Animated Australian Population Projections

Following from the .gif I posted yesterday, I had a request to animate Australia’s projected population.

In 2008, The Australian Bureau of Statistics modeled 72 permutations of fertility, migration and life expectancy in order to project Australia’s population out to 2101. Of the resulting 72 series, the ABS provide us with three of them as annual time series.  ‘Series A’ uses high growth (HG) assumptions, ‘Series B’ follows current trends and ‘Series C’ uses low growth (LG) assumptions. The specific assumptions of each series can be seen in the table below.

The ABS also modeled the population with zero net overseas migration but that was only modeled (thankfully) so it would be apparent what the impact of NOM on the population is.

The assumptions vary significantly under the different scenarios, so much so that the forecast difference in population between the likely scenario and the HG one is more than 15 million by 2101. The difference between HG and LG is enormous with the HG population nearly double that of the LG one in 2101.  More than anything, this shows the uncertainty, sensitivity and danger generated by compounding projections over a long time frame.

2

Below are two versions of the .gif, one with the three growth scenarios plotted and one with just the likely scenario (I like this one more). The ABS estimated the population in yearly age intervals all the way up to 99 years of age then they bundled everyone older into ‘100 or older’. This causes a problem in the charts as is apparent once the projected number of 100+ year old residents start piling up. I stopped the animations at 2080 as I think that is already looking plenty far ahead and that looking too far ahead might detract from the story at hand.

Note that forecasts start from 2011 as I couldn’t find more recent stats in my brief search.

animation

animation1

Apart from the obvious ageing of the population you can also see the increase in the population that occurs either side of the 25 year old mark which is occurring due to NOM.  Australia is essentially getting working age adults for free after we’ve let their home countries bear the cost of raising and educating them – sounds like a good deal to me.

What does it all mean?

As the chart below (implicitly) shows, the percentage of Australians who are of ‘working age’  will fall rapdily into the future. This will put stress on our health system and the public purse. There will be  implications for policy around health, superannuation and the retirement age, as all of these areas will have to accommodate the large shift in the age distribution. As Australia’s population grows we will need to accommodate our new citizens so need the foresight to implement  effective infrastructure spending and urban planning .

4

All data sourced from here – http://www.abs.gov.au/AUSSTATS/abs@.nsf/DetailsPage/3222.02006%20to%202101?OpenDocument

Fun Facts(you deserve it if you got this far):

Under Scenario A, Melbourne overtakes Sydney as Australia’s biggest city in 2039.

Under the likely growth scenario, Tasmania’s population will start declining from 2031

Mapping Australian electoral divisions with ggplot2

I’ve seen some creative visualisations of issues surrounding the Australian election recently though not as many maps as I expected. ‘ggplot2’ is the go-to package for plotting in R so I thought I’d see if I could plot the Australian electoral divisions with ggplot2. By using the Australian Electoral Commission’s GIS mapping coordinates and mutilating Hadley Whickam’s tutorial it was a pretty easy process.

1. Download the AEC boundary GIS data (warning 24mb).

2. Extract the file to your R working directory.

3. Run code…

The data.frame this process creates has 2.5m observations so mapping can take a while. I’m sure there are much more effective ways to map GIS data but I wanted to stick to ggplot2 in this instance.

require("rgdal") # requires sp, will use proj.4 if installed
require("maptools")
require("ggplot2")
require("plyr")
require("rgeos")

#I upped my memory limit as the file we are going to map is pretty large
memory.limit(6000)

australia = readOGR(dsn=".", layer="COM20111216_ELB_region")
australia@data$id = rownames(australia@data)
#This step isn't in the tutorial, need to do this due to a couple of errors in the AEC GIS data.
australia.buffered = gBuffer(australia, width=0, byid=TRUE)
australia.points = fortify(australia.buffered, region="id")
australia.df = join(australia.points, australia@data, by="id")

#This will show you the variables in the dataset
head(australia@data)

ggplot(australia.df) +
aes(long,lat,group=group,fill=ELECT_DIV)+
#Don't want a legend with 150 variables so suppress the legend
geom_polygon(show_guide = FALSE ) +
  geom_path(color="white") +
  #for some reason it maps too much ocean so limit coords (EDIT: due to Christmas Island)
  coord_equal(xlim=c(110,155))

This gives you

austr

While it’s a nice picture, it’s of little use as it is impossible to see small electorates.

State by state mapping. might be more useful Here is some code to map the ACT. I suggest anyone experimenting should play around with mapping the ACT data as it doesn’t take long to process.

ggplot(subset(australia.df, STATE == "ACT")) +
  aes(long,lat,group=group,fill=ELECT_DIV)+
  geom_polygon() +
  geom_path(color="white") +
  #include limits to remove Jervis bay plotting
  coord_equal(xlim=c(148.5,149.5))

Which gives:
act

To include your own data for mapping just add it to the australia@data data.frame, merging by australia@data$ELECT_DIV. The charts look good, but to make them really eye-catching I suggest you take them into inkscape.

Attacking educational disadvantage through school funding

Nicholas Biddle asked me to contribute to a piece written for The Conversation about the funding allocated to tackling educational disadvantage under the National Plan for School Improvement. The piece on The Conversation is yet to be published and will be significantly shorter than this.

“…all students must have access to an acceptable international standard of education, regardless of where they live or the school they attend. …[equity means] differences in educational outcomes are not the result of differences in wealth, income, power or possessions” (pg 105, review of funding for schooling)

Education can be cause or cure for disadvantage within and across societies. The extent to which education reduces rather than exacerbates inequality, however, is largely determined by the quality of education. In Australia, all levels of government and the major political parties recognise the role of the public sector in funding the delivery of education. With regards to school funding, there is debate around three main questions:

  1. What should be the total level of government funding available to school education?
  2. To what extent should governments subsidise the choices made by families to send their children to non-government schools?
  3. How should the characteristics of students and schools impact on the amount of funding received?

The National Plan for School Improvement

The responses to each of these questions are different in the eight Australian States and Territories. While the Federal Government has a minimal direct role in school education, they do provide significant funds to the States and Territories. The National Plan for School Improvement (NPSI) and the Australian Education Bill 2012 represent the Federal Labor Government’s response to the three questions posed above and is an attempts to make funding more standardised across Australia.

Continue reading “Attacking educational disadvantage through school funding”

We can run deficits forever – Australian edition

As Evan Soltas neatly puts, a government can run deficits forever and at the same time reduce their liabilities relative to GDP. All that is needed is for GDP growth to be larger than the growth in government debt (in percentage terms). If we can get this point across to the general public we might avoid some of the misguided backlash caused by governments running deficits. Governments should be running deficits through the troughs of business cycles and funding this by running surpluses through the good times. Consider it the same as dipping into your savings if you’ve been made unemployed while you look for another job. With Australian GDP ≈ $1.4tn and a conservative GDP growth estimate of 2% we can currently run budgets deficits around $25bn without suffering any deterioration in our debt/GDP. Of course our nominal position will be worse but with inflation at 2-3% our real deficit will actually be shrinking as well.

On balance, I see the media engaging the issue of budget deficits as a positive as it increases political accountability, though I think the main point of focus of the media is wide of the mark. Spending too much of the windfall (thereby running smaller surpluses) in the boom years is the biggest crime here. The incentives are such that presiding governments will spend extra money that flows into their coffers during the boom time as showering the public with cash is a fast track to popularity. If the government instead saved the money it reduces their likelihood for re-election (relative to showering the public with cash) and tops up the coffers for the winner of the next election which may well be an opposing political party.

Soltas claims that the US has run deficits in 70 of its last 84 budgets. Even looking at the US Govt debt/GDP line on the chart below it is still surprising to think that 83% (!) of their budgets since 1929 were in deficit.

Fortunately for us, govt debt relative to GDP is much lower in Australia than the US and most other developed countries as the chart below shows*.

Debt-GDP Aus vs US

While the underlying Australian Govt cash position isn’t the current  ‘budget position’ measure it is similar, and there is a bit of historical data on it unlike the current measure which hasn’t been back-cast very far after recent changes. While it would be good to compare like for like, I can’t find data for the last 84 Australian budgets so data from 1970 will have to do. On page 6, this budget summary shows that Australia has run cash deficits for 23 of its last 42 budgets. From 1992 to 2012, only 10 of 21 budgets were in surplus however our debt/GDP ratio was 27% in 1992 and is at 27% now. The deficits were on average, larger than the surpluses. The fact that GDP was constantly expanding is the reason the debt position didn’t deteriorate.

Underlying Cash Position

With positive GDP growth Australia can run small budget deficits for a very long time (theoretically forever). If needed, we have room (especially relative to other developed countries) to allow larger deficits. Australia needs to focus on running surpluses and paying down debt in the during times of economic prosperity.

*Data sources for Debt/GDP are all over the place. US Treasury data on the US budget position is very different to the IMF data. Any pointers on which figures are most trustworthy/comparable across countries would be appreciated.

MOOC dropout rates – Don’t focus on the headline

I listened to this podcast on education and the internet on Econtalk a few weeks ago. The guest speaker Arnold Kling used the high dropout rate in Massive Open Online Courses (MOOCs) as evidence against their effectiveness.

By design, most MOOCs will have a huge dropout rate. This is mainly due to the low-stakes nature of signing up to a course, but also because some courses hide a lot of information behind the ‘sign-up’ button. That enrollment is so easy is great. Why make things harder than they should be? On coursera, an account takes about 30 seconds to set up and to sign up for any individual course takes the click of two buttons.

The fact information is hidden behind the ‘sign-up’ button isn’t ideal. For instance, to see the syllabus on this Combinatorial Game Theory course, you have to sign up to it. There is no reason to hide this information as it helps students with their enrollment decision. While students dropping out of MOOC courses isn’t of great consequence, the more information students have access to before enrolling, the better the data on ‘real’ dropout rates will be. When I say ‘real’ dropouts, I mean students who initially have some non-negligible level of commitment to a course and its content yet don’t complete a course due to its structure, resources, time demands or the like. By figuring out why these students  drop out, courses can be refined to deal with these problems.

I have un-enrolled from courses on coursera and have received little follow up. I suggest that MOOC providers send out an email with a short survey not only at the end of every course but also on each un-enrollment. I’d be happy to say that for one course I was never planning to take it, another that I realised I didn’t meet the prerequisites (limited C++ programming), and another I’d say that I got all I wanted out of the course in the first three weeks.

By setting the sites up a bit better and gathering better data on non-completion rates it should be easier to see what a more ‘real’ MOOC dropout is. Until then, don’t focus on the headline dropout rate.