Public spending and equality

Sunder at Next Left points out that public spending cuts might well bear harder upon the poor than the rich. This raises a question: what exactly is the link between public spending and inequality?
To kick off the enquiry, my chart shows the cross-country relationship for the 23 wealthiest OECD nations in the mid-00s. The Gini coefficient measures inequality after taxes and transfers. Data come from the OECD.
Two things stand our here.
First, there is a negative relationship. between public spending and the Gini coefficient. Higher public spending is associated with greater equality. Across the 23 nations, the correlation is 0.5. Sweden and Denmark have low inequality and high public spending. The US has low public spending and high inequality.
Second, this correlation is not very tight; the R-squared is only 21%. Several countries have both higher public spending and greater inequality than Korea or Australia, for example. This could be because some public spending is indeed a bill for failure; if inequality arises from people being too badly educated to find work, spending on welfare and fighting crime might be large.
Also, big variations in public spending exist with similar levels of inequality: France and Switzerland have the same Gini coefficients, but the French government spends 18 percentage points more of GDP.
For me, this introduces a new perspective. For years, the debate slanging match between left and right has been: should the UK move north-west in my chart to Sweden, or south-east to the US?
But this misses something: why can’t we move south-west, to Korea or Australia? How do they achieve both greater equality than us and lower public spending? They show that smaller government and greater equality are feasible. What conditions must exist for this to be the case?*
If the coalition were sincere about wanting “progressive cuts” - which I doubt - it would answer this question before blindly slashing spending.
* It could well be ones that can’t be created except in the very long-term, such as a highly educated population or an ethos of justice.

Comments

[Jimmy Hill]

"But this misses something: why can’t we move south-east, to Korea or Australia? How do they achieve both greater equality than us and lower public spending?"

Could it be that Korea and Australia and more equal before taxes and transfers?

[william]

Countries with high inequality and high spending just might be spending it all on the well-off.

[Duncan Stott]

(I think you mean "south-*west*, to Korea or Australia".)

[Luis Enrique]

It's not obvious that changes to how the government taxes and spend are what causes countries to move around that chart. I know you are not claiming it is, but one might read this post and think that if the government did things differently, we could move south west.

It may be, for example, that inequality in OECD countries is to do with the industrial mix in each country.

I wonder why Portugal appears to have such an unfavorable mix ... corruption?

[chris]

@Duncan - thats. Change is made.
@ Jimmy. Australia has the same pre-tax and transfer Gini as us; Korea has a much lower one.
One reason for Korea's lower Gini, I suspect, lies in that ethos of justice. The Koreans aren't so stupid as to believe in superstar bosses, so don't pay them masses more than workers.
@ Luis - absolutely. There are loads of reasons for the (lack of) relationship in the chart. I'm just asking what they are.

[Jimmy Hill]

@chris

How does the data take account of benefits in kind? This may distort the level of inequality between countries. Sorry for the dull questions.

[Leigh Caldwell]

Which of the OECD public spending tables have you used? General government total outlays or real public consumption? That is, including or excluding transfers?

Transfers are, almost by definition, going to contribute to the (NW to SE) correlation you identify. Assuming you have included transfers, it would be interesting to see the equivalent graph for public consumption only, as I suspect the correlation would be even weaker - if it is still present at all.

[chris]

Leigh - I used total outlays. I'll check how government consumption changes things. It's notable though that even including transfers in spending gives us only a weakish positive correlation. This could be because lots of transfers are in fact to middle-income groups.
Jimmy - the Gini exclude benefits in kind. But allocating these is very tricky. I suspect, for example, that for many kids from poor backgrounds, education is not a benefit at all, but a burden, as would be spending on the police.

[chris]

Leigh - I've redone the data using only government consumption, and the pattern is exactly the same; the correlation between govt consumption and the Gini is -0.49. The outlying points look the same too: Denmark and Sweden are high spenders, Korea and the US are low spenders.
One reason for this is that the correlation between consumption and total outlays is high (0.81).

[Harry]

@Chris
A podcast I listened to with the author of the book No Shame in My Game: http://www.amazon.co.uk/No-Shame-My-Game-Working/dp/0375703799/ref=sr_1_9?ie=UTF8&s=books&qid=1277992041&sr=8-9 suggested that actually education is highly valued by the working poor - so much so that in America many carry on going to high school in evening classes for years to graduate, and that gaining educational qualifications can and does improve their prospects in at least some cases. In the podcast she suggested that the reason they didn't graduate high school when actually school age was that it was too violent and dangerous and the level of teaching too poor, not a lack of desire. I'm not sure this situation is the same in Britain, and I'd be intrigued to know how the life chances of the working poor and the usefulness of education for them varies in Britain from America (the author was surprised by what she discovered about the working poor in America when she studied the situation).

[Serious Implications]

It would be interesting to look at the composition of the economies of the various countries, e.g., which are more dominated by industrial production vs. financial services? Seems Korea might be more of an industrial-based economy and Australia an agricultural-based economy vs. the US or UK. To what extent are the economies unionized?

[Nick]

Your use of R-square to describe how "tight" the correlation is strange. R-squre = correlation squared. The r in r-square refers to the pearson correlation coefficent. HTH