programs with a Doctorate, one turned theirs into a Masters. (Given the PhD completion rate in Australia is 60-61 per cent, this sample is biased towards survivors.) One works in academia.

PHD completion rates
PhD completion rates. (Source)

Their answers make a very interesting set. But before I let them speak for themselves, a confession: I was accused of asking leading questions. It may be true, but it’s the zeal of the convert.

As an undergrad, surrounded by lecturers with PhDs, I wanted one too. Then I got a job offer at the end of my degree. Starting that job I met plenty of public servants with doctorates. We were doing the same jobs, making the same pay. I came to see my desire for a PhD as the result of sample bias: Most powerful people in universities have PhDs, but most people with PhDs are not powerful people in universities.

With that noted, let’s hear from the experts!


Physics at Australian National University: I’m a researcher/lecturer, which I fundamentally couldn’t do without my PhD of course. I haven’t had any other profession since graduating. On a personal level, I really enjoyed it. I learn’t much more than just the topic of my thesis. I really learnt a lot about problem solving, logic, and life from being in close contact with world class researchers. During my PhD, I had the pleasure of interacting with four Nobel laureates at various times.

Biology at Melbourne University: I became a specialist in a very narrow field of study, possibly even an expert in said field. I also gained valuable generalist skills in experimental design, data analysis, working independently and as part of a team, critical thinking and the effective oral and written presentation of data and ideas. A PhD is also an exercise in persistence and perseverance, not necessarily intelligence, so I consider that most people who have completed aPhD are good at troubleshooting and are quite resourceful.

Computer Science, RMIT:  I (arguably) got my foot in the door of my first job in industry from my PhD, and some of the domain knowledge I acquired there was useful in my job. My written and verbal presentation skills were also improved during my PhD, and I learned to have a healthy disrespect for authority. I also had a lot of fun! I found my PhD years to be quite relaxed (despite the odd moment of stress) and it was actually quite a nice transition between undergraduate spoonfeeding and the “real world” of the workplace.

Chemistry Masters, ANU:  Benefit is a fancier CV. Without a masters I wouldn’t have gotten either my first or second job in consulting. In terms of education I am far more adept at seeing the ulterior motives and ugly manipulative side of people after working for my pitiable (but totally hatable) supervisor.

Economics, Oxford: Time to read and develop deeper / broader domain knowledge. For me, that was Chinese language, China’s economy and history, econometric techniques. Got supervision re: use of econometric analysis and thinking through data, and technical skills. I also played a lot of squash and improved my cooking. Those are the main benefits!

History, Melbourne University: For me, the three years doing it was probably the biggest benefit. I had a scholarship which is a privileged position to be in because it enables you to study something that fascinates you without having to work full time. I had a clear idea of what I needed to achieve each day, and if I got it done more quickly I gave myself the rest of the day off. More recently there have been career benefits I think, but like many people when I finished the thesis I wanted to do something totally different from research. 

graduation day
Graduation day.


Physics at ANU: More than once I felt like quitting. In hindsight, the pay is tiny compared to what you could be earning, but I didn’t really think about that at the time. You have no guarantee that you will finish. There are people who try as hard as they can for over 4 years and come away with nothing. 

Biology at Melbourne University: The downside is that you are very cheap labour at the bottom of a pyramid scheme. You know that chapter in Freakonomics where drug dealing is compared to a pyramid scheme? Well, I’d change the “drug dealing” to “research” and “academia” and the chapter would still be quite accurate (minus the crack cocaine and guns). Everyone’s ego is complimented when they are asked to do a PhD. What hindsight has taught me is that there is very little money in Australian research and really not enough for most senior researchers (lecturers etc) to hire employees (such as postdocs or research assistants) so there is a huge reliance on the cheap PhD labour that the federal government (or uni) pays for. 

Computer Science, RMIT: I didn’t earn that much money (although I found my stipend to be perfectly adequate) and, obviously, my entry to the workforce was delayed by a few years. In terms of climbing the career ladder it was probably not great bang for buck, but I don’t care too much about that sort of thing so don’t consider that a downside.

Chemistry Masters, ANU: Downside was three years of near-suicide misery and lost opportunity. I could have done a law or engineering degree in that time. I still feel that sense of loss – I am always years older than my peers. 

Economics, Oxford: Long distance relationship. I’m not sure Oxford was the best place for me. 3-4 years out of workforce and foregone income and job experience.

History, Melbourne University: Honestly, the main downside was the public speaking obligations and pressure to publish along the way.

academic job market
The academic job market is tight


Physics at ANU: I believe that you shouldn’t do a PhD because you want a piece of paper that says “I have a PhD“. The people who I know who were trying to prove something, either to themselves or to others, always had a terrible time. If you have had a taste of research, and still feel like more, then you should consider a PhD. There’s also the question of talent. There are people who love science, yet just aren’t cut out for research. Fundamentally, if you’re not driven by curiosity, and have a hunger for discovery, specifically towards your project, you probably shouldn’t be there. 

Biology at Melbourne University: I actively discourage most people from starting a PhD. For most professions, the experience you gain in the workplace will be better. For anyone wanting to do scientific research, then permanent positions in Unis or CSIRO are few and far between (see pyramid scheme comments above). With this in mind, I know plenty of people who I did my science PhD with who now work outside of academia, in the public service, as patent attorneys, as technical sales reps, and in private biotech/pharma firms and consulting. Someone contemplating a science PhD should realise that these non-academic jobs are actually where they are most likely to end up and then determine whether a PhD is really still necessary. And realise that their PhD supervisor is always going to encourage the singular academic trajectory (ie they are much happier to be a reference for further academic appointments than anything else). 

Computer Science, RMIT: To consider doing a PhD you will need to be convinced you have the necessary skills and temperament to undertake a prolonged exercise in independent research, and at least one of the following should be true:

-You think you will enjoy doing a PhD for its own sake;
-Your chosen profession requires a PhD (e.g. academia);
-You will acquire **specific** skills that will help your career in your chosen profession.
These are bad reasons to do a PhD:
-You’re not sure what else to do after you finish your undergrad;
-You think it would be cool to get to call yourself “doctor”;
-You expect some ill-specified career advantage from having one;
-Your parents would be proud;
-It’s “prestigious”.

Chemistry Masters, ANU: The only people who should do PhDs are either saintly types who just love their subject and have no expectation of any type of employment related advantage or psychopathic sadists who will enjoy destroying promising young people. 

Economics, Oxford: Almost definitely do one: if you want to become an academic; its paid for; have world leading supervisor who has top track record of placing their students into great academic jobs; at world class research institution, great city, great classmates etc.

Never ever do one: in the opposite scenario: Unsure if academia is the path; not at top ranking place and doing a major where job placements are super tough; where PhDs take 6-7 years; or living in a shit city.

History, Melbourne University: The peers I had that struggled the most were those who set themselves additional challenges such as designing research involving lengthy ethics processes and multiple methods, including undertaking lots of complex interviews and all the responsibilities that come with that. This type of research makes a huge contribution to the evidence base but goes well beyond the three years you sign up for and requires a significant time investment in project management.

There’s more than one way to prove you are a big shot.


Physics at ANU: Only those who tick the boxes in the above question. How many PhD grads do we need? Well, that’s a different story. In the Australian employment market, a PhD isn’t really valued. From an employment point of view, the only reason to do a phd is if you want to become a researcher. The number of permanent research positions that open up every year, divided by the number of phd graduates is bugger all squared.

Biology at Melbourne University: A large proportion of smart undergrads will contemplate it, and even be asked by lecturers to do a PhD. As above, I would suggest that only a very few should take it up. There are too many easily available PhD scholarships in this country and not enough investment in other parts of the research infrastructure, hence the pyramid scheme. If they are smart, they can get the benefits I mentioned above from plenty of other roles.

Chemistry Masters, ANU: Maybe 5 per cent of undergraduates? We need to rethink PhD candidature. There are things to be learnt, but should acknowledge that phd students teach each other and are self taught in what can be an educationally nutritious environment – it all depends on the calibre of your fellow students. With ten students to a professor this will always be the way. The idea of charging tuition for a PhD is outrageous. My only hope is this will decimate the ranks of the fools who sign up for one.

I see PhD years as a type of national service, where you accept below minimum wage for contributing to the R&D pie for the nation. We should acknowledge this and stop pretending it is training for an academic or other research career – those jobs don’t exist. Also we should acknowledge that growing the R&D pie is of national importance. Slave student labour will be cheapest.

Also I would like to see universities sued for misleading students regarding their employment prospects. Or scolded for false advertising. A PhD contract (because you sign a contract) should include a disclaimer with real graduate outcome figures. And then you should have a counselling session to determine if you’re likely to kill yourself.

Economics, Oxford: I’m not going to tell anyone they shouldn’t do it if they’re interested. I just think they should be fully aware of the academic labor market conditions and that it is in supervisor’s interests NOT to be fully candid / upfront about this. By this, I mean how tough the market is, especially in humanities, etc.

History, Melbourne University: Lots. It’s important to have a great topic and a good supervisor and that you like to work alone a lot. I worked part time doing something completely different which was a good balance for me.

I feel like we got a great range of answers there, with evangelism from both sides and plenty of middle ground. No matter your pre-existing biases, there is something to grasp onto!

This is a fascinating topic at the moment, with the Budget proposing charging fees for PhDs. If you have other thoughts please leave a comment below!

One good teacher.

Ask someone the highlight of their school education.

They won’t mention learning cursive writing. They won’t mention timetables. They won’t mention learning cellular respiration, renaissance history or trigonometry. 

They’ll say something like this:

“I remember my grade 9 English teacher. She really believed in me.” 

This is the sort of memory that has inspired a whole genre of films. From To Sir with Love (1967) to Dead Poet’s society (1989) and Freedom Writers (2007), the one good teacher that can change your life is a truism of modern culture. It’s part of the sales pitch for programs like Teach for America and its local knock-off Teach for Australia.

Me, teaching English in China, 2003.

Now a bunch of economists have done the work to measure the influence one good teacher can have.

They had a sample of one million students and found that the ones randomly assigned to high quality teachers were “more likely to attend college, earn higher salaries, live in higher SES neighborhoods, and have higher savings rates” [link to paper]


These graphs tell the story better than words:


There is a catch. The quality of teachers as measured above is in their ability to contribute to results in standardised tests. (they use the term value-added)

That’s not the sort of thing you can make an Oscar™-winning Picture about…”Starring Sean Penn as Mr Brown, who drills his students on multiple choice questions every day in lieu of teaching life lessons.”

I tend to disagree with the merits of using narrow standardised tests to rank teachers.

The saying “if you can’t measure it, you can’t manage it,” has a flipside. People obsess over the things they can put metrics on. You can’t measure the subjective benefits of a good teacher. Nobody denies the importance of all the non-academic work teachers do, but it can’t be measured so it is not addressed. 

Nevertheless there is an important point to make here. Teachers are important. Far more so than society gives them credit for. And teaching is not rocket science, it is harder.